Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6: (44 commits) sparc: Remove Sparc's asm-offsets for sclow.S sparc64: Update defconfig. sparc64: Add Niagara2 RNG driver. sparc64: Add missing hypervisor service group numbers. sparc64: Remove 4MB and 512K base page size options. sparc64: Convert to generic helpers for IPI function calls. sparc: Use new '%pS' infrastructure to print symbols. sparc32: fix init.c allnoconfig build error sparc64: Config category "Processor type and features" absent sparc: arch/sparc/kernel/apc.c to unlocked_ioctl sparc: join the remaining header files sparc: merge header files with trivial differences sparc: when header files are equal use asm-sparc version sparc: copy sparc64 specific files to asm-sparc sparc: Merge asm-sparc{,64}/asi.h sparc: export openprom.h to userspace sparc: Merge asm-sparc{,64}/types.h sparc: Merge asm-sparc{,64}/termios.h sparc: Merge asm-sparc{,64}/termbits.h sparc: Merge asm-sparc{,64}/setup.h ...
author: Linus Torvalds <torvalds@linux-foundation.org> 2008-07-21 09:40:26 -0700
committer: Linus Torvalds <torvalds@linux-foundation.org> 2008-07-21 09:40:26 -0700
commit: f8b71a3a92d69cb1c9c2162f9235d3dd550e4aa0 (patch)
tree: d0853186f33314a6e45903876d0e94d1619fcca9
parent: 42a77a1b8fe4fd78cf8d8fccbaac918bbe888192 (diff)
parent: e4f25060b87a627f5cda84b8134911d43c919458 (diff)
462 files changed, 26902 insertions, 27102 deletions
diff --git a/arch/sparc/kernel/apc.c b/arch/sparc/kernel/apc.c
index 6707422c984..5267d48fb2c 100644
--- a/arch/sparc/kernel/apc.c
+++ b/arch/sparc/kernel/apc.c
@@ -56,7 +56,7 @@ __setup("apc=", apc_setup);
  * CPU idle callback function
  * See .../arch/sparc/kernel/process.c
  */
-void apc_swift_idle(void)
+static void apc_swift_idle(void)
 {
 #ifdef APC_DEBUG_LED
 	set_auxio(0x00, AUXIO_LED); 
@@ -85,54 +85,70 @@ static int apc_release(struct inode *inode, struct file *f)
 	return 0;
 }
 
-static int apc_ioctl(struct inode *inode, struct file *f, 
-		     unsigned int cmd, unsigned long __arg)
+static long apc_ioctl(struct file *f, unsigned int cmd, unsigned long __arg)
 {
 	__u8 inarg, __user *arg;
 
 	arg = (__u8 __user *) __arg;
+
+	lock_kernel();
+
 	switch (cmd) {
 	case APCIOCGFANCTL:
-		if (put_user(apc_readb(APC_FANCTL_REG) & APC_REGMASK, arg))
-				return -EFAULT;
+		if (put_user(apc_readb(APC_FANCTL_REG) & APC_REGMASK, arg)) {
+			unlock_kernel();
+			return -EFAULT;
+		}
 		break;
 
 	case APCIOCGCPWR:
-		if (put_user(apc_readb(APC_CPOWER_REG) & APC_REGMASK, arg))
+		if (put_user(apc_readb(APC_CPOWER_REG) & APC_REGMASK, arg)) {
+			unlock_kernel();
 			return -EFAULT;
+		}
 		break;
 
 	case APCIOCGBPORT:
-		if (put_user(apc_readb(APC_BPORT_REG) & APC_BPMASK, arg))
+		if (put_user(apc_readb(APC_BPORT_REG) & APC_BPMASK, arg)) {
+			unlock_kernel();
 			return -EFAULT;
+		}
 		break;
 
 	case APCIOCSFANCTL:
-		if (get_user(inarg, arg))
+		if (get_user(inarg, arg)) {
+			unlock_kernel();
 			return -EFAULT;
+		}
 		apc_writeb(inarg & APC_REGMASK, APC_FANCTL_REG);
 		break;
 	case APCIOCSCPWR:
-		if (get_user(inarg, arg))
+		if (get_user(inarg, arg)) {
+			unlock_kernel();
 			return -EFAULT;
+		}
 		apc_writeb(inarg & APC_REGMASK, APC_CPOWER_REG);
 		break;
 	case APCIOCSBPORT:
-		if (get_user(inarg, arg))
+		if (get_user(inarg, arg)) {
+			unlock_kernel();
 			return -EFAULT;
+		}
 		apc_writeb(inarg & APC_BPMASK, APC_BPORT_REG);
 		break;
 	default:
+		unlock_kernel();
 		return -EINVAL;
 	};
 
+	unlock_kernel();
 	return 0;
 }
 
 static const struct file_operations apc_fops = {
-	.ioctl =	apc_ioctl,
-	.open =		apc_open,
-	.release =	apc_release,
+	.unlocked_ioctl =	apc_ioctl,
+	.open =			apc_open,
+	.release =		apc_release,
 };
 
 static struct miscdevice apc_miscdev = { APC_MINOR, APC_DEVNAME, &apc_fops };
diff --git a/arch/sparc/kernel/asm-offsets.c b/arch/sparc/kernel/asm-offsets.c
index cd3f7694e9b..b5bb99ed892 100644
--- a/arch/sparc/kernel/asm-offsets.c
+++ b/arch/sparc/kernel/asm-offsets.c
@@ -18,18 +18,6 @@ int foo(void)
 {
 	DEFINE(AOFF_task_thread, offsetof(struct task_struct, thread));
 	BLANK();
-	/* XXX This is the stuff for sclow.S, kill it. */
-	DEFINE(AOFF_task_pid, offsetof(struct task_struct, pid));
-	DEFINE(AOFF_task_uid, offsetof(struct task_struct, uid));
-	DEFINE(AOFF_task_gid, offsetof(struct task_struct, gid));
-	DEFINE(AOFF_task_euid, offsetof(struct task_struct, euid));
-	DEFINE(AOFF_task_egid, offsetof(struct task_struct, egid));
-	/* DEFINE(THREAD_INFO, offsetof(struct task_struct, stack)); */
-	DEFINE(ASIZ_task_uid,	sizeof(current->uid));
-	DEFINE(ASIZ_task_gid,	sizeof(current->gid));
-	DEFINE(ASIZ_task_euid,	sizeof(current->euid));
-	DEFINE(ASIZ_task_egid,	sizeof(current->egid));
-	BLANK();
 	DEFINE(AOFF_thread_fork_kpsr,
 			offsetof(struct thread_struct, fork_kpsr));
 	BLANK();
diff --git a/arch/sparc/kernel/ebus.c b/arch/sparc/kernel/ebus.c
index 92c6fc07e59..97294232259 100644
--- a/arch/sparc/kernel/ebus.c
+++ b/arch/sparc/kernel/ebus.c
@@ -69,7 +69,7 @@ static inline unsigned long ebus_alloc(size_t size)
 
 /*
  */
-int __init ebus_blacklist_irq(const char *name)
+static int __init ebus_blacklist_irq(const char *name)
 {
 	struct ebus_device_irq *dp;
 
@@ -83,8 +83,8 @@ int __init ebus_blacklist_irq(const char *name)
 	return 0;
 }
 
-void __init fill_ebus_child(struct device_node *dp,
-			    struct linux_ebus_child *dev)
+static void __init fill_ebus_child(struct device_node *dp,
+				   struct linux_ebus_child *dev)
 {
 	const int *regs;
 	const int *irqs;
@@ -144,7 +144,8 @@ void __init fill_ebus_child(struct device_node *dp,
 	}
 }
 
-void __init fill_ebus_device(struct device_node *dp, struct linux_ebus_device *dev)
+static void __init fill_ebus_device(struct device_node *dp,
+				    struct linux_ebus_device *dev)
 {
 	const struct linux_prom_registers *regs;
 	struct linux_ebus_child *child;
diff --git a/arch/sparc/kernel/entry.S b/arch/sparc/kernel/entry.S
index 4bcfe54f878..2f96256dc51 100644
--- a/arch/sparc/kernel/entry.S
+++ b/arch/sparc/kernel/entry.S
@@ -19,6 +19,7 @@
 #include <asm/vaddrs.h>
 #include <asm/memreg.h>
 #include <asm/page.h>
+#include <asm/pgtable.h>
 #ifdef CONFIG_SUN4
 #include <asm/pgtsun4.h>
 #else
@@ -1317,7 +1318,6 @@ linux_sparc_syscall:
 	bne	linux_fast_syscall
 	 /* Just do first insn from SAVE_ALL in the delay slot */
 
-	.globl	syscall_is_too_hard
 syscall_is_too_hard:
 	SAVE_ALL_HEAD
 	 rd	%wim, %l3
@@ -1544,8 +1544,7 @@ kgdb_trap_low:
 #endif
 
 	.align	4
-	.globl	__handle_exception, flush_patch_exception
-__handle_exception:
+	.globl	flush_patch_exception
 flush_patch_exception:
 	FLUSH_ALL_KERNEL_WINDOWS;
 	ldd	[%o0], %o6
diff --git a/arch/sparc/kernel/etrap.S b/arch/sparc/kernel/etrap.S
index f37d961d67a..e806fcdc46d 100644
--- a/arch/sparc/kernel/etrap.S
+++ b/arch/sparc/kernel/etrap.S
@@ -228,7 +228,6 @@ tsetup_mmu_patchme:
 	 */
 #define glob_tmp     g1
 
-	.globl	tsetup_sun4c_stackchk
 tsetup_sun4c_stackchk:
 	/* Done by caller: andcc %sp, 0x7, %g0 */
 	bne	trap_setup_user_stack_is_bolixed
diff --git a/arch/sparc/kernel/head.S b/arch/sparc/kernel/head.S
index 3bfd6085a91..50d9a16af79 100644
--- a/arch/sparc/kernel/head.S
+++ b/arch/sparc/kernel/head.S
@@ -32,7 +32,6 @@
  */
 
 	.align 4
-        .globl  cputyp
 cputyp:
         .word   1
 
@@ -1280,7 +1279,6 @@ halt_me:
  * gets initialized in c-code so all routines can use it.
  */
 
-	.globl	prom_vector_p
 prom_vector_p:
 		.word 0
 
diff --git a/arch/sparc/kernel/idprom.c b/arch/sparc/kernel/idprom.c
index 7220562cdb3..fc511f3c4c1 100644
--- a/arch/sparc/kernel/idprom.c
+++ b/arch/sparc/kernel/idprom.c
@@ -24,7 +24,7 @@ static struct idprom idprom_buffer;
  * of the Sparc CPU and have a meaningful IDPROM machtype value that we
  * know about.  See asm-sparc/machines.h for empirical constants.
  */
-struct Sun_Machine_Models Sun_Machines[NUM_SUN_MACHINES] = {
+static struct Sun_Machine_Models Sun_Machines[NUM_SUN_MACHINES] = {
 /* First, Sun4's */
 { "Sun 4/100 Series", (SM_SUN4 | SM_4_110) },
 { "Sun 4/200 Series", (SM_SUN4 | SM_4_260) },
diff --git a/arch/sparc/kernel/ioport.c b/arch/sparc/kernel/ioport.c
index 7b17522f59b..487960919f1 100644
--- a/arch/sparc/kernel/ioport.c
+++ b/arch/sparc/kernel/ioport.c
@@ -49,13 +49,16 @@
 
 #define mmu_inval_dma_area(p, l)	/* Anton pulled it out for 2.4.0-xx */
 
-struct resource *_sparc_find_resource(struct resource *r, unsigned long);
+static struct resource *_sparc_find_resource(struct resource *r,
+					     unsigned long);
 
 static void __iomem *_sparc_ioremap(struct resource *res, u32 bus, u32 pa, int sz);
 static void __iomem *_sparc_alloc_io(unsigned int busno, unsigned long phys,
     unsigned long size, char *name);
 static void _sparc_free_io(struct resource *res);
 
+static void register_proc_sparc_ioport(void);
+
 /* This points to the next to use virtual memory for DVMA mappings */
 static struct resource _sparc_dvma = {
 	.name = "sparc_dvma", .start = DVMA_VADDR, .end = DVMA_END - 1
@@ -539,8 +542,6 @@ void __init sbus_setup_arch_props(struct sbus_bus *sbus, struct device_node *dp)
 
 int __init sbus_arch_preinit(void)
 {
-	extern void register_proc_sparc_ioport(void);
-
 	register_proc_sparc_ioport();
 
 #ifdef CONFIG_SUN4
@@ -853,8 +854,8 @@ _sparc_io_get_info(char *buf, char **start, off_t fpos, int length, int *eof,
  * XXX Too slow. Can have 8192 DVMA pages on sun4m in the worst case.
  * This probably warrants some sort of hashing.
  */
-struct resource *
-_sparc_find_resource(struct resource *root, unsigned long hit)
+static struct resource *_sparc_find_resource(struct resource *root,
+					     unsigned long hit)
 {
         struct resource *tmp;
 
@@ -865,7 +866,7 @@ _sparc_find_resource(struct resource *root, unsigned long hit)
 	return NULL;
 }
 
-void register_proc_sparc_ioport(void)
+static void register_proc_sparc_ioport(void)
 {
 #ifdef CONFIG_PROC_FS
 	create_proc_read_entry("io_map",0,NULL,_sparc_io_get_info,&sparc_iomap);
diff --git a/arch/sparc/kernel/irq.c b/arch/sparc/kernel/irq.c
index 087390b092b..93e1d1c6529 100644
--- a/arch/sparc/kernel/irq.c
+++ b/arch/sparc/kernel/irq.c
@@ -154,7 +154,7 @@ void (*sparc_init_timers)(irq_handler_t ) =
 struct irqaction static_irqaction[MAX_STATIC_ALLOC];
 int static_irq_count;
 
-struct {
+static struct {
 	struct irqaction *action;
 	int flags;
 } sparc_irq[NR_IRQS];
diff --git a/arch/sparc/kernel/process.c b/arch/sparc/kernel/process.c
index da48d248cc1..4bb430940a6 100644
--- a/arch/sparc/kernel/process.c
+++ b/arch/sparc/kernel/process.c
@@ -1,6 +1,6 @@
 /*  linux/arch/sparc/kernel/process.c
  *
- *  Copyright (C) 1995 David S. Miller (davem@davemloft.net)
+ *  Copyright (C) 1995, 2008 David S. Miller (davem@davemloft.net)
  *  Copyright (C) 1996 Eddie C. Dost   (ecd@skynet.be)
  */
 
@@ -14,7 +14,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
-#include <linux/kallsyms.h>
 #include <linux/mm.h>
 #include <linux/stddef.h>
 #include <linux/ptrace.h>
@@ -177,6 +176,8 @@ void machine_power_off(void)
 	machine_halt();
 }
 
+#if 0
+
 static DEFINE_SPINLOCK(sparc_backtrace_lock);
 
 void __show_backtrace(unsigned long fp)
@@ -196,7 +197,7 @@ void __show_backtrace(unsigned long fp)
 		       rw->ins[4], rw->ins[5],
 		       rw->ins[6],
 		       rw->ins[7]);
-		print_symbol("%s\n", rw->ins[7]);
+		printk("%pS\n", (void *) rw->ins[7]);
 		rw = (struct reg_window *) rw->ins[6];
 	}
 	spin_unlock_irqrestore(&sparc_backtrace_lock, flags);
@@ -228,7 +229,6 @@ void smp_show_backtrace_all_cpus(void)
 }
 #endif
 
-#if 0
 void show_stackframe(struct sparc_stackf *sf)
 {
 	unsigned long size;
@@ -264,14 +264,14 @@ void show_regs(struct pt_regs *r)
 
         printk("PSR: %08lx PC: %08lx NPC: %08lx Y: %08lx    %s\n",
 	       r->psr, r->pc, r->npc, r->y, print_tainted());
-	print_symbol("PC: <%s>\n", r->pc);
+	printk("PC: <%pS>\n", (void *) r->pc);
 	printk("%%G: %08lx %08lx  %08lx %08lx  %08lx %08lx  %08lx %08lx\n",
 	       r->u_regs[0], r->u_regs[1], r->u_regs[2], r->u_regs[3],
 	       r->u_regs[4], r->u_regs[5], r->u_regs[6], r->u_regs[7]);
 	printk("%%O: %08lx %08lx  %08lx %08lx  %08lx %08lx  %08lx %08lx\n",
 	       r->u_regs[8], r->u_regs[9], r->u_regs[10], r->u_regs[11],
 	       r->u_regs[12], r->u_regs[13], r->u_regs[14], r->u_regs[15]);
-	print_symbol("RPC: <%s>\n", r->u_regs[15]);
+	printk("RPC: <%pS>\n", (void *) r->u_regs[15]);
 
 	printk("%%L: %08lx %08lx  %08lx %08lx  %08lx %08lx  %08lx %08lx\n",
 	       rw->locals[0], rw->locals[1], rw->locals[2], rw->locals[3],
@@ -306,7 +306,7 @@ void show_stack(struct task_struct *tsk, unsigned long *_ksp)
 		rw = (struct reg_window *) fp;
 		pc = rw->ins[7];
 		printk("[%08lx : ", pc);
-		print_symbol("%s ] ", pc);
+		printk("%pS ] ", (void *) pc);
 		fp = rw->ins[6];
 	} while (++count < 16);
 	printk("\n");
diff --git a/arch/sparc/kernel/rtrap.S b/arch/sparc/kernel/rtrap.S
index ce30082ab26..891f460b7b9 100644
--- a/arch/sparc/kernel/rtrap.S
+++ b/arch/sparc/kernel/rtrap.S
@@ -224,8 +224,6 @@ ret_trap_user_stack_is_bolixed:
 	b	signal_p
 	 ld	[%curptr + TI_FLAGS], %g2
 
-
-	.globl	sun4c_rett_stackchk
 sun4c_rett_stackchk:
 	be	1f
 	 and	%fp, 0xfff, %g1		! delay slot
diff --git a/arch/sparc/kernel/setup.c b/arch/sparc/kernel/setup.c
index a0ea0bc6f47..9e451b21202 100644
--- a/arch/sparc/kernel/setup.c
+++ b/arch/sparc/kernel/setup.c
@@ -67,7 +67,7 @@ struct screen_info screen_info = {
 extern unsigned long trapbase;
 
 /* Pretty sick eh? */
-void prom_sync_me(void)
+static void prom_sync_me(void)
 {
 	unsigned long prom_tbr, flags;
 
@@ -97,7 +97,7 @@ void prom_sync_me(void)
 	return;
 }
 
-unsigned int boot_flags __initdata = 0;
+static unsigned int boot_flags __initdata = 0;
 #define BOOTME_DEBUG  0x1
 
 /* Exported for mm/init.c:paging_init. */
diff --git a/arch/sparc/kernel/smp.c b/arch/sparc/kernel/smp.c
index 6724ab90f82..1619ec15c09 100644
--- a/arch/sparc/kernel/smp.c
+++ b/arch/sparc/kernel/smp.c
@@ -35,13 +35,9 @@
 
 #include "irq.h"
 
-int smp_num_cpus = 1;
 volatile unsigned long cpu_callin_map[NR_CPUS] __initdata = {0,};
 unsigned char boot_cpu_id = 0;
 unsigned char boot_cpu_id4 = 0; /* boot_cpu_id << 2 */
-int smp_activated = 0;
-volatile int __cpu_number_map[NR_CPUS];
-volatile int __cpu_logical_map[NR_CPUS];
 
 cpumask_t cpu_online_map = CPU_MASK_NONE;
 cpumask_t phys_cpu_present_map = CPU_MASK_NONE;
@@ -55,9 +51,6 @@ cpumask_t smp_commenced_mask = CPU_MASK_NONE;
  * instruction which is much better...
  */
 
-/* Used to make bitops atomic */
-unsigned char bitops_spinlock = 0;
-
 void __cpuinit smp_store_cpu_info(int id)
 {
 	int cpu_node;
diff --git a/arch/sparc/kernel/sun4c_irq.c b/arch/sparc/kernel/sun4c_irq.c
index c6ac9fc5256..340fc395fe2 100644
--- a/arch/sparc/kernel/sun4c_irq.c
+++ b/arch/sparc/kernel/sun4c_irq.c
@@ -68,7 +68,8 @@ unsigned char *interrupt_enable = NULL;
 
 static int sun4c_pil_map[] = { 0, 1, 2, 3, 5, 7, 8, 9 };
 
-unsigned int sun4c_sbint_to_irq(struct sbus_dev *sdev, unsigned int sbint)
+static unsigned int sun4c_sbint_to_irq(struct sbus_dev *sdev,
+				       unsigned int sbint)
 {
 	if (sbint >= sizeof(sun4c_pil_map)) {
 		printk(KERN_ERR "%s: bogus SBINT %d\n", sdev->prom_name, sbint);
diff --git a/arch/sparc/kernel/sun4d_irq.c b/arch/sparc/kernel/sun4d_irq.c
index 8ac5661caff..1290b5998f8 100644
--- a/arch/sparc/kernel/sun4d_irq.c
+++ b/arch/sparc/kernel/sun4d_irq.c
@@ -52,13 +52,13 @@ extern struct irqaction static_irqaction[MAX_STATIC_ALLOC];
 extern int static_irq_count;
 unsigned char cpu_leds[32];
 #ifdef CONFIG_SMP
-unsigned char sbus_tid[32];
+static unsigned char sbus_tid[32];
 #endif
 
 static struct irqaction *irq_action[NR_IRQS];
 extern spinlock_t irq_action_lock;
 
-struct sbus_action {
+static struct sbus_action {
 	struct irqaction *action;
 	/* For SMP this needs to be extended */
 } *sbus_actions;
@@ -267,7 +267,8 @@ unsigned int sun4d_build_irq(struct sbus_dev *sdev, int irq)
 		return irq;
 }
 
-unsigned int sun4d_sbint_to_irq(struct sbus_dev *sdev, unsigned int sbint)
+static unsigned int sun4d_sbint_to_irq(struct sbus_dev *sdev,
+				       unsigned int sbint)
 {
 	if (sbint >= sizeof(sbus_to_pil)) {
 		printk(KERN_ERR "%s: bogus SBINT %d\n", sdev->prom_name, sbint);
diff --git a/arch/sparc/kernel/sun4m_irq.c b/arch/sparc/kernel/sun4m_irq.c
index b92d6d2d5b0..94e02de960e 100644
--- a/arch/sparc/kernel/sun4m_irq.c
+++ b/arch/sparc/kernel/sun4m_irq.c
@@ -154,7 +154,8 @@ static unsigned long irq_mask[] = {
 
 static int sun4m_pil_map[] = { 0, 2, 3, 5, 7, 9, 11, 13 };
 
-unsigned int sun4m_sbint_to_irq(struct sbus_dev *sdev, unsigned int sbint) 
+static unsigned int sun4m_sbint_to_irq(struct sbus_dev *sdev,
+				       unsigned int sbint)
 {
 	if (sbint >= sizeof(sun4m_pil_map)) {
 		printk(KERN_ERR "%s: bogus SBINT %d\n", sdev->prom_name, sbint);
@@ -163,7 +164,7 @@ unsigned int sun4m_sbint_to_irq(struct sbus_dev *sdev, unsigned int sbint)
 	return sun4m_pil_map[sbint] | 0x30;
 }
 
-inline unsigned long sun4m_get_irqmask(unsigned int irq)
+static unsigned long sun4m_get_irqmask(unsigned int irq)
 {
 	unsigned long mask;
     
@@ -281,7 +282,7 @@ static void sun4m_set_udt(int cpu)
 #define TIMER_IRQ  	(OBIO_INTR | 10)
 #define PROFILE_IRQ	(OBIO_INTR | 14)
 
-struct sun4m_timer_regs *sun4m_timers;
+static struct sun4m_timer_regs *sun4m_timers;
 unsigned int lvl14_resolution = (((1000000/HZ) + 1) << 10);
 
 static void sun4m_clear_clock_irq(void)
diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c
index ffb875aacb7..406ac1abc83 100644
--- a/arch/sparc/kernel/sun4m_smp.c
+++ b/arch/sparc/kernel/sun4m_smp.c
@@ -244,8 +244,9 @@ static struct smp_funcall {
 static DEFINE_SPINLOCK(cross_call_lock);
 
 /* Cross calls must be serialized, at least currently. */
-void smp4m_cross_call(smpfunc_t func, unsigned long arg1, unsigned long arg2,
-		    unsigned long arg3, unsigned long arg4, unsigned long arg5)
+static void smp4m_cross_call(smpfunc_t func, unsigned long arg1,
+			     unsigned long arg2, unsigned long arg3,
+			     unsigned long arg4, unsigned long arg5)
 {
 		register int ncpus = SUN4M_NCPUS;
 		unsigned long flags;
@@ -344,7 +345,7 @@ static void __init smp_setup_percpu_timer(void)
 		enable_pil_irq(14);
 }
 
-void __init smp4m_blackbox_id(unsigned *addr)
+static void __init smp4m_blackbox_id(unsigned *addr)
 {
 	int rd = *addr & 0x3e000000;
 	int rs1 = rd >> 11;
@@ -354,7 +355,7 @@ void __init smp4m_blackbox_id(unsigned *addr)
 	addr[2] = 0x80082003 | rd | rs1;	/* and reg, 3, reg */
 }
 
-void __init smp4m_blackbox_current(unsigned *addr)
+static void __init smp4m_blackbox_current(unsigned *addr)
 {
 	int rd = *addr & 0x3e000000;
 	int rs1 = rd >> 11;
diff --git a/arch/sparc/kernel/time.c b/arch/sparc/kernel/time.c
index 53caacbb398..ab3dd0b257d 100644
--- a/arch/sparc/kernel/time.c
+++ b/arch/sparc/kernel/time.c
@@ -46,7 +46,7 @@
 #include "irq.h"
 
 DEFINE_SPINLOCK(rtc_lock);
-enum sparc_clock_type sp_clock_typ;
+static enum sparc_clock_type sp_clock_typ;
 DEFINE_SPINLOCK(mostek_lock);
 void __iomem *mstk48t02_regs = NULL;
 static struct mostek48t08 __iomem *mstk48t08_regs = NULL;
@@ -366,7 +366,7 @@ static int __init clock_init(void)
 fs_initcall(clock_init);
 #endif /* !CONFIG_SUN4 */
 
-void __init sbus_time_init(void)
+static void __init sbus_time_init(void)
 {
 
 	BTFIXUPSET_CALL(bus_do_settimeofday, sbus_do_settimeofday, BTFIXUPCALL_NORM);
diff --git a/arch/sparc/kernel/traps.c b/arch/sparc/kernel/traps.c
index 978e9d85949..5d45d5fd8c9 100644
--- a/arch/sparc/kernel/traps.c
+++ b/arch/sparc/kernel/traps.c
@@ -1,7 +1,7 @@
 /*
  * arch/sparc/kernel/traps.c
  *
- * Copyright 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright 1995, 2008 David S. Miller (davem@davemloft.net)
  * Copyright 2000 Jakub Jelinek (jakub@redhat.com)
  */
 
@@ -11,7 +11,6 @@
 
 #include <linux/sched.h>  /* for jiffies */
 #include <linux/kernel.h>
-#include <linux/kallsyms.h>
 #include <linux/signal.h>
 #include <linux/smp.h>
 #include <linux/smp_lock.h>
@@ -33,9 +32,6 @@ struct trap_trace_entry {
 	unsigned long type;
 };
 
-int trap_curbuf = 0;
-struct trap_trace_entry trapbuf[1024];
-
 void syscall_trace_entry(struct pt_regs *regs)
 {
 	printk("%s[%d]: ", current->comm, task_pid_nr(current));
@@ -72,7 +68,7 @@ void sun4d_nmi(struct pt_regs *regs)
 	prom_halt();
 }
 
-void instruction_dump (unsigned long *pc)
+static void instruction_dump(unsigned long *pc)
 {
 	int i;
 	
@@ -119,8 +115,8 @@ void die_if_kernel(char *str, struct pt_regs *regs)
 		      count++ < 30				&&
                       (((unsigned long) rw) >= PAGE_OFFSET)	&&
 		      !(((unsigned long) rw) & 0x7)) {
-			printk("Caller[%08lx]", rw->ins[7]);
-			print_symbol(": %s\n", rw->ins[7]);
+			printk("Caller[%08lx]: %pS\n", rw->ins[7],
+			       (void *) rw->ins[7]);
 			rw = (struct reg_window *)rw->ins[6];
 		}
 	}
@@ -479,10 +475,6 @@ void do_BUG(const char *file, int line)
 
 extern void sparc_cpu_startup(void);
 
-int linux_smp_still_initting;
-unsigned int thiscpus_tbr;
-int thiscpus_mid;
-
 void trap_init(void)
 {
 	extern void thread_info_offsets_are_bolixed_pete(void);
diff --git a/arch/sparc/kernel/wof.S b/arch/sparc/kernel/wof.S
index 4bce38dfe3c..3bbcd8dc9ab 100644
--- a/arch/sparc/kernel/wof.S
+++ b/arch/sparc/kernel/wof.S
@@ -306,7 +306,6 @@ spwin_bad_ustack_from_kernel:
  * As noted above %curptr cannot be touched by this routine at all.
  */
 
-	.globl	spwin_sun4c_stackchk
 spwin_sun4c_stackchk:
 	/* LOCATION: Window to be saved on the stack */
 
diff --git a/arch/sparc/kernel/wuf.S b/arch/sparc/kernel/wuf.S
index 82e5145b0f7..779ff750603 100644
--- a/arch/sparc/kernel/wuf.S
+++ b/arch/sparc/kernel/wuf.S
@@ -243,7 +243,6 @@ fwin_user_finish_up:
 	 */
 
 	.align	4
-	.globl	sun4c_fwin_stackchk
 sun4c_fwin_stackchk:
 	/* LOCATION: Window 'W' */
 
diff --git a/arch/sparc/mm/fault.c b/arch/sparc/mm/fault.c
index 0a3cd8f6cfe..3604c2e8670 100644
--- a/arch/sparc/mm/fault.c
+++ b/arch/sparc/mm/fault.c
@@ -451,7 +451,7 @@ asmlinkage void do_sun4c_fault(struct pt_regs *regs, int text_fault, int write,
 }
 
 /* This always deals with user addresses. */
-inline void force_user_fault(unsigned long address, int write)
+static void force_user_fault(unsigned long address, int write)
 {
 	struct vm_area_struct *vma;
 	struct task_struct *tsk = current;
diff --git a/arch/sparc/mm/init.c b/arch/sparc/mm/init.c
index 7794ecb896e..e103f1bb377 100644
--- a/arch/sparc/mm/init.c
+++ b/arch/sparc/mm/init.c
@@ -22,6 +22,7 @@
 #include <linux/init.h>
 #include <linux/highmem.h>
 #include <linux/bootmem.h>
+#include <linux/pagemap.h>
 
 #include <asm/system.h>
 #include <asm/vac-ops.h>
@@ -128,7 +129,7 @@ unsigned long calc_highpages(void)
 	return nr;
 }
 
-unsigned long calc_max_low_pfn(void)
+static unsigned long calc_max_low_pfn(void)
 {
 	int i;
 	unsigned long tmp = pfn_base + (SRMMU_MAXMEM >> PAGE_SHIFT);
@@ -292,7 +293,7 @@ unsigned long __init bootmem_init(unsigned long *pages_avail)
  *
  * We simply copy the 2.4 implementation for now.
  */
-int pgt_cache_water[2] = { 25, 50 };
+static int pgt_cache_water[2] = { 25, 50 };
 
 void check_pgt_cache(void)
 {
@@ -356,8 +357,6 @@ void __init paging_init(void)
 	device_scan();
 }
 
-struct cache_palias *sparc_aliases;
-
 static void __init taint_real_pages(void)
 {
 	int i;
@@ -375,7 +374,7 @@ static void __init taint_real_pages(void)
 	}
 }
 
-void map_high_region(unsigned long start_pfn, unsigned long end_pfn)
+static void map_high_region(unsigned long start_pfn, unsigned long end_pfn)
 {
 	unsigned long tmp;
 
diff --git a/arch/sparc/mm/srmmu.c b/arch/sparc/mm/srmmu.c
index 23d3291a3e8..c624e04ff03 100644
--- a/arch/sparc/mm/srmmu.c
+++ b/arch/sparc/mm/srmmu.c
@@ -50,7 +50,7 @@
 #include <asm/btfixup.h>
 
 enum mbus_module srmmu_modtype;
-unsigned int hwbug_bitmask;
+static unsigned int hwbug_bitmask;
 int vac_cache_size;
 int vac_line_size;
 
@@ -60,7 +60,7 @@ extern unsigned long last_valid_pfn;
 
 extern unsigned long page_kernel;
 
-pgd_t *srmmu_swapper_pg_dir;
+static pgd_t *srmmu_swapper_pg_dir;
 
 #ifdef CONFIG_SMP
 #define FLUSH_BEGIN(mm)
@@ -83,12 +83,12 @@ BTFIXUPDEF_CALL(void, local_flush_page_for_dma, unsigned long)
 char *srmmu_name;
 
 ctxd_t *srmmu_ctx_table_phys;
-ctxd_t *srmmu_context_table;
+static ctxd_t *srmmu_context_table;
 
 int viking_mxcc_present;
 static DEFINE_SPINLOCK(srmmu_context_spinlock);
 
-int is_hypersparc;
+static int is_hypersparc;
 
 /*
  * In general all page table modifications should use the V8 atomic
@@ -112,11 +112,11 @@ static inline int srmmu_device_memory(unsigned long x)
 	return ((x & 0xF0000000) != 0);
 }
 
-int srmmu_cache_pagetables;
+static int srmmu_cache_pagetables;
 
 /* these will be initialized in srmmu_nocache_calcsize() */
-unsigned long srmmu_nocache_size;
-unsigned long srmmu_nocache_end;
+static unsigned long srmmu_nocache_size;
+static unsigned long srmmu_nocache_end;
 
 /* 1 bit <=> 256 bytes of nocache <=> 64 PTEs */
 #define SRMMU_NOCACHE_BITMAP_SHIFT (PAGE_SHIFT - 4)
@@ -324,7 +324,7 @@ static unsigned long __srmmu_get_nocache(int size, int align)
 	return (SRMMU_NOCACHE_VADDR + (offset << SRMMU_NOCACHE_BITMAP_SHIFT));
 }
 
-unsigned inline long srmmu_get_nocache(int size, int align)
+static unsigned long srmmu_get_nocache(int size, int align)
 {
 	unsigned long tmp;
 
@@ -336,7 +336,7 @@ unsigned inline long srmmu_get_nocache(int size, int align)
 	return tmp;
 }
 
-void srmmu_free_nocache(unsigned long vaddr, int size)
+static void srmmu_free_nocache(unsigned long vaddr, int size)
 {
 	int offset;
 
@@ -369,7 +369,8 @@ void srmmu_free_nocache(unsigned long vaddr, int size)
 	bit_map_clear(&srmmu_nocache_map, offset, size);
 }
 
-void srmmu_early_allocate_ptable_skeleton(unsigned long start, unsigned long end);
+static void srmmu_early_allocate_ptable_skeleton(unsigned long start,
+						 unsigned long end);
 
 extern unsigned long probe_memory(void);	/* in fault.c */
 
@@ -377,7 +378,7 @@ extern unsigned long probe_memory(void);	/* in fault.c */
  * Reserve nocache dynamically proportionally to the amount of
  * system RAM. -- Tomas Szepe <szepe@pinerecords.com>, June 2002
  */
-void srmmu_nocache_calcsize(void)
+static void srmmu_nocache_calcsize(void)
 {
 	unsigned long sysmemavail = probe_memory() / 1024;
 	int srmmu_nocache_npages;
@@ -398,7 +399,7 @@ void srmmu_nocache_calcsize(void)
 	srmmu_nocache_end = SRMMU_NOCACHE_VADDR + srmmu_nocache_size;
 }
 
-void __init srmmu_nocache_init(void)
+static void __init srmmu_nocache_init(void)
 {
 	unsigned int bitmap_bits;
 	pgd_t *pgd;
@@ -645,7 +646,7 @@ static void srmmu_unmapiorange(unsigned long virt_addr, unsigned int len)
  * mappings on the kernel stack without any special code as we did
  * need on the sun4c.
  */
-struct thread_info *srmmu_alloc_thread_info(void)
+static struct thread_info *srmmu_alloc_thread_info(void)
 {
 	struct thread_info *ret;
 
@@ -1045,13 +1046,14 @@ extern void hypersparc_setup_blockops(void);
  *       around 8mb mapped for us.
  */
 
-void __init early_pgtable_allocfail(char *type)
+static void __init early_pgtable_allocfail(char *type)
 {
 	prom_printf("inherit_prom_mappings: Cannot alloc kernel %s.\n", type);
 	prom_halt();
 }
 
-void __init srmmu_early_allocate_ptable_skeleton(unsigned long start, unsigned long end)
+static void __init srmmu_early_allocate_ptable_skeleton(unsigned long start,
+							unsigned long end)
 {
 	pgd_t *pgdp;
 	pmd_t *pmdp;
@@ -1081,7 +1083,8 @@ void __init srmmu_early_allocate_ptable_skeleton(unsigned long start, unsigned l
 	}
 }
 
-void __init srmmu_allocate_ptable_skeleton(unsigned long start, unsigned long end)
+static void __init srmmu_allocate_ptable_skeleton(unsigned long start,
+						  unsigned long end)
 {
 	pgd_t *pgdp;
 	pmd_t *pmdp;
@@ -1116,7 +1119,8 @@ void __init srmmu_allocate_ptable_skeleton(unsigned long start, unsigned long en
  * looking at the prom's page table directly which is what most
  * other OS's do.  Yuck... this is much better.
  */
-void __init srmmu_inherit_prom_mappings(unsigned long start,unsigned long end)
+static void __init srmmu_inherit_prom_mappings(unsigned long start,
+					       unsigned long end)
 {
 	pgd_t *pgdp;
 	pmd_t *pmdp;
diff --git a/arch/sparc/mm/tsunami.S b/arch/sparc/mm/tsunami.S
index db0d6de33a8..4e55e8f7664 100644
--- a/arch/sparc/mm/tsunami.S
+++ b/arch/sparc/mm/tsunami.S
@@ -93,7 +93,6 @@ tsunami_flush_tlb_page_out:
 	ldd	[src + offset + 0x00], t2; \
 	std	t2, [dst + offset + 0x00];
 
-	.globl	tsunami_copy_1page
 tsunami_copy_1page:
 /* NOTE: This routine has to be shorter than 70insns --jj */
 	or	%g0, (PAGE_SIZE >> 8), %g1
diff --git a/arch/sparc64/Kconfig b/arch/sparc64/Kconfig
index fca9246470b..7c88263256a 100644
--- a/arch/sparc64/Kconfig
+++ b/arch/sparc64/Kconfig
@@ -16,6 +16,7 @@ config SPARC64
 	select HAVE_IDE
 	select HAVE_LMB
 	select HAVE_ARCH_KGDB
+	select USE_GENERIC_SMP_HELPERS if SMP
 
 config GENERIC_TIME
 	bool
@@ -81,6 +82,10 @@ config GENERIC_HARDIRQS_NO__DO_IRQ
 	bool
 	def_bool y
 
+source "init/Kconfig"
+
+menu "Processor type and features"
+
 choice
 	prompt "Kernel page size"
 	default SPARC64_PAGE_SIZE_8KB
@@ -93,19 +98,11 @@ config SPARC64_PAGE_SIZE_8KB
 	  8KB and 64KB work quite well, since SPARC ELF sections
 	  provide for up to 64KB alignment.
 
-	  Therefore, 512KB and 4MB are for expert hackers only.
-
 	  If you don't know what to do, choose 8KB.
 
 config SPARC64_PAGE_SIZE_64KB
 	bool "64KB"
 
-config SPARC64_PAGE_SIZE_512KB
-	bool "512KB"
-
-config SPARC64_PAGE_SIZE_4MB
-	bool "4MB"
-
 endchoice
 
 config SECCOMP
@@ -136,14 +133,10 @@ config HOTPLUG_CPU
 	  can be controlled through /sys/devices/system/cpu/cpu#.
 	  Say N if you want to disable CPU hotplug.
 
-source "init/Kconfig"
-
 config GENERIC_HARDIRQS
 	bool
 	default y
 
-menu "General machine setup"
-
 source "kernel/time/Kconfig"
 
 config SMP
@@ -225,11 +218,10 @@ config HUGETLB_PAGE_SIZE_4MB
 	bool "4MB"
 
 config HUGETLB_PAGE_SIZE_512K
-	depends on !SPARC64_PAGE_SIZE_4MB && !SPARC64_PAGE_SIZE_512KB
 	bool "512K"
 
 config HUGETLB_PAGE_SIZE_64K
-	depends on !SPARC64_PAGE_SIZE_4MB && !SPARC64_PAGE_SIZE_512KB && !SPARC64_PAGE_SIZE_64KB
+	depends on !SPARC64_PAGE_SIZE_64KB
 	bool "64K"
 
 endchoice
diff --git a/arch/sparc64/Makefile b/arch/sparc64/Makefile
index 4b8f2b084c2..b785a395b12 100644
--- a/arch/sparc64/Makefile
+++ b/arch/sparc64/Makefile
@@ -9,7 +9,9 @@
 
 CHECKFLAGS	+= -D__sparc__ -D__sparc_v9__ -m64
 
-CPPFLAGS_vmlinux.lds += -Usparc
+# Undefine sparc when processing vmlinux.lds - it is used
+# And teach CPP we are doing 64 bit builds (for this case)
+CPPFLAGS_vmlinux.lds += -m64 -Usparc
 
 LDFLAGS		:= -m elf64_sparc
 
diff --git a/arch/sparc64/defconfig b/arch/sparc64/defconfig
index 76eb832527f..82cab5cc807 100644
--- a/arch/sparc64/defconfig
+++ b/arch/sparc64/defconfig
@@ -1,7 +1,7 @@
 #
 # Automatically generated make config: don't edit
-# Linux kernel version: 2.6.26-rc2
-# Fri May 16 13:36:07 2008
+# Linux kernel version: 2.6.26
+# Fri Jul 18 00:47:07 2008
 #
 CONFIG_SPARC=y
 CONFIG_SPARC64=y
@@ -22,18 +22,6 @@ CONFIG_HAVE_SETUP_PER_CPU_AREA=y
 CONFIG_ARCH_NO_VIRT_TO_BUS=y
 CONFIG_OF=y
 CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ=y
-CONFIG_SPARC64_PAGE_SIZE_8KB=y
-# CONFIG_SPARC64_PAGE_SIZE_64KB is not set
-# CONFIG_SPARC64_PAGE_SIZE_512KB is not set
-# CONFIG_SPARC64_PAGE_SIZE_4MB is not set
-CONFIG_SECCOMP=y
-CONFIG_HZ_100=y
-# CONFIG_HZ_250 is not set
-# CONFIG_HZ_300 is not set
-# CONFIG_HZ_1000 is not set
-CONFIG_HZ=100
-# CONFIG_SCHED_HRTICK is not set
-CONFIG_HOTPLUG_CPU=y
 CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
 
 #
@@ -105,6 +93,7 @@ CONFIG_KRETPROBES=y
 CONFIG_HAVE_KPROBES=y
 CONFIG_HAVE_KRETPROBES=y
 # CONFIG_HAVE_DMA_ATTRS is not set
+CONFIG_USE_GENERIC_SMP_HELPERS=y
 CONFIG_PROC_PAGE_MONITOR=y
 CONFIG_SLABINFO=y
 CONFIG_RT_MUTEXES=y
@@ -121,6 +110,7 @@ CONFIG_STOP_MACHINE=y
 CONFIG_BLOCK=y
 CONFIG_BLK_DEV_IO_TRACE=y
 CONFIG_BLK_DEV_BSG=y
+# CONFIG_BLK_DEV_INTEGRITY is not set
 CONFIG_BLOCK_COMPAT=y
 
 #
@@ -136,11 +126,21 @@ CONFIG_DEFAULT_AS=y
 # CONFIG_DEFAULT_NOOP is not set
 CONFIG_DEFAULT_IOSCHED="anticipatory"
 CONFIG_CLASSIC_RCU=y
-CONFIG_GENERIC_HARDIRQS=y
 
 #
-# General machine setup
+# Processor type and features
 #
+CONFIG_SPARC64_PAGE_SIZE_8KB=y
+# CONFIG_SPARC64_PAGE_SIZE_64KB is not set
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_300 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+# CONFIG_SCHED_HRTICK is not set
+CONFIG_HOTPLUG_CPU=y
+CONFIG_GENERIC_HARDIRQS=y
 CONFIG_TICK_ONESHOT=y
 CONFIG_NO_HZ=y
 CONFIG_HIGH_RES_TIMERS=y
@@ -342,6 +342,8 @@ CONFIG_UEVENT_HELPER_PATH="/sbin/hotplug"
 CONFIG_STANDALONE=y
 # CONFIG_PREVENT_FIRMWARE_BUILD is not set
 CONFIG_FW_LOADER=y
+CONFIG_FIRMWARE_IN_KERNEL=y
+CONFIG_EXTRA_FIRMWARE=""
 # CONFIG_DEBUG_DRIVER is not set
 # CONFIG_DEBUG_DEVRES is not set
 # CONFIG_SYS_HYPERVISOR is not set
@@ -366,6 +368,7 @@ CONFIG_CDROM_PKTCDVD_BUFFERS=8
 CONFIG_CDROM_PKTCDVD_WCACHE=y
 CONFIG_ATA_OVER_ETH=m
 CONFIG_SUNVDC=m
+# CONFIG_BLK_DEV_HD is not set
 CONFIG_MISC_DEVICES=y
 # CONFIG_PHANTOM is not set
 # CONFIG_EEPROM_93CX6 is not set
@@ -379,6 +382,7 @@ CONFIG_BLK_DEV_IDE=y
 #
 # Please see Documentation/ide/ide.txt for help/info on IDE drives
 #
+CONFIG_IDE_TIMINGS=y
 # CONFIG_BLK_DEV_IDE_SATA is not set
 CONFIG_BLK_DEV_IDEDISK=y
 # CONFIG_IDEDISK_MULTI_MODE is not set
@@ -429,8 +433,6 @@ CONFIG_BLK_DEV_ALI15X3=y
 # CONFIG_BLK_DEV_VIA82CXXX is not set
 # CONFIG_BLK_DEV_TC86C001 is not set
 CONFIG_BLK_DEV_IDEDMA=y
-# CONFIG_BLK_DEV_HD_ONLY is not set
-# CONFIG_BLK_DEV_HD is not set
 
 #
 # SCSI device support
@@ -504,6 +506,7 @@ CONFIG_SCSI_LOWLEVEL=y
 # CONFIG_SCSI_DEBUG is not set
 # CONFIG_SCSI_SUNESP is not set
 # CONFIG_SCSI_SRP is not set
+# CONFIG_SCSI_DH is not set
 # CONFIG_ATA is not set
 CONFIG_MD=y
 CONFIG_BLK_DEV_MD=m
@@ -529,6 +532,10 @@ CONFIG_DM_ZERO=m
 #
 # IEEE 1394 (FireWire) support
 #
+
+#
+# Enable only one of the two stacks, unless you know what you are doing
+#
 # CONFIG_FIREWIRE is not set
 # CONFIG_IEEE1394 is not set
 # CONFIG_I2O is not set
@@ -745,7 +752,8 @@ CONFIG_SERIAL_CORE_CONSOLE=y
 CONFIG_UNIX98_PTYS=y
 # CONFIG_LEGACY_PTYS is not set
 # CONFIG_IPMI_HANDLER is not set
-# CONFIG_HW_RANDOM is not set
+CONFIG_HW_RANDOM=m
+CONFIG_HW_RANDOM_N2RNG=m
 # CONFIG_R3964 is not set
 # CONFIG_APPLICOM is not set
 # CONFIG_RAW_DRIVER is not set
@@ -759,38 +767,58 @@ CONFIG_I2C_ALGOBIT=y
 #
 # I2C Hardware Bus support
 #
+
+#
+# PC SMBus host controller drivers
+#
 # CONFIG_I2C_ALI1535 is not set
 # CONFIG_I2C_ALI1563 is not set
 # CONFIG_I2C_ALI15X3 is not set
 # CONFIG_I2C_AMD756 is not set
 # CONFIG_I2C_AMD8111 is not set
 # CONFIG_I2C_I801 is not set
-# CONFIG_I2C_I810 is not set
+# CONFIG_I2C_ISCH is not set
 # CONFIG_I2C_PIIX4 is not set
 # CONFIG_I2C_NFORCE2 is not set
-# CONFIG_I2C_OCORES is not set
-# CONFIG_I2C_PARPORT_LIGHT is not set
-# CONFIG_I2C_PROSAVAGE is not set
-# CONFIG_I2C_SAVAGE4 is not set
-# CONFIG_I2C_SIMTEC is not set
 # CONFIG_I2C_SIS5595 is not set
 # CONFIG_I2C_SIS630 is not set
 # CONFIG_I2C_SIS96X is not set
-# CONFIG_I2C_TAOS_EVM is not set
-# CONFIG_I2C_STUB is not set
-# CONFIG_I2C_TINY_USB is not set
 # CONFIG_I2C_VIA is not set
 # CONFIG_I2C_VIAPRO is not set
+
+#
+# I2C system bus drivers (mostly embedded / system-on-chip)
+#
+# CONFIG_I2C_OCORES is not set
+# CONFIG_I2C_SIMTEC is not set
+
+#
+# External I2C/SMBus adapter drivers
+#
+# CONFIG_I2C_PARPORT_LIGHT is not set
+# CONFIG_I2C_TAOS_EVM is not set
+# CONFIG_I2C_TINY_USB is not set
+
+#
+# Graphics adapter I2C/DDC channel drivers
+#
 # CONFIG_I2C_VOODOO3 is not set
+
+#
+# Other I2C/SMBus bus drivers
+#
 # CONFIG_I2C_PCA_PLATFORM is not set
+# CONFIG_I2C_STUB is not set
 
 #
 # Miscellaneous I2C Chip support
 #
 # CONFIG_DS1682 is not set
+# CONFIG_AT24 is not set
 # CONFIG_SENSORS_EEPROM is not set
 # CONFIG_SENSORS_PCF8574 is not set
 # CONFIG_PCF8575 is not set
+# CONFIG_SENSORS_PCA9539 is not set
 # CONFIG_SENSORS_PCF8591 is not set
 # CONFIG_SENSORS_MAX6875 is not set
 # CONFIG_SENSORS_TSL2550 is not set
@@ -856,6 +884,7 @@ CONFIG_HWMON=y
 # CONFIG_SENSORS_W83627EHF is not set
 # CONFIG_HWMON_DEBUG_CHIP is not set
 # CONFIG_THERMAL is not set
+# CONFIG_THERMAL_HWMON is not set
 # CONFIG_WATCHDOG is not set
 
 #
@@ -985,15 +1014,7 @@ CONFIG_LOGO=y
 # CONFIG_LOGO_LINUX_VGA16 is not set
 # CONFIG_LOGO_LINUX_CLUT224 is not set
 CONFIG_LOGO_SUN_CLUT224=y
-
-#
-# Sound
-#
 CONFIG_SOUND=m
-
-#
-# Advanced Linux Sound Architecture
-#
 CONFIG_SND=m
 CONFIG_SND_TIMER=m
 CONFIG_SND_PCM=m
@@ -1010,21 +1031,17 @@ CONFIG_SND_SUPPORT_OLD_API=y
 CONFIG_SND_VERBOSE_PROCFS=y
 # CONFIG_SND_VERBOSE_PRINTK is not set
 # CONFIG_SND_DEBUG is not set
-
-#
-# Generic devices
-#
+CONFIG_SND_VMASTER=y
 CONFIG_SND_MPU401_UART=m
 CONFIG_SND_AC97_CODEC=m
+CONFIG_SND_DRIVERS=y
 CONFIG_SND_DUMMY=m
 CONFIG_SND_VIRMIDI=m
 CONFIG_SND_MTPAV=m
 # CONFIG_SND_SERIAL_U16550 is not set
 # CONFIG_SND_MPU401 is not set
-
-#
-# PCI devices
-#
+# CONFIG_SND_AC97_POWER_SAVE is not set
+CONFIG_SND_PCI=y
 # CONFIG_SND_AD1889 is not set
 # CONFIG_SND_ALS300 is not set
 CONFIG_SND_ALI5451=m
@@ -1084,37 +1101,14 @@ CONFIG_SND_ALI5451=m
 # CONFIG_SND_VIRTUOSO is not set
 # CONFIG_SND_VX222 is not set
 # CONFIG_SND_YMFPCI is not set
-# CONFIG_SND_AC97_POWER_SAVE is not set
-
-#
-# USB devices
-#
+CONFIG_SND_USB=y
 # CONFIG_SND_USB_AUDIO is not set
 # CONFIG_SND_USB_CAIAQ is not set
-
-#
-# ALSA Sparc devices
-#
+CONFIG_SND_SPARC=y
 # CONFIG_SND_SUN_AMD7930 is not set
 CONFIG_SND_SUN_CS4231=m
 # CONFIG_SND_SUN_DBRI is not set
-
-#
-# System on Chip audio support
-#
 # CONFIG_SND_SOC is not set
-
-#
-# ALSA SoC audio for Freescale SOCs
-#
-
-#
-# SoC Audio for the Texas Instruments OMAP
-#
-
-#
-# Open Sound System
-#
 # CONFIG_SOUND_PRIME is not set
 CONFIG_AC97_BUS=m
 CONFIG_HID_SUPPORT=y
@@ -1167,6 +1161,7 @@ CONFIG_USB_UHCI_HCD=m
 #
 # CONFIG_USB_ACM is not set
 # CONFIG_USB_PRINTER is not set
+# CONFIG_USB_WDM is not set
 
 #
 # NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
@@ -1226,6 +1221,7 @@ CONFIG_USB_STORAGE=m
 # CONFIG_USB_TRANCEVIBRATOR is not set
 # CONFIG_USB_IOWARRIOR is not set
 # CONFIG_USB_TEST is not set
+# CONFIG_USB_ISIGHTFW is not set
 # CONFIG_USB_GADGET is not set
 # CONFIG_MMC is not set
 # CONFIG_MEMSTICK is not set
@@ -1420,6 +1416,12 @@ CONFIG_DEBUG_BUGVERBOSE=y
 # CONFIG_BACKTRACE_SELF_TEST is not set
 # CONFIG_LKDTM is not set
 # CONFIG_FAULT_INJECTION is not set
+CONFIG_HAVE_FTRACE=y
+CONFIG_HAVE_DYNAMIC_FTRACE=y
+# CONFIG_FTRACE is not set
+# CONFIG_IRQSOFF_TRACER is not set
+# CONFIG_SCHED_TRACER is not set
+# CONFIG_CONTEXT_SWITCH_TRACER is not set
 # CONFIG_SAMPLES is not set
 CONFIG_HAVE_ARCH_KGDB=y
 # CONFIG_KGDB is not set
@@ -1486,6 +1488,10 @@ CONFIG_CRYPTO_CRC32C=m
 CONFIG_CRYPTO_MD4=y
 CONFIG_CRYPTO_MD5=y
 CONFIG_CRYPTO_MICHAEL_MIC=m
+# CONFIG_CRYPTO_RMD128 is not set
+# CONFIG_CRYPTO_RMD160 is not set
+# CONFIG_CRYPTO_RMD256 is not set
+# CONFIG_CRYPTO_RMD320 is not set
 CONFIG_CRYPTO_SHA1=y
 CONFIG_CRYPTO_SHA256=m
 CONFIG_CRYPTO_SHA512=m
@@ -1527,6 +1533,7 @@ CONFIG_BITREVERSE=y
 # CONFIG_GENERIC_FIND_FIRST_BIT is not set
 CONFIG_CRC_CCITT=m
 CONFIG_CRC16=m
+# CONFIG_CRC_T10DIF is not set
 # CONFIG_CRC_ITU_T is not set
 CONFIG_CRC32=y
 # CONFIG_CRC7 is not set
diff --git a/arch/sparc64/kernel/central.c b/arch/sparc64/kernel/central.c
index b61b8dfb09c..f2e87d0d7e1 100644
--- a/arch/sparc64/kernel/central.c
+++ b/arch/sparc64/kernel/central.c
@@ -16,8 +16,8 @@
 #include <asm/fhc.h>
 #include <asm/starfire.h>
 
-struct linux_central *central_bus = NULL;
-struct linux_fhc *fhc_list = NULL;
+static struct linux_central *central_bus = NULL;
+static struct linux_fhc *fhc_list = NULL;
 
 #define IS_CENTRAL_FHC(__fhc)	((__fhc) == central_bus->child)
 
@@ -79,9 +79,9 @@ static void adjust_regs(struct linux_prom_registers *regp, int nregs,
 }
 
 /* Apply probed fhc ranges to registers passed, if no ranges return. */
-void apply_fhc_ranges(struct linux_fhc *fhc,
-		      struct linux_prom_registers *regs,
-		      int nregs)
+static void apply_fhc_ranges(struct linux_fhc *fhc,
+			     struct linux_prom_registers *regs,
+			     int nregs)
 {
 	if (fhc->num_fhc_ranges)
 		adjust_regs(regs, nregs, fhc->fhc_ranges,
@@ -89,8 +89,8 @@ void apply_fhc_ranges(struct linux_fhc *fhc,
 }
 
 /* Apply probed central ranges to registers passed, if no ranges return. */
-void apply_central_ranges(struct linux_central *central,
-			  struct linux_prom_registers *regs, int nregs)
+static void apply_central_ranges(struct linux_central *central,
+				 struct linux_prom_registers *regs, int nregs)
 {
 	if (central->num_central_ranges)
 		adjust_regs(regs, nregs, central->central_ranges,
diff --git a/arch/sparc64/kernel/ds.c b/arch/sparc64/kernel/ds.c
index edb74f5a118..d0fa5aa3893 100644
--- a/arch/sparc64/kernel/ds.c
+++ b/arch/sparc64/kernel/ds.c
@@ -159,7 +159,7 @@ static void ds_var_data(struct ds_info *dp,
 			struct ds_cap_state *cp,
 			void *buf, int len);
 
-struct ds_cap_state ds_states_template[] = {
+static struct ds_cap_state ds_states_template[] = {
 	{
 		.service_id	= "md-update",
 		.data		= md_update_data,
diff --git a/arch/sparc64/kernel/hvapi.c b/arch/sparc64/kernel/hvapi.c
index f34f5d6181e..691760b5b01 100644
--- a/arch/sparc64/kernel/hvapi.c
+++ b/arch/sparc64/kernel/hvapi.c
@@ -34,8 +34,12 @@ static struct api_info api_table[] = {
 	{ .group = HV_GRP_LDOM,					},
 	{ .group = HV_GRP_SVC_CHAN,	.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_NCS,		.flags = FLAG_PRE_API	},
+	{ .group = HV_GRP_RNG,					},
 	{ .group = HV_GRP_NIAG_PERF,	.flags = FLAG_PRE_API	},
 	{ .group = HV_GRP_FIRE_PERF,				},
+	{ .group = HV_GRP_N2_CPU,				},
+	{ .group = HV_GRP_NIU,					},
+	{ .group = HV_GRP_VF_CPU,				},
 	{ .group = HV_GRP_DIAG,		.flags = FLAG_PRE_API	},
 };
 
diff --git a/arch/sparc64/kernel/pci_msi.c b/arch/sparc64/kernel/pci_msi.c
index db5e8fd8f67..60c71e35021 100644
--- a/arch/sparc64/kernel/pci_msi.c
+++ b/arch/sparc64/kernel/pci_msi.c
@@ -120,9 +120,9 @@ static struct irq_chip msi_irq = {
 	/* XXX affinity XXX */
 };
 
-int sparc64_setup_msi_irq(unsigned int *virt_irq_p,
-			  struct pci_dev *pdev,
-			  struct msi_desc *entry)
+static int sparc64_setup_msi_irq(unsigned int *virt_irq_p,
+				 struct pci_dev *pdev,
+				 struct msi_desc *entry)
 {
 	struct pci_pbm_info *pbm = pdev->dev.archdata.host_controller;
 	const struct sparc64_msiq_ops *ops = pbm->msi_ops;
@@ -179,8 +179,8 @@ out_err:
 	return err;
 }
 
-void sparc64_teardown_msi_irq(unsigned int virt_irq,
-			      struct pci_dev *pdev)
+static void sparc64_teardown_msi_irq(unsigned int virt_irq,
+				     struct pci_dev *pdev)
 {
 	struct pci_pbm_info *pbm = pdev->dev.archdata.host_controller;
 	const struct sparc64_msiq_ops *ops = pbm->msi_ops;
diff --git a/arch/sparc64/kernel/pci_sun4v.c b/arch/sparc64/kernel/pci_sun4v.c
index e2bb9790039..a104c80d319 100644
--- a/arch/sparc64/kernel/pci_sun4v.c
+++ b/arch/sparc64/kernel/pci_sun4v.c
@@ -531,7 +531,7 @@ static void dma_4v_sync_sg_for_cpu(struct device *dev,
 	/* Nothing to do... */
 }
 
-const struct dma_ops sun4v_dma_ops = {
+static const struct dma_ops sun4v_dma_ops = {
 	.alloc_coherent			= dma_4v_alloc_coherent,
 	.free_coherent			= dma_4v_free_coherent,
 	.map_single			= dma_4v_map_single,
diff --git a/arch/sparc64/kernel/process.c b/arch/sparc64/kernel/process.c
index 2084f81a76e..c1f1ff29e97 100644
--- a/arch/sparc64/kernel/process.c
+++ b/arch/sparc64/kernel/process.c
@@ -15,7 +15,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
-#include <linux/kallsyms.h>
 #include <linux/mm.h>
 #include <linux/fs.h>
 #include <linux/smp.h>
@@ -211,7 +210,7 @@ static void show_regwindow(struct pt_regs *regs)
 	printk("i4: %016lx i5: %016lx i6: %016lx i7: %016lx\n",
 	       rwk->ins[4], rwk->ins[5], rwk->ins[6], rwk->ins[7]);
 	if (regs->tstate & TSTATE_PRIV)
-		print_symbol("I7: <%s>\n", rwk->ins[7]);
+		printk("I7: <%pS>\n", (void *) rwk->ins[7]);
 }
 
 #ifdef CONFIG_SMP
@@ -232,7 +231,7 @@ void __show_regs(struct pt_regs * regs)
 #endif
 	printk("TSTATE: %016lx TPC: %016lx TNPC: %016lx Y: %08x    %s\n", regs->tstate,
 	       regs->tpc, regs->tnpc, regs->y, print_tainted());
-	print_symbol("TPC: <%s>\n", regs->tpc);
+	printk("TPC: <%pS>\n", (void *) regs->tpc);
 	printk("g0: %016lx g1: %016lx g2: %016lx g3: %016lx\n",
 	       regs->u_regs[0], regs->u_regs[1], regs->u_regs[2],
 	       regs->u_regs[3]);
@@ -245,7 +244,7 @@ void __show_regs(struct pt_regs * regs)
 	printk("o4: %016lx o5: %016lx sp: %016lx ret_pc: %016lx\n",
 	       regs->u_regs[12], regs->u_regs[13], regs->u_regs[14],
 	       regs->u_regs[15]);
-	print_symbol("RPC: <%s>\n", regs->u_regs[15]);
+	printk("RPC: <%pS>\n", (void *) regs->u_regs[15]);
 	show_regwindow(regs);
 #ifdef CONFIG_SMP
 	spin_unlock(&regdump_lock);
@@ -346,9 +345,6 @@ static void sysrq_handle_globreg(int key, struct tty_struct *tty)
 {
 	struct thread_info *tp = current_thread_info();
 	struct pt_regs *regs = get_irq_regs();
-#ifdef CONFIG_KALLSYMS
-	char buffer[KSYM_SYMBOL_LEN];
-#endif
 	unsigned long flags;
 	int this_cpu, cpu;
 
@@ -377,17 +373,13 @@ static void sysrq_handle_globreg(int key, struct tty_struct *tty)
 		       gp->tstate, gp->tpc, gp->tnpc,
 		       ((tp && tp->task) ? tp->task->comm : "NULL"),
 		       ((tp && tp->task) ? tp->task->pid : -1));
-#ifdef CONFIG_KALLSYMS
+
 		if (gp->tstate & TSTATE_PRIV) {
-			sprint_symbol(buffer, gp->tpc);
-			printk("             TPC[%s] ", buffer);
-			sprint_symbol(buffer, gp->o7);
-			printk("O7[%s] ", buffer);
-			sprint_symbol(buffer, gp->i7);
-			printk("I7[%s]\n", buffer);
-		} else
-#endif
-		{
+			printk("             TPC[%pS] O7[%pS] I7[%pS]\n",
+			       (void *) gp->tpc,
+			       (void *) gp->o7,
+			       (void *) gp->i7);
+		} else {
 			printk("             TPC[%lx] O7[%lx] I7[%lx]\n",
 			       gp->tpc, gp->o7, gp->i7);
 		}
diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c
index c099d96f123..7cf72b4bb10 100644
--- a/arch/sparc64/kernel/smp.c
+++ b/arch/sparc64/kernel/smp.c
@@ -788,89 +788,36 @@ static void smp_start_sync_tick_client(int cpu)
 			      0, 0, 0, mask);
 }
 
-/* Send cross call to all processors except self. */
-#define smp_cross_call(func, ctx, data1, data2) \
-	smp_cross_call_masked(func, ctx, data1, data2, cpu_online_map)
-
-struct call_data_struct {
-	void (*func) (void *info);
-	void *info;
-	atomic_t finished;
-	int wait;
-};
-
-static struct call_data_struct *call_data;
-
 extern unsigned long xcall_call_function;
 
-/**
- * smp_call_function(): Run a function on all other CPUs.
- * @func: The function to run. This must be fast and non-blocking.
- * @info: An arbitrary pointer to pass to the function.
- * @wait: If true, wait (atomically) until function has completed on other CPUs.
- *
- * Returns 0 on success, else a negative status code. Does not return until
- * remote CPUs are nearly ready to execute <<func>> or are or have executed.
- *
- * You must not call this function with disabled interrupts or from a
- * hardware interrupt handler or from a bottom half handler.
- */
-static int sparc64_smp_call_function_mask(void (*func)(void *info), void *info,
-					  int wait, cpumask_t mask)
+void arch_send_call_function_ipi(cpumask_t mask)
 {
-	struct call_data_struct data;
-	int cpus;
-
-	/* Can deadlock when called with interrupts disabled */
-	WARN_ON(irqs_disabled());
-
-	data.func = func;
-	data.info = info;
-	atomic_set(&data.finished, 0);
-	data.wait = wait;
-
-	spin_lock(&call_lock);
-
-	cpu_clear(smp_processor_id(), mask);
-	cpus = cpus_weight(mask);
-	if (!cpus)
-		goto out_unlock;
-
-	call_data = &data;
-	mb();
-
 	smp_cross_call_masked(&xcall_call_function, 0, 0, 0, mask);
+}
 
-	/* Wait for response */
-	while (atomic_read(&data.finished) != cpus)
-		cpu_relax();
+extern unsigned long xcall_call_function_single;
 
-out_unlock:
-	spin_unlock(&call_lock);
+void arch_send_call_function_single_ipi(int cpu)
+{
+	cpumask_t mask = cpumask_of_cpu(cpu);
 
-	return 0;
+	smp_cross_call_masked(&xcall_call_function_single, 0, 0, 0, mask);
 }
 
-int smp_call_function(void (*func)(void *info), void *info, int wait)
-{
-	return sparc64_smp_call_function_mask(func, info, wait, cpu_online_map);
-}
+/* Send cross call to all processors except self. */
+#define smp_cross_call(func, ctx, data1, data2) \
+	smp_cross_call_masked(func, ctx, data1, data2, cpu_online_map)
 
 void smp_call_function_client(int irq, struct pt_regs *regs)
 {
-	void (*func) (void *info) = call_data->func;
-	void *info = call_data->info;
+	clear_softint(1 << irq);
+	generic_smp_call_function_interrupt();
+}
 
+void smp_call_function_single_client(int irq, struct pt_regs *regs)
+{
 	clear_softint(1 << irq);
-	if (call_data->wait) {
-		/* let initiator proceed only after completion */
-		func(info);
-		atomic_inc(&call_data->finished);
-	} else {
-		/* let initiator proceed after getting data */
-		atomic_inc(&call_data->finished);
-		func(info);
-	}
+	generic_smp_call_function_single_interrupt();
 }
 
 static void tsb_sync(void *info)
@@ -890,7 +837,7 @@ static void tsb_sync(void *info)
 
 void smp_tsb_sync(struct mm_struct *mm)
 {
-	sparc64_smp_call_function_mask(tsb_sync, mm, 1, mm->cpu_vm_mask);
+	smp_call_function_mask(mm->cpu_vm_mask, tsb_sync, mm, 1);
 }
 
 extern unsigned long xcall_flush_tlb_mm;
diff --git a/arch/sparc64/kernel/sparc64_ksyms.c b/arch/sparc64/kernel/sparc64_ksyms.c
index 49d3ea50c24..504e678ee12 100644
--- a/arch/sparc64/kernel/sparc64_ksyms.c
+++ b/arch/sparc64/kernel/sparc64_ksyms.c
@@ -108,8 +108,6 @@ EXPORT_SYMBOL(__read_unlock);
 EXPORT_SYMBOL(__write_lock);
 EXPORT_SYMBOL(__write_unlock);
 EXPORT_SYMBOL(__write_trylock);
-
-EXPORT_SYMBOL(smp_call_function);
 #endif /* CONFIG_SMP */
 
 #ifdef CONFIG_MCOUNT
diff --git a/arch/sparc64/kernel/sys_sparc.c b/arch/sparc64/kernel/sys_sparc.c
index ac1bff58c1a..e1f4eba2e57 100644
--- a/arch/sparc64/kernel/sys_sparc.c
+++ b/arch/sparc64/kernel/sys_sparc.c
@@ -542,7 +542,7 @@ asmlinkage long sparc64_personality(unsigned long personality)
 	return ret;
 }
 
-int sparc64_mmap_check(unsigned long addr, unsigned long len)
+int sparc_mmap_check(unsigned long addr, unsigned long len)
 {
 	if (test_thread_flag(TIF_32BIT)) {
 		if (len >= STACK_TOP32)
@@ -614,9 +614,9 @@ asmlinkage unsigned long sys64_mremap(unsigned long addr,
 		goto out;
 	if (unlikely(new_len >= VA_EXCLUDE_START))
 		goto out;
-	if (unlikely(sparc64_mmap_check(addr, old_len)))
+	if (unlikely(sparc_mmap_check(addr, old_len)))
 		goto out;
-	if (unlikely(sparc64_mmap_check(new_addr, new_len)))
+	if (unlikely(sparc_mmap_check(new_addr, new_len)))
 		goto out;
 
 	down_write(&current->mm->mmap_sem);
diff --git a/arch/sparc64/kernel/sys_sparc32.c b/arch/sparc64/kernel/sys_sparc32.c
index ba5bd626b39..97b77fb5c50 100644
--- a/arch/sparc64/kernel/sys_sparc32.c
+++ b/arch/sparc64/kernel/sys_sparc32.c
@@ -359,7 +359,8 @@ int cp_compat_stat(struct kstat *stat, struct compat_stat __user *statbuf)
 	return err;
 }
 
-int cp_compat_stat64(struct kstat *stat, struct compat_stat64 __user *statbuf)
+static int cp_compat_stat64(struct kstat *stat,
+			    struct compat_stat64 __user *statbuf)
 {
 	int err;
 
@@ -870,9 +871,9 @@ asmlinkage unsigned long sys32_mremap(unsigned long addr,
 	unsigned long ret = -EINVAL;
 	unsigned long new_addr = __new_addr;
 
-	if (unlikely(sparc64_mmap_check(addr, old_len)))
+	if (unlikely(sparc_mmap_check(addr, old_len)))
 		goto out;
-	if (unlikely(sparc64_mmap_check(new_addr, new_len)))
+	if (unlikely(sparc_mmap_check(new_addr, new_len)))
 		goto out;
 	down_write(&current->mm->mmap_sem);
 	ret = do_mremap(addr, old_len, new_len, flags, new_addr);
diff --git a/arch/sparc64/kernel/traps.c b/arch/sparc64/kernel/traps.c
index 36974926265..0aa819c29db 100644
--- a/arch/sparc64/kernel/traps.c
+++ b/arch/sparc64/kernel/traps.c
@@ -1,6 +1,6 @@
 /* arch/sparc64/kernel/traps.c
  *
- * Copyright (C) 1995,1997 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1995,1997,2008 David S. Miller (davem@davemloft.net)
  * Copyright (C) 1997,1999,2000 Jakub Jelinek (jakub@redhat.com)
  */
 
@@ -11,7 +11,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
-#include <linux/kallsyms.h>
 #include <linux/signal.h>
 #include <linux/smp.h>
 #include <linux/mm.h>
@@ -74,7 +73,7 @@ static void dump_tl1_traplog(struct tl1_traplog *p)
 		       i + 1,
 		       p->trapstack[i].tstate, p->trapstack[i].tpc,
 		       p->trapstack[i].tnpc, p->trapstack[i].tt);
-		print_symbol("TRAPLOG: TPC<%s>\n", p->trapstack[i].tpc);
+		printk("TRAPLOG: TPC<%pS>\n", (void *) p->trapstack[i].tpc);
 	}
 }
 
@@ -1081,7 +1080,7 @@ static void cheetah_log_errors(struct pt_regs *regs, struct cheetah_err_info *in
 	       regs->tpc, regs->tnpc, regs->u_regs[UREG_I7], regs->tstate);
 	printk("%s" "ERROR(%d): ",
 	       (recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id());
-	print_symbol("TPC<%s>\n", regs->tpc);
+	printk("TPC<%pS>\n", (void *) regs->tpc);
 	printk("%s" "ERROR(%d): M_SYND(%lx),  E_SYND(%lx)%s%s\n",
 	       (recoverable ? KERN_WARNING : KERN_CRIT), smp_processor_id(),
 	       (afsr & CHAFSR_M_SYNDROME) >> CHAFSR_M_SYNDROME_SHIFT,
@@ -1689,7 +1688,7 @@ void cheetah_plus_parity_error(int type, struct pt_regs *regs)
 		       smp_processor_id(),
 		       (type & 0x1) ? 'I' : 'D',
 		       regs->tpc);
-		print_symbol(KERN_EMERG "TPC<%s>\n", regs->tpc);
+		printk(KERN_EMERG "TPC<%pS>\n", (void *) regs->tpc);
 		panic("Irrecoverable Cheetah+ parity error.");
 	}
 
@@ -1697,7 +1696,7 @@ void cheetah_plus_parity_error(int type, struct pt_regs *regs)
 	       smp_processor_id(),
 	       (type & 0x1) ? 'I' : 'D',
 	       regs->tpc);
-	print_symbol(KERN_WARNING "TPC<%s>\n", regs->tpc);
+	printk(KERN_WARNING "TPC<%pS>\n", (void *) regs->tpc);
 }
 
 struct sun4v_error_entry {
@@ -1904,9 +1903,10 @@ void sun4v_itlb_error_report(struct pt_regs *regs, int tl)
 
 	printk(KERN_EMERG "SUN4V-ITLB: Error at TPC[%lx], tl %d\n",
 	       regs->tpc, tl);
-	print_symbol(KERN_EMERG "SUN4V-ITLB: TPC<%s>\n", regs->tpc);
+	printk(KERN_EMERG "SUN4V-ITLB: TPC<%pS>\n", (void *) regs->tpc);
 	printk(KERN_EMERG "SUN4V-ITLB: O7[%lx]\n", regs->u_regs[UREG_I7]);
-	print_symbol(KERN_EMERG "SUN4V-ITLB: O7<%s>\n", regs->u_regs[UREG_I7]);
+	printk(KERN_EMERG "SUN4V-ITLB: O7<%pS>\n",
+	       (void *) regs->u_regs[UREG_I7]);
 	printk(KERN_EMERG "SUN4V-ITLB: vaddr[%lx] ctx[%lx] "
 	       "pte[%lx] error[%lx]\n",
 	       sun4v_err_itlb_vaddr, sun4v_err_itlb_ctx,
@@ -1927,9 +1927,10 @@ void sun4v_dtlb_error_report(struct pt_regs *regs, int tl)
 
 	printk(KERN_EMERG "SUN4V-DTLB: Error at TPC[%lx], tl %d\n",
 	       regs->tpc, tl);
-	print_symbol(KERN_EMERG "SUN4V-DTLB: TPC<%s>\n", regs->tpc);
+	printk(KERN_EMERG "SUN4V-DTLB: TPC<%pS>\n", (void *) regs->tpc);
 	printk(KERN_EMERG "SUN4V-DTLB: O7[%lx]\n", regs->u_regs[UREG_I7]);
-	print_symbol(KERN_EMERG "SUN4V-DTLB: O7<%s>\n", regs->u_regs[UREG_I7]);
+	printk(KERN_EMERG "SUN4V-DTLB: O7<%pS>\n",
+	       (void *) regs->u_regs[UREG_I7]);
 	printk(KERN_EMERG "SUN4V-DTLB: vaddr[%lx] ctx[%lx] "
 	       "pte[%lx] error[%lx]\n",
 	       sun4v_err_dtlb_vaddr, sun4v_err_dtlb_ctx,
@@ -2111,10 +2112,7 @@ void show_stack(struct task_struct *tsk, unsigned long *_ksp)
 	fp = ksp + STACK_BIAS;
 	thread_base = (unsigned long) tp;
 
-	printk("Call Trace:");
-#ifdef CONFIG_KALLSYMS
-	printk("\n");
-#endif
+	printk("Call Trace:\n");
 	do {
 		struct sparc_stackf *sf;
 		struct pt_regs *regs;
@@ -2137,12 +2135,8 @@ void show_stack(struct task_struct *tsk, unsigned long *_ksp)
 			fp = (unsigned long)sf->fp + STACK_BIAS;
 		}
 
-		printk(" [%016lx] ", pc);
-		print_symbol("%s\n", pc);
+		printk(" [%016lx] %pS\n", pc, (void *) pc);
 	} while (++count < 16);
-#ifndef CONFIG_KALLSYMS
-	printk("\n");
-#endif
 }
 
 void dump_stack(void)
@@ -2211,9 +2205,8 @@ void die_if_kernel(char *str, struct pt_regs *regs)
 		while (rw &&
 		       count++ < 30&&
 		       is_kernel_stack(current, rw)) {
-			printk("Caller[%016lx]", rw->ins[7]);
-			print_symbol(": %s", rw->ins[7]);
-			printk("\n");
+			printk("Caller[%016lx]: %pS\n", rw->ins[7],
+			       (void *) rw->ins[7]);
 
 			rw = kernel_stack_up(rw);
 		}
diff --git a/arch/sparc64/kernel/ttable.S b/arch/sparc64/kernel/ttable.S
index 450053af039..1ade3d6fb7f 100644
--- a/arch/sparc64/kernel/ttable.S
+++ b/arch/sparc64/kernel/ttable.S
@@ -58,7 +58,12 @@ tl0_irq3:	BTRAP(0x43)
 tl0_irq4:	BTRAP(0x44)
 #endif
 tl0_irq5:	TRAP_IRQ(handler_irq, 5)
-tl0_irq6:	BTRAP(0x46) BTRAP(0x47) BTRAP(0x48) BTRAP(0x49)
+#ifdef CONFIG_SMP
+tl0_irq6:	TRAP_IRQ(smp_call_function_single_client, 6)
+#else
+tl0_irq6:	BTRAP(0x46)
+#endif
+tl0_irq7:	BTRAP(0x47) BTRAP(0x48) BTRAP(0x49)
 tl0_irq10:	BTRAP(0x4a) BTRAP(0x4b) BTRAP(0x4c) BTRAP(0x4d)
 tl0_irq14:	TRAP_IRQ(timer_interrupt, 14)
 tl0_irq15:	TRAP_IRQ(handler_irq, 15)
diff --git a/arch/sparc64/kernel/unaligned.c b/arch/sparc64/kernel/unaligned.c
index afa7fc4f519..203ddfad9f2 100644
--- a/arch/sparc64/kernel/unaligned.c
+++ b/arch/sparc64/kernel/unaligned.c
@@ -2,7 +2,7 @@
  * unaligned.c: Unaligned load/store trap handling with special
  *              cases for the kernel to do them more quickly.
  *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996,2008 David S. Miller (davem@davemloft.net)
  * Copyright (C) 1996,1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
  */
 
@@ -20,7 +20,6 @@
 #include <asm/uaccess.h>
 #include <linux/smp.h>
 #include <linux/bitops.h>
-#include <linux/kallsyms.h>
 #include <asm/fpumacro.h>
 
 /* #define DEBUG_MNA */
@@ -289,8 +288,8 @@ static void log_unaligned(struct pt_regs *regs)
 	if (count < 5) {
 		last_time = jiffies;
 		count++;
-		printk("Kernel unaligned access at TPC[%lx] ", regs->tpc);
-		print_symbol("%s\n", regs->tpc);
+		printk("Kernel unaligned access at TPC[%lx] %pS\n",
+		       regs->tpc, (void *) regs->tpc);
 	}
 }
 
diff --git a/arch/sparc64/lib/copy_page.S b/arch/sparc64/lib/copy_page.S
index 37460666a5c..b243d3b606b 100644
--- a/arch/sparc64/lib/copy_page.S
+++ b/arch/sparc64/lib/copy_page.S
@@ -25,9 +25,9 @@
 
 #define	DCACHE_SIZE	(PAGE_SIZE * 2)
 
-#if (PAGE_SHIFT == 13) || (PAGE_SHIFT == 19)
+#if (PAGE_SHIFT == 13)
 #define PAGE_SIZE_REM	0x80
-#elif (PAGE_SHIFT == 16) || (PAGE_SHIFT == 22)
+#elif (PAGE_SHIFT == 16)
 #define PAGE_SIZE_REM	0x100
 #else
 #error Wrong PAGE_SHIFT specified
@@ -198,7 +198,7 @@ cheetah_copy_page_insn:
 	cmp		%o2, PAGE_SIZE_REM
 	bne,pt		%xcc, 1b
 	 add		%o0, 0x40, %o0
-#if (PAGE_SHIFT == 16) || (PAGE_SHIFT == 22)
+#if (PAGE_SHIFT == 16)
 	TOUCH(f0, f2, f4, f6, f8, f10, f12, f14)
 	ldda		[%o1] ASI_BLK_P, %f32
 	stda		%f48, [%o0] %asi
diff --git a/arch/sparc64/mm/fault.c b/arch/sparc64/mm/fault.c
index 236f4d228d2..ea7d7ae76bc 100644
--- a/arch/sparc64/mm/fault.c
+++ b/arch/sparc64/mm/fault.c
@@ -1,7 +1,7 @@
 /*
  * arch/sparc64/mm/fault.c: Page fault handlers for the 64-bit Sparc.
  *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996, 2008 David S. Miller (davem@davemloft.net)
  * Copyright (C) 1997, 1999 Jakub Jelinek (jj@ultra.linux.cz)
  */
 
@@ -18,7 +18,6 @@
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/kprobes.h>
-#include <linux/kallsyms.h>
 #include <linux/kdebug.h>
 
 #include <asm/page.h>
@@ -115,7 +114,7 @@ static void bad_kernel_pc(struct pt_regs *regs, unsigned long vaddr)
 	printk(KERN_CRIT "OOPS: Bogus kernel PC [%016lx] in fault handler\n",
 	       regs->tpc);
 	printk(KERN_CRIT "OOPS: RPC [%016lx]\n", regs->u_regs[15]);
-	print_symbol("RPC: <%s>\n", regs->u_regs[15]);
+	printk("OOPS: RPC <%pS>\n", (void *) regs->u_regs[15]);
 	printk(KERN_CRIT "OOPS: Fault was to vaddr[%lx]\n", vaddr);
 	dump_stack();
 	unhandled_fault(regs->tpc, current, regs);
diff --git a/arch/sparc64/mm/tsb.c b/arch/sparc64/mm/tsb.c
index fe70c8a557b..3547937b17a 100644
--- a/arch/sparc64/mm/tsb.c
+++ b/arch/sparc64/mm/tsb.c
@@ -96,12 +96,6 @@ void flush_tsb_user(struct mmu_gather *mp)
 #elif defined(CONFIG_SPARC64_PAGE_SIZE_64KB)
 #define HV_PGSZ_IDX_BASE	HV_PGSZ_IDX_64K
 #define HV_PGSZ_MASK_BASE	HV_PGSZ_MASK_64K
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_512KB)
-#define HV_PGSZ_IDX_BASE	HV_PGSZ_IDX_512K
-#define HV_PGSZ_MASK_BASE	HV_PGSZ_MASK_512K
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_4MB)
-#define HV_PGSZ_IDX_BASE	HV_PGSZ_IDX_4MB
-#define HV_PGSZ_MASK_BASE	HV_PGSZ_MASK_4MB
 #else
 #error Broken base page size setting...
 #endif
diff --git a/arch/sparc64/mm/ultra.S b/arch/sparc64/mm/ultra.S
index 9bb2d90a9df..4c8ca131ffa 100644
--- a/arch/sparc64/mm/ultra.S
+++ b/arch/sparc64/mm/ultra.S
@@ -688,6 +688,11 @@ xcall_call_function:
 	wr		%g0, (1 << PIL_SMP_CALL_FUNC), %set_softint
 	retry
 
+	.globl		xcall_call_function_single
+xcall_call_function_single:
+	wr		%g0, (1 << PIL_SMP_CALL_FUNC_SNGL), %set_softint
+	retry
+
 	.globl		xcall_receive_signal
 xcall_receive_signal:
 	wr		%g0, (1 << PIL_SMP_RECEIVE_SIGNAL), %set_softint
diff --git a/drivers/char/hw_random/Kconfig b/drivers/char/hw_random/Kconfig
index efd0b4db7c8..8822eca58ff 100644
--- a/drivers/char/hw_random/Kconfig
+++ b/drivers/char/hw_random/Kconfig
@@ -59,6 +59,19 @@ config HW_RANDOM_GEODE
 
 	  If unsure, say Y.
 
+config HW_RANDOM_N2RNG
+	tristate "Niagara2 Random Number Generator support"
+	depends on HW_RANDOM && SPARC64
+	default HW_RANDOM
+	---help---
+	  This driver provides kernel-side support for the Random Number
+	  Generator hardware found on Niagara2 cpus.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called n2-rng.
+
+	  If unsure, say Y.
+
 config HW_RANDOM_VIA
 	tristate "VIA HW Random Number Generator support"
 	depends on HW_RANDOM && X86_32
diff --git a/drivers/char/hw_random/Makefile b/drivers/char/hw_random/Makefile
index b4940ddbb35..b6effb7522c 100644
--- a/drivers/char/hw_random/Makefile
+++ b/drivers/char/hw_random/Makefile
@@ -7,6 +7,8 @@ rng-core-y := core.o
 obj-$(CONFIG_HW_RANDOM_INTEL) += intel-rng.o
 obj-$(CONFIG_HW_RANDOM_AMD) += amd-rng.o
 obj-$(CONFIG_HW_RANDOM_GEODE) += geode-rng.o
+obj-$(CONFIG_HW_RANDOM_N2RNG) += n2-rng.o
+n2-rng-y := n2-drv.o n2-asm.o
 obj-$(CONFIG_HW_RANDOM_VIA) += via-rng.o
 obj-$(CONFIG_HW_RANDOM_IXP4XX) += ixp4xx-rng.o
 obj-$(CONFIG_HW_RANDOM_OMAP) += omap-rng.o
diff --git a/drivers/char/hw_random/n2-asm.S b/drivers/char/hw_random/n2-asm.S
new file mode 100644
index 00000000000..9b6eb5cd59f
--- /dev/null
+++ b/drivers/char/hw_random/n2-asm.S
@@ -0,0 +1,79 @@
+/* n2-asm.S: Niagara2 RNG hypervisor call assembler.
+ *
+ * Copyright (C) 2008 David S. Miller <davem@davemloft.net>
+ */
+#include <linux/linkage.h>
+#include <asm/hypervisor.h>
+#include "n2rng.h"
+
+	.text
+
+ENTRY(sun4v_rng_get_diag_ctl)
+	mov	HV_FAST_RNG_GET_DIAG_CTL, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 nop
+ENDPROC(sun4v_rng_get_diag_ctl)
+
+ENTRY(sun4v_rng_ctl_read_v1)
+	mov	%o1, %o3
+	mov	%o2, %o4
+	mov	HV_FAST_RNG_CTL_READ, %o5
+	ta	HV_FAST_TRAP
+	stx	%o1, [%o3]
+	retl
+	 stx	%o2, [%o4]
+ENDPROC(sun4v_rng_ctl_read_v1)
+
+ENTRY(sun4v_rng_ctl_read_v2)
+	save	%sp, -192, %sp
+	mov	%i0, %o0
+	mov	%i1, %o1
+	mov	HV_FAST_RNG_CTL_READ, %o5
+	ta	HV_FAST_TRAP
+	stx	%o1, [%i2]
+	stx	%o2, [%i3]
+	stx	%o3, [%i4]
+	stx	%o4, [%i5]
+	ret
+	restore	%g0, %o0, %o0
+ENDPROC(sun4v_rng_ctl_read_v2)
+
+ENTRY(sun4v_rng_ctl_write_v1)
+	mov	%o3, %o4
+	mov	HV_FAST_RNG_CTL_WRITE, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 stx	%o1, [%o4]
+ENDPROC(sun4v_rng_ctl_write_v1)
+
+ENTRY(sun4v_rng_ctl_write_v2)
+	mov	HV_FAST_RNG_CTL_WRITE, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 nop
+ENDPROC(sun4v_rng_ctl_write_v2)
+
+ENTRY(sun4v_rng_data_read_diag_v1)
+	mov	%o2, %o4
+	mov	HV_FAST_RNG_DATA_READ_DIAG, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 stx	%o1, [%o4]
+ENDPROC(sun4v_rng_data_read_diag_v1)
+
+ENTRY(sun4v_rng_data_read_diag_v2)
+	mov	%o3, %o4
+	mov	HV_FAST_RNG_DATA_READ_DIAG, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 stx	%o1, [%o4]
+ENDPROC(sun4v_rng_data_read_diag_v2)
+
+ENTRY(sun4v_rng_data_read)
+	mov	%o1, %o4
+	mov	HV_FAST_RNG_DATA_READ, %o5
+	ta	HV_FAST_TRAP
+	retl
+	 stx	%o1, [%o4]
+ENDPROC(sun4v_rng_data_read)
diff --git a/drivers/char/hw_random/n2-drv.c b/drivers/char/hw_random/n2-drv.c
new file mode 100644
index 00000000000..5220f541df2
--- /dev/null
+++ b/drivers/char/hw_random/n2-drv.c
@@ -0,0 +1,771 @@
+/* n2-drv.c: Niagara-2 RNG driver.
+ *
+ * Copyright (C) 2008 David S. Miller <davem@davemloft.net>
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include <linux/preempt.h>
+#include <linux/hw_random.h>
+
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include <asm/hypervisor.h>
+
+#include "n2rng.h"
+
+#define DRV_MODULE_NAME		"n2rng"
+#define PFX DRV_MODULE_NAME	": "
+#define DRV_MODULE_VERSION	"0.1"
+#define DRV_MODULE_RELDATE	"May 15, 2008"
+
+static char version[] __devinitdata =
+	DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
+
+MODULE_AUTHOR("David S. Miller (davem@davemloft.net)");
+MODULE_DESCRIPTION("Niagara2 RNG driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(DRV_MODULE_VERSION);
+
+/* The Niagara2 RNG provides a 64-bit read-only random number
+ * register, plus a control register.  Access to the RNG is
+ * virtualized through the hypervisor so that both guests and control
+ * nodes can access the device.
+ *
+ * The entropy source consists of raw entropy sources, each
+ * constructed from a voltage controlled oscillator whose phase is
+ * jittered by thermal noise sources.
+ *
+ * The oscillator in each of the three raw entropy sources run at
+ * different frequencies.  Normally, all three generator outputs are
+ * gathered, xored together, and fed into a CRC circuit, the output of
+ * which is the 64-bit read-only register.
+ *
+ * Some time is necessary for all the necessary entropy to build up
+ * such that a full 64-bits of entropy are available in the register.
+ * In normal operating mode (RNG_CTL_LFSR is set), the chip implements
+ * an interlock which blocks register reads until sufficient entropy
+ * is available.
+ *
+ * A control register is provided for adjusting various aspects of RNG
+ * operation, and to enable diagnostic modes.  Each of the three raw
+ * entropy sources has an enable bit (RNG_CTL_ES{1,2,3}).  Also
+ * provided are fields for controlling the minimum time in cycles
+ * between read accesses to the register (RNG_CTL_WAIT, this controls
+ * the interlock described in the previous paragraph).
+ *
+ * The standard setting is to have the mode bit (RNG_CTL_LFSR) set,
+ * all three entropy sources enabled, and the interlock time set
+ * appropriately.
+ *
+ * The CRC polynomial used by the chip is:
+ *
+ * P(X) = x64 + x61 + x57 + x56 + x52 + x51 + x50 + x48 + x47 + x46 +
+ *        x43 + x42 + x41 + x39 + x38 + x37 + x35 + x32 + x28 + x25 +
+ *        x22 + x21 + x17 + x15 + x13 + x12 + x11 + x7 + x5 + x + 1
+ *
+ * The RNG_CTL_VCO value of each noise cell must be programmed
+ * seperately.  This is why 4 control register values must be provided
+ * to the hypervisor.  During a write, the hypervisor writes them all,
+ * one at a time, to the actual RNG_CTL register.  The first three
+ * values are used to setup the desired RNG_CTL_VCO for each entropy
+ * source, for example:
+ *
+ *	control 0: (1 << RNG_CTL_VCO_SHIFT) | RNG_CTL_ES1
+ *	control 1: (2 << RNG_CTL_VCO_SHIFT) | RNG_CTL_ES2
+ *	control 2: (3 << RNG_CTL_VCO_SHIFT) | RNG_CTL_ES3
+ *
+ * And then the fourth value sets the final chip state and enables
+ * desired.
+ */
+
+static int n2rng_hv_err_trans(unsigned long hv_err)
+{
+	switch (hv_err) {
+	case HV_EOK:
+		return 0;
+	case HV_EWOULDBLOCK:
+		return -EAGAIN;
+	case HV_ENOACCESS:
+		return -EPERM;
+	case HV_EIO:
+		return -EIO;
+	case HV_EBUSY:
+		return -EBUSY;
+	case HV_EBADALIGN:
+	case HV_ENORADDR:
+		return -EFAULT;
+	default:
+		return -EINVAL;
+	}
+}
+
+static unsigned long n2rng_generic_read_control_v2(unsigned long ra,
+						   unsigned long unit)
+{
+	unsigned long hv_err, state, ticks, watchdog_delta, watchdog_status;
+	int block = 0, busy = 0;
+
+	while (1) {
+		hv_err = sun4v_rng_ctl_read_v2(ra, unit, &state,
+					       &ticks,
+					       &watchdog_delta,
+					       &watchdog_status);
+		if (hv_err == HV_EOK)
+			break;
+
+		if (hv_err == HV_EBUSY) {
+			if (++busy >= N2RNG_BUSY_LIMIT)
+				break;
+
+			udelay(1);
+		} else if (hv_err == HV_EWOULDBLOCK) {
+			if (++block >= N2RNG_BLOCK_LIMIT)
+				break;
+
+			__delay(ticks);
+		} else
+			break;
+	}
+
+	return hv_err;
+}
+
+/* In multi-socket situations, the hypervisor might need to
+ * queue up the RNG control register write if it's for a unit
+ * that is on a cpu socket other than the one we are executing on.
+ *
+ * We poll here waiting for a successful read of that control
+ * register to make sure the write has been actually performed.
+ */
+static unsigned long n2rng_control_settle_v2(struct n2rng *np, int unit)
+{
+	unsigned long ra = __pa(&np->scratch_control[0]);
+
+	return n2rng_generic_read_control_v2(ra, unit);
+}
+
+static unsigned long n2rng_write_ctl_one(struct n2rng *np, int unit,
+					 unsigned long state,
+					 unsigned long control_ra,
+					 unsigned long watchdog_timeout,
+					 unsigned long *ticks)
+{
+	unsigned long hv_err;
+
+	if (np->hvapi_major == 1) {
+		hv_err = sun4v_rng_ctl_write_v1(control_ra, state,
+						watchdog_timeout, ticks);
+	} else {
+		hv_err = sun4v_rng_ctl_write_v2(control_ra, state,
+						watchdog_timeout, unit);
+		if (hv_err == HV_EOK)
+			hv_err = n2rng_control_settle_v2(np, unit);
+		*ticks = N2RNG_ACCUM_CYCLES_DEFAULT;
+	}
+
+	return hv_err;
+}
+
+static int n2rng_generic_read_data(unsigned long data_ra)
+{
+	unsigned long ticks, hv_err;
+	int block = 0, hcheck = 0;
+
+	while (1) {
+		hv_err = sun4v_rng_data_read(data_ra, &ticks);
+		if (hv_err == HV_EOK)
+			return 0;
+
+		if (hv_err == HV_EWOULDBLOCK) {
+			if (++block >= N2RNG_BLOCK_LIMIT)
+				return -EWOULDBLOCK;
+			__delay(ticks);
+		} else if (hv_err == HV_ENOACCESS) {
+			return -EPERM;
+		} else if (hv_err == HV_EIO) {
+			if (++hcheck >= N2RNG_HCHECK_LIMIT)
+				return -EIO;
+			udelay(10000);
+		} else
+			return -ENODEV;
+	}
+}
+
+static unsigned long n2rng_read_diag_data_one(struct n2rng *np,
+					      unsigned long unit,
+					      unsigned long data_ra,
+					      unsigned long data_len,
+					      unsigned long *ticks)
+{
+	unsigned long hv_err;
+
+	if (np->hvapi_major == 1) {
+		hv_err = sun4v_rng_data_read_diag_v1(data_ra, data_len, ticks);
+	} else {
+		hv_err = sun4v_rng_data_read_diag_v2(data_ra, data_len,
+						     unit, ticks);
+		if (!*ticks)
+			*ticks = N2RNG_ACCUM_CYCLES_DEFAULT;
+	}
+	return hv_err;
+}
+
+static int n2rng_generic_read_diag_data(struct n2rng *np,
+					unsigned long unit,
+					unsigned long data_ra,
+					unsigned long data_len)
+{
+	unsigned long ticks, hv_err;
+	int block = 0;
+
+	while (1) {
+		hv_err = n2rng_read_diag_data_one(np, unit,
+						  data_ra, data_len,
+						  &ticks);
+		if (hv_err == HV_EOK)
+			return 0;
+
+		if (hv_err == HV_EWOULDBLOCK) {
+			if (++block >= N2RNG_BLOCK_LIMIT)
+				return -EWOULDBLOCK;
+			__delay(ticks);
+		} else if (hv_err == HV_ENOACCESS) {
+			return -EPERM;
+		} else if (hv_err == HV_EIO) {
+			return -EIO;
+		} else
+			return -ENODEV;
+	}
+}
+
+
+static int n2rng_generic_write_control(struct n2rng *np,
+				       unsigned long control_ra,
+				       unsigned long unit,
+				       unsigned long state)
+{
+	unsigned long hv_err, ticks;
+	int block = 0, busy = 0;
+
+	while (1) {
+		hv_err = n2rng_write_ctl_one(np, unit, state, control_ra,
+					     np->wd_timeo, &ticks);
+		if (hv_err == HV_EOK)
+			return 0;
+
+		if (hv_err == HV_EWOULDBLOCK) {
+			if (++block >= N2RNG_BLOCK_LIMIT)
+				return -EWOULDBLOCK;
+			__delay(ticks);
+		} else if (hv_err == HV_EBUSY) {
+			if (++busy >= N2RNG_BUSY_LIMIT)
+				return -EBUSY;
+			udelay(1);
+		} else
+			return -ENODEV;
+	}
+}
+
+/* Just try to see if we can successfully access the control register
+ * of the RNG on the domain on which we are currently executing.
+ */
+static int n2rng_try_read_ctl(struct n2rng *np)
+{
+	unsigned long hv_err;
+	unsigned long x;
+
+	if (np->hvapi_major == 1) {
+		hv_err = sun4v_rng_get_diag_ctl();
+	} else {
+		/* We purposefully give invalid arguments, HV_NOACCESS
+		 * is higher priority than the errors we'd get from
+		 * these other cases, and that's the error we are
+		 * truly interested in.
+		 */
+		hv_err = sun4v_rng_ctl_read_v2(0UL, ~0UL, &x, &x, &x, &x);
+		switch (hv_err) {
+		case HV_EWOULDBLOCK:
+		case HV_ENOACCESS:
+			break;
+		default:
+			hv_err = HV_EOK;
+			break;
+		}
+	}
+
+	return n2rng_hv_err_trans(hv_err);
+}
+
+#define CONTROL_DEFAULT_BASE		\
+	((2 << RNG_CTL_ASEL_SHIFT) |	\
+	 (N2RNG_ACCUM_CYCLES_DEFAULT << RNG_CTL_WAIT_SHIFT) |	\
+	 RNG_CTL_LFSR)
+
+#define CONTROL_DEFAULT_0		\
+	(CONTROL_DEFAULT_BASE |		\
+	 (1 << RNG_CTL_VCO_SHIFT) |	\
+	 RNG_CTL_ES1)
+#define CONTROL_DEFAULT_1		\
+	(CONTROL_DEFAULT_BASE |		\
+	 (2 << RNG_CTL_VCO_SHIFT) |	\
+	 RNG_CTL_ES2)
+#define CONTROL_DEFAULT_2		\
+	(CONTROL_DEFAULT_BASE |		\
+	 (3 << RNG_CTL_VCO_SHIFT) |	\
+	 RNG_CTL_ES3)
+#define CONTROL_DEFAULT_3		\
+	(CONTROL_DEFAULT_BASE |		\
+	 RNG_CTL_ES1 | RNG_CTL_ES2 | RNG_CTL_ES3)
+
+static void n2rng_control_swstate_init(struct n2rng *np)
+{
+	int i;
+
+	np->flags |= N2RNG_FLAG_CONTROL;
+
+	np->health_check_sec = N2RNG_HEALTH_CHECK_SEC_DEFAULT;
+	np->accum_cycles = N2RNG_ACCUM_CYCLES_DEFAULT;
+	np->wd_timeo = N2RNG_WD_TIMEO_DEFAULT;
+
+	for (i = 0; i < np->num_units; i++) {
+		struct n2rng_unit *up = &np->units[i];
+
+		up->control[0] = CONTROL_DEFAULT_0;
+		up->control[1] = CONTROL_DEFAULT_1;
+		up->control[2] = CONTROL_DEFAULT_2;
+		up->control[3] = CONTROL_DEFAULT_3;
+	}
+
+	np->hv_state = HV_RNG_STATE_UNCONFIGURED;
+}
+
+static int n2rng_grab_diag_control(struct n2rng *np)
+{
+	int i, busy_count, err = -ENODEV;
+
+	busy_count = 0;
+	for (i = 0; i < 100; i++) {
+		err = n2rng_try_read_ctl(np);
+		if (err != -EAGAIN)
+			break;
+
+		if (++busy_count > 100) {
+			dev_err(&np->op->dev,
+				"Grab diag control timeout.\n");
+			return -ENODEV;
+		}
+
+		udelay(1);
+	}
+
+	return err;
+}
+
+static int n2rng_init_control(struct n2rng *np)
+{
+	int err = n2rng_grab_diag_control(np);
+
+	/* Not in the control domain, that's OK we are only a consumer
+	 * of the RNG data, we don't setup and program it.
+	 */
+	if (err == -EPERM)
+		return 0;
+	if (err)
+		return err;
+
+	n2rng_control_swstate_init(np);
+
+	return 0;
+}
+
+static int n2rng_data_read(struct hwrng *rng, u32 *data)
+{
+	struct n2rng *np = (struct n2rng *) rng->priv;
+	unsigned long ra = __pa(&np->test_data);
+	int len;
+
+	if (!(np->flags & N2RNG_FLAG_READY)) {
+		len = 0;
+	} else if (np->flags & N2RNG_FLAG_BUFFER_VALID) {
+		np->flags &= ~N2RNG_FLAG_BUFFER_VALID;
+		*data = np->buffer;
+		len = 4;
+	} else {
+		int err = n2rng_generic_read_data(ra);
+		if (!err) {
+			np->buffer = np->test_data >> 32;
+			*data = np->test_data & 0xffffffff;
+			len = 4;
+		} else {
+			dev_err(&np->op->dev, "RNG error, restesting\n");
+			np->flags &= ~N2RNG_FLAG_READY;
+			if (!(np->flags & N2RNG_FLAG_SHUTDOWN))
+				schedule_delayed_work(&np->work, 0);
+			len = 0;
+		}
+	}
+
+	return len;
+}
+
+/* On a guest node, just make sure we can read random data properly.
+ * If a control node reboots or reloads it's n2rng driver, this won't
+ * work during that time.  So we have to keep probing until the device
+ * becomes usable.
+ */
+static int n2rng_guest_check(struct n2rng *np)
+{
+	unsigned long ra = __pa(&np->test_data);
+
+	return n2rng_generic_read_data(ra);
+}
+
+static int n2rng_entropy_diag_read(struct n2rng *np, unsigned long unit,
+				   u64 *pre_control, u64 pre_state,
+				   u64 *buffer, unsigned long buf_len,
+				   u64 *post_control, u64 post_state)
+{
+	unsigned long post_ctl_ra = __pa(post_control);
+	unsigned long pre_ctl_ra = __pa(pre_control);
+	unsigned long buffer_ra = __pa(buffer);
+	int err;
+
+	err = n2rng_generic_write_control(np, pre_ctl_ra, unit, pre_state);
+	if (err)
+		return err;
+
+	err = n2rng_generic_read_diag_data(np, unit,
+					   buffer_ra, buf_len);
+
+	(void) n2rng_generic_write_control(np, post_ctl_ra, unit,
+					   post_state);
+
+	return err;
+}
+
+static u64 advance_polynomial(u64 poly, u64 val, int count)
+{
+	int i;
+
+	for (i = 0; i < count; i++) {
+		int highbit_set = ((s64)val < 0);
+
+		val <<= 1;
+		if (highbit_set)
+			val ^= poly;
+	}
+
+	return val;
+}
+
+static int n2rng_test_buffer_find(struct n2rng *np, u64 val)
+{
+	int i, count = 0;
+
+	/* Purposefully skip over the first word.  */
+	for (i = 1; i < SELFTEST_BUFFER_WORDS; i++) {
+		if (np->test_buffer[i] == val)
+			count++;
+	}
+	return count;
+}
+
+static void n2rng_dump_test_buffer(struct n2rng *np)
+{
+	int i;
+
+	for (i = 0; i < SELFTEST_BUFFER_WORDS; i++)
+		dev_err(&np->op->dev, "Test buffer slot %d [0x%016lx]\n",
+			i, np->test_buffer[i]);
+}
+
+static int n2rng_check_selftest_buffer(struct n2rng *np, unsigned long unit)
+{
+	u64 val = SELFTEST_VAL;
+	int err, matches, limit;
+
+	matches = 0;
+	for (limit = 0; limit < SELFTEST_LOOPS_MAX; limit++) {
+		matches += n2rng_test_buffer_find(np, val);
+		if (matches >= SELFTEST_MATCH_GOAL)
+			break;
+		val = advance_polynomial(SELFTEST_POLY, val, 1);
+	}
+
+	err = 0;
+	if (limit >= SELFTEST_LOOPS_MAX) {
+		err = -ENODEV;
+		dev_err(&np->op->dev, "Selftest failed on unit %lu\n", unit);
+		n2rng_dump_test_buffer(np);
+	} else
+		dev_info(&np->op->dev, "Selftest passed on unit %lu\n", unit);
+
+	return err;
+}
+
+static int n2rng_control_selftest(struct n2rng *np, unsigned long unit)
+{
+	int err;
+
+	np->test_control[0] = (0x2 << RNG_CTL_ASEL_SHIFT);
+	np->test_control[1] = (0x2 << RNG_CTL_ASEL_SHIFT);
+	np->test_control[2] = (0x2 << RNG_CTL_ASEL_SHIFT);
+	np->test_control[3] = ((0x2 << RNG_CTL_ASEL_SHIFT) |
+			       RNG_CTL_LFSR |
+			       ((SELFTEST_TICKS - 2) << RNG_CTL_WAIT_SHIFT));
+
+
+	err = n2rng_entropy_diag_read(np, unit, np->test_control,
+				      HV_RNG_STATE_HEALTHCHECK,
+				      np->test_buffer,
+				      sizeof(np->test_buffer),
+				      &np->units[unit].control[0],
+				      np->hv_state);
+	if (err)
+		return err;
+
+	return n2rng_check_selftest_buffer(np, unit);
+}
+
+static int n2rng_control_check(struct n2rng *np)
+{
+	int i;
+
+	for (i = 0; i < np->num_units; i++) {
+		int err = n2rng_control_selftest(np, i);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+/* The sanity checks passed, install the final configuration into the
+ * chip, it's ready to use.
+ */
+static int n2rng_control_configure_units(struct n2rng *np)
+{
+	int unit, err;
+
+	err = 0;
+	for (unit = 0; unit < np->num_units; unit++) {
+		struct n2rng_unit *up = &np->units[unit];
+		unsigned long ctl_ra = __pa(&up->control[0]);
+		int esrc;
+		u64 base;
+
+		base = ((np->accum_cycles << RNG_CTL_WAIT_SHIFT) |
+			(2 << RNG_CTL_ASEL_SHIFT) |
+			RNG_CTL_LFSR);
+
+		/* XXX This isn't the best.  We should fetch a bunch
+		 * XXX of words using each entropy source combined XXX
+		 * with each VCO setting, and see which combinations
+		 * XXX give the best random data.
+		 */
+		for (esrc = 0; esrc < 3; esrc++)
+			up->control[esrc] = base |
+				(esrc << RNG_CTL_VCO_SHIFT) |
+				(RNG_CTL_ES1 << esrc);
+
+		up->control[3] = base |
+			(RNG_CTL_ES1 | RNG_CTL_ES2 | RNG_CTL_ES3);
+
+		err = n2rng_generic_write_control(np, ctl_ra, unit,
+						  HV_RNG_STATE_CONFIGURED);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static void n2rng_work(struct work_struct *work)
+{
+	struct n2rng *np = container_of(work, struct n2rng, work.work);
+	int err = 0;
+
+	if (!(np->flags & N2RNG_FLAG_CONTROL)) {
+		err = n2rng_guest_check(np);
+	} else {
+		preempt_disable();
+		err = n2rng_control_check(np);
+		preempt_enable();
+
+		if (!err)
+			err = n2rng_control_configure_units(np);
+	}
+
+	if (!err) {
+		np->flags |= N2RNG_FLAG_READY;
+		dev_info(&np->op->dev, "RNG ready\n");
+	}
+
+	if (err && !(np->flags & N2RNG_FLAG_SHUTDOWN))
+		schedule_delayed_work(&np->work, HZ * 2);
+}
+
+static void __devinit n2rng_driver_version(void)
+{
+	static int n2rng_version_printed;
+
+	if (n2rng_version_printed++ == 0)
+		pr_info("%s", version);
+}
+
+static int __devinit n2rng_probe(struct of_device *op,
+				 const struct of_device_id *match)
+{
+	int victoria_falls = (match->data != NULL);
+	int err = -ENOMEM;
+	struct n2rng *np;
+
+	n2rng_driver_version();
+
+	np = kzalloc(sizeof(*np), GFP_KERNEL);
+	if (!np)
+		goto out;
+	np->op = op;
+
+	INIT_DELAYED_WORK(&np->work, n2rng_work);
+
+	if (victoria_falls)
+		np->flags |= N2RNG_FLAG_VF;
+
+	err = -ENODEV;
+	np->hvapi_major = 2;
+	if (sun4v_hvapi_register(HV_GRP_RNG,
+				 np->hvapi_major,
+				 &np->hvapi_minor)) {
+		np->hvapi_major = 1;
+		if (sun4v_hvapi_register(HV_GRP_RNG,
+					 np->hvapi_major,
+					 &np->hvapi_minor)) {
+			dev_err(&op->dev, "Cannot register suitable "
+				"HVAPI version.\n");
+			goto out_free;
+		}
+	}
+
+	if (np->flags & N2RNG_FLAG_VF) {
+		if (np->hvapi_major < 2) {
+			dev_err(&op->dev, "VF RNG requires HVAPI major "
+				"version 2 or later, got %lu\n",
+				np->hvapi_major);
+			goto out_hvapi_unregister;
+		}
+		np->num_units = of_getintprop_default(op->node,
+						      "rng-#units", 0);
+		if (!np->num_units) {
+			dev_err(&op->dev, "VF RNG lacks rng-#units property\n");
+			goto out_hvapi_unregister;
+		}
+	} else
+		np->num_units = 1;
+
+	dev_info(&op->dev, "Registered RNG HVAPI major %lu minor %lu\n",
+		 np->hvapi_major, np->hvapi_minor);
+
+	np->units = kzalloc(sizeof(struct n2rng_unit) * np->num_units,
+			    GFP_KERNEL);
+	err = -ENOMEM;
+	if (!np->units)
+		goto out_hvapi_unregister;
+
+	err = n2rng_init_control(np);
+	if (err)
+		goto out_free_units;
+
+	dev_info(&op->dev, "Found %s RNG, units: %d\n",
+		 ((np->flags & N2RNG_FLAG_VF) ?
+		  "Victoria Falls" : "Niagara2"),
+		 np->num_units);
+
+	np->hwrng.name = "n2rng";
+	np->hwrng.data_read = n2rng_data_read;
+	np->hwrng.priv = (unsigned long) np;
+
+	err = hwrng_register(&np->hwrng);
+	if (err)
+		goto out_free_units;
+
+	dev_set_drvdata(&op->dev, np);
+
+	schedule_delayed_work(&np->work, 0);
+
+	return 0;
+
+out_free_units:
+	kfree(np->units);
+	np->units = NULL;
+
+out_hvapi_unregister:
+	sun4v_hvapi_unregister(HV_GRP_RNG);
+
+out_free:
+	kfree(np);
+out:
+	return err;
+}
+
+static int __devexit n2rng_remove(struct of_device *op)
+{
+	struct n2rng *np = dev_get_drvdata(&op->dev);
+
+	np->flags |= N2RNG_FLAG_SHUTDOWN;
+
+	cancel_delayed_work_sync(&np->work);
+
+	hwrng_unregister(&np->hwrng);
+
+	sun4v_hvapi_unregister(HV_GRP_RNG);
+
+	kfree(np->units);
+	np->units = NULL;
+
+	kfree(np);
+
+	dev_set_drvdata(&op->dev, NULL);
+
+	return 0;
+}
+
+static struct of_device_id n2rng_match[] = {
+	{
+		.name		= "random-number-generator",
+		.compatible	= "SUNW,n2-rng",
+	},
+	{
+		.name		= "random-number-generator",
+		.compatible	= "SUNW,vf-rng",
+		.data		= (void *) 1,
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, n2rng_match);
+
+static struct of_platform_driver n2rng_driver = {
+	.name		= "n2rng",
+	.match_table	= n2rng_match,
+	.probe		= n2rng_probe,
+	.remove		= __devexit_p(n2rng_remove),
+};
+
+static int __init n2rng_init(void)
+{
+	return of_register_driver(&n2rng_driver, &of_bus_type);
+}
+
+static void __exit n2rng_exit(void)
+{
+	of_unregister_driver(&n2rng_driver);
+}
+
+module_init(n2rng_init);
+module_exit(n2rng_exit);
diff --git a/drivers/char/hw_random/n2rng.h b/drivers/char/hw_random/n2rng.h
new file mode 100644
index 00000000000..a2b81e7bfc1
--- /dev/null
+++ b/drivers/char/hw_random/n2rng.h
@@ -0,0 +1,118 @@
+/* n2rng.h: Niagara2 RNG defines.
+ *
+ * Copyright (C) 2008 David S. Miller <davem@davemloft.net>
+ */
+
+#ifndef _N2RNG_H
+#define _N2RNG_H
+
+#define RNG_CTL_WAIT       0x0000000001fffe00ULL /* Minimum wait time       */
+#define RNG_CTL_WAIT_SHIFT 9
+#define RNG_CTL_BYPASS     0x0000000000000100ULL /* VCO voltage source      */
+#define RNG_CTL_VCO        0x00000000000000c0ULL /* VCO rate control        */
+#define RNG_CTL_VCO_SHIFT  6
+#define RNG_CTL_ASEL       0x0000000000000030ULL /* Analog MUX select       */
+#define RNG_CTL_ASEL_SHIFT 4
+#define RNG_CTL_LFSR       0x0000000000000008ULL /* Use LFSR or plain shift */
+#define RNG_CTL_ES3        0x0000000000000004ULL /* Enable entropy source 3 */
+#define RNG_CTL_ES2        0x0000000000000002ULL /* Enable entropy source 2 */
+#define RNG_CTL_ES1        0x0000000000000001ULL /* Enable entropy source 1 */
+
+#define HV_FAST_RNG_GET_DIAG_CTL	0x130
+#define HV_FAST_RNG_CTL_READ		0x131
+#define HV_FAST_RNG_CTL_WRITE		0x132
+#define HV_FAST_RNG_DATA_READ_DIAG	0x133
+#define HV_FAST_RNG_DATA_READ		0x134
+
+#define HV_RNG_STATE_UNCONFIGURED	0
+#define HV_RNG_STATE_CONFIGURED		1
+#define HV_RNG_STATE_HEALTHCHECK	2
+#define HV_RNG_STATE_ERROR		3
+
+#define HV_RNG_NUM_CONTROL		4
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_rng_get_diag_ctl(void);
+extern unsigned long sun4v_rng_ctl_read_v1(unsigned long ctl_regs_ra,
+					   unsigned long *state,
+					   unsigned long *tick_delta);
+extern unsigned long sun4v_rng_ctl_read_v2(unsigned long ctl_regs_ra,
+					   unsigned long unit,
+					   unsigned long *state,
+					   unsigned long *tick_delta,
+					   unsigned long *watchdog,
+					   unsigned long *write_status);
+extern unsigned long sun4v_rng_ctl_write_v1(unsigned long ctl_regs_ra,
+					    unsigned long state,
+					    unsigned long write_timeout,
+					    unsigned long *tick_delta);
+extern unsigned long sun4v_rng_ctl_write_v2(unsigned long ctl_regs_ra,
+					    unsigned long state,
+					    unsigned long write_timeout,
+					    unsigned long unit);
+extern unsigned long sun4v_rng_data_read_diag_v1(unsigned long data_ra,
+						 unsigned long len,
+						 unsigned long *tick_delta);
+extern unsigned long sun4v_rng_data_read_diag_v2(unsigned long data_ra,
+						 unsigned long len,
+						 unsigned long unit,
+						 unsigned long *tick_delta);
+extern unsigned long sun4v_rng_data_read(unsigned long data_ra,
+					 unsigned long *tick_delta);
+
+struct n2rng_unit {
+	u64			control[HV_RNG_NUM_CONTROL];
+};
+
+struct n2rng {
+	struct of_device	*op;
+
+	unsigned long		flags;
+#define N2RNG_FLAG_VF		0x00000001 /* Victoria Falls RNG, else N2 */
+#define N2RNG_FLAG_CONTROL	0x00000002 /* Operating in control domain */
+#define N2RNG_FLAG_READY	0x00000008 /* Ready for hw-rng layer      */
+#define N2RNG_FLAG_SHUTDOWN	0x00000010 /* Driver unregistering        */
+#define N2RNG_FLAG_BUFFER_VALID	0x00000020 /* u32 buffer holds valid data */
+
+	int			num_units;
+	struct n2rng_unit	*units;
+
+	struct hwrng		hwrng;
+	u32			buffer;
+
+	/* Registered hypervisor group API major and minor version.  */
+	unsigned long		hvapi_major;
+	unsigned long		hvapi_minor;
+
+	struct delayed_work	work;
+
+	unsigned long		hv_state; /* HV_RNG_STATE_foo */
+
+	unsigned long		health_check_sec;
+	unsigned long		accum_cycles;
+	unsigned long		wd_timeo;
+#define N2RNG_HEALTH_CHECK_SEC_DEFAULT	0
+#define N2RNG_ACCUM_CYCLES_DEFAULT	2048
+#define N2RNG_WD_TIMEO_DEFAULT		0
+
+	u64			scratch_control[HV_RNG_NUM_CONTROL];
+
+#define SELFTEST_TICKS		38859
+#define SELFTEST_VAL		((u64)0xB8820C7BD387E32C)
+#define SELFTEST_POLY		((u64)0x231DCEE91262B8A3)
+#define SELFTEST_MATCH_GOAL	6
+#define SELFTEST_LOOPS_MAX	40000
+#define SELFTEST_BUFFER_WORDS	8
+
+	u64			test_data;
+	u64			test_control[HV_RNG_NUM_CONTROL];
+	u64			test_buffer[SELFTEST_BUFFER_WORDS];
+};
+
+#define N2RNG_BLOCK_LIMIT	60000
+#define N2RNG_BUSY_LIMIT	100
+#define N2RNG_HCHECK_LIMIT	100
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* _N2RNG_H */
diff --git a/drivers/sbus/char/uctrl.c b/drivers/sbus/char/uctrl.c
index 513ba61ae96..777637594ac 100644
--- a/drivers/sbus/char/uctrl.c
+++ b/drivers/sbus/char/uctrl.c
@@ -195,8 +195,8 @@ struct uctrl_driver {
 
 static struct uctrl_driver drv;
 
-void uctrl_get_event_status(void);
-void uctrl_get_external_status(void);
+static void uctrl_get_event_status(void);
+static void uctrl_get_external_status(void);
 
 static int
 uctrl_ioctl(struct inode *inode, struct file *file,
@@ -266,12 +266,6 @@ static struct miscdevice uctrl_dev = {
   driver->regs->uctrl_stat = UCTRL_STAT_RXNE_STA; \
 }
 
-void uctrl_set_video(int status)
-{
-	struct uctrl_driver *driver = &drv;
-	
-}
-
 static void uctrl_do_txn(struct uctrl_txn *txn)
 {
 	struct uctrl_driver *driver = &drv;
@@ -311,7 +305,7 @@ static void uctrl_do_txn(struct uctrl_txn *txn)
 	}
 }
 
-void uctrl_get_event_status(void)
+static void uctrl_get_event_status(void)
 {
 	struct uctrl_driver *driver = &drv;
 	struct uctrl_txn txn;
@@ -331,7 +325,7 @@ void uctrl_get_event_status(void)
 	dprintk(("ev is %x\n", driver->status.event_status));
 }
 
-void uctrl_get_external_status(void)
+static void uctrl_get_external_status(void)
 {
 	struct uctrl_driver *driver = &drv;
 	struct uctrl_txn txn;
@@ -363,7 +357,7 @@ void uctrl_get_external_status(void)
 static int __init ts102_uctrl_init(void)
 {
 	struct uctrl_driver *driver = &drv;
-	int len, i;
+	int len;
 	struct linux_prom_irqs tmp_irq[2];
         unsigned int vaddr[2] = { 0, 0 };
 	int tmpnode, uctrlnode = prom_getchild(prom_root_node);
diff --git a/drivers/sbus/char/vfc.h b/drivers/sbus/char/vfc.h
index f1aa1389ea4..a5240c52aa0 100644
--- a/drivers/sbus/char/vfc.h
+++ b/drivers/sbus/char/vfc.h
@@ -133,8 +133,6 @@ struct vfc_dev {
 	unsigned char saa9051_state_array[VFC_SAA9051_NR];
 };
 
-extern struct vfc_dev **vfc_dev_lst;
-
 void captstat_reset(struct vfc_dev *);
 void memptr_reset(struct vfc_dev *);
 
@@ -145,8 +143,6 @@ int vfc_i2c_sendbuf(struct vfc_dev *, unsigned char, char *, int) ;
 int vfc_i2c_recvbuf(struct vfc_dev *, unsigned char, char *, int) ;
 int vfc_i2c_reset_bus(struct vfc_dev *);
 int vfc_init_i2c_bus(struct vfc_dev *);
-void vfc_lock_device(struct vfc_dev *);
-void vfc_unlock_device(struct vfc_dev *);
 
 #define VFC_CONTROL_DIAGMODE  0x10000000
 #define VFC_CONTROL_MEMPTR    0x20000000
diff --git a/drivers/sbus/char/vfc_dev.c b/drivers/sbus/char/vfc_dev.c
index 1f6cb8ae278..25181bb7d62 100644
--- a/drivers/sbus/char/vfc_dev.c
+++ b/drivers/sbus/char/vfc_dev.c
@@ -45,7 +45,7 @@
 #include <asm/vfc_ioctls.h>
 
 static const struct file_operations vfc_fops;
-struct vfc_dev **vfc_dev_lst;
+static struct vfc_dev **vfc_dev_lst;
 static char vfcstr[]="vfc";
 static unsigned char saa9051_init_array[VFC_SAA9051_NR] = {
 	0x00, 0x64, 0x72, 0x52,
@@ -54,18 +54,18 @@ static unsigned char saa9051_init_array[VFC_SAA9051_NR] = {
 	0x3e
 };
 
-void vfc_lock_device(struct vfc_dev *dev)
+static void vfc_lock_device(struct vfc_dev *dev)
 {
 	mutex_lock(&dev->device_lock_mtx);
 }
 
-void vfc_unlock_device(struct vfc_dev *dev)
+static void vfc_unlock_device(struct vfc_dev *dev)
 {
 	mutex_unlock(&dev->device_lock_mtx);
 }
 
 
-void vfc_captstat_reset(struct vfc_dev *dev) 
+static void vfc_captstat_reset(struct vfc_dev *dev)
 {
 	dev->control_reg |= VFC_CONTROL_CAPTRESET;
 	sbus_writel(dev->control_reg, &dev->regs->control);
@@ -75,7 +75,7 @@ void vfc_captstat_reset(struct vfc_dev *dev)
 	sbus_writel(dev->control_reg, &dev->regs->control);
 }
 
-void vfc_memptr_reset(struct vfc_dev *dev) 
+static void vfc_memptr_reset(struct vfc_dev *dev)
 {
 	dev->control_reg |= VFC_CONTROL_MEMPTR;
 	sbus_writel(dev->control_reg, &dev->regs->control);
@@ -85,7 +85,7 @@ void vfc_memptr_reset(struct vfc_dev *dev)
 	sbus_writel(dev->control_reg, &dev->regs->control);
 }
 
-int vfc_csr_init(struct vfc_dev *dev)
+static int vfc_csr_init(struct vfc_dev *dev)
 {
 	dev->control_reg = 0x80000000;
 	sbus_writel(dev->control_reg, &dev->regs->control);
@@ -107,7 +107,7 @@ int vfc_csr_init(struct vfc_dev *dev)
 	return 0;
 }
 
-int vfc_saa9051_init(struct vfc_dev *dev)
+static int vfc_saa9051_init(struct vfc_dev *dev)
 {
 	int i;
 
@@ -119,7 +119,7 @@ int vfc_saa9051_init(struct vfc_dev *dev)
 	return 0;
 }
 
-int init_vfc_hw(struct vfc_dev *dev) 
+static int init_vfc_hw(struct vfc_dev *dev)
 {
 	vfc_lock_device(dev);
 	vfc_csr_init(dev);
@@ -132,7 +132,7 @@ int init_vfc_hw(struct vfc_dev *dev)
 	return 0; 
 }
 
-int init_vfc_devstruct(struct vfc_dev *dev, int instance) 
+static int init_vfc_devstruct(struct vfc_dev *dev, int instance)
 {
 	dev->instance=instance;
 	mutex_init(&dev->device_lock_mtx);
@@ -141,7 +141,8 @@ int init_vfc_devstruct(struct vfc_dev *dev, int instance)
 	return 0;
 }
 
-int init_vfc_device(struct sbus_dev *sdev,struct vfc_dev *dev, int instance)
+static int init_vfc_device(struct sbus_dev *sdev,struct vfc_dev *dev,
+			   int instance)
 {
 	if(dev == NULL) {
 		printk(KERN_ERR "VFC: Bogus pointer passed\n");
@@ -168,7 +169,7 @@ int init_vfc_device(struct sbus_dev *sdev,struct vfc_dev *dev, int instance)
 }
 
 
-struct vfc_dev *vfc_get_dev_ptr(int instance) 
+static struct vfc_dev *vfc_get_dev_ptr(int instance)
 {
 	return vfc_dev_lst[instance];
 }
@@ -292,7 +293,7 @@ static int vfc_debug(struct vfc_dev *dev, int cmd, void __user *argp)
 	return 0;
 }
 
-int vfc_capture_start(struct vfc_dev *dev) 
+static int vfc_capture_start(struct vfc_dev *dev)
 {
 	vfc_captstat_reset(dev);
 	dev->control_reg = sbus_readl(&dev->regs->control);
@@ -314,7 +315,7 @@ int vfc_capture_start(struct vfc_dev *dev)
 	return 0;
 }
 
-int vfc_capture_poll(struct vfc_dev *dev) 
+static int vfc_capture_poll(struct vfc_dev *dev)
 {
 	int timeout = 1000;
 
@@ -390,8 +391,8 @@ static int vfc_set_control_ioctl(struct inode *inode, struct file *file,
 }
 
 
-int vfc_port_change_ioctl(struct inode *inode, struct file *file, 
-			  struct vfc_dev *dev, unsigned long arg) 
+static int vfc_port_change_ioctl(struct inode *inode, struct file *file,
+				 struct vfc_dev *dev, unsigned long arg)
 {
 	int ret = 0;
 	int cmd;
@@ -460,8 +461,8 @@ int vfc_port_change_ioctl(struct inode *inode, struct file *file,
 	return ret;
 }
 
-int vfc_set_video_ioctl(struct inode *inode, struct file *file, 
-			struct vfc_dev *dev, unsigned long arg) 
+static int vfc_set_video_ioctl(struct inode *inode, struct file *file,
+			       struct vfc_dev *dev, unsigned long arg)
 {
 	int ret = 0;
 	int cmd;
@@ -511,8 +512,8 @@ int vfc_set_video_ioctl(struct inode *inode, struct file *file,
 	return ret;
 }
 
-int vfc_get_video_ioctl(struct inode *inode, struct file *file, 
-			struct vfc_dev *dev, unsigned long arg) 
+static int vfc_get_video_ioctl(struct inode *inode, struct file *file,
+			       struct vfc_dev *dev, unsigned long arg)
 {
 	int ret = 0;
 	unsigned int status = NO_LOCK;
diff --git a/drivers/sbus/char/vfc_i2c.c b/drivers/sbus/char/vfc_i2c.c
index 9efed771f6c..32b986e0ed7 100644
--- a/drivers/sbus/char/vfc_i2c.c
+++ b/drivers/sbus/char/vfc_i2c.c
@@ -114,7 +114,7 @@ int vfc_i2c_reset_bus(struct vfc_dev *dev)
 	return 0;
 }
 
-int vfc_i2c_wait_for_bus(struct vfc_dev *dev) 
+static int vfc_i2c_wait_for_bus(struct vfc_dev *dev)
 {
 	int timeout = 1000; 
 
@@ -126,7 +126,7 @@ int vfc_i2c_wait_for_bus(struct vfc_dev *dev)
 	return 0;
 }
 
-int vfc_i2c_wait_for_pin(struct vfc_dev *dev, int ack)
+static int vfc_i2c_wait_for_pin(struct vfc_dev *dev, int ack)
 {
 	int timeout = 1000; 
 	int s1;
@@ -144,7 +144,8 @@ int vfc_i2c_wait_for_pin(struct vfc_dev *dev, int ack)
 }
 
 #define SHIFT(a) ((a) << 24)
-int vfc_i2c_xmit_addr(struct vfc_dev *dev, unsigned char addr, char mode) 
+static int vfc_i2c_xmit_addr(struct vfc_dev *dev, unsigned char addr,
+			     char mode)
 { 
 	int ret, raddr;
 #if 1
@@ -195,7 +196,7 @@ int vfc_i2c_xmit_addr(struct vfc_dev *dev, unsigned char addr, char mode)
 	return 0;
 }
 
-int vfc_i2c_xmit_byte(struct vfc_dev *dev,unsigned char *byte) 
+static int vfc_i2c_xmit_byte(struct vfc_dev *dev,unsigned char *byte)
 {
 	int ret;
 	u32 val = SHIFT((unsigned int)*byte);
@@ -218,7 +219,8 @@ int vfc_i2c_xmit_byte(struct vfc_dev *dev,unsigned char *byte)
 	return ret;
 }
 
-int vfc_i2c_recv_byte(struct vfc_dev *dev, unsigned char *byte, int last) 
+static int vfc_i2c_recv_byte(struct vfc_dev *dev, unsigned char *byte,
+			     int last)
 {
 	int ret;
 
diff --git a/drivers/sbus/dvma.c b/drivers/sbus/dvma.c
index 57e1526746a..ab0d2de3324 100644
--- a/drivers/sbus/dvma.c
+++ b/drivers/sbus/dvma.c
@@ -16,7 +16,7 @@
 
 struct sbus_dma *dma_chain;
 
-void __init init_one_dvma(struct sbus_dma *dma, int num_dma)
+static void __init init_one_dvma(struct sbus_dma *dma, int num_dma)
 {
 	printk("dma%d: ", num_dma);
 	
diff --git a/include/asm-sparc/Kbuild b/include/asm-sparc/Kbuild
index 671223718f0..6cdaf9d33b3 100644
--- a/include/asm-sparc/Kbuild
+++ b/include/asm-sparc/Kbuild
@@ -1,14 +1 @@
-include include/asm-generic/Kbuild.asm
-
-header-y += apc.h
-header-y += asi.h
-header-y += bpp.h
-header-y += jsflash.h
-header-y += openpromio.h
-header-y += reg.h
-header-y += traps.h
-header-y += vfc_ioctls.h
-
-unifdef-y += fbio.h
-unifdef-y += perfctr.h
-unifdef-y += psr.h
+# dummy file to avoid breaking make headers_install
diff --git a/include/asm-sparc/agp.h b/include/asm-sparc/agp.h
new file mode 100644
index 00000000000..c2456870b05
--- /dev/null
+++ b/include/asm-sparc/agp.h
@@ -0,0 +1,20 @@
+#ifndef AGP_H
+#define AGP_H 1
+
+/* dummy for now */
+
+#define map_page_into_agp(page)
+#define unmap_page_from_agp(page)
+#define flush_agp_cache() mb()
+
+/* Convert a physical address to an address suitable for the GART. */
+#define phys_to_gart(x) (x)
+#define gart_to_phys(x) (x)
+
+/* GATT allocation. Returns/accepts GATT kernel virtual address. */
+#define alloc_gatt_pages(order)		\
+	((char *)__get_free_pages(GFP_KERNEL, (order)))
+#define free_gatt_pages(table, order)	\
+	free_pages((unsigned long)(table), (order))
+
+#endif
diff --git a/include/asm-sparc/apb.h b/include/asm-sparc/apb.h
new file mode 100644
index 00000000000..8f3b57db810
--- /dev/null
+++ b/include/asm-sparc/apb.h
@@ -0,0 +1,36 @@
+/*
+ * apb.h: Advanced PCI Bridge Configuration Registers and Bits
+ *
+ * Copyright (C) 1998  Eddie C. Dost  (ecd@skynet.be)
+ */
+
+#ifndef _SPARC64_APB_H
+#define _SPARC64_APB_H
+
+#define APB_TICK_REGISTER			0xb0
+#define APB_INT_ACK				0xb8
+#define APB_PRIMARY_MASTER_RETRY_LIMIT		0xc0
+#define APB_DMA_ASFR				0xc8
+#define APB_DMA_AFAR				0xd0
+#define APB_PIO_TARGET_RETRY_LIMIT		0xd8
+#define APB_PIO_TARGET_LATENCY_TIMER		0xd9
+#define APB_DMA_TARGET_RETRY_LIMIT		0xda
+#define APB_DMA_TARGET_LATENCY_TIMER		0xdb
+#define APB_SECONDARY_MASTER_RETRY_LIMIT	0xdc
+#define APB_SECONDARY_CONTROL			0xdd
+#define APB_IO_ADDRESS_MAP			0xde
+#define APB_MEM_ADDRESS_MAP			0xdf
+
+#define APB_PCI_CONTROL_LOW			0xe0
+#  define APB_PCI_CTL_LOW_ARB_PARK			(1 << 21)
+#  define APB_PCI_CTL_LOW_ERRINT_EN			(1 << 8)
+
+#define APB_PCI_CONTROL_HIGH			0xe4
+#  define APB_PCI_CTL_HIGH_SERR				(1 << 2)
+#  define APB_PCI_CTL_HIGH_ARBITER_EN			(1 << 0)
+
+#define APB_PIO_ASFR				0xe8
+#define APB_PIO_AFAR				0xf0
+#define APB_DIAG_REGISTER			0xf8
+
+#endif /* !(_SPARC64_APB_H) */
diff --git a/include/asm-sparc/asi.h b/include/asm-sparc/asi.h
index 158f9b00d43..74703c5ef98 100644
--- a/include/asm-sparc/asi.h
+++ b/include/asm-sparc/asi.h
@@ -3,7 +3,7 @@
 
 /* asi.h:  Address Space Identifier values for the sparc.
  *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
  *
  * Pioneer work for sun4m: Paul Hatchman (paul@sfe.com.au)
  * Joint edition for sun4c+sun4m: Pete A. Zaitcev <zaitcev@ipmce.su>
@@ -108,4 +108,155 @@
 
 #define ASI_M_ACTION       0x4c   /* Breakpoint Action Register (GNU/Viking) */
 
+/* V9 Architecture mandary ASIs. */
+#define ASI_N			0x04 /* Nucleus				*/
+#define ASI_NL			0x0c /* Nucleus, little endian		*/
+#define ASI_AIUP		0x10 /* Primary, user			*/
+#define ASI_AIUS		0x11 /* Secondary, user			*/
+#define ASI_AIUPL		0x18 /* Primary, user, little endian	*/
+#define ASI_AIUSL		0x19 /* Secondary, user, little endian	*/
+#define ASI_P			0x80 /* Primary, implicit		*/
+#define ASI_S			0x81 /* Secondary, implicit		*/
+#define ASI_PNF			0x82 /* Primary, no fault		*/
+#define ASI_SNF			0x83 /* Secondary, no fault		*/
+#define ASI_PL			0x88 /* Primary, implicit, l-endian	*/
+#define ASI_SL			0x89 /* Secondary, implicit, l-endian	*/
+#define ASI_PNFL		0x8a /* Primary, no fault, l-endian	*/
+#define ASI_SNFL		0x8b /* Secondary, no fault, l-endian	*/
+
+/* SpitFire and later extended ASIs.  The "(III)" marker designates
+ * UltraSparc-III and later specific ASIs.  The "(CMT)" marker designates
+ * Chip Multi Threading specific ASIs.  "(NG)" designates Niagara specific
+ * ASIs, "(4V)" designates SUN4V specific ASIs.
+ */
+#define ASI_PHYS_USE_EC		0x14 /* PADDR, E-cachable		*/
+#define ASI_PHYS_BYPASS_EC_E	0x15 /* PADDR, E-bit			*/
+#define ASI_BLK_AIUP_4V		0x16 /* (4V) Prim, user, block ld/st	*/
+#define ASI_BLK_AIUS_4V		0x17 /* (4V) Sec, user, block ld/st	*/
+#define ASI_PHYS_USE_EC_L	0x1c /* PADDR, E-cachable, little endian*/
+#define ASI_PHYS_BYPASS_EC_E_L	0x1d /* PADDR, E-bit, little endian	*/
+#define ASI_BLK_AIUP_L_4V	0x1e /* (4V) Prim, user, block, l-endian*/
+#define ASI_BLK_AIUS_L_4V	0x1f /* (4V) Sec, user, block, l-endian	*/
+#define ASI_SCRATCHPAD		0x20 /* (4V) Scratch Pad Registers	*/
+#define ASI_MMU			0x21 /* (4V) MMU Context Registers	*/
+#define ASI_BLK_INIT_QUAD_LDD_AIUS 0x23 /* (NG) init-store, twin load,
+					 * secondary, user
+					 */
+#define ASI_NUCLEUS_QUAD_LDD	0x24 /* Cachable, qword load		*/
+#define ASI_QUEUE		0x25 /* (4V) Interrupt Queue Registers	*/
+#define ASI_QUAD_LDD_PHYS_4V	0x26 /* (4V) Physical, qword load	*/
+#define ASI_NUCLEUS_QUAD_LDD_L	0x2c /* Cachable, qword load, l-endian 	*/
+#define ASI_QUAD_LDD_PHYS_L_4V	0x2e /* (4V) Phys, qword load, l-endian	*/
+#define ASI_PCACHE_DATA_STATUS	0x30 /* (III) PCache data stat RAM diag	*/
+#define ASI_PCACHE_DATA		0x31 /* (III) PCache data RAM diag	*/
+#define ASI_PCACHE_TAG		0x32 /* (III) PCache tag RAM diag	*/
+#define ASI_PCACHE_SNOOP_TAG	0x33 /* (III) PCache snoop tag RAM diag	*/
+#define ASI_QUAD_LDD_PHYS	0x34 /* (III+) PADDR, qword load	*/
+#define ASI_WCACHE_VALID_BITS	0x38 /* (III) WCache Valid Bits diag	*/
+#define ASI_WCACHE_DATA		0x39 /* (III) WCache data RAM diag	*/
+#define ASI_WCACHE_TAG		0x3a /* (III) WCache tag RAM diag	*/
+#define ASI_WCACHE_SNOOP_TAG	0x3b /* (III) WCache snoop tag RAM diag	*/
+#define ASI_QUAD_LDD_PHYS_L	0x3c /* (III+) PADDR, qw-load, l-endian	*/
+#define ASI_SRAM_FAST_INIT	0x40 /* (III+) Fast SRAM init		*/
+#define ASI_CORE_AVAILABLE	0x41 /* (CMT) LP Available		*/
+#define ASI_CORE_ENABLE_STAT	0x41 /* (CMT) LP Enable Status		*/
+#define ASI_CORE_ENABLE		0x41 /* (CMT) LP Enable RW		*/
+#define ASI_XIR_STEERING	0x41 /* (CMT) XIR Steering RW		*/
+#define ASI_CORE_RUNNING_RW	0x41 /* (CMT) LP Running RW		*/
+#define ASI_CORE_RUNNING_W1S	0x41 /* (CMT) LP Running Write-One Set	*/
+#define ASI_CORE_RUNNING_W1C	0x41 /* (CMT) LP Running Write-One Clr	*/
+#define ASI_CORE_RUNNING_STAT	0x41 /* (CMT) LP Running Status		*/
+#define ASI_CMT_ERROR_STEERING	0x41 /* (CMT) Error Steering RW		*/
+#define ASI_DCACHE_INVALIDATE	0x42 /* (III) DCache Invalidate diag	*/
+#define ASI_DCACHE_UTAG		0x43 /* (III) DCache uTag diag		*/
+#define ASI_DCACHE_SNOOP_TAG	0x44 /* (III) DCache snoop tag RAM diag	*/
+#define ASI_LSU_CONTROL		0x45 /* Load-store control unit		*/
+#define ASI_DCU_CONTROL_REG	0x45 /* (III) DCache Unit Control reg	*/
+#define ASI_DCACHE_DATA		0x46 /* DCache data-ram diag access	*/
+#define ASI_DCACHE_TAG		0x47 /* Dcache tag/valid ram diag access*/
+#define ASI_INTR_DISPATCH_STAT	0x48 /* IRQ vector dispatch status	*/
+#define ASI_INTR_RECEIVE	0x49 /* IRQ vector receive status	*/
+#define ASI_UPA_CONFIG		0x4a /* UPA config space		*/
+#define ASI_JBUS_CONFIG		0x4a /* (IIIi) JBUS Config Register	*/
+#define ASI_SAFARI_CONFIG	0x4a /* (III) Safari Config Register	*/
+#define ASI_SAFARI_ADDRESS	0x4a /* (III) Safari Address Register	*/
+#define ASI_ESTATE_ERROR_EN	0x4b /* E-cache error enable space	*/
+#define ASI_AFSR		0x4c /* Async fault status register	*/
+#define ASI_AFAR		0x4d /* Async fault address register	*/
+#define ASI_EC_TAG_DATA		0x4e /* E-cache tag/valid ram diag acc	*/
+#define ASI_IMMU		0x50 /* Insn-MMU main register space	*/
+#define ASI_IMMU_TSB_8KB_PTR	0x51 /* Insn-MMU 8KB TSB pointer reg	*/
+#define ASI_IMMU_TSB_64KB_PTR	0x52 /* Insn-MMU 64KB TSB pointer reg	*/
+#define ASI_ITLB_DATA_IN	0x54 /* Insn-MMU TLB data in reg	*/
+#define ASI_ITLB_DATA_ACCESS	0x55 /* Insn-MMU TLB data access reg	*/
+#define ASI_ITLB_TAG_READ	0x56 /* Insn-MMU TLB tag read reg	*/
+#define ASI_IMMU_DEMAP		0x57 /* Insn-MMU TLB demap		*/
+#define ASI_DMMU		0x58 /* Data-MMU main register space	*/
+#define ASI_DMMU_TSB_8KB_PTR	0x59 /* Data-MMU 8KB TSB pointer reg	*/
+#define ASI_DMMU_TSB_64KB_PTR	0x5a /* Data-MMU 16KB TSB pointer reg	*/
+#define ASI_DMMU_TSB_DIRECT_PTR	0x5b /* Data-MMU TSB direct pointer reg	*/
+#define ASI_DTLB_DATA_IN	0x5c /* Data-MMU TLB data in reg	*/
+#define ASI_DTLB_DATA_ACCESS	0x5d /* Data-MMU TLB data access reg	*/
+#define ASI_DTLB_TAG_READ	0x5e /* Data-MMU TLB tag read reg	*/
+#define ASI_DMMU_DEMAP		0x5f /* Data-MMU TLB demap		*/
+#define ASI_IIU_INST_TRAP	0x60 /* (III) Instruction Breakpoint	*/
+#define ASI_INTR_ID		0x63 /* (CMT) Interrupt ID register	*/
+#define ASI_CORE_ID		0x63 /* (CMT) LP ID register		*/
+#define ASI_CESR_ID		0x63 /* (CMT) CESR ID register		*/
+#define ASI_IC_INSTR		0x66 /* Insn cache instrucion ram diag	*/
+#define ASI_IC_TAG		0x67 /* Insn cache tag/valid ram diag 	*/
+#define ASI_IC_STAG		0x68 /* (III) Insn cache snoop tag ram	*/
+#define ASI_IC_PRE_DECODE	0x6e /* Insn cache pre-decode ram diag	*/
+#define ASI_IC_NEXT_FIELD	0x6f /* Insn cache next-field ram diag	*/
+#define ASI_BRPRED_ARRAY	0x6f /* (III) Branch Prediction RAM diag*/
+#define ASI_BLK_AIUP		0x70 /* Primary, user, block load/store	*/
+#define ASI_BLK_AIUS		0x71 /* Secondary, user, block ld/st	*/
+#define ASI_MCU_CTRL_REG	0x72 /* (III) Memory controller regs	*/
+#define ASI_EC_DATA		0x74 /* (III) E-cache data staging reg	*/
+#define ASI_EC_CTRL		0x75 /* (III) E-cache control reg	*/
+#define ASI_EC_W		0x76 /* E-cache diag write access	*/
+#define ASI_UDB_ERROR_W		0x77 /* External UDB error regs W	*/
+#define ASI_UDB_CONTROL_W	0x77 /* External UDB control regs W	*/
+#define ASI_INTR_W		0x77 /* IRQ vector dispatch write	*/
+#define ASI_INTR_DATAN_W	0x77 /* (III) Out irq vector data reg N	*/
+#define ASI_INTR_DISPATCH_W	0x77 /* (III) Interrupt vector dispatch	*/
+#define ASI_BLK_AIUPL		0x78 /* Primary, user, little, blk ld/st*/
+#define ASI_BLK_AIUSL		0x79 /* Secondary, user, little, blk ld/st*/
+#define ASI_EC_R		0x7e /* E-cache diag read access	*/
+#define ASI_UDBH_ERROR_R	0x7f /* External UDB error regs rd hi	*/
+#define ASI_UDBL_ERROR_R	0x7f /* External UDB error regs rd low	*/
+#define ASI_UDBH_CONTROL_R	0x7f /* External UDB control regs rd hi	*/
+#define ASI_UDBL_CONTROL_R	0x7f /* External UDB control regs rd low*/
+#define ASI_INTR_R		0x7f /* IRQ vector dispatch read	*/
+#define ASI_INTR_DATAN_R	0x7f /* (III) In irq vector data reg N	*/
+#define ASI_PST8_P		0xc0 /* Primary, 8 8-bit, partial	*/
+#define ASI_PST8_S		0xc1 /* Secondary, 8 8-bit, partial	*/
+#define ASI_PST16_P		0xc2 /* Primary, 4 16-bit, partial	*/
+#define ASI_PST16_S		0xc3 /* Secondary, 4 16-bit, partial	*/
+#define ASI_PST32_P		0xc4 /* Primary, 2 32-bit, partial	*/
+#define ASI_PST32_S		0xc5 /* Secondary, 2 32-bit, partial	*/
+#define ASI_PST8_PL		0xc8 /* Primary, 8 8-bit, partial, L	*/
+#define ASI_PST8_SL		0xc9 /* Secondary, 8 8-bit, partial, L	*/
+#define ASI_PST16_PL		0xca /* Primary, 4 16-bit, partial, L	*/
+#define ASI_PST16_SL		0xcb /* Secondary, 4 16-bit, partial, L	*/
+#define ASI_PST32_PL		0xcc /* Primary, 2 32-bit, partial, L	*/
+#define ASI_PST32_SL		0xcd /* Secondary, 2 32-bit, partial, L	*/
+#define ASI_FL8_P		0xd0 /* Primary, 1 8-bit, fpu ld/st	*/
+#define ASI_FL8_S		0xd1 /* Secondary, 1 8-bit, fpu ld/st	*/
+#define ASI_FL16_P		0xd2 /* Primary, 1 16-bit, fpu ld/st	*/
+#define ASI_FL16_S		0xd3 /* Secondary, 1 16-bit, fpu ld/st	*/
+#define ASI_FL8_PL		0xd8 /* Primary, 1 8-bit, fpu ld/st, L	*/
+#define ASI_FL8_SL		0xd9 /* Secondary, 1 8-bit, fpu ld/st, L*/
+#define ASI_FL16_PL		0xda /* Primary, 1 16-bit, fpu ld/st, L	*/
+#define ASI_FL16_SL		0xdb /* Secondary, 1 16-bit, fpu ld/st,L*/
+#define ASI_BLK_COMMIT_P	0xe0 /* Primary, blk store commit	*/
+#define ASI_BLK_COMMIT_S	0xe1 /* Secondary, blk store commit	*/
+#define ASI_BLK_INIT_QUAD_LDD_P	0xe2 /* (NG) init-store, twin load,
+				      * primary, implicit
+				      */
+#define ASI_BLK_P		0xf0 /* Primary, blk ld/st		*/
+#define ASI_BLK_S		0xf1 /* Secondary, blk ld/st		*/
+#define ASI_BLK_PL		0xf8 /* Primary, blk ld/st, little	*/
+#define ASI_BLK_SL		0xf9 /* Secondary, blk ld/st, little	*/
+
 #endif /* _SPARC_ASI_H */
diff --git a/include/asm-sparc/atomic.h b/include/asm-sparc/atomic.h
index 5c944b5a804..66d8166ec1d 100644
--- a/include/asm-sparc/atomic.h
+++ b/include/asm-sparc/atomic.h
@@ -1,165 +1,8 @@
-/* atomic.h: These still suck, but the I-cache hit rate is higher.
- *
- * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
- * Copyright (C) 2000 Anton Blanchard (anton@linuxcare.com.au)
- * Copyright (C) 2007 Kyle McMartin (kyle@parisc-linux.org)
- *
- * Additions by Keith M Wesolowski (wesolows@foobazco.org) based
- * on asm-parisc/atomic.h Copyright (C) 2000 Philipp Rumpf <prumpf@tux.org>.
- */
-
-#ifndef __ARCH_SPARC_ATOMIC__
-#define __ARCH_SPARC_ATOMIC__
-
-#include <linux/types.h>
-
-typedef struct { volatile int counter; } atomic_t;
-
-#ifdef __KERNEL__
-
-#define ATOMIC_INIT(i)  { (i) }
-
-extern int __atomic_add_return(int, atomic_t *);
-extern int atomic_cmpxchg(atomic_t *, int, int);
-#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
-extern int atomic_add_unless(atomic_t *, int, int);
-extern void atomic_set(atomic_t *, int);
-
-#define atomic_read(v)          ((v)->counter)
-
-#define atomic_add(i, v)	((void)__atomic_add_return( (int)(i), (v)))
-#define atomic_sub(i, v)	((void)__atomic_add_return(-(int)(i), (v)))
-#define atomic_inc(v)		((void)__atomic_add_return(        1, (v)))
-#define atomic_dec(v)		((void)__atomic_add_return(       -1, (v)))
-
-#define atomic_add_return(i, v)	(__atomic_add_return( (int)(i), (v)))
-#define atomic_sub_return(i, v)	(__atomic_add_return(-(int)(i), (v)))
-#define atomic_inc_return(v)	(__atomic_add_return(        1, (v)))
-#define atomic_dec_return(v)	(__atomic_add_return(       -1, (v)))
-
-#define atomic_add_negative(a, v)	(atomic_add_return((a), (v)) < 0)
-
-/*
- * atomic_inc_and_test - increment and test
- * @v: pointer of type atomic_t
- *
- * Atomically increments @v by 1
- * and returns true if the result is zero, or false for all
- * other cases.
- */
-#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0)
-
-#define atomic_dec_and_test(v) (atomic_dec_return(v) == 0)
-#define atomic_sub_and_test(i, v) (atomic_sub_return(i, v) == 0)
-
-#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
-
-/* This is the old 24-bit implementation.  It's still used internally
- * by some sparc-specific code, notably the semaphore implementation.
- */
-typedef struct { volatile int counter; } atomic24_t;
-
-#ifndef CONFIG_SMP
-
-#define ATOMIC24_INIT(i)  { (i) }
-#define atomic24_read(v)          ((v)->counter)
-#define atomic24_set(v, i)        (((v)->counter) = i)
-
+#ifndef ___ASM_SPARC_ATOMIC_H
+#define ___ASM_SPARC_ATOMIC_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/atomic_64.h>
 #else
-/* We do the bulk of the actual work out of line in two common
- * routines in assembler, see arch/sparc/lib/atomic.S for the
- * "fun" details.
- *
- * For SMP the trick is you embed the spin lock byte within
- * the word, use the low byte so signedness is easily retained
- * via a quick arithmetic shift.  It looks like this:
- *
- *	----------------------------------------
- *	| signed 24-bit counter value |  lock  |  atomic_t
- *	----------------------------------------
- *	 31                          8 7      0
- */
-
-#define ATOMIC24_INIT(i)	{ ((i) << 8) }
-
-static inline int atomic24_read(const atomic24_t *v)
-{
-	int ret = v->counter;
-
-	while(ret & 0xff)
-		ret = v->counter;
-
-	return ret >> 8;
-}
-
-#define atomic24_set(v, i)	(((v)->counter) = ((i) << 8))
+#include <asm-sparc/atomic_32.h>
+#endif
 #endif
-
-static inline int __atomic24_add(int i, atomic24_t *v)
-{
-	register volatile int *ptr asm("g1");
-	register int increment asm("g2");
-	register int tmp1 asm("g3");
-	register int tmp2 asm("g4");
-	register int tmp3 asm("g7");
-
-	ptr = &v->counter;
-	increment = i;
-
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___atomic24_add\n\t"
-	" add	%%o7, 8, %%o7\n"
-	: "=&r" (increment), "=r" (tmp1), "=r" (tmp2), "=r" (tmp3)
-	: "0" (increment), "r" (ptr)
-	: "memory", "cc");
-
-	return increment;
-}
-
-static inline int __atomic24_sub(int i, atomic24_t *v)
-{
-	register volatile int *ptr asm("g1");
-	register int increment asm("g2");
-	register int tmp1 asm("g3");
-	register int tmp2 asm("g4");
-	register int tmp3 asm("g7");
-
-	ptr = &v->counter;
-	increment = i;
-
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___atomic24_sub\n\t"
-	" add	%%o7, 8, %%o7\n"
-	: "=&r" (increment), "=r" (tmp1), "=r" (tmp2), "=r" (tmp3)
-	: "0" (increment), "r" (ptr)
-	: "memory", "cc");
-
-	return increment;
-}
-
-#define atomic24_add(i, v) ((void)__atomic24_add((i), (v)))
-#define atomic24_sub(i, v) ((void)__atomic24_sub((i), (v)))
-
-#define atomic24_dec_return(v) __atomic24_sub(1, (v))
-#define atomic24_inc_return(v) __atomic24_add(1, (v))
-
-#define atomic24_sub_and_test(i, v) (__atomic24_sub((i), (v)) == 0)
-#define atomic24_dec_and_test(v) (__atomic24_sub(1, (v)) == 0)
-
-#define atomic24_inc(v) ((void)__atomic24_add(1, (v)))
-#define atomic24_dec(v) ((void)__atomic24_sub(1, (v)))
-
-#define atomic24_add_negative(i, v) (__atomic24_add((i), (v)) < 0)
-
-/* Atomic operations are already serializing */
-#define smp_mb__before_atomic_dec()	barrier()
-#define smp_mb__after_atomic_dec()	barrier()
-#define smp_mb__before_atomic_inc()	barrier()
-#define smp_mb__after_atomic_inc()	barrier()
-
-#endif /* !(__KERNEL__) */
-
-#include <asm-generic/atomic.h>
-#endif /* !(__ARCH_SPARC_ATOMIC__) */
diff --git a/include/asm-sparc/atomic_32.h b/include/asm-sparc/atomic_32.h
new file mode 100644
index 00000000000..5c944b5a804
--- /dev/null
+++ b/include/asm-sparc/atomic_32.h
@@ -0,0 +1,165 @@
+/* atomic.h: These still suck, but the I-cache hit rate is higher.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 2000 Anton Blanchard (anton@linuxcare.com.au)
+ * Copyright (C) 2007 Kyle McMartin (kyle@parisc-linux.org)
+ *
+ * Additions by Keith M Wesolowski (wesolows@foobazco.org) based
+ * on asm-parisc/atomic.h Copyright (C) 2000 Philipp Rumpf <prumpf@tux.org>.
+ */
+
+#ifndef __ARCH_SPARC_ATOMIC__
+#define __ARCH_SPARC_ATOMIC__
+
+#include <linux/types.h>
+
+typedef struct { volatile int counter; } atomic_t;
+
+#ifdef __KERNEL__
+
+#define ATOMIC_INIT(i)  { (i) }
+
+extern int __atomic_add_return(int, atomic_t *);
+extern int atomic_cmpxchg(atomic_t *, int, int);
+#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
+extern int atomic_add_unless(atomic_t *, int, int);
+extern void atomic_set(atomic_t *, int);
+
+#define atomic_read(v)          ((v)->counter)
+
+#define atomic_add(i, v)	((void)__atomic_add_return( (int)(i), (v)))
+#define atomic_sub(i, v)	((void)__atomic_add_return(-(int)(i), (v)))
+#define atomic_inc(v)		((void)__atomic_add_return(        1, (v)))
+#define atomic_dec(v)		((void)__atomic_add_return(       -1, (v)))
+
+#define atomic_add_return(i, v)	(__atomic_add_return( (int)(i), (v)))
+#define atomic_sub_return(i, v)	(__atomic_add_return(-(int)(i), (v)))
+#define atomic_inc_return(v)	(__atomic_add_return(        1, (v)))
+#define atomic_dec_return(v)	(__atomic_add_return(       -1, (v)))
+
+#define atomic_add_negative(a, v)	(atomic_add_return((a), (v)) < 0)
+
+/*
+ * atomic_inc_and_test - increment and test
+ * @v: pointer of type atomic_t
+ *
+ * Atomically increments @v by 1
+ * and returns true if the result is zero, or false for all
+ * other cases.
+ */
+#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0)
+
+#define atomic_dec_and_test(v) (atomic_dec_return(v) == 0)
+#define atomic_sub_and_test(i, v) (atomic_sub_return(i, v) == 0)
+
+#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
+
+/* This is the old 24-bit implementation.  It's still used internally
+ * by some sparc-specific code, notably the semaphore implementation.
+ */
+typedef struct { volatile int counter; } atomic24_t;
+
+#ifndef CONFIG_SMP
+
+#define ATOMIC24_INIT(i)  { (i) }
+#define atomic24_read(v)          ((v)->counter)
+#define atomic24_set(v, i)        (((v)->counter) = i)
+
+#else
+/* We do the bulk of the actual work out of line in two common
+ * routines in assembler, see arch/sparc/lib/atomic.S for the
+ * "fun" details.
+ *
+ * For SMP the trick is you embed the spin lock byte within
+ * the word, use the low byte so signedness is easily retained
+ * via a quick arithmetic shift.  It looks like this:
+ *
+ *	----------------------------------------
+ *	| signed 24-bit counter value |  lock  |  atomic_t
+ *	----------------------------------------
+ *	 31                          8 7      0
+ */
+
+#define ATOMIC24_INIT(i)	{ ((i) << 8) }
+
+static inline int atomic24_read(const atomic24_t *v)
+{
+	int ret = v->counter;
+
+	while(ret & 0xff)
+		ret = v->counter;
+
+	return ret >> 8;
+}
+
+#define atomic24_set(v, i)	(((v)->counter) = ((i) << 8))
+#endif
+
+static inline int __atomic24_add(int i, atomic24_t *v)
+{
+	register volatile int *ptr asm("g1");
+	register int increment asm("g2");
+	register int tmp1 asm("g3");
+	register int tmp2 asm("g4");
+	register int tmp3 asm("g7");
+
+	ptr = &v->counter;
+	increment = i;
+
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___atomic24_add\n\t"
+	" add	%%o7, 8, %%o7\n"
+	: "=&r" (increment), "=r" (tmp1), "=r" (tmp2), "=r" (tmp3)
+	: "0" (increment), "r" (ptr)
+	: "memory", "cc");
+
+	return increment;
+}
+
+static inline int __atomic24_sub(int i, atomic24_t *v)
+{
+	register volatile int *ptr asm("g1");
+	register int increment asm("g2");
+	register int tmp1 asm("g3");
+	register int tmp2 asm("g4");
+	register int tmp3 asm("g7");
+
+	ptr = &v->counter;
+	increment = i;
+
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___atomic24_sub\n\t"
+	" add	%%o7, 8, %%o7\n"
+	: "=&r" (increment), "=r" (tmp1), "=r" (tmp2), "=r" (tmp3)
+	: "0" (increment), "r" (ptr)
+	: "memory", "cc");
+
+	return increment;
+}
+
+#define atomic24_add(i, v) ((void)__atomic24_add((i), (v)))
+#define atomic24_sub(i, v) ((void)__atomic24_sub((i), (v)))
+
+#define atomic24_dec_return(v) __atomic24_sub(1, (v))
+#define atomic24_inc_return(v) __atomic24_add(1, (v))
+
+#define atomic24_sub_and_test(i, v) (__atomic24_sub((i), (v)) == 0)
+#define atomic24_dec_and_test(v) (__atomic24_sub(1, (v)) == 0)
+
+#define atomic24_inc(v) ((void)__atomic24_add(1, (v)))
+#define atomic24_dec(v) ((void)__atomic24_sub(1, (v)))
+
+#define atomic24_add_negative(i, v) (__atomic24_add((i), (v)) < 0)
+
+/* Atomic operations are already serializing */
+#define smp_mb__before_atomic_dec()	barrier()
+#define smp_mb__after_atomic_dec()	barrier()
+#define smp_mb__before_atomic_inc()	barrier()
+#define smp_mb__after_atomic_inc()	barrier()
+
+#endif /* !(__KERNEL__) */
+
+#include <asm-generic/atomic.h>
+#endif /* !(__ARCH_SPARC_ATOMIC__) */
diff --git a/include/asm-sparc/atomic_64.h b/include/asm-sparc/atomic_64.h
new file mode 100644
index 00000000000..2c71ec4a3b1
--- /dev/null
+++ b/include/asm-sparc/atomic_64.h
@@ -0,0 +1,128 @@
+/* atomic.h: Thankfully the V9 is at least reasonable for this
+ *           stuff.
+ *
+ * Copyright (C) 1996, 1997, 2000 David S. Miller (davem@redhat.com)
+ */
+
+#ifndef __ARCH_SPARC64_ATOMIC__
+#define __ARCH_SPARC64_ATOMIC__
+
+#include <linux/types.h>
+#include <asm/system.h>
+
+typedef struct { volatile int counter; } atomic_t;
+typedef struct { volatile __s64 counter; } atomic64_t;
+
+#define ATOMIC_INIT(i)		{ (i) }
+#define ATOMIC64_INIT(i)	{ (i) }
+
+#define atomic_read(v)		((v)->counter)
+#define atomic64_read(v)	((v)->counter)
+
+#define atomic_set(v, i)	(((v)->counter) = i)
+#define atomic64_set(v, i)	(((v)->counter) = i)
+
+extern void atomic_add(int, atomic_t *);
+extern void atomic64_add(int, atomic64_t *);
+extern void atomic_sub(int, atomic_t *);
+extern void atomic64_sub(int, atomic64_t *);
+
+extern int atomic_add_ret(int, atomic_t *);
+extern int atomic64_add_ret(int, atomic64_t *);
+extern int atomic_sub_ret(int, atomic_t *);
+extern int atomic64_sub_ret(int, atomic64_t *);
+
+#define atomic_dec_return(v) atomic_sub_ret(1, v)
+#define atomic64_dec_return(v) atomic64_sub_ret(1, v)
+
+#define atomic_inc_return(v) atomic_add_ret(1, v)
+#define atomic64_inc_return(v) atomic64_add_ret(1, v)
+
+#define atomic_sub_return(i, v) atomic_sub_ret(i, v)
+#define atomic64_sub_return(i, v) atomic64_sub_ret(i, v)
+
+#define atomic_add_return(i, v) atomic_add_ret(i, v)
+#define atomic64_add_return(i, v) atomic64_add_ret(i, v)
+
+/*
+ * atomic_inc_and_test - increment and test
+ * @v: pointer of type atomic_t
+ *
+ * Atomically increments @v by 1
+ * and returns true if the result is zero, or false for all
+ * other cases.
+ */
+#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0)
+#define atomic64_inc_and_test(v) (atomic64_inc_return(v) == 0)
+
+#define atomic_sub_and_test(i, v) (atomic_sub_ret(i, v) == 0)
+#define atomic64_sub_and_test(i, v) (atomic64_sub_ret(i, v) == 0)
+
+#define atomic_dec_and_test(v) (atomic_sub_ret(1, v) == 0)
+#define atomic64_dec_and_test(v) (atomic64_sub_ret(1, v) == 0)
+
+#define atomic_inc(v) atomic_add(1, v)
+#define atomic64_inc(v) atomic64_add(1, v)
+
+#define atomic_dec(v) atomic_sub(1, v)
+#define atomic64_dec(v) atomic64_sub(1, v)
+
+#define atomic_add_negative(i, v) (atomic_add_ret(i, v) < 0)
+#define atomic64_add_negative(i, v) (atomic64_add_ret(i, v) < 0)
+
+#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n)))
+#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
+
+static inline int atomic_add_unless(atomic_t *v, int a, int u)
+{
+	int c, old;
+	c = atomic_read(v);
+	for (;;) {
+		if (unlikely(c == (u)))
+			break;
+		old = atomic_cmpxchg((v), c, c + (a));
+		if (likely(old == c))
+			break;
+		c = old;
+	}
+	return c != (u);
+}
+
+#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
+
+#define atomic64_cmpxchg(v, o, n) \
+	((__typeof__((v)->counter))cmpxchg(&((v)->counter), (o), (n)))
+#define atomic64_xchg(v, new) (xchg(&((v)->counter), new))
+
+static inline int atomic64_add_unless(atomic64_t *v, long a, long u)
+{
+	long c, old;
+	c = atomic64_read(v);
+	for (;;) {
+		if (unlikely(c == (u)))
+			break;
+		old = atomic64_cmpxchg((v), c, c + (a));
+		if (likely(old == c))
+			break;
+		c = old;
+	}
+	return c != (u);
+}
+
+#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
+
+/* Atomic operations are already serializing */
+#ifdef CONFIG_SMP
+#define smp_mb__before_atomic_dec()	membar_storeload_loadload();
+#define smp_mb__after_atomic_dec()	membar_storeload_storestore();
+#define smp_mb__before_atomic_inc()	membar_storeload_loadload();
+#define smp_mb__after_atomic_inc()	membar_storeload_storestore();
+#else
+#define smp_mb__before_atomic_dec()	barrier()
+#define smp_mb__after_atomic_dec()	barrier()
+#define smp_mb__before_atomic_inc()	barrier()
+#define smp_mb__after_atomic_inc()	barrier()
+#endif
+
+#include <asm-generic/atomic.h>
+#endif /* !(__ARCH_SPARC64_ATOMIC__) */
diff --git a/include/asm-sparc/auxio.h b/include/asm-sparc/auxio.h
index e552b8d6845..24c6f3c0f57 100644
--- a/include/asm-sparc/auxio.h
+++ b/include/asm-sparc/auxio.h
@@ -1,89 +1,8 @@
-/*
- * auxio.h:  Definitions and code for the Auxiliary I/O register.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-#ifndef _SPARC_AUXIO_H
-#define _SPARC_AUXIO_H
-
-#include <asm/system.h>
-#include <asm/vaddrs.h>
-
-/* This register is an unsigned char in IO space.  It does two things.
- * First, it is used to control the front panel LED light on machines
- * that have it (good for testing entry points to trap handlers and irq's)
- * Secondly, it controls various floppy drive parameters.
- */
-#define AUXIO_ORMEIN      0xf0    /* All writes must set these bits. */
-#define AUXIO_ORMEIN4M    0xc0    /* sun4m - All writes must set these bits. */
-#define AUXIO_FLPY_DENS   0x20    /* Floppy density, high if set. Read only. */
-#define AUXIO_FLPY_DCHG   0x10    /* A disk change occurred.  Read only. */
-#define AUXIO_EDGE_ON     0x10    /* sun4m - On means Jumper block is in. */
-#define AUXIO_FLPY_DSEL   0x08    /* Drive select/start-motor. Write only. */
-#define AUXIO_LINK_TEST   0x08    /* sun4m - On means TPE Carrier detect. */
-
-/* Set the following to one, then zero, after doing a pseudo DMA transfer. */
-#define AUXIO_FLPY_TCNT   0x04    /* Floppy terminal count. Write only. */
-
-/* Set the following to zero to eject the floppy. */
-#define AUXIO_FLPY_EJCT   0x02    /* Eject floppy disk.  Write only. */
-#define AUXIO_LED         0x01    /* On if set, off if unset. Read/Write */
-
-#ifndef __ASSEMBLY__
-
-/* 
- * NOTE: these routines are implementation dependent-- 
- * understand the hardware you are querying! 
- */
-extern void set_auxio(unsigned char bits_on, unsigned char bits_off);
-extern unsigned char get_auxio(void); /* .../asm-sparc/floppy.h */
-
-/*
- * The following routines are provided for driver-compatibility
- * with sparc64 (primarily sunlance.c)
- */
-
-#define AUXIO_LTE_ON    1
-#define AUXIO_LTE_OFF   0
-
-/* auxio_set_lte - Set Link Test Enable (TPE Link Detect)
- *
- * on - AUXIO_LTE_ON or AUXIO_LTE_OFF
- */
-#define auxio_set_lte(on) \
-do { \
-	if(on) { \
-		set_auxio(AUXIO_LINK_TEST, 0); \
-	} else { \
-		set_auxio(0, AUXIO_LINK_TEST); \
-	} \
-} while (0)
-
-#define AUXIO_LED_ON    1
-#define AUXIO_LED_OFF   0
-
-/* auxio_set_led - Set system front panel LED
- *
- * on - AUXIO_LED_ON or AUXIO_LED_OFF
- */
-#define auxio_set_led(on) \
-do { \
-	if(on) { \
-		set_auxio(AUXIO_LED, 0); \
-	} else { \
-		set_auxio(0, AUXIO_LED); \
-	} \
-} while (0)
-
-#endif /* !(__ASSEMBLY__) */
-
-
-/* AUXIO2 (Power Off Control) */
-extern __volatile__ unsigned char * auxio_power_register;
-
-#define	AUXIO_POWER_DETECT_FAILURE	32
-#define	AUXIO_POWER_CLEAR_FAILURE	2
-#define	AUXIO_POWER_OFF			1
-
-
-#endif /* !(_SPARC_AUXIO_H) */
+#ifndef ___ASM_SPARC_AUXIO_H
+#define ___ASM_SPARC_AUXIO_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/auxio_64.h>
+#else
+#include <asm-sparc/auxio_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/auxio_32.h b/include/asm-sparc/auxio_32.h
new file mode 100644
index 00000000000..4db8f23db20
--- /dev/null
+++ b/include/asm-sparc/auxio_32.h
@@ -0,0 +1,89 @@
+/*
+ * auxio.h:  Definitions and code for the Auxiliary I/O register.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+#ifndef _SPARC_AUXIO_H
+#define _SPARC_AUXIO_H
+
+#include <asm/system.h>
+#include <asm/vaddrs.h>
+
+/* This register is an unsigned char in IO space.  It does two things.
+ * First, it is used to control the front panel LED light on machines
+ * that have it (good for testing entry points to trap handlers and irq's)
+ * Secondly, it controls various floppy drive parameters.
+ */
+#define AUXIO_ORMEIN      0xf0    /* All writes must set these bits. */
+#define AUXIO_ORMEIN4M    0xc0    /* sun4m - All writes must set these bits. */
+#define AUXIO_FLPY_DENS   0x20    /* Floppy density, high if set. Read only. */
+#define AUXIO_FLPY_DCHG   0x10    /* A disk change occurred.  Read only. */
+#define AUXIO_EDGE_ON     0x10    /* sun4m - On means Jumper block is in. */
+#define AUXIO_FLPY_DSEL   0x08    /* Drive select/start-motor. Write only. */
+#define AUXIO_LINK_TEST   0x08    /* sun4m - On means TPE Carrier detect. */
+
+/* Set the following to one, then zero, after doing a pseudo DMA transfer. */
+#define AUXIO_FLPY_TCNT   0x04    /* Floppy terminal count. Write only. */
+
+/* Set the following to zero to eject the floppy. */
+#define AUXIO_FLPY_EJCT   0x02    /* Eject floppy disk.  Write only. */
+#define AUXIO_LED         0x01    /* On if set, off if unset. Read/Write */
+
+#ifndef __ASSEMBLY__
+
+/*
+ * NOTE: these routines are implementation dependent--
+ * understand the hardware you are querying!
+ */
+extern void set_auxio(unsigned char bits_on, unsigned char bits_off);
+extern unsigned char get_auxio(void); /* .../asm-sparc/floppy.h */
+
+/*
+ * The following routines are provided for driver-compatibility
+ * with sparc64 (primarily sunlance.c)
+ */
+
+#define AUXIO_LTE_ON    1
+#define AUXIO_LTE_OFF   0
+
+/* auxio_set_lte - Set Link Test Enable (TPE Link Detect)
+ *
+ * on - AUXIO_LTE_ON or AUXIO_LTE_OFF
+ */
+#define auxio_set_lte(on) \
+do { \
+	if(on) { \
+		set_auxio(AUXIO_LINK_TEST, 0); \
+	} else { \
+		set_auxio(0, AUXIO_LINK_TEST); \
+	} \
+} while (0)
+
+#define AUXIO_LED_ON    1
+#define AUXIO_LED_OFF   0
+
+/* auxio_set_led - Set system front panel LED
+ *
+ * on - AUXIO_LED_ON or AUXIO_LED_OFF
+ */
+#define auxio_set_led(on) \
+do { \
+	if(on) { \
+		set_auxio(AUXIO_LED, 0); \
+	} else { \
+		set_auxio(0, AUXIO_LED); \
+	} \
+} while (0)
+
+#endif /* !(__ASSEMBLY__) */
+
+
+/* AUXIO2 (Power Off Control) */
+extern __volatile__ unsigned char * auxio_power_register;
+
+#define	AUXIO_POWER_DETECT_FAILURE	32
+#define	AUXIO_POWER_CLEAR_FAILURE	2
+#define	AUXIO_POWER_OFF			1
+
+
+#endif /* !(_SPARC_AUXIO_H) */
diff --git a/include/asm-sparc/auxio_64.h b/include/asm-sparc/auxio_64.h
new file mode 100644
index 00000000000..f61cd1e3e39
--- /dev/null
+++ b/include/asm-sparc/auxio_64.h
@@ -0,0 +1,100 @@
+/*
+ * auxio.h:  Definitions and code for the Auxiliary I/O registers.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ *
+ * Refactoring for unified NCR/PCIO support 2002 Eric Brower (ebrower@usa.net)
+ */
+#ifndef _SPARC64_AUXIO_H
+#define _SPARC64_AUXIO_H
+
+/* AUXIO implementations:
+ * sbus-based NCR89C105 "Slavio"
+ *	LED/Floppy (AUX1) register
+ *	Power (AUX2) register
+ *
+ * ebus-based auxio on PCIO
+ *	LED Auxio Register
+ *	Power Auxio Register
+ *
+ * Register definitions from NCR _NCR89C105 Chip Specification_
+ *
+ * SLAVIO AUX1 @ 0x1900000
+ * -------------------------------------------------
+ * | (R) | (R) |  D  | (R) |  E  |  M  |  T  |  L  |
+ * -------------------------------------------------
+ * (R) - bit 7:6,4 are reserved and should be masked in s/w
+ *  D  - Floppy Density Sense (1=high density) R/O
+ *  E  - Link Test Enable, directly reflected on AT&T 7213 LTE pin
+ *  M  - Monitor/Mouse Mux, directly reflected on MON_MSE_MUX pin
+ *  T  - Terminal Count: sends TC pulse to 82077 floppy controller
+ *  L  - System LED on front panel (0=off, 1=on)
+ */
+#define AUXIO_AUX1_MASK		0xc0 /* Mask bits 		*/
+#define AUXIO_AUX1_FDENS	0x20 /* Floppy Density Sense	*/
+#define AUXIO_AUX1_LTE 		0x08 /* Link Test Enable 	*/
+#define AUXIO_AUX1_MMUX		0x04 /* Monitor/Mouse Mux	*/
+#define AUXIO_AUX1_FTCNT	0x02 /* Terminal Count, 	*/
+#define AUXIO_AUX1_LED		0x01 /* System LED		*/
+
+/* SLAVIO AUX2 @ 0x1910000
+ * -------------------------------------------------
+ * | (R) | (R) |  D  | (R) | (R) | (R) |  C  |  F  |
+ * -------------------------------------------------
+ * (R) - bits 7:6,4:2 are reserved and should be masked in s/w
+ *  D  - Power Failure Detect (1=power fail)
+ *  C  - Clear Power Failure Detect Int (1=clear)
+ *  F  - Power Off (1=power off)
+ */
+#define AUXIO_AUX2_MASK		0xdc /* Mask Bits		*/
+#define AUXIO_AUX2_PFAILDET	0x20 /* Power Fail Detect	*/
+#define AUXIO_AUX2_PFAILCLR 	0x02 /* Clear Pwr Fail Det Intr	*/
+#define AUXIO_AUX2_PWR_OFF	0x01 /* Power Off		*/
+
+/* Register definitions from Sun Microsystems _PCIO_ p/n 802-7837
+ *
+ * PCIO LED Auxio @ 0x726000
+ * -------------------------------------------------
+ * |             31:1 Unused                 | LED |
+ * -------------------------------------------------
+ * Bits 31:1 unused
+ * LED - System LED on front panel (0=off, 1=on)
+ */
+#define AUXIO_PCIO_LED		0x01 /* System LED 		*/
+
+/* PCIO Power Auxio @ 0x724000
+ * -------------------------------------------------
+ * |             31:2 Unused           | CPO | SPO |
+ * -------------------------------------------------
+ * Bits 31:2 unused
+ * CPO - Courtesy Power Off (1=off)
+ * SPO - System Power Off   (1=off)
+ */
+#define AUXIO_PCIO_CPWR_OFF	0x02 /* Courtesy Power Off	*/
+#define AUXIO_PCIO_SPWR_OFF	0x01 /* System Power Off	*/
+
+#ifndef __ASSEMBLY__
+
+extern void __iomem *auxio_register;
+
+#define AUXIO_LTE_ON	1
+#define AUXIO_LTE_OFF	0
+
+/* auxio_set_lte - Set Link Test Enable (TPE Link Detect)
+ *
+ * on - AUXIO_LTE_ON or AUXIO_LTE_OFF
+ */
+extern void auxio_set_lte(int on);
+
+#define AUXIO_LED_ON	1
+#define AUXIO_LED_OFF	0
+
+/* auxio_set_led - Set system front panel LED
+ *
+ * on - AUXIO_LED_ON or AUXIO_LED_OFF
+ */
+extern void auxio_set_led(int on);
+
+#endif /* ifndef __ASSEMBLY__ */
+
+#endif /* !(_SPARC64_AUXIO_H) */
diff --git a/include/asm-sparc/backoff.h b/include/asm-sparc/backoff.h
new file mode 100644
index 00000000000..fa1fdf67e35
--- /dev/null
+++ b/include/asm-sparc/backoff.h
@@ -0,0 +1,31 @@
+#ifndef _SPARC64_BACKOFF_H
+#define _SPARC64_BACKOFF_H
+
+#define BACKOFF_LIMIT	(4 * 1024)
+
+#ifdef CONFIG_SMP
+
+#define BACKOFF_SETUP(reg)	\
+	mov	1, reg
+
+#define BACKOFF_SPIN(reg, tmp, label)	\
+	mov	reg, tmp; \
+88:	brnz,pt	tmp, 88b; \
+	 sub	tmp, 1, tmp; \
+	set	BACKOFF_LIMIT, tmp; \
+	cmp	reg, tmp; \
+	bg,pn	%xcc, label; \
+	 nop; \
+	ba,pt	%xcc, label; \
+	 sllx	reg, 1, reg;
+
+#else
+
+#define BACKOFF_SETUP(reg)
+#define BACKOFF_SPIN(reg, tmp, label) \
+	ba,pt	%xcc, label; \
+	 nop;
+
+#endif
+
+#endif /* _SPARC64_BACKOFF_H */
diff --git a/include/asm-sparc/bbc.h b/include/asm-sparc/bbc.h
new file mode 100644
index 00000000000..423a85800aa
--- /dev/null
+++ b/include/asm-sparc/bbc.h
@@ -0,0 +1,225 @@
+/*
+ * bbc.h: Defines for BootBus Controller found on UltraSPARC-III
+ *        systems.
+ *
+ * Copyright (C) 2000 David S. Miller (davem@redhat.com)
+ */
+
+#ifndef _SPARC64_BBC_H
+#define _SPARC64_BBC_H
+
+/* Register sizes are indicated by "B" (Byte, 1-byte),
+ * "H" (Half-word, 2 bytes), "W" (Word, 4 bytes) or
+ * "Q" (Quad, 8 bytes) inside brackets.
+ */
+
+#define BBC_AID		0x00	/* [B] Agent ID			*/
+#define BBC_DEVP	0x01	/* [B] Device Present		*/
+#define BBC_ARB		0x02	/* [B] Arbitration		*/
+#define BBC_QUIESCE	0x03	/* [B] Quiesce			*/
+#define BBC_WDACTION	0x04	/* [B] Watchdog Action		*/
+#define BBC_SPG		0x06	/* [B] Soft POR Gen		*/
+#define BBC_SXG		0x07	/* [B] Soft XIR Gen		*/
+#define BBC_PSRC	0x08	/* [W] POR Source		*/
+#define BBC_XSRC	0x0c	/* [B] XIR Source		*/
+#define BBC_CSC		0x0d	/* [B] Clock Synthesizers Control*/
+#define BBC_ES_CTRL	0x0e	/* [H] Energy Star Control	*/
+#define BBC_ES_ACT	0x10	/* [W] E* Assert Change Time	*/
+#define BBC_ES_DACT	0x14	/* [B] E* De-Assert Change Time	*/
+#define BBC_ES_DABT	0x15	/* [B] E* De-Assert Bypass Time	*/
+#define BBC_ES_ABT	0x16	/* [H] E* Assert Bypass Time	*/
+#define BBC_ES_PST	0x18	/* [W] E* PLL Settle Time	*/
+#define BBC_ES_FSL	0x1c	/* [W] E* Frequency Switch Latency*/
+#define BBC_EBUST	0x20	/* [Q] EBUS Timing		*/
+#define BBC_JTAG_CMD	0x28	/* [W] JTAG+ Command		*/
+#define BBC_JTAG_CTRL	0x2c	/* [B] JTAG+ Control		*/
+#define BBC_I2C_SEL	0x2d	/* [B] I2C Selection		*/
+#define BBC_I2C_0_S1	0x2e	/* [B] I2C ctrlr-0 reg S1	*/
+#define BBC_I2C_0_S0	0x2f	/* [B] I2C ctrlr-0 regs S0,S0',S2,S3*/
+#define BBC_I2C_1_S1	0x30	/* [B] I2C ctrlr-1 reg S1	*/
+#define BBC_I2C_1_S0	0x31	/* [B] I2C ctrlr-1 regs S0,S0',S2,S3*/
+#define BBC_KBD_BEEP	0x32	/* [B] Keyboard Beep		*/
+#define BBC_KBD_BCNT	0x34	/* [W] Keyboard Beep Counter	*/
+
+#define BBC_REGS_SIZE	0x40
+
+/* There is a 2K scratch ram area at offset 0x80000 but I doubt
+ * we will use it for anything.
+ */
+
+/* Agent ID register.  This register shows the Safari Agent ID
+ * for the processors.  The value returned depends upon which
+ * cpu is reading the register.
+ */
+#define BBC_AID_ID	0x07	/* Safari ID		*/
+#define BBC_AID_RESV	0xf8	/* Reserved		*/
+
+/* Device Present register.  One can determine which cpus are actually
+ * present in the machine by interrogating this register.
+ */
+#define BBC_DEVP_CPU0	0x01	/* Processor 0 present	*/
+#define BBC_DEVP_CPU1	0x02	/* Processor 1 present	*/
+#define BBC_DEVP_CPU2	0x04	/* Processor 2 present	*/
+#define BBC_DEVP_CPU3	0x08	/* Processor 3 present	*/
+#define BBC_DEVP_RESV	0xf0	/* Reserved		*/
+
+/* Arbitration register.  This register is used to block access to
+ * the BBC from a particular cpu.
+ */
+#define BBC_ARB_CPU0	0x01	/* Enable cpu 0 BBC arbitratrion */
+#define BBC_ARB_CPU1	0x02	/* Enable cpu 1 BBC arbitratrion */
+#define BBC_ARB_CPU2	0x04	/* Enable cpu 2 BBC arbitratrion */
+#define BBC_ARB_CPU3	0x08	/* Enable cpu 3 BBC arbitratrion */
+#define BBC_ARB_RESV	0xf0	/* Reserved			 */
+
+/* Quiesce register.  Bus and BBC segments for cpus can be disabled
+ * with this register, ie. for hot plugging.
+ */
+#define BBC_QUIESCE_S02	0x01	/* Quiesce Safari segment for cpu 0 and 2 */
+#define BBC_QUIESCE_S13	0x02	/* Quiesce Safari segment for cpu 1 and 3 */
+#define BBC_QUIESCE_B02	0x04	/* Quiesce BBC segment for cpu 0 and 2    */
+#define BBC_QUIESCE_B13	0x08	/* Quiesce BBC segment for cpu 1 and 3    */
+#define BBC_QUIESCE_FD0 0x10	/* Disable Fatal_Error[0] reporting	  */
+#define BBC_QUIESCE_FD1 0x20	/* Disable Fatal_Error[1] reporting	  */
+#define BBC_QUIESCE_FD2 0x40	/* Disable Fatal_Error[2] reporting	  */
+#define BBC_QUIESCE_FD3 0x80	/* Disable Fatal_Error[3] reporting	  */
+
+/* Watchdog Action register.  When the watchdog device timer expires
+ * a line is enabled to the BBC.  The action BBC takes when this line
+ * is asserted can be controlled by this regiser.
+ */
+#define BBC_WDACTION_RST  0x01	/* When set, watchdog causes system reset.
+				 * When clear, BBC ignores watchdog signal.
+				 */
+#define BBC_WDACTION_RESV 0xfe	/* Reserved */
+
+/* Soft_POR_GEN register.  The POR (Power On Reset) signal may be asserted
+ * for specific processors or all processors via this register.
+ */
+#define BBC_SPG_CPU0	0x01 /* Assert POR for processor 0	*/
+#define BBC_SPG_CPU1	0x02 /* Assert POR for processor 1	*/
+#define BBC_SPG_CPU2	0x04 /* Assert POR for processor 2	*/
+#define BBC_SPG_CPU3	0x08 /* Assert POR for processor 3	*/
+#define BBC_SPG_CPUALL	0x10 /* Reset all processors and reset
+			      * the entire system.
+			      */
+#define BBC_SPG_RESV	0xe0 /* Reserved			*/
+
+/* Soft_XIR_GEN register.  The XIR (eXternally Initiated Reset) signal
+ * may be asserted to specific processors via this register.
+ */
+#define BBC_SXG_CPU0	0x01 /* Assert XIR for processor 0	*/
+#define BBC_SXG_CPU1	0x02 /* Assert XIR for processor 1	*/
+#define BBC_SXG_CPU2	0x04 /* Assert XIR for processor 2	*/
+#define BBC_SXG_CPU3	0x08 /* Assert XIR for processor 3	*/
+#define BBC_SXG_RESV	0xf0 /* Reserved			*/
+
+/* POR Source register.  One may identify the cause of the most recent
+ * reset by reading this register.
+ */
+#define BBC_PSRC_SPG0	0x0001 /* CPU 0 reset via BBC_SPG register	*/
+#define BBC_PSRC_SPG1	0x0002 /* CPU 1 reset via BBC_SPG register	*/
+#define BBC_PSRC_SPG2	0x0004 /* CPU 2 reset via BBC_SPG register	*/
+#define BBC_PSRC_SPG3	0x0008 /* CPU 3 reset via BBC_SPG register	*/
+#define BBC_PSRC_SPGSYS	0x0010 /* System reset via BBC_SPG register	*/
+#define BBC_PSRC_JTAG	0x0020 /* System reset via JTAG+		*/
+#define BBC_PSRC_BUTTON	0x0040 /* System reset via push-button dongle	*/
+#define BBC_PSRC_PWRUP	0x0080 /* System reset via power-up		*/
+#define BBC_PSRC_FE0	0x0100 /* CPU 0 reported Fatal_Error		*/
+#define BBC_PSRC_FE1	0x0200 /* CPU 1 reported Fatal_Error		*/
+#define BBC_PSRC_FE2	0x0400 /* CPU 2 reported Fatal_Error		*/
+#define BBC_PSRC_FE3	0x0800 /* CPU 3 reported Fatal_Error		*/
+#define BBC_PSRC_FE4	0x1000 /* Schizo reported Fatal_Error		*/
+#define BBC_PSRC_FE5	0x2000 /* Safari device 5 reported Fatal_Error	*/
+#define BBC_PSRC_FE6	0x4000 /* CPMS reported Fatal_Error		*/
+#define BBC_PSRC_SYNTH	0x8000 /* System reset when on-board clock synthesizers
+				* were updated.
+				*/
+#define BBC_PSRC_WDT   0x10000 /* System reset via Super I/O watchdog	*/
+#define BBC_PSRC_RSC   0x20000 /* System reset via RSC remote monitoring
+				* device
+				*/
+
+/* XIR Source register.  The source of an XIR event sent to a processor may
+ * be determined via this register.
+ */
+#define BBC_XSRC_SXG0	0x01	/* CPU 0 received XIR via Soft_XIR_GEN reg */
+#define BBC_XSRC_SXG1	0x02	/* CPU 1 received XIR via Soft_XIR_GEN reg */
+#define BBC_XSRC_SXG2	0x04	/* CPU 2 received XIR via Soft_XIR_GEN reg */
+#define BBC_XSRC_SXG3	0x08	/* CPU 3 received XIR via Soft_XIR_GEN reg */
+#define BBC_XSRC_JTAG	0x10	/* All CPUs received XIR via JTAG+         */
+#define BBC_XSRC_W_OR_B	0x20	/* All CPUs received XIR either because:
+				 * a) Super I/O watchdog fired, or
+				 * b) XIR push button was activated
+				 */
+#define BBC_XSRC_RESV	0xc0	/* Reserved				   */
+
+/* Clock Synthesizers Control register.  This register provides the big-bang
+ * programming interface to the two clock synthesizers of the machine.
+ */
+#define BBC_CSC_SLOAD	0x01	/* Directly connected to S_LOAD pins	*/
+#define BBC_CSC_SDATA	0x02	/* Directly connected to S_DATA pins	*/
+#define BBC_CSC_SCLOCK	0x04	/* Directly connected to S_CLOCK pins	*/
+#define BBC_CSC_RESV	0x78	/* Reserved				*/
+#define BBC_CSC_RST	0x80	/* Generate system reset when S_LOAD==1	*/
+
+/* Energy Star Control register.  This register is used to generate the
+ * clock frequency change trigger to the main system devices (Schizo and
+ * the processors).  The transition occurs when bits in this register
+ * go from 0 to 1, only one bit must be set at once else no action
+ * occurs.  Basically the sequence of events is:
+ * a) Choose new frequency: full, 1/2 or 1/32
+ * b) Program this desired frequency into the cpus and Schizo.
+ * c) Set the same value in this register.
+ * d) 16 system clocks later, clear this register.
+ */
+#define BBC_ES_CTRL_1_1		0x01	/* Full frequency	*/
+#define BBC_ES_CTRL_1_2		0x02	/* 1/2 frequency	*/
+#define BBC_ES_CTRL_1_32	0x20	/* 1/32 frequency	*/
+#define BBC_ES_RESV		0xdc	/* Reserved		*/
+
+/* Energy Star Assert Change Time register.  This determines the number
+ * of BBC clock cycles (which is half the system frequency) between
+ * the detection of FREEZE_ACK being asserted and the assertion of
+ * the CLK_CHANGE_L[2:0] signals.
+ */
+#define BBC_ES_ACT_VAL	0xff
+
+/* Energy Star Assert Bypass Time register.  This determines the number
+ * of BBC clock cycles (which is half the system frequency) between
+ * the assertion of the CLK_CHANGE_L[2:0] signals and the assertion of
+ * the ESTAR_PLL_BYPASS signal.
+ */
+#define BBC_ES_ABT_VAL	0xffff
+
+/* Energy Star PLL Settle Time register.  This determines the number of
+ * BBC clock cycles (which is half the system frequency) between the
+ * de-assertion of CLK_CHANGE_L[2:0] and the de-assertion of the FREEZE_L
+ * signal.
+ */
+#define BBC_ES_PST_VAL	0xffffffff
+
+/* Energy Star Frequency Switch Latency register.  This is the number of
+ * BBC clocks between the de-assertion of CLK_CHANGE_L[2:0] and the first
+ * edge of the Safari clock at the new frequency.
+ */
+#define BBC_ES_FSL_VAL	0xffffffff
+
+/* Keyboard Beep control register.  This is a simple enabler for the audio
+ * beep sound.
+ */
+#define BBC_KBD_BEEP_ENABLE	0x01 /* Enable beep	*/
+#define BBC_KBD_BEEP_RESV	0xfe /* Reserved	*/
+
+/* Keyboard Beep Counter register.  There is a free-running counter inside
+ * the BBC which runs at half the system clock.  The bit set in this register
+ * determines when the audio sound is generated.  So for example if bit
+ * 10 is set, the audio beep will oscillate at 1/(2**12).  The keyboard beep
+ * generator automatically selects a different bit to use if the system clock
+ * is changed via Energy Star.
+ */
+#define BBC_KBD_BCNT_BITS	0x0007fc00
+#define BBC_KBC_BCNT_RESV	0xfff803ff
+
+#endif /* _SPARC64_BBC_H */
+
diff --git a/include/asm-sparc/bitops.h b/include/asm-sparc/bitops.h
index 68b98a7e645..1a2949d0193 100644
--- a/include/asm-sparc/bitops.h
+++ b/include/asm-sparc/bitops.h
@@ -1,111 +1,8 @@
-/*
- * bitops.h: Bit string operations on the Sparc.
- *
- * Copyright 1995 David S. Miller (davem@caip.rutgers.edu)
- * Copyright 1996 Eddie C. Dost   (ecd@skynet.be)
- * Copyright 2001 Anton Blanchard (anton@samba.org)
- */
-
-#ifndef _SPARC_BITOPS_H
-#define _SPARC_BITOPS_H
-
-#include <linux/compiler.h>
-#include <asm/byteorder.h>
-
-#ifdef __KERNEL__
-
-#ifndef _LINUX_BITOPS_H
-#error only <linux/bitops.h> can be included directly
+#ifndef ___ASM_SPARC_BITOPS_H
+#define ___ASM_SPARC_BITOPS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/bitops_64.h>
+#else
+#include <asm-sparc/bitops_32.h>
+#endif
 #endif
-
-extern unsigned long ___set_bit(unsigned long *addr, unsigned long mask);
-extern unsigned long ___clear_bit(unsigned long *addr, unsigned long mask);
-extern unsigned long ___change_bit(unsigned long *addr, unsigned long mask);
-
-/*
- * Set bit 'nr' in 32-bit quantity at address 'addr' where bit '0'
- * is in the highest of the four bytes and bit '31' is the high bit
- * within the first byte. Sparc is BIG-Endian. Unless noted otherwise
- * all bit-ops return 0 if bit was previously clear and != 0 otherwise.
- */
-static inline int test_and_set_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	return ___set_bit(ADDR, mask) != 0;
-}
-
-static inline void set_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	(void) ___set_bit(ADDR, mask);
-}
-
-static inline int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	return ___clear_bit(ADDR, mask) != 0;
-}
-
-static inline void clear_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	(void) ___clear_bit(ADDR, mask);
-}
-
-static inline int test_and_change_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	return ___change_bit(ADDR, mask) != 0;
-}
-
-static inline void change_bit(unsigned long nr, volatile unsigned long *addr)
-{
-	unsigned long *ADDR, mask;
-
-	ADDR = ((unsigned long *) addr) + (nr >> 5);
-	mask = 1 << (nr & 31);
-
-	(void) ___change_bit(ADDR, mask);
-}
-
-#include <asm-generic/bitops/non-atomic.h>
-
-#define smp_mb__before_clear_bit()	do { } while(0)
-#define smp_mb__after_clear_bit()	do { } while(0)
-
-#include <asm-generic/bitops/ffz.h>
-#include <asm-generic/bitops/__ffs.h>
-#include <asm-generic/bitops/sched.h>
-#include <asm-generic/bitops/ffs.h>
-#include <asm-generic/bitops/fls.h>
-#include <asm-generic/bitops/fls64.h>
-#include <asm-generic/bitops/hweight.h>
-#include <asm-generic/bitops/lock.h>
-#include <asm-generic/bitops/find.h>
-#include <asm-generic/bitops/ext2-non-atomic.h>
-#include <asm-generic/bitops/ext2-atomic.h>
-#include <asm-generic/bitops/minix.h>
-
-#endif /* __KERNEL__ */
-
-#endif /* defined(_SPARC_BITOPS_H) */
diff --git a/include/asm-sparc/bitops_32.h b/include/asm-sparc/bitops_32.h
new file mode 100644
index 00000000000..68b98a7e645
--- /dev/null
+++ b/include/asm-sparc/bitops_32.h
@@ -0,0 +1,111 @@
+/*
+ * bitops.h: Bit string operations on the Sparc.
+ *
+ * Copyright 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright 1996 Eddie C. Dost   (ecd@skynet.be)
+ * Copyright 2001 Anton Blanchard (anton@samba.org)
+ */
+
+#ifndef _SPARC_BITOPS_H
+#define _SPARC_BITOPS_H
+
+#include <linux/compiler.h>
+#include <asm/byteorder.h>
+
+#ifdef __KERNEL__
+
+#ifndef _LINUX_BITOPS_H
+#error only <linux/bitops.h> can be included directly
+#endif
+
+extern unsigned long ___set_bit(unsigned long *addr, unsigned long mask);
+extern unsigned long ___clear_bit(unsigned long *addr, unsigned long mask);
+extern unsigned long ___change_bit(unsigned long *addr, unsigned long mask);
+
+/*
+ * Set bit 'nr' in 32-bit quantity at address 'addr' where bit '0'
+ * is in the highest of the four bytes and bit '31' is the high bit
+ * within the first byte. Sparc is BIG-Endian. Unless noted otherwise
+ * all bit-ops return 0 if bit was previously clear and != 0 otherwise.
+ */
+static inline int test_and_set_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	return ___set_bit(ADDR, mask) != 0;
+}
+
+static inline void set_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	(void) ___set_bit(ADDR, mask);
+}
+
+static inline int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	return ___clear_bit(ADDR, mask) != 0;
+}
+
+static inline void clear_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	(void) ___clear_bit(ADDR, mask);
+}
+
+static inline int test_and_change_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	return ___change_bit(ADDR, mask) != 0;
+}
+
+static inline void change_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	unsigned long *ADDR, mask;
+
+	ADDR = ((unsigned long *) addr) + (nr >> 5);
+	mask = 1 << (nr & 31);
+
+	(void) ___change_bit(ADDR, mask);
+}
+
+#include <asm-generic/bitops/non-atomic.h>
+
+#define smp_mb__before_clear_bit()	do { } while(0)
+#define smp_mb__after_clear_bit()	do { } while(0)
+
+#include <asm-generic/bitops/ffz.h>
+#include <asm-generic/bitops/__ffs.h>
+#include <asm-generic/bitops/sched.h>
+#include <asm-generic/bitops/ffs.h>
+#include <asm-generic/bitops/fls.h>
+#include <asm-generic/bitops/fls64.h>
+#include <asm-generic/bitops/hweight.h>
+#include <asm-generic/bitops/lock.h>
+#include <asm-generic/bitops/find.h>
+#include <asm-generic/bitops/ext2-non-atomic.h>
+#include <asm-generic/bitops/ext2-atomic.h>
+#include <asm-generic/bitops/minix.h>
+
+#endif /* __KERNEL__ */
+
+#endif /* defined(_SPARC_BITOPS_H) */
diff --git a/include/asm-sparc/bitops_64.h b/include/asm-sparc/bitops_64.h
new file mode 100644
index 00000000000..bb87b808022
--- /dev/null
+++ b/include/asm-sparc/bitops_64.h
@@ -0,0 +1,107 @@
+/*
+ * bitops.h: Bit string operations on the V9.
+ *
+ * Copyright 1996, 1997 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _SPARC64_BITOPS_H
+#define _SPARC64_BITOPS_H
+
+#ifndef _LINUX_BITOPS_H
+#error only <linux/bitops.h> can be included directly
+#endif
+
+#include <linux/compiler.h>
+#include <asm/byteorder.h>
+
+extern int test_and_set_bit(unsigned long nr, volatile unsigned long *addr);
+extern int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr);
+extern int test_and_change_bit(unsigned long nr, volatile unsigned long *addr);
+extern void set_bit(unsigned long nr, volatile unsigned long *addr);
+extern void clear_bit(unsigned long nr, volatile unsigned long *addr);
+extern void change_bit(unsigned long nr, volatile unsigned long *addr);
+
+#include <asm-generic/bitops/non-atomic.h>
+
+#ifdef CONFIG_SMP
+#define smp_mb__before_clear_bit()	membar_storeload_loadload()
+#define smp_mb__after_clear_bit()	membar_storeload_storestore()
+#else
+#define smp_mb__before_clear_bit()	barrier()
+#define smp_mb__after_clear_bit()	barrier()
+#endif
+
+#include <asm-generic/bitops/ffz.h>
+#include <asm-generic/bitops/__ffs.h>
+#include <asm-generic/bitops/fls.h>
+#include <asm-generic/bitops/__fls.h>
+#include <asm-generic/bitops/fls64.h>
+
+#ifdef __KERNEL__
+
+#include <asm-generic/bitops/sched.h>
+#include <asm-generic/bitops/ffs.h>
+
+/*
+ * hweightN: returns the hamming weight (i.e. the number
+ * of bits set) of a N-bit word
+ */
+
+#ifdef ULTRA_HAS_POPULATION_COUNT
+
+static inline unsigned int hweight64(unsigned long w)
+{
+	unsigned int res;
+
+	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w));
+	return res;
+}
+
+static inline unsigned int hweight32(unsigned int w)
+{
+	unsigned int res;
+
+	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xffffffff));
+	return res;
+}
+
+static inline unsigned int hweight16(unsigned int w)
+{
+	unsigned int res;
+
+	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xffff));
+	return res;
+}
+
+static inline unsigned int hweight8(unsigned int w)
+{
+	unsigned int res;
+
+	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xff));
+	return res;
+}
+
+#else
+
+#include <asm-generic/bitops/hweight.h>
+
+#endif
+#include <asm-generic/bitops/lock.h>
+#endif /* __KERNEL__ */
+
+#include <asm-generic/bitops/find.h>
+
+#ifdef __KERNEL__
+
+#include <asm-generic/bitops/ext2-non-atomic.h>
+
+#define ext2_set_bit_atomic(lock,nr,addr) \
+	test_and_set_bit((nr) ^ 0x38,(unsigned long *)(addr))
+#define ext2_clear_bit_atomic(lock,nr,addr) \
+	test_and_clear_bit((nr) ^ 0x38,(unsigned long *)(addr))
+
+#include <asm-generic/bitops/minix.h>
+
+#endif /* __KERNEL__ */
+
+#endif /* defined(_SPARC64_BITOPS_H) */
diff --git a/include/asm-sparc/cacheflush.h b/include/asm-sparc/cacheflush.h
index 68ac1091027..2b6a37957c2 100644
--- a/include/asm-sparc/cacheflush.h
+++ b/include/asm-sparc/cacheflush.h
@@ -1,85 +1,8 @@
-#ifndef _SPARC_CACHEFLUSH_H
-#define _SPARC_CACHEFLUSH_H
-
-#include <linux/mm.h>		/* Common for other includes */
-// #include <linux/kernel.h> from pgalloc.h
-// #include <linux/sched.h>  from pgalloc.h
-
-// #include <asm/page.h>
-#include <asm/btfixup.h>
-
-/*
- * Fine grained cache flushing.
- */
-#ifdef CONFIG_SMP
-
-BTFIXUPDEF_CALL(void, local_flush_cache_all, void)
-BTFIXUPDEF_CALL(void, local_flush_cache_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, local_flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_cache_page, struct vm_area_struct *, unsigned long)
-
-#define local_flush_cache_all() BTFIXUP_CALL(local_flush_cache_all)()
-#define local_flush_cache_mm(mm) BTFIXUP_CALL(local_flush_cache_mm)(mm)
-#define local_flush_cache_range(vma,start,end) BTFIXUP_CALL(local_flush_cache_range)(vma,start,end)
-#define local_flush_cache_page(vma,addr) BTFIXUP_CALL(local_flush_cache_page)(vma,addr)
-
-BTFIXUPDEF_CALL(void, local_flush_page_to_ram, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_sig_insns, struct mm_struct *, unsigned long)
-
-#define local_flush_page_to_ram(addr) BTFIXUP_CALL(local_flush_page_to_ram)(addr)
-#define local_flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(local_flush_sig_insns)(mm,insn_addr)
-
-extern void smp_flush_cache_all(void);
-extern void smp_flush_cache_mm(struct mm_struct *mm);
-extern void smp_flush_cache_range(struct vm_area_struct *vma,
-				  unsigned long start,
-				  unsigned long end);
-extern void smp_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
-
-extern void smp_flush_page_to_ram(unsigned long page);
-extern void smp_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
-
-#endif /* CONFIG_SMP */
-
-BTFIXUPDEF_CALL(void, flush_cache_all, void)
-BTFIXUPDEF_CALL(void, flush_cache_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, flush_cache_page, struct vm_area_struct *, unsigned long)
-
-#define flush_cache_all() BTFIXUP_CALL(flush_cache_all)()
-#define flush_cache_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
-#define flush_cache_dup_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
-#define flush_cache_range(vma,start,end) BTFIXUP_CALL(flush_cache_range)(vma,start,end)
-#define flush_cache_page(vma,addr,pfn) BTFIXUP_CALL(flush_cache_page)(vma,addr)
-#define flush_icache_range(start, end)		do { } while (0)
-#define flush_icache_page(vma, pg)		do { } while (0)
-
-#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
-
-#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
-	do {							\
-		flush_cache_page(vma, vaddr, page_to_pfn(page));\
-		memcpy(dst, src, len);				\
-	} while (0)
-#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
-	do {							\
-		flush_cache_page(vma, vaddr, page_to_pfn(page));\
-		memcpy(dst, src, len);				\
-	} while (0)
-
-BTFIXUPDEF_CALL(void, __flush_page_to_ram, unsigned long)
-BTFIXUPDEF_CALL(void, flush_sig_insns, struct mm_struct *, unsigned long)
-
-#define __flush_page_to_ram(addr) BTFIXUP_CALL(__flush_page_to_ram)(addr)
-#define flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(flush_sig_insns)(mm,insn_addr)
-
-extern void sparc_flush_page_to_ram(struct page *page);
-
-#define flush_dcache_page(page)			sparc_flush_page_to_ram(page)
-#define flush_dcache_mmap_lock(mapping)		do { } while (0)
-#define flush_dcache_mmap_unlock(mapping)	do { } while (0)
-
-#define flush_cache_vmap(start, end)		flush_cache_all()
-#define flush_cache_vunmap(start, end)		flush_cache_all()
-
-#endif /* _SPARC_CACHEFLUSH_H */
+#ifndef ___ASM_SPARC_CACHEFLUSH_H
+#define ___ASM_SPARC_CACHEFLUSH_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/cacheflush_64.h>
+#else
+#include <asm-sparc/cacheflush_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/cacheflush_32.h b/include/asm-sparc/cacheflush_32.h
new file mode 100644
index 00000000000..68ac1091027
--- /dev/null
+++ b/include/asm-sparc/cacheflush_32.h
@@ -0,0 +1,85 @@
+#ifndef _SPARC_CACHEFLUSH_H
+#define _SPARC_CACHEFLUSH_H
+
+#include <linux/mm.h>		/* Common for other includes */
+// #include <linux/kernel.h> from pgalloc.h
+// #include <linux/sched.h>  from pgalloc.h
+
+// #include <asm/page.h>
+#include <asm/btfixup.h>
+
+/*
+ * Fine grained cache flushing.
+ */
+#ifdef CONFIG_SMP
+
+BTFIXUPDEF_CALL(void, local_flush_cache_all, void)
+BTFIXUPDEF_CALL(void, local_flush_cache_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, local_flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_cache_page, struct vm_area_struct *, unsigned long)
+
+#define local_flush_cache_all() BTFIXUP_CALL(local_flush_cache_all)()
+#define local_flush_cache_mm(mm) BTFIXUP_CALL(local_flush_cache_mm)(mm)
+#define local_flush_cache_range(vma,start,end) BTFIXUP_CALL(local_flush_cache_range)(vma,start,end)
+#define local_flush_cache_page(vma,addr) BTFIXUP_CALL(local_flush_cache_page)(vma,addr)
+
+BTFIXUPDEF_CALL(void, local_flush_page_to_ram, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_sig_insns, struct mm_struct *, unsigned long)
+
+#define local_flush_page_to_ram(addr) BTFIXUP_CALL(local_flush_page_to_ram)(addr)
+#define local_flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(local_flush_sig_insns)(mm,insn_addr)
+
+extern void smp_flush_cache_all(void);
+extern void smp_flush_cache_mm(struct mm_struct *mm);
+extern void smp_flush_cache_range(struct vm_area_struct *vma,
+				  unsigned long start,
+				  unsigned long end);
+extern void smp_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
+
+extern void smp_flush_page_to_ram(unsigned long page);
+extern void smp_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
+
+#endif /* CONFIG_SMP */
+
+BTFIXUPDEF_CALL(void, flush_cache_all, void)
+BTFIXUPDEF_CALL(void, flush_cache_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, flush_cache_page, struct vm_area_struct *, unsigned long)
+
+#define flush_cache_all() BTFIXUP_CALL(flush_cache_all)()
+#define flush_cache_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
+#define flush_cache_dup_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
+#define flush_cache_range(vma,start,end) BTFIXUP_CALL(flush_cache_range)(vma,start,end)
+#define flush_cache_page(vma,addr,pfn) BTFIXUP_CALL(flush_cache_page)(vma,addr)
+#define flush_icache_range(start, end)		do { } while (0)
+#define flush_icache_page(vma, pg)		do { } while (0)
+
+#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
+
+#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
+	do {							\
+		flush_cache_page(vma, vaddr, page_to_pfn(page));\
+		memcpy(dst, src, len);				\
+	} while (0)
+#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
+	do {							\
+		flush_cache_page(vma, vaddr, page_to_pfn(page));\
+		memcpy(dst, src, len);				\
+	} while (0)
+
+BTFIXUPDEF_CALL(void, __flush_page_to_ram, unsigned long)
+BTFIXUPDEF_CALL(void, flush_sig_insns, struct mm_struct *, unsigned long)
+
+#define __flush_page_to_ram(addr) BTFIXUP_CALL(__flush_page_to_ram)(addr)
+#define flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(flush_sig_insns)(mm,insn_addr)
+
+extern void sparc_flush_page_to_ram(struct page *page);
+
+#define flush_dcache_page(page)			sparc_flush_page_to_ram(page)
+#define flush_dcache_mmap_lock(mapping)		do { } while (0)
+#define flush_dcache_mmap_unlock(mapping)	do { } while (0)
+
+#define flush_cache_vmap(start, end)		flush_cache_all()
+#define flush_cache_vunmap(start, end)		flush_cache_all()
+
+#endif /* _SPARC_CACHEFLUSH_H */
diff --git a/include/asm-sparc/cacheflush_64.h b/include/asm-sparc/cacheflush_64.h
new file mode 100644
index 00000000000..c43321729b3
--- /dev/null
+++ b/include/asm-sparc/cacheflush_64.h
@@ -0,0 +1,76 @@
+#ifndef _SPARC64_CACHEFLUSH_H
+#define _SPARC64_CACHEFLUSH_H
+
+#include <asm/page.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/mm.h>
+
+/* Cache flush operations. */
+
+/* These are the same regardless of whether this is an SMP kernel or not. */
+#define flush_cache_mm(__mm) \
+	do { if ((__mm) == current->mm) flushw_user(); } while(0)
+#define flush_cache_dup_mm(mm) flush_cache_mm(mm)
+#define flush_cache_range(vma, start, end) \
+	flush_cache_mm((vma)->vm_mm)
+#define flush_cache_page(vma, page, pfn) \
+	flush_cache_mm((vma)->vm_mm)
+
+/*
+ * On spitfire, the icache doesn't snoop local stores and we don't
+ * use block commit stores (which invalidate icache lines) during
+ * module load, so we need this.
+ */
+extern void flush_icache_range(unsigned long start, unsigned long end);
+extern void __flush_icache_page(unsigned long);
+
+extern void __flush_dcache_page(void *addr, int flush_icache);
+extern void flush_dcache_page_impl(struct page *page);
+#ifdef CONFIG_SMP
+extern void smp_flush_dcache_page_impl(struct page *page, int cpu);
+extern void flush_dcache_page_all(struct mm_struct *mm, struct page *page);
+#else
+#define smp_flush_dcache_page_impl(page,cpu) flush_dcache_page_impl(page)
+#define flush_dcache_page_all(mm,page) flush_dcache_page_impl(page)
+#endif
+
+extern void __flush_dcache_range(unsigned long start, unsigned long end);
+extern void flush_dcache_page(struct page *page);
+
+#define flush_icache_page(vma, pg)	do { } while(0)
+#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
+
+extern void flush_ptrace_access(struct vm_area_struct *, struct page *,
+				unsigned long uaddr, void *kaddr,
+				unsigned long len, int write);
+
+#define copy_to_user_page(vma, page, vaddr, dst, src, len)		\
+	do {								\
+		flush_cache_page(vma, vaddr, page_to_pfn(page));	\
+		memcpy(dst, src, len);					\
+		flush_ptrace_access(vma, page, vaddr, src, len, 0);	\
+	} while (0)
+
+#define copy_from_user_page(vma, page, vaddr, dst, src, len) 		\
+	do {								\
+		flush_cache_page(vma, vaddr, page_to_pfn(page));	\
+		memcpy(dst, src, len);					\
+		flush_ptrace_access(vma, page, vaddr, dst, len, 1);	\
+	} while (0)
+
+#define flush_dcache_mmap_lock(mapping)		do { } while (0)
+#define flush_dcache_mmap_unlock(mapping)	do { } while (0)
+
+#define flush_cache_vmap(start, end)		do { } while (0)
+#define flush_cache_vunmap(start, end)		do { } while (0)
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+/* internal debugging function */
+void kernel_map_pages(struct page *page, int numpages, int enable);
+#endif
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* _SPARC64_CACHEFLUSH_H */
diff --git a/include/asm-sparc/chafsr.h b/include/asm-sparc/chafsr.h
new file mode 100644
index 00000000000..85c69b38220
--- /dev/null
+++ b/include/asm-sparc/chafsr.h
@@ -0,0 +1,241 @@
+#ifndef _SPARC64_CHAFSR_H
+#define _SPARC64_CHAFSR_H
+
+/* Cheetah Asynchronous Fault Status register, ASI=0x4C VA<63:0>=0x0 */
+
+/* Comments indicate which processor variants on which the bit definition
+ * is valid.  Codes are:
+ * ch	-->	cheetah
+ * ch+	-->	cheetah plus
+ * jp	-->	jalapeno
+ */
+
+/* All bits of this register except M_SYNDROME and E_SYNDROME are
+ * read, write 1 to clear.  M_SYNDROME and E_SYNDROME are read-only.
+ */
+
+/* Software bit set by linux trap handlers to indicate that the trap was
+ * signalled at %tl >= 1.
+ */
+#define CHAFSR_TL1		(1UL << 63UL) /* n/a */
+
+/* Unmapped error from system bus for prefetch queue or
+ * store queue read operation
+ */
+#define CHPAFSR_DTO		(1UL << 59UL) /* ch+ */
+
+/* Bus error from system bus for prefetch queue or store queue
+ * read operation
+ */
+#define CHPAFSR_DBERR		(1UL << 58UL) /* ch+ */
+
+/* Hardware corrected E-cache Tag ECC error */
+#define CHPAFSR_THCE		(1UL << 57UL) /* ch+ */
+/* System interface protocol error, hw timeout caused */
+#define JPAFSR_JETO		(1UL << 57UL) /* jp */
+
+/* SW handled correctable E-cache Tag ECC error */
+#define CHPAFSR_TSCE		(1UL << 56UL) /* ch+ */
+/* Parity error on system snoop results */
+#define JPAFSR_SCE		(1UL << 56UL) /* jp */
+
+/* Uncorrectable E-cache Tag ECC error */
+#define CHPAFSR_TUE		(1UL << 55UL) /* ch+ */
+/* System interface protocol error, illegal command detected */
+#define JPAFSR_JEIC		(1UL << 55UL) /* jp */
+
+/* Uncorrectable system bus data ECC error due to prefetch
+ * or store fill request
+ */
+#define CHPAFSR_DUE		(1UL << 54UL) /* ch+ */
+/* System interface protocol error, illegal ADTYPE detected */
+#define JPAFSR_JEIT		(1UL << 54UL) /* jp */
+
+/* Multiple errors of the same type have occurred.  This bit is set when
+ * an uncorrectable error or a SW correctable error occurs and the status
+ * bit to report that error is already set.  When multiple errors of
+ * different types are indicated by setting multiple status bits.
+ *
+ * This bit is not set if multiple HW corrected errors with the same
+ * status bit occur, only uncorrectable and SW correctable ones have
+ * this behavior.
+ *
+ * This bit is not set when multiple ECC errors happen within a single
+ * 64-byte system bus transaction.  Only the first ECC error in a 16-byte
+ * subunit will be logged.  All errors in subsequent 16-byte subunits
+ * from the same 64-byte transaction are ignored.
+ */
+#define CHAFSR_ME		(1UL << 53UL) /* ch,ch+,jp */
+
+/* Privileged state error has occurred.  This is a capture of PSTATE.PRIV
+ * at the time the error is detected.
+ */
+#define CHAFSR_PRIV		(1UL << 52UL) /* ch,ch+,jp */
+
+/* The following bits 51 (CHAFSR_PERR) to 33 (CHAFSR_CE) are sticky error
+ * bits and record the most recently detected errors.  Bits accumulate
+ * errors that have been detected since the last write to clear the bit.
+ */
+
+/* System interface protocol error.  The processor asserts its' ERROR
+ * pin when this event occurs and it also logs a specific cause code
+ * into a JTAG scannable flop.
+ */
+#define CHAFSR_PERR		(1UL << 51UL) /* ch,ch+,jp */
+
+/* Internal processor error.  The processor asserts its' ERROR
+ * pin when this event occurs and it also logs a specific cause code
+ * into a JTAG scannable flop.
+ */
+#define CHAFSR_IERR		(1UL << 50UL) /* ch,ch+,jp */
+
+/* System request parity error on incoming address */
+#define CHAFSR_ISAP		(1UL << 49UL) /* ch,ch+,jp */
+
+/* HW Corrected system bus MTAG ECC error */
+#define CHAFSR_EMC		(1UL << 48UL) /* ch,ch+ */
+/* Parity error on L2 cache tag SRAM */
+#define JPAFSR_ETP		(1UL << 48UL) /* jp */
+
+/* Uncorrectable system bus MTAG ECC error */
+#define CHAFSR_EMU		(1UL << 47UL) /* ch,ch+ */
+/* Out of range memory error has occurred */
+#define JPAFSR_OM		(1UL << 47UL) /* jp */
+
+/* HW Corrected system bus data ECC error for read of interrupt vector */
+#define CHAFSR_IVC		(1UL << 46UL) /* ch,ch+ */
+/* Error due to unsupported store */
+#define JPAFSR_UMS		(1UL << 46UL) /* jp */
+
+/* Uncorrectable system bus data ECC error for read of interrupt vector */
+#define CHAFSR_IVU		(1UL << 45UL) /* ch,ch+,jp */
+
+/* Unmapped error from system bus */
+#define CHAFSR_TO		(1UL << 44UL) /* ch,ch+,jp */
+
+/* Bus error response from system bus */
+#define CHAFSR_BERR		(1UL << 43UL) /* ch,ch+,jp */
+
+/* SW Correctable E-cache ECC error for instruction fetch or data access
+ * other than block load.
+ */
+#define CHAFSR_UCC		(1UL << 42UL) /* ch,ch+,jp */
+
+/* Uncorrectable E-cache ECC error for instruction fetch or data access
+ * other than block load.
+ */
+#define CHAFSR_UCU		(1UL << 41UL) /* ch,ch+,jp */
+
+/* Copyout HW Corrected ECC error */
+#define CHAFSR_CPC		(1UL << 40UL) /* ch,ch+,jp */
+
+/* Copyout Uncorrectable ECC error */
+#define CHAFSR_CPU		(1UL << 39UL) /* ch,ch+,jp */
+
+/* HW Corrected ECC error from E-cache for writeback */
+#define CHAFSR_WDC		(1UL << 38UL) /* ch,ch+,jp */
+
+/* Uncorrectable ECC error from E-cache for writeback */
+#define CHAFSR_WDU		(1UL << 37UL) /* ch,ch+,jp */
+
+/* HW Corrected ECC error from E-cache for store merge or block load */
+#define CHAFSR_EDC		(1UL << 36UL) /* ch,ch+,jp */
+
+/* Uncorrectable ECC error from E-cache for store merge or block load */
+#define CHAFSR_EDU		(1UL << 35UL) /* ch,ch+,jp */
+
+/* Uncorrectable system bus data ECC error for read of memory or I/O */
+#define CHAFSR_UE		(1UL << 34UL) /* ch,ch+,jp */
+
+/* HW Corrected system bus data ECC error for read of memory or I/O */
+#define CHAFSR_CE		(1UL << 33UL) /* ch,ch+,jp */
+
+/* Uncorrectable ECC error from remote cache/memory */
+#define JPAFSR_RUE		(1UL << 32UL) /* jp */
+
+/* Correctable ECC error from remote cache/memory */
+#define JPAFSR_RCE		(1UL << 31UL) /* jp */
+
+/* JBUS parity error on returned read data */
+#define JPAFSR_BP		(1UL << 30UL) /* jp */
+
+/* JBUS parity error on data for writeback or block store */
+#define JPAFSR_WBP		(1UL << 29UL) /* jp */
+
+/* Foreign read to DRAM incurring correctable ECC error */
+#define JPAFSR_FRC		(1UL << 28UL) /* jp */
+
+/* Foreign read to DRAM incurring uncorrectable ECC error */
+#define JPAFSR_FRU		(1UL << 27UL) /* jp */
+
+#define CHAFSR_ERRORS		(CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP | CHAFSR_EMC | \
+				 CHAFSR_EMU | CHAFSR_IVC | CHAFSR_IVU | CHAFSR_TO | \
+				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | CHAFSR_CPC | \
+				 CHAFSR_CPU | CHAFSR_WDC | CHAFSR_WDU | CHAFSR_EDC | \
+				 CHAFSR_EDU | CHAFSR_UE | CHAFSR_CE)
+#define CHPAFSR_ERRORS		(CHPAFSR_DTO | CHPAFSR_DBERR | CHPAFSR_THCE | \
+				 CHPAFSR_TSCE | CHPAFSR_TUE | CHPAFSR_DUE | \
+				 CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP | CHAFSR_EMC | \
+				 CHAFSR_EMU | CHAFSR_IVC | CHAFSR_IVU | CHAFSR_TO | \
+				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | CHAFSR_CPC | \
+				 CHAFSR_CPU | CHAFSR_WDC | CHAFSR_WDU | CHAFSR_EDC | \
+				 CHAFSR_EDU | CHAFSR_UE | CHAFSR_CE)
+#define JPAFSR_ERRORS		(JPAFSR_JETO | JPAFSR_SCE | JPAFSR_JEIC | \
+				 JPAFSR_JEIT | CHAFSR_PERR | CHAFSR_IERR | \
+				 CHAFSR_ISAP | JPAFSR_ETP | JPAFSR_OM | \
+				 JPAFSR_UMS | CHAFSR_IVU | CHAFSR_TO | \
+				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | \
+				 CHAFSR_CPC | CHAFSR_CPU | CHAFSR_WDC | \
+				 CHAFSR_WDU | CHAFSR_EDC | CHAFSR_EDU | \
+				 CHAFSR_UE | CHAFSR_CE | JPAFSR_RUE | \
+				 JPAFSR_RCE | JPAFSR_BP | JPAFSR_WBP | \
+				 JPAFSR_FRC | JPAFSR_FRU)
+
+/* Active JBUS request signal when error occurred */
+#define JPAFSR_JBREQ		(0x7UL << 24UL) /* jp */
+#define JPAFSR_JBREQ_SHIFT	24UL
+
+/* L2 cache way information */
+#define JPAFSR_ETW		(0x3UL << 22UL) /* jp */
+#define JPAFSR_ETW_SHIFT	22UL
+
+/* System bus MTAG ECC syndrome.  This field captures the status of the
+ * first occurrence of the highest-priority error according to the M_SYND
+ * overwrite policy.  After the AFSR sticky bit, corresponding to the error
+ * for which the M_SYND is reported, is cleared, the contents of the M_SYND
+ * field will be unchanged by will be unfrozen for further error capture.
+ */
+#define CHAFSR_M_SYNDROME	(0xfUL << 16UL) /* ch,ch+,jp */
+#define CHAFSR_M_SYNDROME_SHIFT	16UL
+
+/* Agenid Id of the foreign device causing the UE/CE errors */
+#define JPAFSR_AID		(0x1fUL << 9UL) /* jp */
+#define JPAFSR_AID_SHIFT	9UL
+
+/* System bus or E-cache data ECC syndrome.  This field captures the status
+ * of the first occurrence of the highest-priority error according to the
+ * E_SYND overwrite policy.  After the AFSR sticky bit, corresponding to the
+ * error for which the E_SYND is reported, is cleare, the contents of the E_SYND
+ * field will be unchanged but will be unfrozen for further error capture.
+ */
+#define CHAFSR_E_SYNDROME	(0x1ffUL << 0UL) /* ch,ch+,jp */
+#define CHAFSR_E_SYNDROME_SHIFT	0UL
+
+/* The AFSR must be explicitly cleared by software, it is not cleared automatically
+ * by a read.  Writes to bits <51:33> with bits set will clear the corresponding
+ * bits in the AFSR.  Bits associated with disrupting traps must be cleared before
+ * interrupts are re-enabled to prevent multiple traps for the same error.  I.e.
+ * PSTATE.IE and AFSR bits control delivery of disrupting traps.
+ *
+ * Since there is only one AFAR, when multiple events have been logged by the
+ * bits in the AFSR, at most one of these events will have its status captured
+ * in the AFAR.  The highest priority of those event bits will get AFAR logging.
+ * The AFAR will be unlocked and available to capture the address of another event
+ * as soon as the one bit in AFSR that corresponds to the event logged in AFAR is
+ * cleared.  For example, if AFSR.CE is detected, then AFSR.UE (which overwrites
+ * the AFAR), and AFSR.UE is cleared by not AFSR.CE, then the AFAR will be unlocked
+ * and ready for another event, even though AFSR.CE is still set.  The same rules
+ * also apply to the M_SYNDROME and E_SYNDROME fields of the AFSR.
+ */
+
+#endif /* _SPARC64_CHAFSR_H */
diff --git a/include/asm-sparc/checksum.h b/include/asm-sparc/checksum.h
index d044ddb5a3c..4e3553d4f6e 100644
--- a/include/asm-sparc/checksum.h
+++ b/include/asm-sparc/checksum.h
@@ -1,241 +1,8 @@
-#ifndef __SPARC_CHECKSUM_H
-#define __SPARC_CHECKSUM_H
-
-/*  checksum.h:  IP/UDP/TCP checksum routines on the Sparc.
- *
- *  Copyright(C) 1995 Linus Torvalds
- *  Copyright(C) 1995 Miguel de Icaza
- *  Copyright(C) 1996 David S. Miller
- *  Copyright(C) 1996 Eddie C. Dost
- *  Copyright(C) 1997 Jakub Jelinek
- *
- * derived from:
- *	Alpha checksum c-code
- *      ix86 inline assembly
- *      RFC1071 Computing the Internet Checksum
- */
- 
-#include <linux/in6.h>
-#include <asm/uaccess.h>
-
-/* computes the checksum of a memory block at buff, length len,
- * and adds in "sum" (32-bit)
- *
- * returns a 32-bit number suitable for feeding into itself
- * or csum_tcpudp_magic
- *
- * this function must be called with even lengths, except
- * for the last fragment, which may be odd
- *
- * it's best to have buff aligned on a 32-bit boundary
- */
-extern __wsum csum_partial(const void *buff, int len, __wsum sum);
-
-/* the same as csum_partial, but copies from fs:src while it
- * checksums
- *
- * here even more important to align src and dst on a 32-bit (or even
- * better 64-bit) boundary
- */
-
-extern unsigned int __csum_partial_copy_sparc_generic (const unsigned char *, unsigned char *);
-
-static inline __wsum
-csum_partial_copy_nocheck(const void *src, void *dst, int len, __wsum sum)
-{
-	register unsigned int ret asm("o0") = (unsigned int)src;
-	register char *d asm("o1") = dst;
-	register int l asm("g1") = len;
-
-	__asm__ __volatile__ (
-		"call __csum_partial_copy_sparc_generic\n\t"
-		" mov %6, %%g7\n"
-	: "=&r" (ret), "=&r" (d), "=&r" (l)
-	: "0" (ret), "1" (d), "2" (l), "r" (sum)
-	: "o2", "o3", "o4", "o5", "o7",
-	  "g2", "g3", "g4", "g5", "g7",
-	  "memory", "cc");
-	return (__force __wsum)ret;
-}
-
-static inline __wsum
-csum_partial_copy_from_user(const void __user *src, void *dst, int len,
-			    __wsum sum, int *err)
-  {
-	register unsigned long ret asm("o0") = (unsigned long)src;
-	register char *d asm("o1") = dst;
-	register int l asm("g1") = len;
-	register __wsum s asm("g7") = sum;
-
-	__asm__ __volatile__ (
-	".section __ex_table,#alloc\n\t"
-	".align 4\n\t"
-	".word 1f,2\n\t"
-	".previous\n"
-	"1:\n\t"
-	"call __csum_partial_copy_sparc_generic\n\t"
-	" st %8, [%%sp + 64]\n"
-	: "=&r" (ret), "=&r" (d), "=&r" (l), "=&r" (s)
-	: "0" (ret), "1" (d), "2" (l), "3" (s), "r" (err)
-	: "o2", "o3", "o4", "o5", "o7", "g2", "g3", "g4", "g5",
-	  "cc", "memory");
-	return (__force __wsum)ret;
-}
-  
-static inline __wsum
-csum_partial_copy_to_user(const void *src, void __user *dst, int len,
-			  __wsum sum, int *err)
-{
-	if (!access_ok (VERIFY_WRITE, dst, len)) {
-		*err = -EFAULT;
-		return sum;
-	} else {
-		register unsigned long ret asm("o0") = (unsigned long)src;
-		register char __user *d asm("o1") = dst;
-		register int l asm("g1") = len;
-		register __wsum s asm("g7") = sum;
-
-		__asm__ __volatile__ (
-		".section __ex_table,#alloc\n\t"
-		".align 4\n\t"
-		".word 1f,1\n\t"
-		".previous\n"
-		"1:\n\t"
-		"call __csum_partial_copy_sparc_generic\n\t"
-		" st %8, [%%sp + 64]\n"
-		: "=&r" (ret), "=&r" (d), "=&r" (l), "=&r" (s)
-		: "0" (ret), "1" (d), "2" (l), "3" (s), "r" (err)
-		: "o2", "o3", "o4", "o5", "o7",
-		  "g2", "g3", "g4", "g5",
-		  "cc", "memory");
-		return (__force __wsum)ret;
-	}
-}
-
-#define HAVE_CSUM_COPY_USER
-#define csum_and_copy_to_user csum_partial_copy_to_user
-
-/* ihl is always 5 or greater, almost always is 5, and iph is word aligned
- * the majority of the time.
- */
-static inline __sum16 ip_fast_csum(const void *iph, unsigned int ihl)
-{
-	__sum16 sum;
-
-	/* Note: We must read %2 before we touch %0 for the first time,
-	 *       because GCC can legitimately use the same register for
-	 *       both operands.
-	 */
-	__asm__ __volatile__("sub\t%2, 4, %%g4\n\t"
-			     "ld\t[%1 + 0x00], %0\n\t"
-			     "ld\t[%1 + 0x04], %%g2\n\t"
-			     "ld\t[%1 + 0x08], %%g3\n\t"
-			     "addcc\t%%g2, %0, %0\n\t"
-			     "addxcc\t%%g3, %0, %0\n\t"
-			     "ld\t[%1 + 0x0c], %%g2\n\t"
-			     "ld\t[%1 + 0x10], %%g3\n\t"
-			     "addxcc\t%%g2, %0, %0\n\t"
-			     "addx\t%0, %%g0, %0\n"
-			     "1:\taddcc\t%%g3, %0, %0\n\t"
-			     "add\t%1, 4, %1\n\t"
-			     "addxcc\t%0, %%g0, %0\n\t"
-			     "subcc\t%%g4, 1, %%g4\n\t"
-			     "be,a\t2f\n\t"
-			     "sll\t%0, 16, %%g2\n\t"
-			     "b\t1b\n\t"
-			     "ld\t[%1 + 0x10], %%g3\n"
-			     "2:\taddcc\t%0, %%g2, %%g2\n\t"
-			     "srl\t%%g2, 16, %0\n\t"
-			     "addx\t%0, %%g0, %0\n\t"
-			     "xnor\t%%g0, %0, %0"
-			     : "=r" (sum), "=&r" (iph)
-			     : "r" (ihl), "1" (iph)
-			     : "g2", "g3", "g4", "cc", "memory");
-	return sum;
-}
-
-/* Fold a partial checksum without adding pseudo headers. */
-static inline __sum16 csum_fold(__wsum sum)
-{
-	unsigned int tmp;
-
-	__asm__ __volatile__("addcc\t%0, %1, %1\n\t"
-			     "srl\t%1, 16, %1\n\t"
-			     "addx\t%1, %%g0, %1\n\t"
-			     "xnor\t%%g0, %1, %0"
-			     : "=&r" (sum), "=r" (tmp)
-			     : "0" (sum), "1" ((__force u32)sum<<16)
-			     : "cc");
-	return (__force __sum16)sum;
-}
-
-static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
-					       unsigned short len,
-					       unsigned short proto,
-					       __wsum sum)
-{
-	__asm__ __volatile__("addcc\t%1, %0, %0\n\t"
-			     "addxcc\t%2, %0, %0\n\t"
-			     "addxcc\t%3, %0, %0\n\t"
-			     "addx\t%0, %%g0, %0\n\t"
-			     : "=r" (sum), "=r" (saddr)
-			     : "r" (daddr), "r" (proto + len), "0" (sum),
-			       "1" (saddr)
-			     : "cc");
-	return sum;
-}
-
-/*
- * computes the checksum of the TCP/UDP pseudo-header
- * returns a 16-bit checksum, already complemented
- */
-static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr,
-						   unsigned short len,
-						   unsigned short proto,
-						   __wsum sum)
-{
-	return csum_fold(csum_tcpudp_nofold(saddr,daddr,len,proto,sum));
-}
-
-#define _HAVE_ARCH_IPV6_CSUM
-
-static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
-				      const struct in6_addr *daddr,
-				      __u32 len, unsigned short proto,
-				      __wsum sum)
-{
-	__asm__ __volatile__ (
-		"addcc	%3, %4, %%g4\n\t"
-		"addxcc	%5, %%g4, %%g4\n\t"
-		"ld	[%2 + 0x0c], %%g2\n\t"
-		"ld	[%2 + 0x08], %%g3\n\t"
-		"addxcc	%%g2, %%g4, %%g4\n\t"
-		"ld	[%2 + 0x04], %%g2\n\t"
-		"addxcc	%%g3, %%g4, %%g4\n\t"
-		"ld	[%2 + 0x00], %%g3\n\t"
-		"addxcc	%%g2, %%g4, %%g4\n\t"
-		"ld	[%1 + 0x0c], %%g2\n\t"
-		"addxcc	%%g3, %%g4, %%g4\n\t"
-		"ld	[%1 + 0x08], %%g3\n\t"
-		"addxcc	%%g2, %%g4, %%g4\n\t"
-		"ld	[%1 + 0x04], %%g2\n\t"
-		"addxcc	%%g3, %%g4, %%g4\n\t"
-		"ld	[%1 + 0x00], %%g3\n\t"
-		"addxcc	%%g2, %%g4, %%g4\n\t"
-		"addxcc	%%g3, %%g4, %0\n\t"
-		"addx	0, %0, %0\n"
-		: "=&r" (sum)
-		: "r" (saddr), "r" (daddr), 
-		  "r"(htonl(len)), "r"(htonl(proto)), "r"(sum)
-		: "g2", "g3", "g4", "cc");
-
-	return csum_fold(sum);
-}
-
-/* this routine is used for miscellaneous IP-like checksums, mainly in icmp.c */
-static inline __sum16 ip_compute_csum(const void *buff, int len)
-{
-	return csum_fold(csum_partial(buff, len, 0));
-}
-
-#endif /* !(__SPARC_CHECKSUM_H) */
+#ifndef ___ASM_SPARC_CHECKSUM_H
+#define ___ASM_SPARC_CHECKSUM_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/checksum_64.h>
+#else
+#include <asm-sparc/checksum_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/checksum_32.h b/include/asm-sparc/checksum_32.h
new file mode 100644
index 00000000000..bdbda1453aa
--- /dev/null
+++ b/include/asm-sparc/checksum_32.h
@@ -0,0 +1,241 @@
+#ifndef __SPARC_CHECKSUM_H
+#define __SPARC_CHECKSUM_H
+
+/*  checksum.h:  IP/UDP/TCP checksum routines on the Sparc.
+ *
+ *  Copyright(C) 1995 Linus Torvalds
+ *  Copyright(C) 1995 Miguel de Icaza
+ *  Copyright(C) 1996 David S. Miller
+ *  Copyright(C) 1996 Eddie C. Dost
+ *  Copyright(C) 1997 Jakub Jelinek
+ *
+ * derived from:
+ *	Alpha checksum c-code
+ *      ix86 inline assembly
+ *      RFC1071 Computing the Internet Checksum
+ */
+
+#include <linux/in6.h>
+#include <asm/uaccess.h>
+
+/* computes the checksum of a memory block at buff, length len,
+ * and adds in "sum" (32-bit)
+ *
+ * returns a 32-bit number suitable for feeding into itself
+ * or csum_tcpudp_magic
+ *
+ * this function must be called with even lengths, except
+ * for the last fragment, which may be odd
+ *
+ * it's best to have buff aligned on a 32-bit boundary
+ */
+extern __wsum csum_partial(const void *buff, int len, __wsum sum);
+
+/* the same as csum_partial, but copies from fs:src while it
+ * checksums
+ *
+ * here even more important to align src and dst on a 32-bit (or even
+ * better 64-bit) boundary
+ */
+
+extern unsigned int __csum_partial_copy_sparc_generic (const unsigned char *, unsigned char *);
+
+static inline __wsum
+csum_partial_copy_nocheck(const void *src, void *dst, int len, __wsum sum)
+{
+	register unsigned int ret asm("o0") = (unsigned int)src;
+	register char *d asm("o1") = dst;
+	register int l asm("g1") = len;
+
+	__asm__ __volatile__ (
+		"call __csum_partial_copy_sparc_generic\n\t"
+		" mov %6, %%g7\n"
+	: "=&r" (ret), "=&r" (d), "=&r" (l)
+	: "0" (ret), "1" (d), "2" (l), "r" (sum)
+	: "o2", "o3", "o4", "o5", "o7",
+	  "g2", "g3", "g4", "g5", "g7",
+	  "memory", "cc");
+	return (__force __wsum)ret;
+}
+
+static inline __wsum
+csum_partial_copy_from_user(const void __user *src, void *dst, int len,
+			    __wsum sum, int *err)
+  {
+	register unsigned long ret asm("o0") = (unsigned long)src;
+	register char *d asm("o1") = dst;
+	register int l asm("g1") = len;
+	register __wsum s asm("g7") = sum;
+
+	__asm__ __volatile__ (
+	".section __ex_table,#alloc\n\t"
+	".align 4\n\t"
+	".word 1f,2\n\t"
+	".previous\n"
+	"1:\n\t"
+	"call __csum_partial_copy_sparc_generic\n\t"
+	" st %8, [%%sp + 64]\n"
+	: "=&r" (ret), "=&r" (d), "=&r" (l), "=&r" (s)
+	: "0" (ret), "1" (d), "2" (l), "3" (s), "r" (err)
+	: "o2", "o3", "o4", "o5", "o7", "g2", "g3", "g4", "g5",
+	  "cc", "memory");
+	return (__force __wsum)ret;
+}
+
+static inline __wsum
+csum_partial_copy_to_user(const void *src, void __user *dst, int len,
+			  __wsum sum, int *err)
+{
+	if (!access_ok (VERIFY_WRITE, dst, len)) {
+		*err = -EFAULT;
+		return sum;
+	} else {
+		register unsigned long ret asm("o0") = (unsigned long)src;
+		register char __user *d asm("o1") = dst;
+		register int l asm("g1") = len;
+		register __wsum s asm("g7") = sum;
+
+		__asm__ __volatile__ (
+		".section __ex_table,#alloc\n\t"
+		".align 4\n\t"
+		".word 1f,1\n\t"
+		".previous\n"
+		"1:\n\t"
+		"call __csum_partial_copy_sparc_generic\n\t"
+		" st %8, [%%sp + 64]\n"
+		: "=&r" (ret), "=&r" (d), "=&r" (l), "=&r" (s)
+		: "0" (ret), "1" (d), "2" (l), "3" (s), "r" (err)
+		: "o2", "o3", "o4", "o5", "o7",
+		  "g2", "g3", "g4", "g5",
+		  "cc", "memory");
+		return (__force __wsum)ret;
+	}
+}
+
+#define HAVE_CSUM_COPY_USER
+#define csum_and_copy_to_user csum_partial_copy_to_user
+
+/* ihl is always 5 or greater, almost always is 5, and iph is word aligned
+ * the majority of the time.
+ */
+static inline __sum16 ip_fast_csum(const void *iph, unsigned int ihl)
+{
+	__sum16 sum;
+
+	/* Note: We must read %2 before we touch %0 for the first time,
+	 *       because GCC can legitimately use the same register for
+	 *       both operands.
+	 */
+	__asm__ __volatile__("sub\t%2, 4, %%g4\n\t"
+			     "ld\t[%1 + 0x00], %0\n\t"
+			     "ld\t[%1 + 0x04], %%g2\n\t"
+			     "ld\t[%1 + 0x08], %%g3\n\t"
+			     "addcc\t%%g2, %0, %0\n\t"
+			     "addxcc\t%%g3, %0, %0\n\t"
+			     "ld\t[%1 + 0x0c], %%g2\n\t"
+			     "ld\t[%1 + 0x10], %%g3\n\t"
+			     "addxcc\t%%g2, %0, %0\n\t"
+			     "addx\t%0, %%g0, %0\n"
+			     "1:\taddcc\t%%g3, %0, %0\n\t"
+			     "add\t%1, 4, %1\n\t"
+			     "addxcc\t%0, %%g0, %0\n\t"
+			     "subcc\t%%g4, 1, %%g4\n\t"
+			     "be,a\t2f\n\t"
+			     "sll\t%0, 16, %%g2\n\t"
+			     "b\t1b\n\t"
+			     "ld\t[%1 + 0x10], %%g3\n"
+			     "2:\taddcc\t%0, %%g2, %%g2\n\t"
+			     "srl\t%%g2, 16, %0\n\t"
+			     "addx\t%0, %%g0, %0\n\t"
+			     "xnor\t%%g0, %0, %0"
+			     : "=r" (sum), "=&r" (iph)
+			     : "r" (ihl), "1" (iph)
+			     : "g2", "g3", "g4", "cc", "memory");
+	return sum;
+}
+
+/* Fold a partial checksum without adding pseudo headers. */
+static inline __sum16 csum_fold(__wsum sum)
+{
+	unsigned int tmp;
+
+	__asm__ __volatile__("addcc\t%0, %1, %1\n\t"
+			     "srl\t%1, 16, %1\n\t"
+			     "addx\t%1, %%g0, %1\n\t"
+			     "xnor\t%%g0, %1, %0"
+			     : "=&r" (sum), "=r" (tmp)
+			     : "0" (sum), "1" ((__force u32)sum<<16)
+			     : "cc");
+	return (__force __sum16)sum;
+}
+
+static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+					       unsigned short len,
+					       unsigned short proto,
+					       __wsum sum)
+{
+	__asm__ __volatile__("addcc\t%1, %0, %0\n\t"
+			     "addxcc\t%2, %0, %0\n\t"
+			     "addxcc\t%3, %0, %0\n\t"
+			     "addx\t%0, %%g0, %0\n\t"
+			     : "=r" (sum), "=r" (saddr)
+			     : "r" (daddr), "r" (proto + len), "0" (sum),
+			       "1" (saddr)
+			     : "cc");
+	return sum;
+}
+
+/*
+ * computes the checksum of the TCP/UDP pseudo-header
+ * returns a 16-bit checksum, already complemented
+ */
+static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr,
+						   unsigned short len,
+						   unsigned short proto,
+						   __wsum sum)
+{
+	return csum_fold(csum_tcpudp_nofold(saddr,daddr,len,proto,sum));
+}
+
+#define _HAVE_ARCH_IPV6_CSUM
+
+static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
+				      const struct in6_addr *daddr,
+				      __u32 len, unsigned short proto,
+				      __wsum sum)
+{
+	__asm__ __volatile__ (
+		"addcc	%3, %4, %%g4\n\t"
+		"addxcc	%5, %%g4, %%g4\n\t"
+		"ld	[%2 + 0x0c], %%g2\n\t"
+		"ld	[%2 + 0x08], %%g3\n\t"
+		"addxcc	%%g2, %%g4, %%g4\n\t"
+		"ld	[%2 + 0x04], %%g2\n\t"
+		"addxcc	%%g3, %%g4, %%g4\n\t"
+		"ld	[%2 + 0x00], %%g3\n\t"
+		"addxcc	%%g2, %%g4, %%g4\n\t"
+		"ld	[%1 + 0x0c], %%g2\n\t"
+		"addxcc	%%g3, %%g4, %%g4\n\t"
+		"ld	[%1 + 0x08], %%g3\n\t"
+		"addxcc	%%g2, %%g4, %%g4\n\t"
+		"ld	[%1 + 0x04], %%g2\n\t"
+		"addxcc	%%g3, %%g4, %%g4\n\t"
+		"ld	[%1 + 0x00], %%g3\n\t"
+		"addxcc	%%g2, %%g4, %%g4\n\t"
+		"addxcc	%%g3, %%g4, %0\n\t"
+		"addx	0, %0, %0\n"
+		: "=&r" (sum)
+		: "r" (saddr), "r" (daddr),
+		  "r"(htonl(len)), "r"(htonl(proto)), "r"(sum)
+		: "g2", "g3", "g4", "cc");
+
+	return csum_fold(sum);
+}
+
+/* this routine is used for miscellaneous IP-like checksums, mainly in icmp.c */
+static inline __sum16 ip_compute_csum(const void *buff, int len)
+{
+	return csum_fold(csum_partial(buff, len, 0));
+}
+
+#endif /* !(__SPARC_CHECKSUM_H) */
diff --git a/include/asm-sparc/checksum_64.h b/include/asm-sparc/checksum_64.h
new file mode 100644
index 00000000000..019b9615e43
--- /dev/null
+++ b/include/asm-sparc/checksum_64.h
@@ -0,0 +1,167 @@
+#ifndef __SPARC64_CHECKSUM_H
+#define __SPARC64_CHECKSUM_H
+
+/*  checksum.h:  IP/UDP/TCP checksum routines on the V9.
+ *
+ *  Copyright(C) 1995 Linus Torvalds
+ *  Copyright(C) 1995 Miguel de Icaza
+ *  Copyright(C) 1996 David S. Miller
+ *  Copyright(C) 1996 Eddie C. Dost
+ *  Copyright(C) 1997 Jakub Jelinek
+ *
+ * derived from:
+ *	Alpha checksum c-code
+ *      ix86 inline assembly
+ *      RFC1071 Computing the Internet Checksum
+ */
+
+#include <linux/in6.h>
+#include <asm/uaccess.h>
+
+/* computes the checksum of a memory block at buff, length len,
+ * and adds in "sum" (32-bit)
+ *
+ * returns a 32-bit number suitable for feeding into itself
+ * or csum_tcpudp_magic
+ *
+ * this function must be called with even lengths, except
+ * for the last fragment, which may be odd
+ *
+ * it's best to have buff aligned on a 32-bit boundary
+ */
+extern __wsum csum_partial(const void * buff, int len, __wsum sum);
+
+/* the same as csum_partial, but copies from user space while it
+ * checksums
+ *
+ * here even more important to align src and dst on a 32-bit (or even
+ * better 64-bit) boundary
+ */
+extern __wsum csum_partial_copy_nocheck(const void *src, void *dst,
+					      int len, __wsum sum);
+
+extern long __csum_partial_copy_from_user(const void __user *src,
+					  void *dst, int len,
+					  __wsum sum);
+
+static inline __wsum
+csum_partial_copy_from_user(const void __user *src,
+			    void *dst, int len,
+			    __wsum sum, int *err)
+{
+	long ret = __csum_partial_copy_from_user(src, dst, len, sum);
+	if (ret < 0)
+		*err = -EFAULT;
+	return (__force __wsum) ret;
+}
+
+/*
+ *	Copy and checksum to user
+ */
+#define HAVE_CSUM_COPY_USER
+extern long __csum_partial_copy_to_user(const void *src,
+					void __user *dst, int len,
+					  __wsum sum);
+
+static inline __wsum
+csum_and_copy_to_user(const void *src,
+		      void __user *dst, int len,
+		      __wsum sum, int *err)
+{
+	long ret = __csum_partial_copy_to_user(src, dst, len, sum);
+	if (ret < 0)
+		*err = -EFAULT;
+	return (__force __wsum) ret;
+}
+
+/* ihl is always 5 or greater, almost always is 5, and iph is word aligned
+ * the majority of the time.
+ */
+extern __sum16 ip_fast_csum(const void *iph, unsigned int ihl);
+
+/* Fold a partial checksum without adding pseudo headers. */
+static inline __sum16 csum_fold(__wsum sum)
+{
+	unsigned int tmp;
+
+	__asm__ __volatile__(
+"	addcc		%0, %1, %1\n"
+"	srl		%1, 16, %1\n"
+"	addc		%1, %%g0, %1\n"
+"	xnor		%%g0, %1, %0\n"
+	: "=&r" (sum), "=r" (tmp)
+	: "0" (sum), "1" ((__force u32)sum<<16)
+	: "cc");
+	return (__force __sum16)sum;
+}
+
+static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+					       unsigned int len,
+					       unsigned short proto,
+					       __wsum sum)
+{
+	__asm__ __volatile__(
+"	addcc		%1, %0, %0\n"
+"	addccc		%2, %0, %0\n"
+"	addccc		%3, %0, %0\n"
+"	addc		%0, %%g0, %0\n"
+	: "=r" (sum), "=r" (saddr)
+	: "r" (daddr), "r" (proto + len), "0" (sum), "1" (saddr)
+	: "cc");
+	return sum;
+}
+
+/*
+ * computes the checksum of the TCP/UDP pseudo-header
+ * returns a 16-bit checksum, already complemented
+ */
+static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr,
+						   unsigned short len,
+						   unsigned short proto,
+						   __wsum sum)
+{
+	return csum_fold(csum_tcpudp_nofold(saddr,daddr,len,proto,sum));
+}
+
+#define _HAVE_ARCH_IPV6_CSUM
+
+static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
+				      const struct in6_addr *daddr,
+				      __u32 len, unsigned short proto,
+				      __wsum sum)
+{
+	__asm__ __volatile__ (
+"	addcc		%3, %4, %%g7\n"
+"	addccc		%5, %%g7, %%g7\n"
+"	lduw		[%2 + 0x0c], %%g2\n"
+"	lduw		[%2 + 0x08], %%g3\n"
+"	addccc		%%g2, %%g7, %%g7\n"
+"	lduw		[%2 + 0x04], %%g2\n"
+"	addccc		%%g3, %%g7, %%g7\n"
+"	lduw		[%2 + 0x00], %%g3\n"
+"	addccc		%%g2, %%g7, %%g7\n"
+"	lduw		[%1 + 0x0c], %%g2\n"
+"	addccc		%%g3, %%g7, %%g7\n"
+"	lduw		[%1 + 0x08], %%g3\n"
+"	addccc		%%g2, %%g7, %%g7\n"
+"	lduw		[%1 + 0x04], %%g2\n"
+"	addccc		%%g3, %%g7, %%g7\n"
+"	lduw		[%1 + 0x00], %%g3\n"
+"	addccc		%%g2, %%g7, %%g7\n"
+"	addccc		%%g3, %%g7, %0\n"
+"	addc		0, %0, %0\n"
+	: "=&r" (sum)
+	: "r" (saddr), "r" (daddr), "r"(htonl(len)),
+	  "r"(htonl(proto)), "r"(sum)
+	: "g2", "g3", "g7", "cc");
+
+	return csum_fold(sum);
+}
+
+/* this routine is used for miscellaneous IP-like checksums, mainly in icmp.c */
+static inline __sum16 ip_compute_csum(const void *buff, int len)
+{
+	return csum_fold(csum_partial(buff, len, 0));
+}
+
+#endif /* !(__SPARC64_CHECKSUM_H) */
diff --git a/include/asm-sparc/chmctrl.h b/include/asm-sparc/chmctrl.h
new file mode 100644
index 00000000000..859b4a4b0d3
--- /dev/null
+++ b/include/asm-sparc/chmctrl.h
@@ -0,0 +1,183 @@
+#ifndef _SPARC64_CHMCTRL_H
+#define _SPARC64_CHMCTRL_H
+
+/* Cheetah memory controller programmable registers. */
+#define CHMCTRL_TCTRL1		0x00 /* Memory Timing Control I		*/
+#define CHMCTRL_TCTRL2		0x08 /* Memory Timing Control II	*/
+#define CHMCTRL_TCTRL3		0x38 /* Memory Timing Control III	*/
+#define CHMCTRL_TCTRL4		0x40 /* Memory Timing Control IV	*/
+#define CHMCTRL_DECODE1		0x10 /* Memory Address Decode I		*/
+#define CHMCTRL_DECODE2		0x18 /* Memory Address Decode II	*/
+#define CHMCTRL_DECODE3		0x20 /* Memory Address Decode III	*/
+#define CHMCTRL_DECODE4		0x28 /* Memory Address Decode IV	*/
+#define CHMCTRL_MACTRL		0x30 /* Memory Address Control		*/
+
+/* Memory Timing Control I */
+#define TCTRL1_SDRAMCTL_DLY	0xf000000000000000UL
+#define TCTRL1_SDRAMCTL_DLY_SHIFT     60
+#define TCTRL1_SDRAMCLK_DLY	0x0e00000000000000UL
+#define TCTRL1_SDRAMCLK_DLY_SHIFT     57
+#define TCTRL1_R		0x0100000000000000UL
+#define TCTRL1_R_SHIFT 		      56
+#define TCTRL1_AUTORFR_CYCLE	0x00fe000000000000UL
+#define TCTRL1_AUTORFR_CYCLE_SHIFT    49
+#define TCTRL1_RD_WAIT		0x0001f00000000000UL
+#define TCTRL1_RD_WAIT_SHIFT	      44
+#define TCTRL1_PC_CYCLE		0x00000fc000000000UL
+#define TCTRL1_PC_CYCLE_SHIFT	      38
+#define TCTRL1_WR_MORE_RAS_PW	0x0000003f00000000UL
+#define TCTRL1_WR_MORE_RAS_PW_SHIFT   32
+#define TCTRL1_RD_MORE_RAW_PW	0x00000000fc000000UL
+#define TCTRL1_RD_MORE_RAS_PW_SHIFT   26
+#define TCTRL1_ACT_WR_DLY	0x0000000003f00000UL
+#define TCTRL1_ACT_WR_DLY_SHIFT	      20
+#define TCTRL1_ACT_RD_DLY	0x00000000000fc000UL
+#define TCTRL1_ACT_RD_DLY_SHIFT	      14
+#define TCTRL1_BANK_PRESENT	0x0000000000003000UL
+#define TCTRL1_BANK_PRESENT_SHIFT     12
+#define TCTRL1_RFR_INT		0x0000000000000ff8UL
+#define TCTRL1_RFR_INT_SHIFT	      3
+#define TCTRL1_SET_MODE_REG	0x0000000000000004UL
+#define TCTRL1_SET_MODE_REG_SHIFT     2
+#define TCTRL1_RFR_ENABLE	0x0000000000000002UL
+#define TCTRL1_RFR_ENABLE_SHIFT	      1
+#define TCTRL1_PRECHG_ALL	0x0000000000000001UL
+#define TCTRL1_PRECHG_ALL_SHIFT	      0
+
+/* Memory Timing Control II */
+#define TCTRL2_WR_MSEL_DLY	0xfc00000000000000UL
+#define TCTRL2_WR_MSEL_DLY_SHIFT      58
+#define TCTRL2_RD_MSEL_DLY	0x03f0000000000000UL
+#define TCTRL2_RD_MSEL_DLY_SHIFT      52
+#define TCTRL2_WRDATA_THLD	0x000c000000000000UL
+#define TCTRL2_WRDATA_THLD_SHIFT      50
+#define TCTRL2_RDWR_RD_TI_DLY	0x0003f00000000000UL
+#define TCTRL2_RDWR_RD_TI_DLY_SHIFT   44
+#define TCTRL2_AUTOPRECHG_ENBL	0x0000080000000000UL
+#define TCTRL2_AUTOPRECHG_ENBL_SHIFT  43
+#define TCTRL2_RDWR_PI_MORE_DLY	0x000007c000000000UL
+#define TCTRL2_RDWR_PI_MORE_DLY_SHIFT 38
+#define TCTRL2_RDWR_1_DLY	0x0000003f00000000UL
+#define TCTRL2_RDWR_1_DLY_SHIFT       32
+#define TCTRL2_WRWR_PI_MORE_DLY	0x00000000f8000000UL
+#define TCTRL2_WRWR_PI_MORE_DLY_SHIFT 27
+#define TCTRL2_WRWR_1_DLY	0x0000000007e00000UL
+#define TCTRL2_WRWR_1_DLY_SHIFT       21
+#define TCTRL2_RDWR_RD_PI_MORE_DLY 0x00000000001f0000UL
+#define TCTRL2_RDWR_RD_PI_MORE_DLY_SHIFT 16
+#define TCTRL2_R		0x0000000000008000UL
+#define TCTRL2_R_SHIFT		      15
+#define TCTRL2_SDRAM_MODE_REG_DATA 0x0000000000007fffUL
+#define TCTRL2_SDRAM_MODE_REG_DATA_SHIFT 0
+
+/* Memory Timing Control III */
+#define TCTRL3_SDRAM_CTL_DLY	0xf000000000000000UL
+#define TCTRL3_SDRAM_CTL_DLY_SHIFT    60
+#define TCTRL3_SDRAM_CLK_DLY	0x0e00000000000000UL
+#define TCTRL3_SDRAM_CLK_DLY_SHIFT    57
+#define TCTRL3_R		0x0100000000000000UL
+#define TCTRL3_R_SHIFT		      56
+#define TCTRL3_AUTO_RFR_CYCLE	0x00fe000000000000UL
+#define TCTRL3_AUTO_RFR_CYCLE_SHIFT   49
+#define TCTRL3_RD_WAIT		0x0001f00000000000UL
+#define TCTRL3_RD_WAIT_SHIFT	      44
+#define TCTRL3_PC_CYCLE		0x00000fc000000000UL
+#define TCTRL3_PC_CYCLE_SHIFT	      38
+#define TCTRL3_WR_MORE_RAW_PW	0x0000003f00000000UL
+#define TCTRL3_WR_MORE_RAW_PW_SHIFT   32
+#define TCTRL3_RD_MORE_RAW_PW	0x00000000fc000000UL
+#define TCTRL3_RD_MORE_RAW_PW_SHIFT   26
+#define TCTRL3_ACT_WR_DLY	0x0000000003f00000UL
+#define TCTRL3_ACT_WR_DLY_SHIFT       20
+#define TCTRL3_ACT_RD_DLY	0x00000000000fc000UL
+#define TCTRL3_ACT_RD_DLY_SHIFT       14
+#define TCTRL3_BANK_PRESENT	0x0000000000003000UL
+#define TCTRL3_BANK_PRESENT_SHIFT     12
+#define TCTRL3_RFR_INT		0x0000000000000ff8UL
+#define TCTRL3_RFR_INT_SHIFT	      3
+#define TCTRL3_SET_MODE_REG	0x0000000000000004UL
+#define TCTRL3_SET_MODE_REG_SHIFT     2
+#define TCTRL3_RFR_ENABLE	0x0000000000000002UL
+#define TCTRL3_RFR_ENABLE_SHIFT       1
+#define TCTRL3_PRECHG_ALL	0x0000000000000001UL
+#define TCTRL3_PRECHG_ALL_SHIFT	      0
+
+/* Memory Timing Control IV */
+#define TCTRL4_WR_MSEL_DLY	0xfc00000000000000UL
+#define TCTRL4_WR_MSEL_DLY_SHIFT      58
+#define TCTRL4_RD_MSEL_DLY	0x03f0000000000000UL
+#define TCTRL4_RD_MSEL_DLY_SHIFT      52
+#define TCTRL4_WRDATA_THLD	0x000c000000000000UL
+#define TCTRL4_WRDATA_THLD_SHIFT      50
+#define TCTRL4_RDWR_RD_RI_DLY	0x0003f00000000000UL
+#define TCTRL4_RDWR_RD_RI_DLY_SHIFT   44
+#define TCTRL4_AUTO_PRECHG_ENBL	0x0000080000000000UL
+#define TCTRL4_AUTO_PRECHG_ENBL_SHIFT 43
+#define TCTRL4_RD_WR_PI_MORE_DLY 0x000007c000000000UL
+#define TCTRL4_RD_WR_PI_MORE_DLY_SHIFT 38
+#define TCTRL4_RD_WR_TI_DLY	0x0000003f00000000UL
+#define TCTRL4_RD_WR_TI_DLY_SHIFT     32
+#define TCTRL4_WR_WR_PI_MORE_DLY 0x00000000f8000000UL
+#define TCTRL4_WR_WR_PI_MORE_DLY_SHIFT 27
+#define TCTRL4_WR_WR_TI_DLY	0x0000000007e00000UL
+#define TCTRL4_WR_WR_TI_DLY_SHIFT     21
+#define TCTRL4_RDWR_RD_PI_MORE_DLY 0x00000000001f000UL0
+#define TCTRL4_RDWR_RD_PI_MORE_DLY_SHIFT 16
+#define TCTRL4_R		0x0000000000008000UL
+#define TCTRL4_R_SHIFT		      15
+#define TCTRL4_SDRAM_MODE_REG_DATA 0x0000000000007fffUL
+#define TCTRL4_SDRAM_MODE_REG_DATA_SHIFT 0
+
+/* All 4 memory address decoding registers have the
+ * same layout.
+ */
+#define MEM_DECODE_VALID	0x8000000000000000UL /* Valid */
+#define MEM_DECODE_VALID_SHIFT	      63
+#define MEM_DECODE_UK		0x001ffe0000000000UL /* Upper mask */
+#define MEM_DECODE_UK_SHIFT	      41
+#define MEM_DECODE_UM		0x0000001ffff00000UL /* Upper match */
+#define MEM_DECODE_UM_SHIFT	      20
+#define MEM_DECODE_LK		0x000000000003c000UL /* Lower mask */
+#define MEM_DECODE_LK_SHIFT	      14
+#define MEM_DECODE_LM		0x0000000000000f00UL /* Lower match */
+#define MEM_DECODE_LM_SHIFT           8
+
+#define PA_UPPER_BITS		0x000007fffc000000UL
+#define PA_UPPER_BITS_SHIFT	26
+#define PA_LOWER_BITS		0x00000000000003c0UL
+#define PA_LOWER_BITS_SHIFT	6
+
+#define MACTRL_R0		         0x8000000000000000UL
+#define MACTRL_R0_SHIFT		         63
+#define MACTRL_ADDR_LE_PW                0x7000000000000000UL
+#define MACTRL_ADDR_LE_PW_SHIFT		 60
+#define MACTRL_CMD_PW                    0x0f00000000000000UL
+#define MACTRL_CMD_PW_SHIFT		 56
+#define MACTRL_HALF_MODE_WR_MSEL_DLY     0x00fc000000000000UL
+#define MACTRL_HALF_MODE_WR_MSEL_DLY_SHIFT 50
+#define MACTRL_HALF_MODE_RD_MSEL_DLY     0x0003f00000000000UL
+#define MACTRL_HALF_MODE_RD_MSEL_DLY_SHIFT 44
+#define MACTRL_HALF_MODE_SDRAM_CTL_DLY   0x00000f0000000000UL
+#define MACTRL_HALF_MODE_SDRAM_CTL_DLY_SHIFT 40
+#define MACTRL_HALF_MODE_SDRAM_CLK_DLY   0x000000e000000000UL
+#define MACTRL_HALF_MODE_SDRAM_CLK_DLY_SHIFT 37
+#define MACTRL_R1                        0x0000001000000000UL
+#define MACTRL_R1_SHIFT                      36
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B3 0x0000000f00000000UL
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B3_SHIFT 32
+#define MACTRL_ENC_INTLV_B3              0x00000000f8000000UL
+#define MACTRL_ENC_INTLV_B3_SHIFT              27
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B2 0x0000000007800000UL
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B2_SHIFT 23
+#define MACTRL_ENC_INTLV_B2              0x00000000007c0000UL
+#define MACTRL_ENC_INTLV_B2_SHIFT              18
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B1 0x000000000003c000UL
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B1_SHIFT 14
+#define MACTRL_ENC_INTLV_B1              0x0000000000003e00UL
+#define MACTRL_ENC_INTLV_B1_SHIFT               9
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B0 0x00000000000001e0UL
+#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B0_SHIFT  5
+#define MACTRL_ENC_INTLV_B0              0x000000000000001fUL
+#define MACTRL_ENC_INTLV_B0_SHIFT               0
+
+#endif /* _SPARC64_CHMCTRL_H */
diff --git a/include/asm-sparc/cmt.h b/include/asm-sparc/cmt.h
new file mode 100644
index 00000000000..870db592857
--- /dev/null
+++ b/include/asm-sparc/cmt.h
@@ -0,0 +1,59 @@
+#ifndef _SPARC64_CMT_H
+#define _SPARC64_CMT_H
+
+/* cmt.h: Chip Multi-Threading register definitions
+ *
+ * Copyright (C) 2004 David S. Miller (davem@redhat.com)
+ */
+
+/* ASI_CORE_ID - private */
+#define LP_ID		0x0000000000000010UL
+#define  LP_ID_MAX	0x00000000003f0000UL
+#define  LP_ID_ID	0x000000000000003fUL
+
+/* ASI_INTR_ID - private */
+#define LP_INTR_ID	0x0000000000000000UL
+#define  LP_INTR_ID_ID	0x00000000000003ffUL
+
+/* ASI_CESR_ID - private */
+#define CESR_ID		0x0000000000000040UL
+#define  CESR_ID_ID	0x00000000000000ffUL
+
+/* ASI_CORE_AVAILABLE - shared */
+#define LP_AVAIL	0x0000000000000000UL
+#define  LP_AVAIL_1	0x0000000000000002UL
+#define  LP_AVAIL_0	0x0000000000000001UL
+
+/* ASI_CORE_ENABLE_STATUS - shared */
+#define LP_ENAB_STAT	0x0000000000000010UL
+#define  LP_ENAB_STAT_1	0x0000000000000002UL
+#define  LP_ENAB_STAT_0	0x0000000000000001UL
+
+/* ASI_CORE_ENABLE - shared */
+#define LP_ENAB		0x0000000000000020UL
+#define  LP_ENAB_1	0x0000000000000002UL
+#define  LP_ENAB_0	0x0000000000000001UL
+
+/* ASI_CORE_RUNNING - shared */
+#define LP_RUNNING_RW	0x0000000000000050UL
+#define LP_RUNNING_W1S	0x0000000000000060UL
+#define LP_RUNNING_W1C	0x0000000000000068UL
+#define  LP_RUNNING_1	0x0000000000000002UL
+#define  LP_RUNNING_0	0x0000000000000001UL
+
+/* ASI_CORE_RUNNING_STAT - shared */
+#define LP_RUN_STAT	0x0000000000000058UL
+#define  LP_RUN_STAT_1	0x0000000000000002UL
+#define  LP_RUN_STAT_0	0x0000000000000001UL
+
+/* ASI_XIR_STEERING - shared */
+#define LP_XIR_STEER	0x0000000000000030UL
+#define  LP_XIR_STEER_1	0x0000000000000002UL
+#define  LP_XIR_STEER_0	0x0000000000000001UL
+
+/* ASI_CMT_ERROR_STEERING - shared */
+#define CMT_ER_STEER	0x0000000000000040UL
+#define  CMT_ER_STEER_1	0x0000000000000002UL
+#define  CMT_ER_STEER_0	0x0000000000000001UL
+
+#endif /* _SPARC64_CMT_H */
diff --git a/include/asm-sparc/compat.h b/include/asm-sparc/compat.h
new file mode 100644
index 00000000000..f260b58f5ce
--- /dev/null
+++ b/include/asm-sparc/compat.h
@@ -0,0 +1,243 @@
+#ifndef _ASM_SPARC64_COMPAT_H
+#define _ASM_SPARC64_COMPAT_H
+/*
+ * Architecture specific compatibility types
+ */
+#include <linux/types.h>
+
+#define COMPAT_USER_HZ	100
+
+typedef u32		compat_size_t;
+typedef s32		compat_ssize_t;
+typedef s32		compat_time_t;
+typedef s32		compat_clock_t;
+typedef s32		compat_pid_t;
+typedef u16		__compat_uid_t;
+typedef u16		__compat_gid_t;
+typedef u32		__compat_uid32_t;
+typedef u32		__compat_gid32_t;
+typedef u16		compat_mode_t;
+typedef u32		compat_ino_t;
+typedef u16		compat_dev_t;
+typedef s32		compat_off_t;
+typedef s64		compat_loff_t;
+typedef s16		compat_nlink_t;
+typedef u16		compat_ipc_pid_t;
+typedef s32		compat_daddr_t;
+typedef u32		compat_caddr_t;
+typedef __kernel_fsid_t	compat_fsid_t;
+typedef s32		compat_key_t;
+typedef s32		compat_timer_t;
+
+typedef s32		compat_int_t;
+typedef s32		compat_long_t;
+typedef s64		compat_s64;
+typedef u32		compat_uint_t;
+typedef u32		compat_ulong_t;
+typedef u64		compat_u64;
+
+struct compat_timespec {
+	compat_time_t	tv_sec;
+	s32		tv_nsec;
+};
+
+struct compat_timeval {
+	compat_time_t	tv_sec;
+	s32		tv_usec;
+};
+
+struct compat_stat {
+	compat_dev_t	st_dev;
+	compat_ino_t	st_ino;
+	compat_mode_t	st_mode;
+	compat_nlink_t	st_nlink;
+	__compat_uid_t	st_uid;
+	__compat_gid_t	st_gid;
+	compat_dev_t	st_rdev;
+	compat_off_t	st_size;
+	compat_time_t	st_atime;
+	compat_ulong_t	st_atime_nsec;
+	compat_time_t	st_mtime;
+	compat_ulong_t	st_mtime_nsec;
+	compat_time_t	st_ctime;
+	compat_ulong_t	st_ctime_nsec;
+	compat_off_t	st_blksize;
+	compat_off_t	st_blocks;
+	u32		__unused4[2];
+};
+
+struct compat_stat64 {
+	unsigned long long	st_dev;
+
+	unsigned long long	st_ino;
+
+	unsigned int	st_mode;
+	unsigned int	st_nlink;
+
+	unsigned int	st_uid;
+	unsigned int	st_gid;
+
+	unsigned long long	st_rdev;
+
+	unsigned char	__pad3[8];
+
+	long long	st_size;
+	unsigned int	st_blksize;
+
+	unsigned char	__pad4[8];
+	unsigned int	st_blocks;
+
+	unsigned int	st_atime;
+	unsigned int	st_atime_nsec;
+
+	unsigned int	st_mtime;
+	unsigned int	st_mtime_nsec;
+
+	unsigned int	st_ctime;
+	unsigned int	st_ctime_nsec;
+
+	unsigned int	__unused4;
+	unsigned int	__unused5;
+};
+
+struct compat_flock {
+	short		l_type;
+	short		l_whence;
+	compat_off_t	l_start;
+	compat_off_t	l_len;
+	compat_pid_t	l_pid;
+	short		__unused;
+};
+
+#define F_GETLK64	12
+#define F_SETLK64	13
+#define F_SETLKW64	14
+
+struct compat_flock64 {
+	short		l_type;
+	short		l_whence;
+	compat_loff_t	l_start;
+	compat_loff_t	l_len;
+	compat_pid_t	l_pid;
+	short		__unused;
+};
+
+struct compat_statfs {
+	int		f_type;
+	int		f_bsize;
+	int		f_blocks;
+	int		f_bfree;
+	int		f_bavail;
+	int		f_files;
+	int		f_ffree;
+	compat_fsid_t	f_fsid;
+	int		f_namelen;	/* SunOS ignores this field. */
+	int		f_frsize;
+	int		f_spare[5];
+};
+
+#define COMPAT_RLIM_INFINITY 0x7fffffff
+
+typedef u32		compat_old_sigset_t;
+
+#define _COMPAT_NSIG		64
+#define _COMPAT_NSIG_BPW	32
+
+typedef u32		compat_sigset_word;
+
+#define COMPAT_OFF_T_MAX	0x7fffffff
+#define COMPAT_LOFF_T_MAX	0x7fffffffffffffffL
+
+/*
+ * A pointer passed in from user mode. This should not
+ * be used for syscall parameters, just declare them
+ * as pointers because the syscall entry code will have
+ * appropriately converted them already.
+ */
+typedef	u32		compat_uptr_t;
+
+static inline void __user *compat_ptr(compat_uptr_t uptr)
+{
+	return (void __user *)(unsigned long)uptr;
+}
+
+static inline compat_uptr_t ptr_to_compat(void __user *uptr)
+{
+	return (u32)(unsigned long)uptr;
+}
+
+static inline void __user *compat_alloc_user_space(long len)
+{
+	struct pt_regs *regs = current_thread_info()->kregs;
+	unsigned long usp = regs->u_regs[UREG_I6];
+
+	if (!(test_thread_flag(TIF_32BIT)))
+		usp += STACK_BIAS;
+	else
+		usp &= 0xffffffffUL;
+
+	usp -= len;
+	usp &= ~0x7UL;
+
+	return (void __user *) usp;
+}
+
+struct compat_ipc64_perm {
+	compat_key_t key;
+	__compat_uid32_t uid;
+	__compat_gid32_t gid;
+	__compat_uid32_t cuid;
+	__compat_gid32_t cgid;
+	unsigned short __pad1;
+	compat_mode_t mode;
+	unsigned short __pad2;
+	unsigned short seq;
+	unsigned long __unused1;	/* yes they really are 64bit pads */
+	unsigned long __unused2;
+};
+
+struct compat_semid64_ds {
+	struct compat_ipc64_perm sem_perm;
+	unsigned int	__pad1;
+	compat_time_t	sem_otime;
+	unsigned int	__pad2;
+	compat_time_t	sem_ctime;
+	u32		sem_nsems;
+	u32		__unused1;
+	u32		__unused2;
+};
+
+struct compat_msqid64_ds {
+	struct compat_ipc64_perm msg_perm;
+	unsigned int	__pad1;
+	compat_time_t	msg_stime;
+	unsigned int	__pad2;
+	compat_time_t	msg_rtime;
+	unsigned int	__pad3;
+	compat_time_t	msg_ctime;
+	unsigned int	msg_cbytes;
+	unsigned int	msg_qnum;
+	unsigned int	msg_qbytes;
+	compat_pid_t	msg_lspid;
+	compat_pid_t	msg_lrpid;
+	unsigned int	__unused1;
+	unsigned int	__unused2;
+};
+
+struct compat_shmid64_ds {
+	struct compat_ipc64_perm shm_perm;
+	unsigned int	__pad1;
+	compat_time_t	shm_atime;
+	unsigned int	__pad2;
+	compat_time_t	shm_dtime;
+	unsigned int	__pad3;
+	compat_time_t	shm_ctime;
+	compat_size_t	shm_segsz;
+	compat_pid_t	shm_cpid;
+	compat_pid_t	shm_lpid;
+	unsigned int	shm_nattch;
+	unsigned int	__unused1;
+	unsigned int	__unused2;
+};
+
+#endif /* _ASM_SPARC64_COMPAT_H */
diff --git a/include/asm-sparc/compat_signal.h b/include/asm-sparc/compat_signal.h
new file mode 100644
index 00000000000..b759eab9b51
--- /dev/null
+++ b/include/asm-sparc/compat_signal.h
@@ -0,0 +1,29 @@
+#ifndef _COMPAT_SIGNAL_H
+#define _COMPAT_SIGNAL_H
+
+#include <linux/compat.h>
+#include <asm/signal.h>
+
+#ifdef CONFIG_COMPAT
+struct __new_sigaction32 {
+	unsigned		sa_handler;
+	unsigned int    	sa_flags;
+	unsigned		sa_restorer;     /* not used by Linux/SPARC yet */
+	compat_sigset_t 	sa_mask;
+};
+
+struct __old_sigaction32 {
+	unsigned		sa_handler;
+	compat_old_sigset_t  	sa_mask;
+	unsigned int    	sa_flags;
+	unsigned		sa_restorer;     /* not used by Linux/SPARC yet */
+};
+
+typedef struct sigaltstack32 {
+	u32			ss_sp;
+	int			ss_flags;
+	compat_size_t		ss_size;
+} stack_t32;
+#endif
+
+#endif /* !(_COMPAT_SIGNAL_H) */
diff --git a/include/asm-sparc/cpudata.h b/include/asm-sparc/cpudata.h
index a2c4d51d36c..b76fac0c8d8 100644
--- a/include/asm-sparc/cpudata.h
+++ b/include/asm-sparc/cpudata.h
@@ -1,27 +1,8 @@
-/* cpudata.h: Per-cpu parameters.
- *
- * Copyright (C) 2004 Keith M Wesolowski (wesolows@foobazco.org)
- *
- * Based on include/asm-sparc64/cpudata.h and Linux 2.4 smp.h
- * both (C) David S. Miller.
- */
-
-#ifndef _SPARC_CPUDATA_H
-#define _SPARC_CPUDATA_H
-
-#include <linux/percpu.h>
-
-typedef struct {
-	unsigned long udelay_val;
-	unsigned long clock_tick;
-	unsigned int multiplier;
-	unsigned int counter;
-	int prom_node;
-	int mid;
-	int next;
-} cpuinfo_sparc;
-
-DECLARE_PER_CPU(cpuinfo_sparc, __cpu_data);
-#define cpu_data(__cpu) per_cpu(__cpu_data, (__cpu))
-
-#endif /* _SPARC_CPUDATA_H */
+#ifndef ___ASM_SPARC_CPUDATA_H
+#define ___ASM_SPARC_CPUDATA_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/cpudata_64.h>
+#else
+#include <asm-sparc/cpudata_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/cpudata_32.h b/include/asm-sparc/cpudata_32.h
new file mode 100644
index 00000000000..a2c4d51d36c
--- /dev/null
+++ b/include/asm-sparc/cpudata_32.h
@@ -0,0 +1,27 @@
+/* cpudata.h: Per-cpu parameters.
+ *
+ * Copyright (C) 2004 Keith M Wesolowski (wesolows@foobazco.org)
+ *
+ * Based on include/asm-sparc64/cpudata.h and Linux 2.4 smp.h
+ * both (C) David S. Miller.
+ */
+
+#ifndef _SPARC_CPUDATA_H
+#define _SPARC_CPUDATA_H
+
+#include <linux/percpu.h>
+
+typedef struct {
+	unsigned long udelay_val;
+	unsigned long clock_tick;
+	unsigned int multiplier;
+	unsigned int counter;
+	int prom_node;
+	int mid;
+	int next;
+} cpuinfo_sparc;
+
+DECLARE_PER_CPU(cpuinfo_sparc, __cpu_data);
+#define cpu_data(__cpu) per_cpu(__cpu_data, (__cpu))
+
+#endif /* _SPARC_CPUDATA_H */
diff --git a/include/asm-sparc/cpudata_64.h b/include/asm-sparc/cpudata_64.h
new file mode 100644
index 00000000000..532975ecfe1
--- /dev/null
+++ b/include/asm-sparc/cpudata_64.h
@@ -0,0 +1,240 @@
+/* cpudata.h: Per-cpu parameters.
+ *
+ * Copyright (C) 2003, 2005, 2006 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC64_CPUDATA_H
+#define _SPARC64_CPUDATA_H
+
+#include <asm/hypervisor.h>
+#include <asm/asi.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/percpu.h>
+#include <linux/threads.h>
+
+typedef struct {
+	/* Dcache line 1 */
+	unsigned int	__softirq_pending; /* must be 1st, see rtrap.S */
+	unsigned int	__pad0;
+	unsigned long	clock_tick;	/* %tick's per second */
+	unsigned long	__pad;
+	unsigned int	__pad1;
+	unsigned int	__pad2;
+
+	/* Dcache line 2, rarely used */
+	unsigned int	dcache_size;
+	unsigned int	dcache_line_size;
+	unsigned int	icache_size;
+	unsigned int	icache_line_size;
+	unsigned int	ecache_size;
+	unsigned int	ecache_line_size;
+	int		core_id;
+	int		proc_id;
+} cpuinfo_sparc;
+
+DECLARE_PER_CPU(cpuinfo_sparc, __cpu_data);
+#define cpu_data(__cpu)		per_cpu(__cpu_data, (__cpu))
+#define local_cpu_data()	__get_cpu_var(__cpu_data)
+
+/* Trap handling code needs to get at a few critical values upon
+ * trap entry and to process TSB misses.  These cannot be in the
+ * per_cpu() area as we really need to lock them into the TLB and
+ * thus make them part of the main kernel image.  As a result we
+ * try to make this as small as possible.
+ *
+ * This is padded out and aligned to 64-bytes to avoid false sharing
+ * on SMP.
+ */
+
+/* If you modify the size of this structure, please update
+ * TRAP_BLOCK_SZ_SHIFT below.
+ */
+struct thread_info;
+struct trap_per_cpu {
+/* D-cache line 1: Basic thread information, cpu and device mondo queues */
+	struct thread_info	*thread;
+	unsigned long		pgd_paddr;
+	unsigned long		cpu_mondo_pa;
+	unsigned long		dev_mondo_pa;
+
+/* D-cache line 2: Error Mondo Queue and kernel buffer pointers */
+	unsigned long		resum_mondo_pa;
+	unsigned long		resum_kernel_buf_pa;
+	unsigned long		nonresum_mondo_pa;
+	unsigned long		nonresum_kernel_buf_pa;
+
+/* Dcache lines 3, 4, 5, and 6: Hypervisor Fault Status */
+	struct hv_fault_status	fault_info;
+
+/* Dcache line 7: Physical addresses of CPU send mondo block and CPU list.  */
+	unsigned long		cpu_mondo_block_pa;
+	unsigned long		cpu_list_pa;
+	unsigned long		tsb_huge;
+	unsigned long		tsb_huge_temp;
+
+/* Dcache line 8: IRQ work list, and keep trap_block a power-of-2 in size.  */
+	unsigned long		irq_worklist_pa;
+	unsigned int		cpu_mondo_qmask;
+	unsigned int		dev_mondo_qmask;
+	unsigned int		resum_qmask;
+	unsigned int		nonresum_qmask;
+	void			*hdesc;
+} __attribute__((aligned(64)));
+extern struct trap_per_cpu trap_block[NR_CPUS];
+extern void init_cur_cpu_trap(struct thread_info *);
+extern void setup_tba(void);
+extern int ncpus_probed;
+extern void __init cpu_probe(void);
+extern const struct seq_operations cpuinfo_op;
+
+extern unsigned long real_hard_smp_processor_id(void);
+
+struct cpuid_patch_entry {
+	unsigned int	addr;
+	unsigned int	cheetah_safari[4];
+	unsigned int	cheetah_jbus[4];
+	unsigned int	starfire[4];
+	unsigned int	sun4v[4];
+};
+extern struct cpuid_patch_entry __cpuid_patch, __cpuid_patch_end;
+
+struct sun4v_1insn_patch_entry {
+	unsigned int	addr;
+	unsigned int	insn;
+};
+extern struct sun4v_1insn_patch_entry __sun4v_1insn_patch,
+	__sun4v_1insn_patch_end;
+
+struct sun4v_2insn_patch_entry {
+	unsigned int	addr;
+	unsigned int	insns[2];
+};
+extern struct sun4v_2insn_patch_entry __sun4v_2insn_patch,
+	__sun4v_2insn_patch_end;
+
+#endif /* !(__ASSEMBLY__) */
+
+#define TRAP_PER_CPU_THREAD		0x00
+#define TRAP_PER_CPU_PGD_PADDR		0x08
+#define TRAP_PER_CPU_CPU_MONDO_PA	0x10
+#define TRAP_PER_CPU_DEV_MONDO_PA	0x18
+#define TRAP_PER_CPU_RESUM_MONDO_PA	0x20
+#define TRAP_PER_CPU_RESUM_KBUF_PA	0x28
+#define TRAP_PER_CPU_NONRESUM_MONDO_PA	0x30
+#define TRAP_PER_CPU_NONRESUM_KBUF_PA	0x38
+#define TRAP_PER_CPU_FAULT_INFO		0x40
+#define TRAP_PER_CPU_CPU_MONDO_BLOCK_PA	0xc0
+#define TRAP_PER_CPU_CPU_LIST_PA	0xc8
+#define TRAP_PER_CPU_TSB_HUGE		0xd0
+#define TRAP_PER_CPU_TSB_HUGE_TEMP	0xd8
+#define TRAP_PER_CPU_IRQ_WORKLIST_PA	0xe0
+#define TRAP_PER_CPU_CPU_MONDO_QMASK	0xe8
+#define TRAP_PER_CPU_DEV_MONDO_QMASK	0xec
+#define TRAP_PER_CPU_RESUM_QMASK	0xf0
+#define TRAP_PER_CPU_NONRESUM_QMASK	0xf4
+
+#define TRAP_BLOCK_SZ_SHIFT		8
+
+#include <asm/scratchpad.h>
+
+#define __GET_CPUID(REG)				\
+	/* Spitfire implementation (default). */	\
+661:	ldxa		[%g0] ASI_UPA_CONFIG, REG;	\
+	srlx		REG, 17, REG;			\
+	 and		REG, 0x1f, REG;			\
+	nop;						\
+	.section	.cpuid_patch, "ax";		\
+	/* Instruction location. */			\
+	.word		661b;				\
+	/* Cheetah Safari implementation. */		\
+	ldxa		[%g0] ASI_SAFARI_CONFIG, REG;	\
+	srlx		REG, 17, REG;			\
+	and		REG, 0x3ff, REG;		\
+	nop;						\
+	/* Cheetah JBUS implementation. */		\
+	ldxa		[%g0] ASI_JBUS_CONFIG, REG;	\
+	srlx		REG, 17, REG;			\
+	and		REG, 0x1f, REG;			\
+	nop;						\
+	/* Starfire implementation. */			\
+	sethi		%hi(0x1fff40000d0 >> 9), REG;	\
+	sllx		REG, 9, REG;			\
+	or		REG, 0xd0, REG;			\
+	lduwa		[REG] ASI_PHYS_BYPASS_EC_E, REG;\
+	/* sun4v implementation. */			\
+	mov		SCRATCHPAD_CPUID, REG;		\
+	ldxa		[REG] ASI_SCRATCHPAD, REG;	\
+	nop;						\
+	nop;						\
+	.previous;
+
+#ifdef CONFIG_SMP
+
+#define TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	__GET_CPUID(TMP)			\
+	sethi	%hi(trap_block), DEST;		\
+	sllx	TMP, TRAP_BLOCK_SZ_SHIFT, TMP;	\
+	or	DEST, %lo(trap_block), DEST;	\
+	add	DEST, TMP, DEST;		\
+
+/* Clobbers TMP, current address space PGD phys address into DEST.  */
+#define TRAP_LOAD_PGD_PHYS(DEST, TMP)		\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	ldx	[DEST + TRAP_PER_CPU_PGD_PADDR], DEST;
+
+/* Clobbers TMP, loads local processor's IRQ work area into DEST.  */
+#define TRAP_LOAD_IRQ_WORK_PA(DEST, TMP)	\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	add	DEST, TRAP_PER_CPU_IRQ_WORKLIST_PA, DEST;
+
+/* Clobbers TMP, loads DEST with current thread info pointer.  */
+#define TRAP_LOAD_THREAD_REG(DEST, TMP)		\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	ldx	[DEST + TRAP_PER_CPU_THREAD], DEST;
+
+/* Given the current thread info pointer in THR, load the per-cpu
+ * area base of the current processor into DEST.  REG1, REG2, and REG3 are
+ * clobbered.
+ *
+ * You absolutely cannot use DEST as a temporary in this code.  The
+ * reason is that traps can happen during execution, and return from
+ * trap will load the fully resolved DEST per-cpu base.  This can corrupt
+ * the calculations done by the macro mid-stream.
+ */
+#define LOAD_PER_CPU_BASE(DEST, THR, REG1, REG2, REG3)	\
+	lduh	[THR + TI_CPU], REG1;			\
+	sethi	%hi(__per_cpu_shift), REG3;		\
+	sethi	%hi(__per_cpu_base), REG2;		\
+	ldx	[REG3 + %lo(__per_cpu_shift)], REG3;	\
+	ldx	[REG2 + %lo(__per_cpu_base)], REG2;	\
+	sllx	REG1, REG3, REG3;			\
+	add	REG3, REG2, DEST;
+
+#else
+
+#define TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	sethi	%hi(trap_block), DEST;		\
+	or	DEST, %lo(trap_block), DEST;	\
+
+/* Uniprocessor versions, we know the cpuid is zero.  */
+#define TRAP_LOAD_PGD_PHYS(DEST, TMP)		\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	ldx	[DEST + TRAP_PER_CPU_PGD_PADDR], DEST;
+
+/* Clobbers TMP, loads local processor's IRQ work area into DEST.  */
+#define TRAP_LOAD_IRQ_WORK_PA(DEST, TMP)	\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	add	DEST, TRAP_PER_CPU_IRQ_WORKLIST_PA, DEST;
+
+#define TRAP_LOAD_THREAD_REG(DEST, TMP)		\
+	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
+	ldx	[DEST + TRAP_PER_CPU_THREAD], DEST;
+
+/* No per-cpu areas on uniprocessor, so no need to load DEST.  */
+#define LOAD_PER_CPU_BASE(DEST, THR, REG1, REG2, REG3)
+
+#endif /* !(CONFIG_SMP) */
+
+#endif /* _SPARC64_CPUDATA_H */
diff --git a/include/asm-sparc/dcr.h b/include/asm-sparc/dcr.h
new file mode 100644
index 00000000000..620c9ba642e
--- /dev/null
+++ b/include/asm-sparc/dcr.h
@@ -0,0 +1,14 @@
+#ifndef _SPARC64_DCR_H
+#define _SPARC64_DCR_H
+
+/* UltraSparc-III/III+ Dispatch Control Register, ASR 0x12 */
+#define DCR_DPE		0x0000000000001000 /* III+: D$ Parity Error Enable	*/
+#define DCR_OBS		0x0000000000000fc0 /* Observability Bus Controls	*/
+#define DCR_BPE		0x0000000000000020 /* Branch Predict Enable		*/
+#define DCR_RPE		0x0000000000000010 /* Return Address Prediction Enable	*/
+#define DCR_SI		0x0000000000000008 /* Single Instruction Disable	*/
+#define DCR_IPE		0x0000000000000004 /* III+: I$ Parity Error Enable	*/
+#define DCR_IFPOE	0x0000000000000002 /* IRQ FP Operation Enable		*/
+#define DCR_MS		0x0000000000000001 /* Multi-Scalar dispatch		*/
+
+#endif /* _SPARC64_DCR_H */
diff --git a/include/asm-sparc/dcu.h b/include/asm-sparc/dcu.h
new file mode 100644
index 00000000000..0f704e106a1
--- /dev/null
+++ b/include/asm-sparc/dcu.h
@@ -0,0 +1,27 @@
+#ifndef _SPARC64_DCU_H
+#define _SPARC64_DCU_H
+
+#include <linux/const.h>
+
+/* UltraSparc-III Data Cache Unit Control Register */
+#define DCU_CP	_AC(0x0002000000000000,UL) /* Phys Cache Enable w/o mmu	*/
+#define DCU_CV	_AC(0x0001000000000000,UL) /* Virt Cache Enable w/o mmu	*/
+#define DCU_ME	_AC(0x0000800000000000,UL) /* NC-store Merging Enable	*/
+#define DCU_RE	_AC(0x0000400000000000,UL) /* RAW bypass Enable		*/
+#define DCU_PE	_AC(0x0000200000000000,UL) /* PCache Enable		*/
+#define DCU_HPE	_AC(0x0000100000000000,UL) /* HW prefetch Enable	*/
+#define DCU_SPE	_AC(0x0000080000000000,UL) /* SW prefetch Enable	*/
+#define DCU_SL	_AC(0x0000040000000000,UL) /* Secondary ld-steering Enab*/
+#define DCU_WE	_AC(0x0000020000000000,UL) /* WCache enable		*/
+#define DCU_PM	_AC(0x000001fe00000000,UL) /* PA Watchpoint Byte Mask	*/
+#define DCU_VM	_AC(0x00000001fe000000,UL) /* VA Watchpoint Byte Mask	*/
+#define DCU_PR	_AC(0x0000000001000000,UL) /* PA Watchpoint Read Enable	*/
+#define DCU_PW	_AC(0x0000000000800000,UL) /* PA Watchpoint Write Enable*/
+#define DCU_VR	_AC(0x0000000000400000,UL) /* VA Watchpoint Read Enable	*/
+#define DCU_VW	_AC(0x0000000000200000,UL) /* VA Watchpoint Write Enable*/
+#define DCU_DM	_AC(0x0000000000000008,UL) /* DMMU Enable		*/
+#define DCU_IM	_AC(0x0000000000000004,UL) /* IMMU Enable		*/
+#define DCU_DC	_AC(0x0000000000000002,UL) /* Data Cache Enable		*/
+#define DCU_IC	_AC(0x0000000000000001,UL) /* Instruction Cache Enable	*/
+
+#endif /* _SPARC64_DCU_H */
diff --git a/include/asm-sparc/delay.h b/include/asm-sparc/delay.h
index bc9aba2bead..6210a3ce975 100644
--- a/include/asm-sparc/delay.h
+++ b/include/asm-sparc/delay.h
@@ -1,34 +1,8 @@
-/*
- * delay.h: Linux delay routines on the Sparc.
- *
- * Copyright (C) 1994 David S. Miller (davem@caip.rutgers.edu).
- */
-
-#ifndef __SPARC_DELAY_H
-#define __SPARC_DELAY_H
-
-#include <asm/cpudata.h>
-
-static inline void __delay(unsigned long loops)
-{
-	__asm__ __volatile__("cmp %0, 0\n\t"
-			     "1: bne 1b\n\t"
-			     "subcc %0, 1, %0\n" :
-			     "=&r" (loops) :
-			     "0" (loops) :
-			     "cc");
-}
-
-/* This is too messy with inline asm on the Sparc. */
-extern void __udelay(unsigned long usecs, unsigned long lpj);
-extern void __ndelay(unsigned long nsecs, unsigned long lpj);
-
-#ifdef CONFIG_SMP
-#define __udelay_val	cpu_data(smp_processor_id()).udelay_val
-#else /* SMP */
-#define __udelay_val	loops_per_jiffy
-#endif /* SMP */
-#define udelay(__usecs)	__udelay(__usecs, __udelay_val)
-#define ndelay(__nsecs)	__ndelay(__nsecs, __udelay_val)
-
-#endif /* defined(__SPARC_DELAY_H) */
+#ifndef ___ASM_SPARC_DELAY_H
+#define ___ASM_SPARC_DELAY_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/delay_64.h>
+#else
+#include <asm-sparc/delay_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/delay_32.h b/include/asm-sparc/delay_32.h
new file mode 100644
index 00000000000..bc9aba2bead
--- /dev/null
+++ b/include/asm-sparc/delay_32.h
@@ -0,0 +1,34 @@
+/*
+ * delay.h: Linux delay routines on the Sparc.
+ *
+ * Copyright (C) 1994 David S. Miller (davem@caip.rutgers.edu).
+ */
+
+#ifndef __SPARC_DELAY_H
+#define __SPARC_DELAY_H
+
+#include <asm/cpudata.h>
+
+static inline void __delay(unsigned long loops)
+{
+	__asm__ __volatile__("cmp %0, 0\n\t"
+			     "1: bne 1b\n\t"
+			     "subcc %0, 1, %0\n" :
+			     "=&r" (loops) :
+			     "0" (loops) :
+			     "cc");
+}
+
+/* This is too messy with inline asm on the Sparc. */
+extern void __udelay(unsigned long usecs, unsigned long lpj);
+extern void __ndelay(unsigned long nsecs, unsigned long lpj);
+
+#ifdef CONFIG_SMP
+#define __udelay_val	cpu_data(smp_processor_id()).udelay_val
+#else /* SMP */
+#define __udelay_val	loops_per_jiffy
+#endif /* SMP */
+#define udelay(__usecs)	__udelay(__usecs, __udelay_val)
+#define ndelay(__nsecs)	__ndelay(__nsecs, __udelay_val)
+
+#endif /* defined(__SPARC_DELAY_H) */
diff --git a/include/asm-sparc/delay_64.h b/include/asm-sparc/delay_64.h
new file mode 100644
index 00000000000..a77aa622d76
--- /dev/null
+++ b/include/asm-sparc/delay_64.h
@@ -0,0 +1,17 @@
+/* delay.h: Linux delay routines on sparc64.
+ *
+ * Copyright (C) 1996, 2004, 2007 David S. Miller (davem@davemloft.net).
+ */
+
+#ifndef _SPARC64_DELAY_H
+#define _SPARC64_DELAY_H
+
+#ifndef __ASSEMBLY__
+
+extern void __delay(unsigned long loops);
+extern void udelay(unsigned long usecs);
+#define mdelay(n)	udelay((n) * 1000)
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* _SPARC64_DELAY_H */
diff --git a/include/asm-sparc/display7seg.h b/include/asm-sparc/display7seg.h
new file mode 100644
index 00000000000..86d4a901df2
--- /dev/null
+++ b/include/asm-sparc/display7seg.h
@@ -0,0 +1,79 @@
+/*
+ *
+ * display7seg - Driver interface for the 7-segment display
+ * present on Sun Microsystems CP1400 and CP1500
+ *
+ * Copyright (c) 2000 Eric Brower <ebrower@usa.net>
+ *
+ */
+
+#ifndef __display7seg_h__
+#define __display7seg_h__
+
+#define D7S_IOC	'p'
+
+#define D7SIOCRD _IOR(D7S_IOC, 0x45, int)	/* Read device state	*/
+#define D7SIOCWR _IOW(D7S_IOC, 0x46, int)	/* Write device state	*/
+#define D7SIOCTM _IO (D7S_IOC, 0x47)		/* Translate mode (FLIP)*/
+
+/*
+ * ioctl flag definitions
+ *
+ * POINT	- Toggle decimal point	(0=absent 1=present)
+ * ALARM	- Toggle alarm LED 		(0=green  1=red)
+ * FLIP		- Toggle inverted mode 	(0=normal 1=flipped)
+ * bits 0-4	- Character displayed	(see definitions below)
+ *
+ * Display segments are defined as follows,
+ * subject to D7S_FLIP register state:
+ *
+ *    a
+ *   ---
+ * f|   |b
+ *   -g-
+ * e|   |c
+ *   ---
+ *    d
+ */
+
+#define D7S_POINT	(1 << 7)	/* Decimal point*/
+#define D7S_ALARM	(1 << 6)	/* Alarm LED 	*/
+#define D7S_FLIP	(1 << 5)	/* Flip display */
+
+#define D7S_0		0x00		/* Numerals 0-9 */
+#define D7S_1		0x01
+#define D7S_2		0x02
+#define D7S_3		0x03
+#define D7S_4		0x04
+#define D7S_5		0x05
+#define D7S_6		0x06
+#define D7S_7		0x07
+#define D7S_8		0x08
+#define D7S_9		0x09
+#define D7S_A		0x0A		/* Letters A-F, H, L, P */
+#define D7S_B		0x0B
+#define D7S_C		0x0C
+#define D7S_D		0x0D
+#define D7S_E		0x0E
+#define D7S_F		0x0F
+#define D7S_H		0x10
+#define D7S_E2		0x11
+#define D7S_L		0x12
+#define D7S_P		0x13
+#define D7S_SEGA	0x14		/* Individual segments */
+#define D7S_SEGB	0x15
+#define D7S_SEGC	0x16
+#define D7S_SEGD	0x17
+#define D7S_SEGE	0x18
+#define D7S_SEGF	0x19
+#define D7S_SEGG	0x1A
+#define D7S_SEGABFG 0x1B		/* Segment groupings */
+#define D7S_SEGCDEG	0x1C
+#define D7S_SEGBCEF 0x1D
+#define D7S_SEGADG	0x1E
+#define D7S_BLANK	0x1F		/* Clear all segments */
+
+#define D7S_MIN_VAL	0x0
+#define D7S_MAX_VAL	0x1F
+
+#endif /* ifndef __display7seg_h__ */
diff --git a/include/asm-sparc/dma-mapping.h b/include/asm-sparc/dma-mapping.h
index f3a641e6b2c..7483504259c 100644
--- a/include/asm-sparc/dma-mapping.h
+++ b/include/asm-sparc/dma-mapping.h
@@ -1,11 +1,8 @@
-#ifndef _ASM_SPARC_DMA_MAPPING_H
-#define _ASM_SPARC_DMA_MAPPING_H
-
-
-#ifdef CONFIG_PCI
-#include <asm-generic/dma-mapping.h>
+#ifndef ___ASM_SPARC_DMA_MAPPING_H
+#define ___ASM_SPARC_DMA_MAPPING_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/dma-mapping_64.h>
 #else
-#include <asm-generic/dma-mapping-broken.h>
-#endif /* PCI */
-
-#endif /* _ASM_SPARC_DMA_MAPPING_H */
+#include <asm-sparc/dma-mapping_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/dma-mapping_32.h b/include/asm-sparc/dma-mapping_32.h
new file mode 100644
index 00000000000..f3a641e6b2c
--- /dev/null
+++ b/include/asm-sparc/dma-mapping_32.h
@@ -0,0 +1,11 @@
+#ifndef _ASM_SPARC_DMA_MAPPING_H
+#define _ASM_SPARC_DMA_MAPPING_H
+
+
+#ifdef CONFIG_PCI
+#include <asm-generic/dma-mapping.h>
+#else
+#include <asm-generic/dma-mapping-broken.h>
+#endif /* PCI */
+
+#endif /* _ASM_SPARC_DMA_MAPPING_H */
diff --git a/include/asm-sparc/dma-mapping_64.h b/include/asm-sparc/dma-mapping_64.h
new file mode 100644
index 00000000000..38cbec76a33
--- /dev/null
+++ b/include/asm-sparc/dma-mapping_64.h
@@ -0,0 +1,154 @@
+#ifndef _ASM_SPARC64_DMA_MAPPING_H
+#define _ASM_SPARC64_DMA_MAPPING_H
+
+#include <linux/scatterlist.h>
+#include <linux/mm.h>
+
+#define DMA_ERROR_CODE	(~(dma_addr_t)0x0)
+
+struct dma_ops {
+	void *(*alloc_coherent)(struct device *dev, size_t size,
+				dma_addr_t *dma_handle, gfp_t flag);
+	void (*free_coherent)(struct device *dev, size_t size,
+			      void *cpu_addr, dma_addr_t dma_handle);
+	dma_addr_t (*map_single)(struct device *dev, void *cpu_addr,
+				 size_t size,
+				 enum dma_data_direction direction);
+	void (*unmap_single)(struct device *dev, dma_addr_t dma_addr,
+			     size_t size,
+			     enum dma_data_direction direction);
+	int (*map_sg)(struct device *dev, struct scatterlist *sg, int nents,
+		      enum dma_data_direction direction);
+	void (*unmap_sg)(struct device *dev, struct scatterlist *sg,
+			 int nhwentries,
+			 enum dma_data_direction direction);
+	void (*sync_single_for_cpu)(struct device *dev,
+				    dma_addr_t dma_handle, size_t size,
+				    enum dma_data_direction direction);
+	void (*sync_sg_for_cpu)(struct device *dev, struct scatterlist *sg,
+				int nelems,
+				enum dma_data_direction direction);
+};
+extern const struct dma_ops *dma_ops;
+
+extern int dma_supported(struct device *dev, u64 mask);
+extern int dma_set_mask(struct device *dev, u64 dma_mask);
+
+static inline void *dma_alloc_coherent(struct device *dev, size_t size,
+				       dma_addr_t *dma_handle, gfp_t flag)
+{
+	return dma_ops->alloc_coherent(dev, size, dma_handle, flag);
+}
+
+static inline void dma_free_coherent(struct device *dev, size_t size,
+				     void *cpu_addr, dma_addr_t dma_handle)
+{
+	dma_ops->free_coherent(dev, size, cpu_addr, dma_handle);
+}
+
+static inline dma_addr_t dma_map_single(struct device *dev, void *cpu_addr,
+					size_t size,
+					enum dma_data_direction direction)
+{
+	return dma_ops->map_single(dev, cpu_addr, size, direction);
+}
+
+static inline void dma_unmap_single(struct device *dev, dma_addr_t dma_addr,
+				    size_t size,
+				    enum dma_data_direction direction)
+{
+	dma_ops->unmap_single(dev, dma_addr, size, direction);
+}
+
+static inline dma_addr_t dma_map_page(struct device *dev, struct page *page,
+				      unsigned long offset, size_t size,
+				      enum dma_data_direction direction)
+{
+	return dma_ops->map_single(dev, page_address(page) + offset,
+				   size, direction);
+}
+
+static inline void dma_unmap_page(struct device *dev, dma_addr_t dma_address,
+				  size_t size,
+				  enum dma_data_direction direction)
+{
+	dma_ops->unmap_single(dev, dma_address, size, direction);
+}
+
+static inline int dma_map_sg(struct device *dev, struct scatterlist *sg,
+			     int nents, enum dma_data_direction direction)
+{
+	return dma_ops->map_sg(dev, sg, nents, direction);
+}
+
+static inline void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
+				int nents, enum dma_data_direction direction)
+{
+	dma_ops->unmap_sg(dev, sg, nents, direction);
+}
+
+static inline void dma_sync_single_for_cpu(struct device *dev,
+					   dma_addr_t dma_handle, size_t size,
+					   enum dma_data_direction direction)
+{
+	dma_ops->sync_single_for_cpu(dev, dma_handle, size, direction);
+}
+
+static inline void dma_sync_single_for_device(struct device *dev,
+					      dma_addr_t dma_handle,
+					      size_t size,
+					      enum dma_data_direction direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+static inline void dma_sync_single_range_for_cpu(struct device *dev,
+						 dma_addr_t dma_handle,
+						 unsigned long offset,
+						 size_t size,
+						 enum dma_data_direction direction)
+{
+	dma_sync_single_for_cpu(dev, dma_handle+offset, size, direction);
+}
+
+static inline void dma_sync_single_range_for_device(struct device *dev,
+						    dma_addr_t dma_handle,
+						    unsigned long offset,
+						    size_t size,
+						    enum dma_data_direction direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+
+static inline void dma_sync_sg_for_cpu(struct device *dev,
+				       struct scatterlist *sg, int nelems,
+				       enum dma_data_direction direction)
+{
+	dma_ops->sync_sg_for_cpu(dev, sg, nelems, direction);
+}
+
+static inline void dma_sync_sg_for_device(struct device *dev,
+					  struct scatterlist *sg, int nelems,
+					  enum dma_data_direction direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+static inline int dma_mapping_error(dma_addr_t dma_addr)
+{
+	return (dma_addr == DMA_ERROR_CODE);
+}
+
+static inline int dma_get_cache_alignment(void)
+{
+	/* no easy way to get cache size on all processors, so return
+	 * the maximum possible, to be safe */
+	return (1 << INTERNODE_CACHE_SHIFT);
+}
+
+#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
+#define dma_free_noncoherent(d, s, v, h) dma_free_coherent(d, s, v, h)
+#define dma_is_consistent(d, h)	(1)
+
+#endif /* _ASM_SPARC64_DMA_MAPPING_H */
diff --git a/include/asm-sparc/dma.h b/include/asm-sparc/dma.h
index 959d6c8a71a..8cc69bfaae2 100644
--- a/include/asm-sparc/dma.h
+++ b/include/asm-sparc/dma.h
@@ -1,288 +1,8 @@
-/* include/asm-sparc/dma.h
- *
- * Copyright 1995 (C) David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _ASM_SPARC_DMA_H
-#define _ASM_SPARC_DMA_H
-
-#include <linux/kernel.h>
-#include <linux/types.h>
-
-#include <asm/vac-ops.h>  /* for invalidate's, etc. */
-#include <asm/sbus.h>
-#include <asm/delay.h>
-#include <asm/oplib.h>
-#include <asm/system.h>
-#include <asm/io.h>
-#include <linux/spinlock.h>
-
-struct page;
-extern spinlock_t  dma_spin_lock;
-
-static inline unsigned long claim_dma_lock(void)
-{
-	unsigned long flags;
-	spin_lock_irqsave(&dma_spin_lock, flags);
-	return flags;
-}
-
-static inline void release_dma_lock(unsigned long flags)
-{
-	spin_unlock_irqrestore(&dma_spin_lock, flags);
-}
-
-/* These are irrelevant for Sparc DMA, but we leave it in so that
- * things can compile.
- */
-#define MAX_DMA_CHANNELS 8
-#define MAX_DMA_ADDRESS  (~0UL)
-#define DMA_MODE_READ    1
-#define DMA_MODE_WRITE   2
-
-/* Useful constants */
-#define SIZE_16MB      (16*1024*1024)
-#define SIZE_64K       (64*1024)
-
-/* SBUS DMA controller reg offsets */
-#define DMA_CSR		0x00UL		/* rw  DMA control/status register    0x00   */
-#define DMA_ADDR	0x04UL		/* rw  DMA transfer address register  0x04   */
-#define DMA_COUNT	0x08UL		/* rw  DMA transfer count register    0x08   */
-#define DMA_TEST	0x0cUL		/* rw  DMA test/debug register        0x0c   */
-
-/* DVMA chip revisions */
-enum dvma_rev {
-	dvmarev0,
-	dvmaesc1,
-	dvmarev1,
-	dvmarev2,
-	dvmarev3,
-	dvmarevplus,
-	dvmahme
-};
-
-#define DMA_HASCOUNT(rev)  ((rev)==dvmaesc1)
-
-/* Linux DMA information structure, filled during probe. */
-struct sbus_dma {
-	struct sbus_dma *next;
-	struct sbus_dev *sdev;
-	void __iomem *regs;
-
-	/* Status, misc info */
-	int node;                /* Prom node for this DMA device */
-	int running;             /* Are we doing DMA now? */
-	int allocated;           /* Are we "owned" by anyone yet? */
-
-	/* Transfer information. */
-	unsigned long addr;      /* Start address of current transfer */
-	int nbytes;              /* Size of current transfer */
-	int realbytes;           /* For splitting up large transfers, etc. */
-
-	/* DMA revision */
-	enum dvma_rev revision;
-};
-
-extern struct sbus_dma *dma_chain;
-
-/* Broken hardware... */
-#ifdef CONFIG_SUN4
-/* Have to sort this out. Does rev0 work fine on sun4[cmd] without isbroken?
- * Or is rev0 present only on sun4 boxes? -jj */
-#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev0 || (dma)->revision == dvmarev1)
+#ifndef ___ASM_SPARC_DMA_H
+#define ___ASM_SPARC_DMA_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/dma_64.h>
 #else
-#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev1)
+#include <asm-sparc/dma_32.h>
 #endif
-#define DMA_ISESC1(dma)      ((dma)->revision == dvmaesc1)
-
-/* Main routines in dma.c */
-extern void dvma_init(struct sbus_bus *);
-
-/* Fields in the cond_reg register */
-/* First, the version identification bits */
-#define DMA_DEVICE_ID    0xf0000000        /* Device identification bits */
-#define DMA_VERS0        0x00000000        /* Sunray DMA version */
-#define DMA_ESCV1        0x40000000        /* DMA ESC Version 1 */
-#define DMA_VERS1        0x80000000        /* DMA rev 1 */
-#define DMA_VERS2        0xa0000000        /* DMA rev 2 */
-#define DMA_VERHME       0xb0000000        /* DMA hme gate array */
-#define DMA_VERSPLUS     0x90000000        /* DMA rev 1 PLUS */
-
-#define DMA_HNDL_INTR    0x00000001        /* An IRQ needs to be handled */
-#define DMA_HNDL_ERROR   0x00000002        /* We need to take an error */
-#define DMA_FIFO_ISDRAIN 0x0000000c        /* The DMA FIFO is draining */
-#define DMA_INT_ENAB     0x00000010        /* Turn on interrupts */
-#define DMA_FIFO_INV     0x00000020        /* Invalidate the FIFO */
-#define DMA_ACC_SZ_ERR   0x00000040        /* The access size was bad */
-#define DMA_FIFO_STDRAIN 0x00000040        /* DMA_VERS1 Drain the FIFO */
-#define DMA_RST_SCSI     0x00000080        /* Reset the SCSI controller */
-#define DMA_RST_ENET     DMA_RST_SCSI      /* Reset the ENET controller */
-#define DMA_RST_BPP      DMA_RST_SCSI      /* Reset the BPP controller */
-#define DMA_ST_WRITE     0x00000100        /* write from device to memory */
-#define DMA_ENABLE       0x00000200        /* Fire up DMA, handle requests */
-#define DMA_PEND_READ    0x00000400        /* DMA_VERS1/0/PLUS Pending Read */
-#define DMA_ESC_BURST    0x00000800        /* 1=16byte 0=32byte */
-#define DMA_READ_AHEAD   0x00001800        /* DMA read ahead partial longword */
-#define DMA_DSBL_RD_DRN  0x00001000        /* No EC drain on slave reads */
-#define DMA_BCNT_ENAB    0x00002000        /* If on, use the byte counter */
-#define DMA_TERM_CNTR    0x00004000        /* Terminal counter */
-#define DMA_SCSI_SBUS64  0x00008000        /* HME: Enable 64-bit SBUS mode. */
-#define DMA_CSR_DISAB    0x00010000        /* No FIFO drains during csr */
-#define DMA_SCSI_DISAB   0x00020000        /* No FIFO drains during reg */
-#define DMA_DSBL_WR_INV  0x00020000        /* No EC inval. on slave writes */
-#define DMA_ADD_ENABLE   0x00040000        /* Special ESC DVMA optimization */
-#define DMA_E_BURSTS	 0x000c0000	   /* ENET: SBUS r/w burst mask */
-#define DMA_E_BURST32	 0x00040000	   /* ENET: SBUS 32 byte r/w burst */
-#define DMA_E_BURST16	 0x00000000	   /* ENET: SBUS 16 byte r/w burst */
-#define DMA_BRST_SZ      0x000c0000        /* SCSI: SBUS r/w burst size */
-#define DMA_BRST64       0x00080000        /* SCSI: 64byte bursts (HME on UltraSparc only) */
-#define DMA_BRST32       0x00040000        /* SCSI/BPP: 32byte bursts */
-#define DMA_BRST16       0x00000000        /* SCSI/BPP: 16byte bursts */
-#define DMA_BRST0        0x00080000        /* SCSI: no bursts (non-HME gate arrays) */
-#define DMA_ADDR_DISAB   0x00100000        /* No FIFO drains during addr */
-#define DMA_2CLKS        0x00200000        /* Each transfer = 2 clock ticks */
-#define DMA_3CLKS        0x00400000        /* Each transfer = 3 clock ticks */
-#define DMA_EN_ENETAUI   DMA_3CLKS         /* Put lance into AUI-cable mode */
-#define DMA_CNTR_DISAB   0x00800000        /* No IRQ when DMA_TERM_CNTR set */
-#define DMA_AUTO_NADDR   0x01000000        /* Use "auto nxt addr" feature */
-#define DMA_SCSI_ON      0x02000000        /* Enable SCSI dma */
-#define DMA_BPP_ON       DMA_SCSI_ON       /* Enable BPP dma */
-#define DMA_PARITY_OFF   0x02000000        /* HME: disable parity checking */
-#define DMA_LOADED_ADDR  0x04000000        /* Address has been loaded */
-#define DMA_LOADED_NADDR 0x08000000        /* Next address has been loaded */
-#define DMA_RESET_FAS366 0x08000000        /* HME: Assert RESET to FAS366 */
-
-/* Values describing the burst-size property from the PROM */
-#define DMA_BURST1       0x01
-#define DMA_BURST2       0x02
-#define DMA_BURST4       0x04
-#define DMA_BURST8       0x08
-#define DMA_BURST16      0x10
-#define DMA_BURST32      0x20
-#define DMA_BURST64      0x40
-#define DMA_BURSTBITS    0x7f
-
-/* Determine highest possible final transfer address given a base */
-#define DMA_MAXEND(addr) (0x01000000UL-(((unsigned long)(addr))&0x00ffffffUL))
-
-/* Yes, I hack a lot of elisp in my spare time... */
-#define DMA_ERROR_P(regs)  ((((regs)->cond_reg) & DMA_HNDL_ERROR))
-#define DMA_IRQ_P(regs)    ((((regs)->cond_reg) & (DMA_HNDL_INTR | DMA_HNDL_ERROR)))
-#define DMA_WRITE_P(regs)  ((((regs)->cond_reg) & DMA_ST_WRITE))
-#define DMA_OFF(regs)      ((((regs)->cond_reg) &= (~DMA_ENABLE)))
-#define DMA_INTSOFF(regs)  ((((regs)->cond_reg) &= (~DMA_INT_ENAB)))
-#define DMA_INTSON(regs)   ((((regs)->cond_reg) |= (DMA_INT_ENAB)))
-#define DMA_PUNTFIFO(regs) ((((regs)->cond_reg) |= DMA_FIFO_INV))
-#define DMA_SETSTART(regs, addr)  ((((regs)->st_addr) = (char *) addr))
-#define DMA_BEGINDMA_W(regs) \
-        ((((regs)->cond_reg |= (DMA_ST_WRITE|DMA_ENABLE|DMA_INT_ENAB))))
-#define DMA_BEGINDMA_R(regs) \
-        ((((regs)->cond_reg |= ((DMA_ENABLE|DMA_INT_ENAB)&(~DMA_ST_WRITE)))))
-
-/* For certain DMA chips, we need to disable ints upon irq entry
- * and turn them back on when we are done.  So in any ESP interrupt
- * handler you *must* call DMA_IRQ_ENTRY upon entry and DMA_IRQ_EXIT
- * when leaving the handler.  You have been warned...
- */
-#define DMA_IRQ_ENTRY(dma, dregs) do { \
-        if(DMA_ISBROKEN(dma)) DMA_INTSOFF(dregs); \
-   } while (0)
-
-#define DMA_IRQ_EXIT(dma, dregs) do { \
-	if(DMA_ISBROKEN(dma)) DMA_INTSON(dregs); \
-   } while(0)
-
-#if 0	/* P3 this stuff is inline in ledma.c:init_restart_ledma() */
-/* Pause until counter runs out or BIT isn't set in the DMA condition
- * register.
- */
-static inline void sparc_dma_pause(struct sparc_dma_registers *regs,
-				       unsigned long bit)
-{
-	int ctr = 50000;   /* Let's find some bugs ;) */
-
-	/* Busy wait until the bit is not set any more */
-	while((regs->cond_reg&bit) && (ctr>0)) {
-		ctr--;
-		__delay(5);
-	}
-
-	/* Check for bogus outcome. */
-	if(!ctr)
-		panic("DMA timeout");
-}
-
-/* Reset the friggin' thing... */
-#define DMA_RESET(dma) do { \
-	struct sparc_dma_registers *regs = dma->regs;                      \
-	/* Let the current FIFO drain itself */                            \
-	sparc_dma_pause(regs, (DMA_FIFO_ISDRAIN));                         \
-	/* Reset the logic */                                              \
-	regs->cond_reg |= (DMA_RST_SCSI);     /* assert */                 \
-	__delay(400);                         /* let the bits set ;) */    \
-	regs->cond_reg &= ~(DMA_RST_SCSI);    /* de-assert */              \
-	sparc_dma_enable_interrupts(regs);    /* Re-enable interrupts */   \
-	/* Enable FAST transfers if available */                           \
-	if(dma->revision>dvmarev1) regs->cond_reg |= DMA_3CLKS;            \
-	dma->running = 0;                                                  \
-} while(0)
 #endif
-
-#define for_each_dvma(dma) \
-        for((dma) = dma_chain; (dma); (dma) = (dma)->next)
-
-extern int get_dma_list(char *);
-extern int request_dma(unsigned int, __const__ char *);
-extern void free_dma(unsigned int);
-
-/* From PCI */
-
-#ifdef CONFIG_PCI
-extern int isa_dma_bridge_buggy;
-#else
-#define isa_dma_bridge_buggy	(0)
-#endif
-
-/* Routines for data transfer buffers. */
-BTFIXUPDEF_CALL(char *, mmu_lockarea, char *, unsigned long)
-BTFIXUPDEF_CALL(void,   mmu_unlockarea, char *, unsigned long)
-
-#define mmu_lockarea(vaddr,len) BTFIXUP_CALL(mmu_lockarea)(vaddr,len)
-#define mmu_unlockarea(vaddr,len) BTFIXUP_CALL(mmu_unlockarea)(vaddr,len)
-
-/* These are implementations for sbus_map_sg/sbus_unmap_sg... collapse later */
-BTFIXUPDEF_CALL(__u32, mmu_get_scsi_one, char *, unsigned long, struct sbus_bus *sbus)
-BTFIXUPDEF_CALL(void,  mmu_get_scsi_sgl, struct scatterlist *, int, struct sbus_bus *sbus)
-BTFIXUPDEF_CALL(void,  mmu_release_scsi_one, __u32, unsigned long, struct sbus_bus *sbus)
-BTFIXUPDEF_CALL(void,  mmu_release_scsi_sgl, struct scatterlist *, int, struct sbus_bus *sbus)
-
-#define mmu_get_scsi_one(vaddr,len,sbus) BTFIXUP_CALL(mmu_get_scsi_one)(vaddr,len,sbus)
-#define mmu_get_scsi_sgl(sg,sz,sbus) BTFIXUP_CALL(mmu_get_scsi_sgl)(sg,sz,sbus)
-#define mmu_release_scsi_one(vaddr,len,sbus) BTFIXUP_CALL(mmu_release_scsi_one)(vaddr,len,sbus)
-#define mmu_release_scsi_sgl(sg,sz,sbus) BTFIXUP_CALL(mmu_release_scsi_sgl)(sg,sz,sbus)
-
-/*
- * mmu_map/unmap are provided by iommu/iounit; Invalid to call on IIep.
- *
- * The mmu_map_dma_area establishes two mappings in one go.
- * These mappings point to pages normally mapped at 'va' (linear address).
- * First mapping is for CPU visible address at 'a', uncached.
- * This is an alias, but it works because it is an uncached mapping.
- * Second mapping is for device visible address, or "bus" address.
- * The bus address is returned at '*pba'.
- *
- * These functions seem distinct, but are hard to split. On sun4c,
- * at least for now, 'a' is equal to bus address, and retured in *pba.
- * On sun4m, page attributes depend on the CPU type, so we have to
- * know if we are mapping RAM or I/O, so it has to be an additional argument
- * to a separate mapping function for CPU visible mappings.
- */
-BTFIXUPDEF_CALL(int,  mmu_map_dma_area, dma_addr_t *, unsigned long, unsigned long, int len)
-BTFIXUPDEF_CALL(struct page *, mmu_translate_dvma, unsigned long busa)
-BTFIXUPDEF_CALL(void,  mmu_unmap_dma_area, unsigned long busa, int len)
-
-#define mmu_map_dma_area(pba,va,a,len) BTFIXUP_CALL(mmu_map_dma_area)(pba,va,a,len)
-#define mmu_unmap_dma_area(ba,len) BTFIXUP_CALL(mmu_unmap_dma_area)(ba,len)
-#define mmu_translate_dvma(ba)     BTFIXUP_CALL(mmu_translate_dvma)(ba)
-
-#endif /* !(_ASM_SPARC_DMA_H) */
diff --git a/include/asm-sparc/dma_32.h b/include/asm-sparc/dma_32.h
new file mode 100644
index 00000000000..959d6c8a71a
--- /dev/null
+++ b/include/asm-sparc/dma_32.h
@@ -0,0 +1,288 @@
+/* include/asm-sparc/dma.h
+ *
+ * Copyright 1995 (C) David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _ASM_SPARC_DMA_H
+#define _ASM_SPARC_DMA_H
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+
+#include <asm/vac-ops.h>  /* for invalidate's, etc. */
+#include <asm/sbus.h>
+#include <asm/delay.h>
+#include <asm/oplib.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <linux/spinlock.h>
+
+struct page;
+extern spinlock_t  dma_spin_lock;
+
+static inline unsigned long claim_dma_lock(void)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&dma_spin_lock, flags);
+	return flags;
+}
+
+static inline void release_dma_lock(unsigned long flags)
+{
+	spin_unlock_irqrestore(&dma_spin_lock, flags);
+}
+
+/* These are irrelevant for Sparc DMA, but we leave it in so that
+ * things can compile.
+ */
+#define MAX_DMA_CHANNELS 8
+#define MAX_DMA_ADDRESS  (~0UL)
+#define DMA_MODE_READ    1
+#define DMA_MODE_WRITE   2
+
+/* Useful constants */
+#define SIZE_16MB      (16*1024*1024)
+#define SIZE_64K       (64*1024)
+
+/* SBUS DMA controller reg offsets */
+#define DMA_CSR		0x00UL		/* rw  DMA control/status register    0x00   */
+#define DMA_ADDR	0x04UL		/* rw  DMA transfer address register  0x04   */
+#define DMA_COUNT	0x08UL		/* rw  DMA transfer count register    0x08   */
+#define DMA_TEST	0x0cUL		/* rw  DMA test/debug register        0x0c   */
+
+/* DVMA chip revisions */
+enum dvma_rev {
+	dvmarev0,
+	dvmaesc1,
+	dvmarev1,
+	dvmarev2,
+	dvmarev3,
+	dvmarevplus,
+	dvmahme
+};
+
+#define DMA_HASCOUNT(rev)  ((rev)==dvmaesc1)
+
+/* Linux DMA information structure, filled during probe. */
+struct sbus_dma {
+	struct sbus_dma *next;
+	struct sbus_dev *sdev;
+	void __iomem *regs;
+
+	/* Status, misc info */
+	int node;                /* Prom node for this DMA device */
+	int running;             /* Are we doing DMA now? */
+	int allocated;           /* Are we "owned" by anyone yet? */
+
+	/* Transfer information. */
+	unsigned long addr;      /* Start address of current transfer */
+	int nbytes;              /* Size of current transfer */
+	int realbytes;           /* For splitting up large transfers, etc. */
+
+	/* DMA revision */
+	enum dvma_rev revision;
+};
+
+extern struct sbus_dma *dma_chain;
+
+/* Broken hardware... */
+#ifdef CONFIG_SUN4
+/* Have to sort this out. Does rev0 work fine on sun4[cmd] without isbroken?
+ * Or is rev0 present only on sun4 boxes? -jj */
+#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev0 || (dma)->revision == dvmarev1)
+#else
+#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev1)
+#endif
+#define DMA_ISESC1(dma)      ((dma)->revision == dvmaesc1)
+
+/* Main routines in dma.c */
+extern void dvma_init(struct sbus_bus *);
+
+/* Fields in the cond_reg register */
+/* First, the version identification bits */
+#define DMA_DEVICE_ID    0xf0000000        /* Device identification bits */
+#define DMA_VERS0        0x00000000        /* Sunray DMA version */
+#define DMA_ESCV1        0x40000000        /* DMA ESC Version 1 */
+#define DMA_VERS1        0x80000000        /* DMA rev 1 */
+#define DMA_VERS2        0xa0000000        /* DMA rev 2 */
+#define DMA_VERHME       0xb0000000        /* DMA hme gate array */
+#define DMA_VERSPLUS     0x90000000        /* DMA rev 1 PLUS */
+
+#define DMA_HNDL_INTR    0x00000001        /* An IRQ needs to be handled */
+#define DMA_HNDL_ERROR   0x00000002        /* We need to take an error */
+#define DMA_FIFO_ISDRAIN 0x0000000c        /* The DMA FIFO is draining */
+#define DMA_INT_ENAB     0x00000010        /* Turn on interrupts */
+#define DMA_FIFO_INV     0x00000020        /* Invalidate the FIFO */
+#define DMA_ACC_SZ_ERR   0x00000040        /* The access size was bad */
+#define DMA_FIFO_STDRAIN 0x00000040        /* DMA_VERS1 Drain the FIFO */
+#define DMA_RST_SCSI     0x00000080        /* Reset the SCSI controller */
+#define DMA_RST_ENET     DMA_RST_SCSI      /* Reset the ENET controller */
+#define DMA_RST_BPP      DMA_RST_SCSI      /* Reset the BPP controller */
+#define DMA_ST_WRITE     0x00000100        /* write from device to memory */
+#define DMA_ENABLE       0x00000200        /* Fire up DMA, handle requests */
+#define DMA_PEND_READ    0x00000400        /* DMA_VERS1/0/PLUS Pending Read */
+#define DMA_ESC_BURST    0x00000800        /* 1=16byte 0=32byte */
+#define DMA_READ_AHEAD   0x00001800        /* DMA read ahead partial longword */
+#define DMA_DSBL_RD_DRN  0x00001000        /* No EC drain on slave reads */
+#define DMA_BCNT_ENAB    0x00002000        /* If on, use the byte counter */
+#define DMA_TERM_CNTR    0x00004000        /* Terminal counter */
+#define DMA_SCSI_SBUS64  0x00008000        /* HME: Enable 64-bit SBUS mode. */
+#define DMA_CSR_DISAB    0x00010000        /* No FIFO drains during csr */
+#define DMA_SCSI_DISAB   0x00020000        /* No FIFO drains during reg */
+#define DMA_DSBL_WR_INV  0x00020000        /* No EC inval. on slave writes */
+#define DMA_ADD_ENABLE   0x00040000        /* Special ESC DVMA optimization */
+#define DMA_E_BURSTS	 0x000c0000	   /* ENET: SBUS r/w burst mask */
+#define DMA_E_BURST32	 0x00040000	   /* ENET: SBUS 32 byte r/w burst */
+#define DMA_E_BURST16	 0x00000000	   /* ENET: SBUS 16 byte r/w burst */
+#define DMA_BRST_SZ      0x000c0000        /* SCSI: SBUS r/w burst size */
+#define DMA_BRST64       0x00080000        /* SCSI: 64byte bursts (HME on UltraSparc only) */
+#define DMA_BRST32       0x00040000        /* SCSI/BPP: 32byte bursts */
+#define DMA_BRST16       0x00000000        /* SCSI/BPP: 16byte bursts */
+#define DMA_BRST0        0x00080000        /* SCSI: no bursts (non-HME gate arrays) */
+#define DMA_ADDR_DISAB   0x00100000        /* No FIFO drains during addr */
+#define DMA_2CLKS        0x00200000        /* Each transfer = 2 clock ticks */
+#define DMA_3CLKS        0x00400000        /* Each transfer = 3 clock ticks */
+#define DMA_EN_ENETAUI   DMA_3CLKS         /* Put lance into AUI-cable mode */
+#define DMA_CNTR_DISAB   0x00800000        /* No IRQ when DMA_TERM_CNTR set */
+#define DMA_AUTO_NADDR   0x01000000        /* Use "auto nxt addr" feature */
+#define DMA_SCSI_ON      0x02000000        /* Enable SCSI dma */
+#define DMA_BPP_ON       DMA_SCSI_ON       /* Enable BPP dma */
+#define DMA_PARITY_OFF   0x02000000        /* HME: disable parity checking */
+#define DMA_LOADED_ADDR  0x04000000        /* Address has been loaded */
+#define DMA_LOADED_NADDR 0x08000000        /* Next address has been loaded */
+#define DMA_RESET_FAS366 0x08000000        /* HME: Assert RESET to FAS366 */
+
+/* Values describing the burst-size property from the PROM */
+#define DMA_BURST1       0x01
+#define DMA_BURST2       0x02
+#define DMA_BURST4       0x04
+#define DMA_BURST8       0x08
+#define DMA_BURST16      0x10
+#define DMA_BURST32      0x20
+#define DMA_BURST64      0x40
+#define DMA_BURSTBITS    0x7f
+
+/* Determine highest possible final transfer address given a base */
+#define DMA_MAXEND(addr) (0x01000000UL-(((unsigned long)(addr))&0x00ffffffUL))
+
+/* Yes, I hack a lot of elisp in my spare time... */
+#define DMA_ERROR_P(regs)  ((((regs)->cond_reg) & DMA_HNDL_ERROR))
+#define DMA_IRQ_P(regs)    ((((regs)->cond_reg) & (DMA_HNDL_INTR | DMA_HNDL_ERROR)))
+#define DMA_WRITE_P(regs)  ((((regs)->cond_reg) & DMA_ST_WRITE))
+#define DMA_OFF(regs)      ((((regs)->cond_reg) &= (~DMA_ENABLE)))
+#define DMA_INTSOFF(regs)  ((((regs)->cond_reg) &= (~DMA_INT_ENAB)))
+#define DMA_INTSON(regs)   ((((regs)->cond_reg) |= (DMA_INT_ENAB)))
+#define DMA_PUNTFIFO(regs) ((((regs)->cond_reg) |= DMA_FIFO_INV))
+#define DMA_SETSTART(regs, addr)  ((((regs)->st_addr) = (char *) addr))
+#define DMA_BEGINDMA_W(regs) \
+        ((((regs)->cond_reg |= (DMA_ST_WRITE|DMA_ENABLE|DMA_INT_ENAB))))
+#define DMA_BEGINDMA_R(regs) \
+        ((((regs)->cond_reg |= ((DMA_ENABLE|DMA_INT_ENAB)&(~DMA_ST_WRITE)))))
+
+/* For certain DMA chips, we need to disable ints upon irq entry
+ * and turn them back on when we are done.  So in any ESP interrupt
+ * handler you *must* call DMA_IRQ_ENTRY upon entry and DMA_IRQ_EXIT
+ * when leaving the handler.  You have been warned...
+ */
+#define DMA_IRQ_ENTRY(dma, dregs) do { \
+        if(DMA_ISBROKEN(dma)) DMA_INTSOFF(dregs); \
+   } while (0)
+
+#define DMA_IRQ_EXIT(dma, dregs) do { \
+	if(DMA_ISBROKEN(dma)) DMA_INTSON(dregs); \
+   } while(0)
+
+#if 0	/* P3 this stuff is inline in ledma.c:init_restart_ledma() */
+/* Pause until counter runs out or BIT isn't set in the DMA condition
+ * register.
+ */
+static inline void sparc_dma_pause(struct sparc_dma_registers *regs,
+				       unsigned long bit)
+{
+	int ctr = 50000;   /* Let's find some bugs ;) */
+
+	/* Busy wait until the bit is not set any more */
+	while((regs->cond_reg&bit) && (ctr>0)) {
+		ctr--;
+		__delay(5);
+	}
+
+	/* Check for bogus outcome. */
+	if(!ctr)
+		panic("DMA timeout");
+}
+
+/* Reset the friggin' thing... */
+#define DMA_RESET(dma) do { \
+	struct sparc_dma_registers *regs = dma->regs;                      \
+	/* Let the current FIFO drain itself */                            \
+	sparc_dma_pause(regs, (DMA_FIFO_ISDRAIN));                         \
+	/* Reset the logic */                                              \
+	regs->cond_reg |= (DMA_RST_SCSI);     /* assert */                 \
+	__delay(400);                         /* let the bits set ;) */    \
+	regs->cond_reg &= ~(DMA_RST_SCSI);    /* de-assert */              \
+	sparc_dma_enable_interrupts(regs);    /* Re-enable interrupts */   \
+	/* Enable FAST transfers if available */                           \
+	if(dma->revision>dvmarev1) regs->cond_reg |= DMA_3CLKS;            \
+	dma->running = 0;                                                  \
+} while(0)
+#endif
+
+#define for_each_dvma(dma) \
+        for((dma) = dma_chain; (dma); (dma) = (dma)->next)
+
+extern int get_dma_list(char *);
+extern int request_dma(unsigned int, __const__ char *);
+extern void free_dma(unsigned int);
+
+/* From PCI */
+
+#ifdef CONFIG_PCI
+extern int isa_dma_bridge_buggy;
+#else
+#define isa_dma_bridge_buggy	(0)
+#endif
+
+/* Routines for data transfer buffers. */
+BTFIXUPDEF_CALL(char *, mmu_lockarea, char *, unsigned long)
+BTFIXUPDEF_CALL(void,   mmu_unlockarea, char *, unsigned long)
+
+#define mmu_lockarea(vaddr,len) BTFIXUP_CALL(mmu_lockarea)(vaddr,len)
+#define mmu_unlockarea(vaddr,len) BTFIXUP_CALL(mmu_unlockarea)(vaddr,len)
+
+/* These are implementations for sbus_map_sg/sbus_unmap_sg... collapse later */
+BTFIXUPDEF_CALL(__u32, mmu_get_scsi_one, char *, unsigned long, struct sbus_bus *sbus)
+BTFIXUPDEF_CALL(void,  mmu_get_scsi_sgl, struct scatterlist *, int, struct sbus_bus *sbus)
+BTFIXUPDEF_CALL(void,  mmu_release_scsi_one, __u32, unsigned long, struct sbus_bus *sbus)
+BTFIXUPDEF_CALL(void,  mmu_release_scsi_sgl, struct scatterlist *, int, struct sbus_bus *sbus)
+
+#define mmu_get_scsi_one(vaddr,len,sbus) BTFIXUP_CALL(mmu_get_scsi_one)(vaddr,len,sbus)
+#define mmu_get_scsi_sgl(sg,sz,sbus) BTFIXUP_CALL(mmu_get_scsi_sgl)(sg,sz,sbus)
+#define mmu_release_scsi_one(vaddr,len,sbus) BTFIXUP_CALL(mmu_release_scsi_one)(vaddr,len,sbus)
+#define mmu_release_scsi_sgl(sg,sz,sbus) BTFIXUP_CALL(mmu_release_scsi_sgl)(sg,sz,sbus)
+
+/*
+ * mmu_map/unmap are provided by iommu/iounit; Invalid to call on IIep.
+ *
+ * The mmu_map_dma_area establishes two mappings in one go.
+ * These mappings point to pages normally mapped at 'va' (linear address).
+ * First mapping is for CPU visible address at 'a', uncached.
+ * This is an alias, but it works because it is an uncached mapping.
+ * Second mapping is for device visible address, or "bus" address.
+ * The bus address is returned at '*pba'.
+ *
+ * These functions seem distinct, but are hard to split. On sun4c,
+ * at least for now, 'a' is equal to bus address, and retured in *pba.
+ * On sun4m, page attributes depend on the CPU type, so we have to
+ * know if we are mapping RAM or I/O, so it has to be an additional argument
+ * to a separate mapping function for CPU visible mappings.
+ */
+BTFIXUPDEF_CALL(int,  mmu_map_dma_area, dma_addr_t *, unsigned long, unsigned long, int len)
+BTFIXUPDEF_CALL(struct page *, mmu_translate_dvma, unsigned long busa)
+BTFIXUPDEF_CALL(void,  mmu_unmap_dma_area, unsigned long busa, int len)
+
+#define mmu_map_dma_area(pba,va,a,len) BTFIXUP_CALL(mmu_map_dma_area)(pba,va,a,len)
+#define mmu_unmap_dma_area(ba,len) BTFIXUP_CALL(mmu_unmap_dma_area)(ba,len)
+#define mmu_translate_dvma(ba)     BTFIXUP_CALL(mmu_translate_dvma)(ba)
+
+#endif /* !(_ASM_SPARC_DMA_H) */
diff --git a/include/asm-sparc/dma_64.h b/include/asm-sparc/dma_64.h
new file mode 100644
index 00000000000..9d4c024bd3b
--- /dev/null
+++ b/include/asm-sparc/dma_64.h
@@ -0,0 +1,205 @@
+/*
+ * include/asm-sparc64/dma.h
+ *
+ * Copyright 1996 (C) David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _ASM_SPARC64_DMA_H
+#define _ASM_SPARC64_DMA_H
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/spinlock.h>
+
+#include <asm/sbus.h>
+#include <asm/delay.h>
+#include <asm/oplib.h>
+
+/* These are irrelevant for Sparc DMA, but we leave it in so that
+ * things can compile.
+ */
+#define MAX_DMA_CHANNELS 8
+#define DMA_MODE_READ    1
+#define DMA_MODE_WRITE   2
+#define MAX_DMA_ADDRESS  (~0UL)
+
+/* Useful constants */
+#define SIZE_16MB      (16*1024*1024)
+#define SIZE_64K       (64*1024)
+
+/* SBUS DMA controller reg offsets */
+#define DMA_CSR		0x00UL		/* rw  DMA control/status register    0x00   */
+#define DMA_ADDR	0x04UL		/* rw  DMA transfer address register  0x04   */
+#define DMA_COUNT	0x08UL		/* rw  DMA transfer count register    0x08   */
+#define DMA_TEST	0x0cUL		/* rw  DMA test/debug register        0x0c   */
+
+/* DVMA chip revisions */
+enum dvma_rev {
+	dvmarev0,
+	dvmaesc1,
+	dvmarev1,
+	dvmarev2,
+	dvmarev3,
+	dvmarevplus,
+	dvmahme
+};
+
+#define DMA_HASCOUNT(rev)  ((rev)==dvmaesc1)
+
+/* Linux DMA information structure, filled during probe. */
+struct sbus_dma {
+	struct sbus_dma *next;
+	struct sbus_dev *sdev;
+	void __iomem *regs;
+
+	/* Status, misc info */
+	int node;                /* Prom node for this DMA device */
+	int running;             /* Are we doing DMA now? */
+	int allocated;           /* Are we "owned" by anyone yet? */
+
+	/* Transfer information. */
+	u32 addr;                /* Start address of current transfer */
+	int nbytes;              /* Size of current transfer */
+	int realbytes;           /* For splitting up large transfers, etc. */
+
+	/* DMA revision */
+	enum dvma_rev revision;
+};
+
+extern struct sbus_dma *dma_chain;
+
+/* Broken hardware... */
+#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev1)
+#define DMA_ISESC1(dma)      ((dma)->revision == dvmaesc1)
+
+/* Main routines in dma.c */
+extern void dvma_init(struct sbus_bus *);
+
+/* Fields in the cond_reg register */
+/* First, the version identification bits */
+#define DMA_DEVICE_ID    0xf0000000        /* Device identification bits */
+#define DMA_VERS0        0x00000000        /* Sunray DMA version */
+#define DMA_ESCV1        0x40000000        /* DMA ESC Version 1 */
+#define DMA_VERS1        0x80000000        /* DMA rev 1 */
+#define DMA_VERS2        0xa0000000        /* DMA rev 2 */
+#define DMA_VERHME       0xb0000000        /* DMA hme gate array */
+#define DMA_VERSPLUS     0x90000000        /* DMA rev 1 PLUS */
+
+#define DMA_HNDL_INTR    0x00000001        /* An IRQ needs to be handled */
+#define DMA_HNDL_ERROR   0x00000002        /* We need to take an error */
+#define DMA_FIFO_ISDRAIN 0x0000000c        /* The DMA FIFO is draining */
+#define DMA_INT_ENAB     0x00000010        /* Turn on interrupts */
+#define DMA_FIFO_INV     0x00000020        /* Invalidate the FIFO */
+#define DMA_ACC_SZ_ERR   0x00000040        /* The access size was bad */
+#define DMA_FIFO_STDRAIN 0x00000040        /* DMA_VERS1 Drain the FIFO */
+#define DMA_RST_SCSI     0x00000080        /* Reset the SCSI controller */
+#define DMA_RST_ENET     DMA_RST_SCSI      /* Reset the ENET controller */
+#define DMA_ST_WRITE     0x00000100        /* write from device to memory */
+#define DMA_ENABLE       0x00000200        /* Fire up DMA, handle requests */
+#define DMA_PEND_READ    0x00000400        /* DMA_VERS1/0/PLUS Pending Read */
+#define DMA_ESC_BURST    0x00000800        /* 1=16byte 0=32byte */
+#define DMA_READ_AHEAD   0x00001800        /* DMA read ahead partial longword */
+#define DMA_DSBL_RD_DRN  0x00001000        /* No EC drain on slave reads */
+#define DMA_BCNT_ENAB    0x00002000        /* If on, use the byte counter */
+#define DMA_TERM_CNTR    0x00004000        /* Terminal counter */
+#define DMA_SCSI_SBUS64  0x00008000        /* HME: Enable 64-bit SBUS mode. */
+#define DMA_CSR_DISAB    0x00010000        /* No FIFO drains during csr */
+#define DMA_SCSI_DISAB   0x00020000        /* No FIFO drains during reg */
+#define DMA_DSBL_WR_INV  0x00020000        /* No EC inval. on slave writes */
+#define DMA_ADD_ENABLE   0x00040000        /* Special ESC DVMA optimization */
+#define DMA_E_BURSTS	 0x000c0000	   /* ENET: SBUS r/w burst mask */
+#define DMA_E_BURST32	 0x00040000	   /* ENET: SBUS 32 byte r/w burst */
+#define DMA_E_BURST16	 0x00000000	   /* ENET: SBUS 16 byte r/w burst */
+#define DMA_BRST_SZ      0x000c0000        /* SCSI: SBUS r/w burst size */
+#define DMA_BRST64       0x000c0000        /* SCSI: 64byte bursts (HME on UltraSparc only) */
+#define DMA_BRST32       0x00040000        /* SCSI: 32byte bursts */
+#define DMA_BRST16       0x00000000        /* SCSI: 16byte bursts */
+#define DMA_BRST0        0x00080000        /* SCSI: no bursts (non-HME gate arrays) */
+#define DMA_ADDR_DISAB   0x00100000        /* No FIFO drains during addr */
+#define DMA_2CLKS        0x00200000        /* Each transfer = 2 clock ticks */
+#define DMA_3CLKS        0x00400000        /* Each transfer = 3 clock ticks */
+#define DMA_EN_ENETAUI   DMA_3CLKS         /* Put lance into AUI-cable mode */
+#define DMA_CNTR_DISAB   0x00800000        /* No IRQ when DMA_TERM_CNTR set */
+#define DMA_AUTO_NADDR   0x01000000        /* Use "auto nxt addr" feature */
+#define DMA_SCSI_ON      0x02000000        /* Enable SCSI dma */
+#define DMA_PARITY_OFF   0x02000000        /* HME: disable parity checking */
+#define DMA_LOADED_ADDR  0x04000000        /* Address has been loaded */
+#define DMA_LOADED_NADDR 0x08000000        /* Next address has been loaded */
+#define DMA_RESET_FAS366 0x08000000        /* HME: Assert RESET to FAS366 */
+
+/* Values describing the burst-size property from the PROM */
+#define DMA_BURST1       0x01
+#define DMA_BURST2       0x02
+#define DMA_BURST4       0x04
+#define DMA_BURST8       0x08
+#define DMA_BURST16      0x10
+#define DMA_BURST32      0x20
+#define DMA_BURST64      0x40
+#define DMA_BURSTBITS    0x7f
+
+/* Determine highest possible final transfer address given a base */
+#define DMA_MAXEND(addr) (0x01000000UL-(((unsigned long)(addr))&0x00ffffffUL))
+
+/* Yes, I hack a lot of elisp in my spare time... */
+#define DMA_ERROR_P(regs)  ((sbus_readl((regs) + DMA_CSR) & DMA_HNDL_ERROR))
+#define DMA_IRQ_P(regs)    ((sbus_readl((regs) + DMA_CSR)) & (DMA_HNDL_INTR | DMA_HNDL_ERROR))
+#define DMA_WRITE_P(regs)  ((sbus_readl((regs) + DMA_CSR) & DMA_ST_WRITE))
+#define DMA_OFF(__regs)		\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp &= ~DMA_ENABLE; \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+#define DMA_INTSOFF(__regs)	\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp &= ~DMA_INT_ENAB; \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+#define DMA_INTSON(__regs)	\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp |= DMA_INT_ENAB; \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+#define DMA_PUNTFIFO(__regs)	\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp |= DMA_FIFO_INV; \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+#define DMA_SETSTART(__regs, __addr)	\
+	sbus_writel((u32)(__addr), (__regs) + DMA_ADDR);
+#define DMA_BEGINDMA_W(__regs)	\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp |= (DMA_ST_WRITE|DMA_ENABLE|DMA_INT_ENAB); \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+#define DMA_BEGINDMA_R(__regs)	\
+do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
+	tmp |= (DMA_ENABLE|DMA_INT_ENAB); \
+	tmp &= ~DMA_ST_WRITE; \
+	sbus_writel(tmp, (__regs) + DMA_CSR); \
+} while(0)
+
+/* For certain DMA chips, we need to disable ints upon irq entry
+ * and turn them back on when we are done.  So in any ESP interrupt
+ * handler you *must* call DMA_IRQ_ENTRY upon entry and DMA_IRQ_EXIT
+ * when leaving the handler.  You have been warned...
+ */
+#define DMA_IRQ_ENTRY(dma, dregs) do { \
+        if(DMA_ISBROKEN(dma)) DMA_INTSOFF(dregs); \
+   } while (0)
+
+#define DMA_IRQ_EXIT(dma, dregs) do { \
+	if(DMA_ISBROKEN(dma)) DMA_INTSON(dregs); \
+   } while(0)
+
+#define for_each_dvma(dma) \
+        for((dma) = dma_chain; (dma); (dma) = (dma)->next)
+
+/* From PCI */
+
+#ifdef CONFIG_PCI
+extern int isa_dma_bridge_buggy;
+#else
+#define isa_dma_bridge_buggy 	(0)
+#endif
+
+#endif /* !(_ASM_SPARC64_DMA_H) */
diff --git a/include/asm-sparc/ebus.h b/include/asm-sparc/ebus.h
index 491f85d662d..a5da2d00cd1 100644
--- a/include/asm-sparc/ebus.h
+++ b/include/asm-sparc/ebus.h
@@ -1,99 +1,8 @@
-/*
- * ebus.h: PCI to Ebus pseudo driver software state.
- *
- * Copyright (C) 1997 Eddie C. Dost (ecd@skynet.be) 
- *
- * Adopted for sparc by V. Roganov and G. Raiko.
- */
-
-#ifndef __SPARC_EBUS_H
-#define __SPARC_EBUS_H
-
-#ifndef _LINUX_IOPORT_H
-#include <linux/ioport.h>
+#ifndef ___ASM_SPARC_EBUS_H
+#define ___ASM_SPARC_EBUS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/ebus_64.h>
+#else
+#include <asm-sparc/ebus_32.h>
+#endif
 #endif
-#include <asm/oplib.h>
-#include <asm/prom.h>
-#include <asm/of_device.h>
-
-struct linux_ebus_child {
-	struct linux_ebus_child		*next;
-	struct linux_ebus_device	*parent;
-	struct linux_ebus		*bus;
-	struct device_node		*prom_node;
-	struct resource			 resource[PROMREG_MAX];
-	int				 num_addrs;
-	unsigned int			 irqs[PROMINTR_MAX];
-	int				 num_irqs;
-};
-
-struct linux_ebus_device {
-	struct of_device		ofdev;
-	struct linux_ebus_device	*next;
-	struct linux_ebus_child		*children;
-	struct linux_ebus		*bus;
-	struct device_node		*prom_node;
-	struct resource			 resource[PROMREG_MAX];
-	int				 num_addrs;
-	unsigned int			 irqs[PROMINTR_MAX];
-	int				 num_irqs;
-};
-#define to_ebus_device(d) container_of(d, struct linux_ebus_device, ofdev.dev)
-
-struct linux_ebus {
-	struct of_device		ofdev;
-	struct linux_ebus		*next;
-	struct linux_ebus_device	*devices;
-	struct linux_pbm_info		*parent;
-	struct pci_dev			*self;
-	struct device_node		*prom_node;
-};
-#define to_ebus(d) container_of(d, struct linux_ebus, ofdev.dev)
-
-struct linux_ebus_dma {
-	unsigned int dcsr;
-	unsigned int dacr;
-	unsigned int dbcr;
-};
-
-#define EBUS_DCSR_INT_PEND	0x00000001
-#define EBUS_DCSR_ERR_PEND	0x00000002
-#define EBUS_DCSR_DRAIN		0x00000004
-#define EBUS_DCSR_INT_EN	0x00000010
-#define EBUS_DCSR_RESET		0x00000080
-#define EBUS_DCSR_WRITE		0x00000100
-#define EBUS_DCSR_EN_DMA	0x00000200
-#define EBUS_DCSR_CYC_PEND	0x00000400
-#define EBUS_DCSR_DIAG_RD_DONE	0x00000800
-#define EBUS_DCSR_DIAG_WR_DONE	0x00001000
-#define EBUS_DCSR_EN_CNT	0x00002000
-#define EBUS_DCSR_TC		0x00004000
-#define EBUS_DCSR_DIS_CSR_DRN	0x00010000
-#define EBUS_DCSR_BURST_SZ_MASK	0x000c0000
-#define EBUS_DCSR_BURST_SZ_1	0x00080000
-#define EBUS_DCSR_BURST_SZ_4	0x00000000
-#define EBUS_DCSR_BURST_SZ_8	0x00040000
-#define EBUS_DCSR_BURST_SZ_16	0x000c0000
-#define EBUS_DCSR_DIAG_EN	0x00100000
-#define EBUS_DCSR_DIS_ERR_PEND	0x00400000
-#define EBUS_DCSR_TCI_DIS	0x00800000
-#define EBUS_DCSR_EN_NEXT	0x01000000
-#define EBUS_DCSR_DMA_ON	0x02000000
-#define EBUS_DCSR_A_LOADED	0x04000000
-#define EBUS_DCSR_NA_LOADED	0x08000000
-#define EBUS_DCSR_DEV_ID_MASK	0xf0000000
-
-extern struct linux_ebus		*ebus_chain;
-
-extern void ebus_init(void);
-
-#define for_each_ebus(bus)						\
-        for((bus) = ebus_chain; (bus); (bus) = (bus)->next)
-
-#define for_each_ebusdev(dev, bus)					\
-        for((dev) = (bus)->devices; (dev); (dev) = (dev)->next)
-
-#define for_each_edevchild(dev, child)					\
-        for((child) = (dev)->children; (child); (child) = (child)->next)
-
-#endif /* !(__SPARC_EBUS_H) */
diff --git a/include/asm-sparc/ebus_32.h b/include/asm-sparc/ebus_32.h
new file mode 100644
index 00000000000..29cb7dfc6b7
--- /dev/null
+++ b/include/asm-sparc/ebus_32.h
@@ -0,0 +1,99 @@
+/*
+ * ebus.h: PCI to Ebus pseudo driver software state.
+ *
+ * Copyright (C) 1997 Eddie C. Dost (ecd@skynet.be)
+ *
+ * Adopted for sparc by V. Roganov and G. Raiko.
+ */
+
+#ifndef __SPARC_EBUS_H
+#define __SPARC_EBUS_H
+
+#ifndef _LINUX_IOPORT_H
+#include <linux/ioport.h>
+#endif
+#include <asm/oplib.h>
+#include <asm/prom.h>
+#include <asm/of_device.h>
+
+struct linux_ebus_child {
+	struct linux_ebus_child		*next;
+	struct linux_ebus_device	*parent;
+	struct linux_ebus		*bus;
+	struct device_node		*prom_node;
+	struct resource			 resource[PROMREG_MAX];
+	int				 num_addrs;
+	unsigned int			 irqs[PROMINTR_MAX];
+	int				 num_irqs;
+};
+
+struct linux_ebus_device {
+	struct of_device		ofdev;
+	struct linux_ebus_device	*next;
+	struct linux_ebus_child		*children;
+	struct linux_ebus		*bus;
+	struct device_node		*prom_node;
+	struct resource			 resource[PROMREG_MAX];
+	int				 num_addrs;
+	unsigned int			 irqs[PROMINTR_MAX];
+	int				 num_irqs;
+};
+#define to_ebus_device(d) container_of(d, struct linux_ebus_device, ofdev.dev)
+
+struct linux_ebus {
+	struct of_device		ofdev;
+	struct linux_ebus		*next;
+	struct linux_ebus_device	*devices;
+	struct linux_pbm_info		*parent;
+	struct pci_dev			*self;
+	struct device_node		*prom_node;
+};
+#define to_ebus(d) container_of(d, struct linux_ebus, ofdev.dev)
+
+struct linux_ebus_dma {
+	unsigned int dcsr;
+	unsigned int dacr;
+	unsigned int dbcr;
+};
+
+#define EBUS_DCSR_INT_PEND	0x00000001
+#define EBUS_DCSR_ERR_PEND	0x00000002
+#define EBUS_DCSR_DRAIN		0x00000004
+#define EBUS_DCSR_INT_EN	0x00000010
+#define EBUS_DCSR_RESET		0x00000080
+#define EBUS_DCSR_WRITE		0x00000100
+#define EBUS_DCSR_EN_DMA	0x00000200
+#define EBUS_DCSR_CYC_PEND	0x00000400
+#define EBUS_DCSR_DIAG_RD_DONE	0x00000800
+#define EBUS_DCSR_DIAG_WR_DONE	0x00001000
+#define EBUS_DCSR_EN_CNT	0x00002000
+#define EBUS_DCSR_TC		0x00004000
+#define EBUS_DCSR_DIS_CSR_DRN	0x00010000
+#define EBUS_DCSR_BURST_SZ_MASK	0x000c0000
+#define EBUS_DCSR_BURST_SZ_1	0x00080000
+#define EBUS_DCSR_BURST_SZ_4	0x00000000
+#define EBUS_DCSR_BURST_SZ_8	0x00040000
+#define EBUS_DCSR_BURST_SZ_16	0x000c0000
+#define EBUS_DCSR_DIAG_EN	0x00100000
+#define EBUS_DCSR_DIS_ERR_PEND	0x00400000
+#define EBUS_DCSR_TCI_DIS	0x00800000
+#define EBUS_DCSR_EN_NEXT	0x01000000
+#define EBUS_DCSR_DMA_ON	0x02000000
+#define EBUS_DCSR_A_LOADED	0x04000000
+#define EBUS_DCSR_NA_LOADED	0x08000000
+#define EBUS_DCSR_DEV_ID_MASK	0xf0000000
+
+extern struct linux_ebus		*ebus_chain;
+
+extern void ebus_init(void);
+
+#define for_each_ebus(bus)						\
+        for((bus) = ebus_chain; (bus); (bus) = (bus)->next)
+
+#define for_each_ebusdev(dev, bus)					\
+        for((dev) = (bus)->devices; (dev); (dev) = (dev)->next)
+
+#define for_each_edevchild(dev, child)					\
+        for((child) = (dev)->children; (child); (child) = (child)->next)
+
+#endif /* !(__SPARC_EBUS_H) */
diff --git a/include/asm-sparc/ebus_64.h b/include/asm-sparc/ebus_64.h
new file mode 100644
index 00000000000..fcc62b97ced
--- /dev/null
+++ b/include/asm-sparc/ebus_64.h
@@ -0,0 +1,94 @@
+/*
+ * ebus.h: PCI to Ebus pseudo driver software state.
+ *
+ * Copyright (C) 1997 Eddie C. Dost (ecd@skynet.be)
+ * Copyright (C) 1999 David S. Miller (davem@redhat.com)
+ */
+
+#ifndef __SPARC64_EBUS_H
+#define __SPARC64_EBUS_H
+
+#include <asm/oplib.h>
+#include <asm/prom.h>
+#include <asm/of_device.h>
+
+struct linux_ebus_child {
+	struct linux_ebus_child		*next;
+	struct linux_ebus_device	*parent;
+	struct linux_ebus		*bus;
+	struct device_node		*prom_node;
+	struct resource			 resource[PROMREG_MAX];
+	int				 num_addrs;
+	unsigned int			 irqs[PROMINTR_MAX];
+	int				 num_irqs;
+};
+
+struct linux_ebus_device {
+	struct of_device		ofdev;
+	struct linux_ebus_device	*next;
+	struct linux_ebus_child		*children;
+	struct linux_ebus		*bus;
+	struct device_node		*prom_node;
+	struct resource			 resource[PROMREG_MAX];
+	int				 num_addrs;
+	unsigned int			 irqs[PROMINTR_MAX];
+	int				 num_irqs;
+};
+#define to_ebus_device(d) container_of(d, struct linux_ebus_device, ofdev.dev)
+
+struct linux_ebus {
+	struct of_device		ofdev;
+	struct linux_ebus		*next;
+	struct linux_ebus_device	*devices;
+	struct pci_dev			*self;
+	int				 index;
+	int				 is_rio;
+	struct device_node		*prom_node;
+};
+#define to_ebus(d) container_of(d, struct linux_ebus, ofdev.dev)
+
+struct ebus_dma_info {
+	spinlock_t	lock;
+	void __iomem	*regs;
+
+	unsigned int	flags;
+#define EBUS_DMA_FLAG_USE_EBDMA_HANDLER		0x00000001
+#define EBUS_DMA_FLAG_TCI_DISABLE		0x00000002
+
+	/* These are only valid is EBUS_DMA_FLAG_USE_EBDMA_HANDLER is
+	 * set.
+	 */
+	void (*callback)(struct ebus_dma_info *p, int event, void *cookie);
+	void *client_cookie;
+	unsigned int	irq;
+#define EBUS_DMA_EVENT_ERROR	1
+#define EBUS_DMA_EVENT_DMA	2
+#define EBUS_DMA_EVENT_DEVICE	4
+
+	unsigned char	name[64];
+};
+
+extern int ebus_dma_register(struct ebus_dma_info *p);
+extern int ebus_dma_irq_enable(struct ebus_dma_info *p, int on);
+extern void ebus_dma_unregister(struct ebus_dma_info *p);
+extern int ebus_dma_request(struct ebus_dma_info *p, dma_addr_t bus_addr,
+			    size_t len);
+extern void ebus_dma_prepare(struct ebus_dma_info *p, int write);
+extern unsigned int ebus_dma_residue(struct ebus_dma_info *p);
+extern unsigned int ebus_dma_addr(struct ebus_dma_info *p);
+extern void ebus_dma_enable(struct ebus_dma_info *p, int on);
+
+extern struct linux_ebus		*ebus_chain;
+
+extern void ebus_init(void);
+
+#define for_each_ebus(bus)						\
+        for((bus) = ebus_chain; (bus); (bus) = (bus)->next)
+
+#define for_each_ebusdev(dev, bus)					\
+        for((dev) = (bus)->devices; (dev); (dev) = (dev)->next)
+
+#define for_each_edevchild(dev, child)					\
+        for((child) = (dev)->children; (child); (child) = (child)->next)
+
+#endif /* !(__SPARC64_EBUS_H) */
diff --git a/include/asm-sparc/elf.h b/include/asm-sparc/elf.h
index d043f80bc2f..f035c45d7b5 100644
--- a/include/asm-sparc/elf.h
+++ b/include/asm-sparc/elf.h
@@ -1,145 +1,8 @@
-#ifndef __ASMSPARC_ELF_H
-#define __ASMSPARC_ELF_H
-
-/*
- * ELF register definitions..
- */
-
-#include <asm/ptrace.h>
-
-/*
- * Sparc section types
- */
-#define STT_REGISTER		13
-
-/*
- * Sparc ELF relocation types
- */
-#define	R_SPARC_NONE		0
-#define	R_SPARC_8		1
-#define	R_SPARC_16		2
-#define	R_SPARC_32		3
-#define	R_SPARC_DISP8		4
-#define	R_SPARC_DISP16		5
-#define	R_SPARC_DISP32		6
-#define	R_SPARC_WDISP30		7
-#define	R_SPARC_WDISP22		8
-#define	R_SPARC_HI22		9
-#define	R_SPARC_22		10
-#define	R_SPARC_13		11
-#define	R_SPARC_LO10		12
-#define	R_SPARC_GOT10		13
-#define	R_SPARC_GOT13		14
-#define	R_SPARC_GOT22		15
-#define	R_SPARC_PC10		16
-#define	R_SPARC_PC22		17
-#define	R_SPARC_WPLT30		18
-#define	R_SPARC_COPY		19
-#define	R_SPARC_GLOB_DAT	20
-#define	R_SPARC_JMP_SLOT	21
-#define	R_SPARC_RELATIVE	22
-#define	R_SPARC_UA32		23
-#define R_SPARC_PLT32		24
-#define R_SPARC_HIPLT22		25
-#define R_SPARC_LOPLT10		26
-#define R_SPARC_PCPLT32		27
-#define R_SPARC_PCPLT22		28
-#define R_SPARC_PCPLT10		29
-#define R_SPARC_10		30
-#define R_SPARC_11		31
-#define R_SPARC_64		32
-#define R_SPARC_OLO10		33
-#define R_SPARC_WDISP16		40
-#define R_SPARC_WDISP19		41
-#define R_SPARC_7		43
-#define R_SPARC_5		44
-#define R_SPARC_6		45
-
-/* Bits present in AT_HWCAP, primarily for Sparc32.  */
-
-#define HWCAP_SPARC_FLUSH       1    /* CPU supports flush instruction. */
-#define HWCAP_SPARC_STBAR       2
-#define HWCAP_SPARC_SWAP        4
-#define HWCAP_SPARC_MULDIV      8
-#define HWCAP_SPARC_V9		16
-#define HWCAP_SPARC_ULTRA3	32
-
-#define CORE_DUMP_USE_REGSET
-
-/* Format is:
- * 	G0 --> G7
- *	O0 --> O7
- *	L0 --> L7
- *	I0 --> I7
- *	PSR, PC, nPC, Y, WIM, TBR
- */
-typedef unsigned long elf_greg_t;
-#define ELF_NGREG 38
-typedef elf_greg_t elf_gregset_t[ELF_NGREG];
-
-typedef struct {
-	union {
-		unsigned long	pr_regs[32];
-		double		pr_dregs[16];
-	} pr_fr;
-	unsigned long __unused;
-	unsigned long	pr_fsr;
-	unsigned char	pr_qcnt;
-	unsigned char	pr_q_entrysize;
-	unsigned char	pr_en;
-	unsigned int	pr_q[64];
-} elf_fpregset_t;
-
-#include <asm/mbus.h>
-
-/*
- * This is used to ensure we don't load something for the wrong architecture.
- */
-#define elf_check_arch(x) ((x)->e_machine == EM_SPARC)
-
-/*
- * These are used to set parameters in the core dumps.
- */
-#define ELF_ARCH	EM_SPARC
-#define ELF_CLASS	ELFCLASS32
-#define ELF_DATA	ELFDATA2MSB
-
-#define USE_ELF_CORE_DUMP
-#ifndef CONFIG_SUN4
-#define ELF_EXEC_PAGESIZE	4096
+#ifndef ___ASM_SPARC_ELF_H
+#define ___ASM_SPARC_ELF_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/elf_64.h>
 #else
-#define ELF_EXEC_PAGESIZE	8192
+#include <asm-sparc/elf_32.h>
+#endif
 #endif
-
-
-/* This is the location that an ET_DYN program is loaded if exec'ed.  Typical
-   use of this is to invoke "./ld.so someprog" to test out a new version of
-   the loader.  We need to make sure that it is out of the way of the program
-   that it will "exec", and that there is sufficient room for the brk.  */
-
-#define ELF_ET_DYN_BASE         (TASK_UNMAPPED_BASE)
-
-/* This yields a mask that user programs can use to figure out what
-   instruction set this cpu supports.  This can NOT be done in userspace
-   on Sparc.  */
-
-/* Sun4c has none of the capabilities, most sun4m's have them all.
- * XXX This is gross, set some global variable at boot time. -DaveM
- */
-#define ELF_HWCAP	((ARCH_SUN4C_SUN4) ? 0 : \
-			 (HWCAP_SPARC_FLUSH | HWCAP_SPARC_STBAR | \
-			  HWCAP_SPARC_SWAP | \
-			  ((srmmu_modtype != Cypress && \
-			    srmmu_modtype != Cypress_vE && \
-			    srmmu_modtype != Cypress_vD) ? \
-			   HWCAP_SPARC_MULDIV : 0)))
-
-/* This yields a string that ld.so will use to load implementation
-   specific libraries for optimization.  This is more specific in
-   intent than poking at uname or /proc/cpuinfo. */
-
-#define ELF_PLATFORM	(NULL)
-
-#define SET_PERSONALITY(ex, ibcs2) set_personality((ibcs2)?PER_SVR4:PER_LINUX)
-
-#endif /* !(__ASMSPARC_ELF_H) */
diff --git a/include/asm-sparc/elf_32.h b/include/asm-sparc/elf_32.h
new file mode 100644
index 00000000000..d043f80bc2f
--- /dev/null
+++ b/include/asm-sparc/elf_32.h
@@ -0,0 +1,145 @@
+#ifndef __ASMSPARC_ELF_H
+#define __ASMSPARC_ELF_H
+
+/*
+ * ELF register definitions..
+ */
+
+#include <asm/ptrace.h>
+
+/*
+ * Sparc section types
+ */
+#define STT_REGISTER		13
+
+/*
+ * Sparc ELF relocation types
+ */
+#define	R_SPARC_NONE		0
+#define	R_SPARC_8		1
+#define	R_SPARC_16		2
+#define	R_SPARC_32		3
+#define	R_SPARC_DISP8		4
+#define	R_SPARC_DISP16		5
+#define	R_SPARC_DISP32		6
+#define	R_SPARC_WDISP30		7
+#define	R_SPARC_WDISP22		8
+#define	R_SPARC_HI22		9
+#define	R_SPARC_22		10
+#define	R_SPARC_13		11
+#define	R_SPARC_LO10		12
+#define	R_SPARC_GOT10		13
+#define	R_SPARC_GOT13		14
+#define	R_SPARC_GOT22		15
+#define	R_SPARC_PC10		16
+#define	R_SPARC_PC22		17
+#define	R_SPARC_WPLT30		18
+#define	R_SPARC_COPY		19
+#define	R_SPARC_GLOB_DAT	20
+#define	R_SPARC_JMP_SLOT	21
+#define	R_SPARC_RELATIVE	22
+#define	R_SPARC_UA32		23
+#define R_SPARC_PLT32		24
+#define R_SPARC_HIPLT22		25
+#define R_SPARC_LOPLT10		26
+#define R_SPARC_PCPLT32		27
+#define R_SPARC_PCPLT22		28
+#define R_SPARC_PCPLT10		29
+#define R_SPARC_10		30
+#define R_SPARC_11		31
+#define R_SPARC_64		32
+#define R_SPARC_OLO10		33
+#define R_SPARC_WDISP16		40
+#define R_SPARC_WDISP19		41
+#define R_SPARC_7		43
+#define R_SPARC_5		44
+#define R_SPARC_6		45
+
+/* Bits present in AT_HWCAP, primarily for Sparc32.  */
+
+#define HWCAP_SPARC_FLUSH       1    /* CPU supports flush instruction. */
+#define HWCAP_SPARC_STBAR       2
+#define HWCAP_SPARC_SWAP        4
+#define HWCAP_SPARC_MULDIV      8
+#define HWCAP_SPARC_V9		16
+#define HWCAP_SPARC_ULTRA3	32
+
+#define CORE_DUMP_USE_REGSET
+
+/* Format is:
+ * 	G0 --> G7
+ *	O0 --> O7
+ *	L0 --> L7
+ *	I0 --> I7
+ *	PSR, PC, nPC, Y, WIM, TBR
+ */
+typedef unsigned long elf_greg_t;
+#define ELF_NGREG 38
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+
+typedef struct {
+	union {
+		unsigned long	pr_regs[32];
+		double		pr_dregs[16];
+	} pr_fr;
+	unsigned long __unused;
+	unsigned long	pr_fsr;
+	unsigned char	pr_qcnt;
+	unsigned char	pr_q_entrysize;
+	unsigned char	pr_en;
+	unsigned int	pr_q[64];
+} elf_fpregset_t;
+
+#include <asm/mbus.h>
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x) ((x)->e_machine == EM_SPARC)
+
+/*
+ * These are used to set parameters in the core dumps.
+ */
+#define ELF_ARCH	EM_SPARC
+#define ELF_CLASS	ELFCLASS32
+#define ELF_DATA	ELFDATA2MSB
+
+#define USE_ELF_CORE_DUMP
+#ifndef CONFIG_SUN4
+#define ELF_EXEC_PAGESIZE	4096
+#else
+#define ELF_EXEC_PAGESIZE	8192
+#endif
+
+
+/* This is the location that an ET_DYN program is loaded if exec'ed.  Typical
+   use of this is to invoke "./ld.so someprog" to test out a new version of
+   the loader.  We need to make sure that it is out of the way of the program
+   that it will "exec", and that there is sufficient room for the brk.  */
+
+#define ELF_ET_DYN_BASE         (TASK_UNMAPPED_BASE)
+
+/* This yields a mask that user programs can use to figure out what
+   instruction set this cpu supports.  This can NOT be done in userspace
+   on Sparc.  */
+
+/* Sun4c has none of the capabilities, most sun4m's have them all.
+ * XXX This is gross, set some global variable at boot time. -DaveM
+ */
+#define ELF_HWCAP	((ARCH_SUN4C_SUN4) ? 0 : \
+			 (HWCAP_SPARC_FLUSH | HWCAP_SPARC_STBAR | \
+			  HWCAP_SPARC_SWAP | \
+			  ((srmmu_modtype != Cypress && \
+			    srmmu_modtype != Cypress_vE && \
+			    srmmu_modtype != Cypress_vD) ? \
+			   HWCAP_SPARC_MULDIV : 0)))
+
+/* This yields a string that ld.so will use to load implementation
+   specific libraries for optimization.  This is more specific in
+   intent than poking at uname or /proc/cpuinfo. */
+
+#define ELF_PLATFORM	(NULL)
+
+#define SET_PERSONALITY(ex, ibcs2) set_personality((ibcs2)?PER_SVR4:PER_LINUX)
+
+#endif /* !(__ASMSPARC_ELF_H) */
diff --git a/include/asm-sparc/elf_64.h b/include/asm-sparc/elf_64.h
new file mode 100644
index 00000000000..0818a1308f4
--- /dev/null
+++ b/include/asm-sparc/elf_64.h
@@ -0,0 +1,217 @@
+#ifndef __ASM_SPARC64_ELF_H
+#define __ASM_SPARC64_ELF_H
+
+/*
+ * ELF register definitions..
+ */
+
+#include <asm/ptrace.h>
+#include <asm/processor.h>
+#include <asm/uaccess.h>
+#include <asm/spitfire.h>
+
+/*
+ * Sparc section types
+ */
+#define STT_REGISTER		13
+
+/*
+ * Sparc ELF relocation types
+ */
+#define	R_SPARC_NONE		0
+#define	R_SPARC_8		1
+#define	R_SPARC_16		2
+#define	R_SPARC_32		3
+#define	R_SPARC_DISP8		4
+#define	R_SPARC_DISP16		5
+#define	R_SPARC_DISP32		6
+#define	R_SPARC_WDISP30		7
+#define	R_SPARC_WDISP22		8
+#define	R_SPARC_HI22		9
+#define	R_SPARC_22		10
+#define	R_SPARC_13		11
+#define	R_SPARC_LO10		12
+#define	R_SPARC_GOT10		13
+#define	R_SPARC_GOT13		14
+#define	R_SPARC_GOT22		15
+#define	R_SPARC_PC10		16
+#define	R_SPARC_PC22		17
+#define	R_SPARC_WPLT30		18
+#define	R_SPARC_COPY		19
+#define	R_SPARC_GLOB_DAT	20
+#define	R_SPARC_JMP_SLOT	21
+#define	R_SPARC_RELATIVE	22
+#define	R_SPARC_UA32		23
+#define R_SPARC_PLT32		24
+#define R_SPARC_HIPLT22		25
+#define R_SPARC_LOPLT10		26
+#define R_SPARC_PCPLT32		27
+#define R_SPARC_PCPLT22		28
+#define R_SPARC_PCPLT10		29
+#define R_SPARC_10		30
+#define R_SPARC_11		31
+#define R_SPARC_64		32
+#define R_SPARC_OLO10		33
+#define R_SPARC_WDISP16		40
+#define R_SPARC_WDISP19		41
+#define R_SPARC_7		43
+#define R_SPARC_5		44
+#define R_SPARC_6		45
+
+/* Bits present in AT_HWCAP, primarily for Sparc32.  */
+
+#define HWCAP_SPARC_FLUSH       1    /* CPU supports flush instruction. */
+#define HWCAP_SPARC_STBAR       2
+#define HWCAP_SPARC_SWAP        4
+#define HWCAP_SPARC_MULDIV      8
+#define HWCAP_SPARC_V9		16
+#define HWCAP_SPARC_ULTRA3	32
+#define HWCAP_SPARC_BLKINIT	64
+#define HWCAP_SPARC_N2		128
+
+#define CORE_DUMP_USE_REGSET
+
+/*
+ * These are used to set parameters in the core dumps.
+ */
+#define ELF_ARCH		EM_SPARCV9
+#define ELF_CLASS		ELFCLASS64
+#define ELF_DATA		ELFDATA2MSB
+
+/* Format of 64-bit elf_gregset_t is:
+ * 	G0 --> G7
+ * 	O0 --> O7
+ * 	L0 --> L7
+ * 	I0 --> I7
+ *	TSTATE
+ *	TPC
+ *	TNPC
+ *	Y
+ */
+typedef unsigned long elf_greg_t;
+#define ELF_NGREG 36
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+
+typedef struct {
+	unsigned long	pr_regs[32];
+	unsigned long	pr_fsr;
+	unsigned long	pr_gsr;
+	unsigned long	pr_fprs;
+} elf_fpregset_t;
+
+/* Format of 32-bit elf_gregset_t is:
+ * 	G0 --> G7
+ *	O0 --> O7
+ *	L0 --> L7
+ *	I0 --> I7
+ *	PSR, PC, nPC, Y, WIM, TBR
+ */
+typedef unsigned int compat_elf_greg_t;
+#define COMPAT_ELF_NGREG 38
+typedef compat_elf_greg_t compat_elf_gregset_t[COMPAT_ELF_NGREG];
+
+typedef struct {
+	union {
+		unsigned int	pr_regs[32];
+		unsigned long	pr_dregs[16];
+	} pr_fr;
+	unsigned int __unused;
+	unsigned int	pr_fsr;
+	unsigned char	pr_qcnt;
+	unsigned char	pr_q_entrysize;
+	unsigned char	pr_en;
+	unsigned int	pr_q[64];
+} compat_elf_fpregset_t;
+
+/* UltraSparc extensions.  Still unused, but will be eventually.  */
+typedef struct {
+	unsigned int pr_type;
+	unsigned int pr_align;
+	union {
+		struct {
+			union {
+				unsigned int	pr_regs[32];
+				unsigned long	pr_dregs[16];
+				long double	pr_qregs[8];
+			} pr_xfr;
+		} pr_v8p;
+		unsigned int	pr_xfsr;
+		unsigned int	pr_fprs;
+		unsigned int	pr_xg[8];
+		unsigned int	pr_xo[8];
+		unsigned long	pr_tstate;
+		unsigned int	pr_filler[8];
+	} pr_un;
+} elf_xregset_t;
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x)		((x)->e_machine == ELF_ARCH)
+#define compat_elf_check_arch(x)	((x)->e_machine == EM_SPARC || \
+					 (x)->e_machine == EM_SPARC32PLUS)
+#define compat_start_thread		start_thread32
+
+#define USE_ELF_CORE_DUMP
+#define ELF_EXEC_PAGESIZE	PAGE_SIZE
+
+/* This is the location that an ET_DYN program is loaded if exec'ed.  Typical
+   use of this is to invoke "./ld.so someprog" to test out a new version of
+   the loader.  We need to make sure that it is out of the way of the program
+   that it will "exec", and that there is sufficient room for the brk.  */
+
+#define ELF_ET_DYN_BASE		0x0000010000000000UL
+#define COMPAT_ELF_ET_DYN_BASE	0x0000000070000000UL
+
+
+/* This yields a mask that user programs can use to figure out what
+   instruction set this cpu supports.  */
+
+/* On Ultra, we support all of the v8 capabilities. */
+static inline unsigned int sparc64_elf_hwcap(void)
+{
+	unsigned int cap = (HWCAP_SPARC_FLUSH | HWCAP_SPARC_STBAR |
+			    HWCAP_SPARC_SWAP | HWCAP_SPARC_MULDIV |
+			    HWCAP_SPARC_V9);
+
+	if (tlb_type == cheetah || tlb_type == cheetah_plus)
+		cap |= HWCAP_SPARC_ULTRA3;
+	else if (tlb_type == hypervisor) {
+		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 ||
+		    sun4v_chip_type == SUN4V_CHIP_NIAGARA2)
+			cap |= HWCAP_SPARC_BLKINIT;
+		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2)
+			cap |= HWCAP_SPARC_N2;
+	}
+
+	return cap;
+}
+
+#define ELF_HWCAP	sparc64_elf_hwcap();
+
+/* This yields a string that ld.so will use to load implementation
+   specific libraries for optimization.  This is more specific in
+   intent than poking at uname or /proc/cpuinfo.  */
+
+#define ELF_PLATFORM	(NULL)
+
+#define SET_PERSONALITY(ex, ibcs2)			\
+do {	unsigned long new_flags = current_thread_info()->flags; \
+	new_flags &= _TIF_32BIT;			\
+	if ((ex).e_ident[EI_CLASS] == ELFCLASS32)	\
+		new_flags |= _TIF_32BIT;		\
+	else						\
+		new_flags &= ~_TIF_32BIT;		\
+	if ((current_thread_info()->flags & _TIF_32BIT) \
+	    != new_flags)				\
+		set_thread_flag(TIF_ABI_PENDING);	\
+	else						\
+		clear_thread_flag(TIF_ABI_PENDING);	\
+	/* flush_thread will update pgd cache */	\
+	if (ibcs2)					\
+		set_personality(PER_SVR4);		\
+	else if (current->personality != PER_LINUX32)	\
+		set_personality(PER_LINUX);		\
+} while (0)
+
+#endif /* !(__ASM_SPARC64_ELF_H) */
diff --git a/include/asm-sparc/envctrl.h b/include/asm-sparc/envctrl.h
new file mode 100644
index 00000000000..624fa7e2da8
--- /dev/null
+++ b/include/asm-sparc/envctrl.h
@@ -0,0 +1,103 @@
+/*
+ *
+ * envctrl.h: Definitions for access to the i2c environment
+ *            monitoring on Ultrasparc systems.
+ *
+ * Copyright (C) 1998  Eddie C. Dost  (ecd@skynet.be)
+ * Copyright (C) 2000  Vinh Truong  (vinh.truong@eng.sun.com)
+ * VT - Add all ioctl commands and environment status definitions
+ * VT - Add application note
+ */
+#ifndef _SPARC64_ENVCTRL_H
+#define _SPARC64_ENVCTRL_H 1
+
+#include <linux/ioctl.h>
+
+/* Application note:
+ *
+ * The driver supports 4 operations: open(), close(), ioctl(), read()
+ * The device name is /dev/envctrl.
+ * Below is sample usage:
+ *
+ *	fd = open("/dev/envtrl", O_RDONLY);
+ *	if (ioctl(fd, ENVCTRL_READ_SHUTDOWN_TEMPERATURE, 0) < 0)
+ *		printf("error\n");
+ *	ret = read(fd, buf, 10);
+ *	close(fd);
+ *
+ * Notice in the case of cpu voltage and temperature, the default is
+ * cpu0.  If we need to know the info of cpu1, cpu2, cpu3, we need to
+ * pass in cpu number in ioctl() last parameter.  For example, to
+ * get the voltage of cpu2:
+ *
+ *	ioctlbuf[0] = 2;
+ *	if (ioctl(fd, ENVCTRL_READ_CPU_VOLTAGE, ioctlbuf) < 0)
+ *		printf("error\n");
+ *	ret = read(fd, buf, 10);
+ *
+ * All the return values are in ascii.  So check read return value
+ * and do appropriate conversions in your application.
+ */
+
+/* IOCTL commands */
+
+/* Note: these commands reflect possible monitor features.
+ * Some boards choose to support some of the features only.
+ */
+#define ENVCTRL_RD_CPU_TEMPERATURE	_IOR('p', 0x40, int)
+#define ENVCTRL_RD_CPU_VOLTAGE		_IOR('p', 0x41, int)
+#define ENVCTRL_RD_FAN_STATUS		_IOR('p', 0x42, int)
+#define ENVCTRL_RD_WARNING_TEMPERATURE	_IOR('p', 0x43, int)
+#define ENVCTRL_RD_SHUTDOWN_TEMPERATURE	_IOR('p', 0x44, int)
+#define ENVCTRL_RD_VOLTAGE_STATUS	_IOR('p', 0x45, int)
+#define ENVCTRL_RD_SCSI_TEMPERATURE	_IOR('p', 0x46, int)
+#define ENVCTRL_RD_ETHERNET_TEMPERATURE	_IOR('p', 0x47, int)
+#define ENVCTRL_RD_MTHRBD_TEMPERATURE	_IOR('p', 0x48, int)
+
+#define ENVCTRL_RD_GLOBALADDRESS	_IOR('p', 0x49, int)
+
+/* Read return values for a voltage status request. */
+#define ENVCTRL_VOLTAGE_POWERSUPPLY_GOOD	0x01
+#define ENVCTRL_VOLTAGE_BAD			0x02
+#define ENVCTRL_POWERSUPPLY_BAD			0x03
+#define ENVCTRL_VOLTAGE_POWERSUPPLY_BAD		0x04
+
+/* Read return values for a fan status request.
+ * A failure match means either the fan fails or
+ * the fan is not connected.  Some boards have optional
+ * connectors to connect extra fans.
+ *
+ * There are maximum 8 monitor fans.  Some are cpu fans
+ * some are system fans.  The mask below only indicates
+ * fan by order number.
+ * Below is a sample application:
+ *
+ *	if (ioctl(fd, ENVCTRL_READ_FAN_STATUS, 0) < 0) {
+ *		printf("ioctl fan failed\n");
+ *	}
+ *	if (read(fd, rslt, 1) <= 0) {
+ *		printf("error or fan not monitored\n");
+ *	} else {
+ *		if (rslt[0] == ENVCTRL_ALL_FANS_GOOD) {
+ *			printf("all fans good\n");
+ *	} else if (rslt[0] == ENVCTRL_ALL_FANS_BAD) {
+ *		printf("all fans bad\n");
+ *	} else {
+ *		if (rslt[0] & ENVCTRL_FAN0_FAILURE_MASK) {
+ *			printf("fan 0 failed or not connected\n");
+ *	}
+ *	......
+ */
+
+#define ENVCTRL_ALL_FANS_GOOD			0x00
+#define ENVCTRL_FAN0_FAILURE_MASK		0x01
+#define ENVCTRL_FAN1_FAILURE_MASK		0x02
+#define ENVCTRL_FAN2_FAILURE_MASK		0x04
+#define ENVCTRL_FAN3_FAILURE_MASK		0x08
+#define ENVCTRL_FAN4_FAILURE_MASK		0x10
+#define ENVCTRL_FAN5_FAILURE_MASK		0x20
+#define ENVCTRL_FAN6_FAILURE_MASK		0x40
+#define ENVCTRL_FAN7_FAILURE_MASK		0x80
+#define ENVCTRL_ALL_FANS_BAD 			0xFF
+
+#endif /* !(_SPARC64_ENVCTRL_H) */
diff --git a/include/asm-sparc/estate.h b/include/asm-sparc/estate.h
new file mode 100644
index 00000000000..520c08560d1
--- /dev/null
+++ b/include/asm-sparc/estate.h
@@ -0,0 +1,49 @@
+#ifndef _SPARC64_ESTATE_H
+#define _SPARC64_ESTATE_H
+
+/* UltraSPARC-III E-cache Error Enable */
+#define ESTATE_ERROR_FMT	0x0000000000040000 /* Force MTAG ECC		*/
+#define ESTATE_ERROR_FMESS	0x000000000003c000 /* Forced MTAG ECC val	*/
+#define ESTATE_ERROR_FMD	0x0000000000002000 /* Force DATA ECC		*/
+#define ESTATE_ERROR_FDECC	0x0000000000001ff0 /* Forced DATA ECC val	*/
+#define ESTATE_ERROR_UCEEN	0x0000000000000008 /* See below			*/
+#define ESTATE_ERROR_NCEEN	0x0000000000000002 /* See below			*/
+#define ESTATE_ERROR_CEEN	0x0000000000000001 /* See below			*/
+
+/* UCEEN enables the fast_ECC_error trap for: 1) software correctable E-cache
+ * errors 2) uncorrectable E-cache errors.  Such events only occur on reads
+ * of the E-cache by the local processor for: 1) data loads 2) instruction
+ * fetches 3) atomic operations.  Such events _cannot_ occur for: 1) merge
+ * 2) writeback 2) copyout.  The AFSR bits associated with these traps are
+ * UCC and UCU.
+ */
+
+/* NCEEN enables instruction_access_error, data_access_error, and ECC_error traps
+ * for uncorrectable ECC errors and system errors.
+ *
+ * Uncorrectable system bus data error or MTAG ECC error, system bus TimeOUT,
+ * or system bus BusERR:
+ * 1) As the result of an instruction fetch, will generate instruction_access_error
+ * 2) As the result of a load etc. will generate data_access_error.
+ * 3) As the result of store merge completion, writeback, or copyout will
+ *    generate a disrupting ECC_error trap.
+ * 4) As the result of such errors on instruction vector fetch can generate any
+ *    of the 3 trap types.
+ *
+ * The AFSR bits associated with these traps are EMU, EDU, WDU, CPU, IVU, UE,
+ * BERR, and TO.
+ */
+
+/* CEEN enables the ECC_error trap for hardware corrected ECC errors.  System bus
+ * reads resulting in a hardware corrected data or MTAG ECC error will generate an
+ * ECC_error disrupting trap with this bit enabled.
+ *
+ * This same trap will also be generated when a hardware corrected ECC error results
+ * during store merge, writeback, and copyout operations.
+ */
+
+/* In general, if the trap enable bits above are disabled the AFSR bits will still
+ * log the events even though the trap will not be generated by the processor.
+ */
+
+#endif /* _SPARC64_ESTATE_H */
diff --git a/include/asm-sparc/fbio.h b/include/asm-sparc/fbio.h
index c2b27e7a7ca..b9215a0907d 100644
--- a/include/asm-sparc/fbio.h
+++ b/include/asm-sparc/fbio.h
@@ -1,6 +1,9 @@
 #ifndef __LINUX_FBIO_H
 #define __LINUX_FBIO_H
 
+#include <linux/compiler.h>
+#include <linux/types.h>
+
 /* Constants used for fbio SunOS compatibility */
 /* (C) 1996 Miguel de Icaza */
 
@@ -38,6 +41,9 @@
 #define FBTYPE_PCI_IGA1682	23
 #define FBTYPE_P9100COLOR	24
 
+#define FBTYPE_PCI_GENERIC	1000
+#define FBTYPE_PCI_MACH64	1001
+
 /* fbio ioctls */
 /* Returned by FBIOGTYPE */
 struct  fbtype {
@@ -97,8 +103,8 @@ struct fbcursor {
         struct fbcurpos hot;    /* cursor hot spot */
         struct fbcmap cmap;     /* color map info */
         struct fbcurpos size;   /* cursor bit map size */
-        char *image;            /* cursor image bits */
-        char *mask;             /* cursor mask bits */
+        char __user *image;     /* cursor image bits */
+        char __user *mask;      /* cursor mask bits */
 };
 
 /* set/get cursor attributes/shape */
@@ -294,4 +300,31 @@ struct fb_clut32 {
 #define LEO_LD_GBL_MAP         0x01009000
 #define LEO_UNK2_MAP           0x0100a000
 
+#ifdef __KERNEL__
+struct  fbcmap32 {
+	int             index;          /* first element (0 origin) */
+	int             count;
+	u32		red;
+	u32		green;
+	u32		blue;
+};
+
+#define FBIOPUTCMAP32	_IOW('F', 3, struct fbcmap32)
+#define FBIOGETCMAP32	_IOW('F', 4, struct fbcmap32)
+
+struct fbcursor32 {
+	short set;		/* what to set, choose from the list above */
+	short enable;		/* cursor on/off */
+	struct fbcurpos pos;	/* cursor position */
+	struct fbcurpos hot;	/* cursor hot spot */
+	struct fbcmap32 cmap;	/* color map info */
+	struct fbcurpos size;	/* cursor bit map size */
+	u32	image;		/* cursor image bits */
+	u32	mask;		/* cursor mask bits */
+};
+
+#define FBIOSCURSOR32	_IOW('F', 24, struct fbcursor32)
+#define FBIOGCURSOR32	_IOW('F', 25, struct fbcursor32)
+#endif
+
 #endif /* __LINUX_FBIO_H */
diff --git a/include/asm-sparc/fcntl.h b/include/asm-sparc/fcntl.h
index 07bd2d80257..d4d9c9d852c 100644
--- a/include/asm-sparc/fcntl.h
+++ b/include/asm-sparc/fcntl.h
@@ -10,7 +10,11 @@
 #define O_EXCL		0x0800	/* not fcntl */
 #define O_SYNC		0x2000
 #define O_NONBLOCK	0x4000
+#if defined(__sparc__) && defined(__arch64__)
+#define O_NDELAY	0x0004
+#else
 #define O_NDELAY	(0x0004 | O_NONBLOCK)
+#endif
 #define O_NOCTTY	0x8000	/* not fcntl */
 #define O_LARGEFILE	0x40000
 #define O_DIRECT        0x100000 /* direct disk access hint */
diff --git a/include/asm-sparc/fhc.h b/include/asm-sparc/fhc.h
new file mode 100644
index 00000000000..788cbc46a11
--- /dev/null
+++ b/include/asm-sparc/fhc.h
@@ -0,0 +1,121 @@
+/*
+ * fhc.h: Structures for central/fhc pseudo driver on Sunfire/Starfire/Wildfire.
+ *
+ * Copyright (C) 1997, 1999 David S. Miller (davem@redhat.com)
+ */
+
+#ifndef _SPARC64_FHC_H
+#define _SPARC64_FHC_H
+
+#include <linux/timer.h>
+
+#include <asm/oplib.h>
+#include <asm/prom.h>
+#include <asm/upa.h>
+
+struct linux_fhc;
+
+/* Clock board register offsets. */
+#define CLOCK_CTRL	0x00UL	/* Main control */
+#define CLOCK_STAT1	0x10UL	/* Status one */
+#define CLOCK_STAT2	0x20UL	/* Status two */
+#define CLOCK_PWRSTAT	0x30UL	/* Power status */
+#define CLOCK_PWRPRES	0x40UL	/* Power presence */
+#define CLOCK_TEMP	0x50UL	/* Temperature */
+#define CLOCK_IRQDIAG	0x60UL	/* IRQ diagnostics */
+#define CLOCK_PWRSTAT2	0x70UL	/* Power status two */
+
+#define CLOCK_CTRL_LLED		0x04	/* Left LED, 0 == on */
+#define CLOCK_CTRL_MLED		0x02	/* Mid LED, 1 == on */
+#define CLOCK_CTRL_RLED		0x01	/* RIght LED, 1 == on */
+
+struct linux_central {
+	struct linux_fhc		*child;
+	unsigned long			cfreg;
+	unsigned long			clkregs;
+	unsigned long			clkver;
+	int				slots;
+	struct device_node		*prom_node;
+
+	struct linux_prom_ranges	central_ranges[PROMREG_MAX];
+	int				num_central_ranges;
+};
+
+/* Firehose controller register offsets */
+struct fhc_regs {
+	unsigned long			pregs;	/* FHC internal regs */
+#define FHC_PREGS_ID	0x00UL	/* FHC ID */
+#define  FHC_ID_VERS		0xf0000000 /* Version of this FHC		*/
+#define  FHC_ID_PARTID		0x0ffff000 /* Part ID code (0x0f9f == FHC)	*/
+#define  FHC_ID_MANUF		0x0000007e /* Manufacturer (0x3e == SUN's JEDEC)*/
+#define  FHC_ID_RESV		0x00000001 /* Read as one			*/
+#define FHC_PREGS_RCS	0x10UL	/* FHC Reset Control/Status Register */
+#define  FHC_RCS_POR		0x80000000 /* Last reset was a power cycle	*/
+#define  FHC_RCS_SPOR		0x40000000 /* Last reset was sw power on reset	*/
+#define  FHC_RCS_SXIR		0x20000000 /* Last reset was sw XIR reset	*/
+#define  FHC_RCS_BPOR		0x10000000 /* Last reset was due to POR button	*/
+#define  FHC_RCS_BXIR		0x08000000 /* Last reset was due to XIR button	*/
+#define  FHC_RCS_WEVENT		0x04000000 /* CPU reset was due to wakeup event	*/
+#define  FHC_RCS_CFATAL		0x02000000 /* Centerplane Fatal Error signalled	*/
+#define  FHC_RCS_FENAB		0x01000000 /* Fatal errors elicit system reset	*/
+#define FHC_PREGS_CTRL	0x20UL	/* FHC Control Register */
+#define  FHC_CONTROL_ICS	0x00100000 /* Ignore Centerplane Signals	*/
+#define  FHC_CONTROL_FRST	0x00080000 /* Fatal Error Reset Enable		*/
+#define  FHC_CONTROL_LFAT	0x00040000 /* AC/DC signalled a local error	*/
+#define  FHC_CONTROL_SLINE	0x00010000 /* Firmware Synchronization Line	*/
+#define  FHC_CONTROL_DCD	0x00008000 /* DC-->DC Converter Disable		*/
+#define  FHC_CONTROL_POFF	0x00004000 /* AC/DC Controller PLL Disable	*/
+#define  FHC_CONTROL_FOFF	0x00002000 /* FHC Controller PLL Disable	*/
+#define  FHC_CONTROL_AOFF	0x00001000 /* CPU A SRAM/SBD Low Power Mode	*/
+#define  FHC_CONTROL_BOFF	0x00000800 /* CPU B SRAM/SBD Low Power Mode	*/
+#define  FHC_CONTROL_PSOFF	0x00000400 /* Turns off this FHC's power supply	*/
+#define  FHC_CONTROL_IXIST	0x00000200 /* 0=FHC tells clock board it exists	*/
+#define  FHC_CONTROL_XMSTR	0x00000100 /* 1=Causes this FHC to be XIR master*/
+#define  FHC_CONTROL_LLED	0x00000040 /* 0=Left LED ON			*/
+#define  FHC_CONTROL_MLED	0x00000020 /* 1=Middle LED ON			*/
+#define  FHC_CONTROL_RLED	0x00000010 /* 1=Right LED			*/
+#define  FHC_CONTROL_BPINS	0x00000003 /* Spare Bidirectional Pins		*/
+#define FHC_PREGS_BSR	0x30UL	/* FHC Board Status Register */
+#define  FHC_BSR_DA64		0x00040000 /* Port A: 0=128bit 1=64bit data path */
+#define  FHC_BSR_DB64		0x00020000 /* Port B: 0=128bit 1=64bit data path */
+#define  FHC_BSR_BID		0x0001e000 /* Board ID                           */
+#define  FHC_BSR_SA		0x00001c00 /* Port A UPA Speed (from the pins)   */
+#define  FHC_BSR_SB		0x00000380 /* Port B UPA Speed (from the pins)   */
+#define  FHC_BSR_NDIAG		0x00000040 /* Not in Diag Mode                   */
+#define  FHC_BSR_NTBED		0x00000020 /* Not in TestBED Mode                */
+#define  FHC_BSR_NIA		0x0000001c /* Jumper, bit 18 in PROM space       */
+#define  FHC_BSR_SI		0x00000001 /* Spare input pin value              */
+#define FHC_PREGS_ECC	0x40UL	/* FHC ECC Control Register (16 bits) */
+#define FHC_PREGS_JCTRL	0xf0UL	/* FHC JTAG Control Register */
+#define  FHC_JTAG_CTRL_MENAB	0x80000000 /* Indicates this is JTAG Master	 */
+#define  FHC_JTAG_CTRL_MNONE	0x40000000 /* Indicates no JTAG Master present	 */
+#define FHC_PREGS_JCMD	0x100UL	/* FHC JTAG Command Register */
+	unsigned long			ireg;	/* FHC IGN reg */
+#define FHC_IREG_IGN	0x00UL	/* This FHC's IGN */
+	unsigned long			ffregs;	/* FHC fanfail regs */
+#define FHC_FFREGS_IMAP	0x00UL	/* FHC Fanfail IMAP */
+#define FHC_FFREGS_ICLR	0x10UL	/* FHC Fanfail ICLR */
+	unsigned long			sregs;	/* FHC system regs */
+#define FHC_SREGS_IMAP	0x00UL	/* FHC System IMAP */
+#define FHC_SREGS_ICLR	0x10UL	/* FHC System ICLR */
+	unsigned long			uregs;	/* FHC uart regs */
+#define FHC_UREGS_IMAP	0x00UL	/* FHC Uart IMAP */
+#define FHC_UREGS_ICLR	0x10UL	/* FHC Uart ICLR */
+	unsigned long			tregs;	/* FHC TOD regs */
+#define FHC_TREGS_IMAP	0x00UL	/* FHC TOD IMAP */
+#define FHC_TREGS_ICLR	0x10UL	/* FHC TOD ICLR */
+};
+
+struct linux_fhc {
+	struct linux_fhc		*next;
+	struct linux_central		*parent;	/* NULL if not central FHC */
+	struct fhc_regs			fhc_regs;
+	int				board;
+	int				jtag_master;
+	struct device_node		*prom_node;
+
+	struct linux_prom_ranges	fhc_ranges[PROMREG_MAX];
+	int				num_fhc_ranges;
+};
+
+#endif /* !(_SPARC64_FHC_H) */
diff --git a/include/asm-sparc/floppy.h b/include/asm-sparc/floppy.h
index d3978e068e2..6c628ba15a8 100644
--- a/include/asm-sparc/floppy.h
+++ b/include/asm-sparc/floppy.h
@@ -1,388 +1,8 @@
-/* asm-sparc/floppy.h: Sparc specific parts of the Floppy driver.
- *
- * Copyright (C) 1995 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef __ASM_SPARC_FLOPPY_H
-#define __ASM_SPARC_FLOPPY_H
-
-#include <asm/page.h>
-#include <asm/pgtable.h>
-#include <asm/system.h>
-#include <asm/idprom.h>
-#include <asm/machines.h>
-#include <asm/oplib.h>
-#include <asm/auxio.h>
-#include <asm/irq.h>
-
-/* We don't need no stinkin' I/O port allocation crap. */
-#undef release_region
-#undef request_region
-#define release_region(X, Y)	do { } while(0)
-#define request_region(X, Y, Z)	(1)
-
-/* References:
- * 1) Netbsd Sun floppy driver.
- * 2) NCR 82077 controller manual
- * 3) Intel 82077 controller manual
- */
-struct sun_flpy_controller {
-	volatile unsigned char status_82072;  /* Main Status reg. */
-#define dcr_82072              status_82072   /* Digital Control reg. */
-#define status1_82077          status_82072   /* Auxiliary Status reg. 1 */
-
-	volatile unsigned char data_82072;    /* Data fifo. */
-#define status2_82077          data_82072     /* Auxiliary Status reg. 2 */
-
-	volatile unsigned char dor_82077;     /* Digital Output reg. */
-	volatile unsigned char tapectl_82077; /* What the? Tape control reg? */
-
-	volatile unsigned char status_82077;  /* Main Status Register. */
-#define drs_82077              status_82077   /* Digital Rate Select reg. */
-
-	volatile unsigned char data_82077;    /* Data fifo. */
-	volatile unsigned char ___unused;
-	volatile unsigned char dir_82077;     /* Digital Input reg. */
-#define dcr_82077              dir_82077      /* Config Control reg. */
-};
-
-/* You'll only ever find one controller on a SparcStation anyways. */
-static struct sun_flpy_controller *sun_fdc = NULL;
-extern volatile unsigned char *fdc_status;
-
-struct sun_floppy_ops {
-	unsigned char (*fd_inb)(int port);
-	void (*fd_outb)(unsigned char value, int port);
-};
-
-static struct sun_floppy_ops sun_fdops;
-
-#define fd_inb(port)              sun_fdops.fd_inb(port)
-#define fd_outb(value,port)       sun_fdops.fd_outb(value,port)
-#define fd_enable_dma()           sun_fd_enable_dma()
-#define fd_disable_dma()          sun_fd_disable_dma()
-#define fd_request_dma()          (0) /* nothing... */
-#define fd_free_dma()             /* nothing... */
-#define fd_clear_dma_ff()         /* nothing... */
-#define fd_set_dma_mode(mode)     sun_fd_set_dma_mode(mode)
-#define fd_set_dma_addr(addr)     sun_fd_set_dma_addr(addr)
-#define fd_set_dma_count(count)   sun_fd_set_dma_count(count)
-#define fd_enable_irq()           /* nothing... */
-#define fd_disable_irq()          /* nothing... */
-#define fd_cacheflush(addr, size) /* nothing... */
-#define fd_request_irq()          sun_fd_request_irq()
-#define fd_free_irq()             /* nothing... */
-#if 0  /* P3: added by Alain, these cause a MMU corruption. 19960524 XXX */
-#define fd_dma_mem_alloc(size)    ((unsigned long) vmalloc(size))
-#define fd_dma_mem_free(addr,size) (vfree((void *)(addr)))
+#ifndef ___ASM_SPARC_FLOPPY_H
+#define ___ASM_SPARC_FLOPPY_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/floppy_64.h>
+#else
+#include <asm-sparc/floppy_32.h>
+#endif
 #endif
-
-/* XXX This isn't really correct. XXX */
-#define get_dma_residue(x)        (0)
-
-#define FLOPPY0_TYPE  4
-#define FLOPPY1_TYPE  0
-
-/* Super paranoid... */
-#undef HAVE_DISABLE_HLT
-
-/* Here is where we catch the floppy driver trying to initialize,
- * therefore this is where we call the PROM device tree probing
- * routine etc. on the Sparc.
- */
-#define FDC1                      sun_floppy_init()
-
-#define N_FDC    1
-#define N_DRIVE  8
-
-/* No 64k boundary crossing problems on the Sparc. */
-#define CROSS_64KB(a,s) (0)
-
-/* Routines unique to each controller type on a Sun. */
-static void sun_set_dor(unsigned char value, int fdc_82077)
-{
-	if (sparc_cpu_model == sun4c) {
-		unsigned int bits = 0;
-		if (value & 0x10)
-			bits |= AUXIO_FLPY_DSEL;
-		if ((value & 0x80) == 0)
-			bits |= AUXIO_FLPY_EJCT;
-		set_auxio(bits, (~bits) & (AUXIO_FLPY_DSEL|AUXIO_FLPY_EJCT));
-	}
-	if (fdc_82077) {
-		sun_fdc->dor_82077 = value;
-	}
-}
-
-static unsigned char sun_read_dir(void)
-{
-	if (sparc_cpu_model == sun4c)
-		return (get_auxio() & AUXIO_FLPY_DCHG) ? 0x80 : 0;
-	else
-		return sun_fdc->dir_82077;
-}
-
-static unsigned char sun_82072_fd_inb(int port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to read unknown port %d\n", port);
-		panic("floppy: Port bolixed.");
-	case 4: /* FD_STATUS */
-		return sun_fdc->status_82072 & ~STATUS_DMA;
-	case 5: /* FD_DATA */
-		return sun_fdc->data_82072;
-	case 7: /* FD_DIR */
-		return sun_read_dir();
-	};
-	panic("sun_82072_fd_inb: How did I get here?");
-}
-
-static void sun_82072_fd_outb(unsigned char value, int port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to write to unknown port %d\n", port);
-		panic("floppy: Port bolixed.");
-	case 2: /* FD_DOR */
-		sun_set_dor(value, 0);
-		break;
-	case 5: /* FD_DATA */
-		sun_fdc->data_82072 = value;
-		break;
-	case 7: /* FD_DCR */
-		sun_fdc->dcr_82072 = value;
-		break;
-	case 4: /* FD_STATUS */
-		sun_fdc->status_82072 = value;
-		break;
-	};
-	return;
-}
-
-static unsigned char sun_82077_fd_inb(int port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to read unknown port %d\n", port);
-		panic("floppy: Port bolixed.");
-	case 0: /* FD_STATUS_0 */
-		return sun_fdc->status1_82077;
-	case 1: /* FD_STATUS_1 */
-		return sun_fdc->status2_82077;
-	case 2: /* FD_DOR */
-		return sun_fdc->dor_82077;
-	case 3: /* FD_TDR */
-		return sun_fdc->tapectl_82077;
-	case 4: /* FD_STATUS */
-		return sun_fdc->status_82077 & ~STATUS_DMA;
-	case 5: /* FD_DATA */
-		return sun_fdc->data_82077;
-	case 7: /* FD_DIR */
-		return sun_read_dir();
-	};
-	panic("sun_82077_fd_inb: How did I get here?");
-}
-
-static void sun_82077_fd_outb(unsigned char value, int port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to write to unknown port %d\n", port);
-		panic("floppy: Port bolixed.");
-	case 2: /* FD_DOR */
-		sun_set_dor(value, 1);
-		break;
-	case 5: /* FD_DATA */
-		sun_fdc->data_82077 = value;
-		break;
-	case 7: /* FD_DCR */
-		sun_fdc->dcr_82077 = value;
-		break;
-	case 4: /* FD_STATUS */
-		sun_fdc->status_82077 = value;
-		break;
-	case 3: /* FD_TDR */
-		sun_fdc->tapectl_82077 = value;
-		break;
-	};
-	return;
-}
-
-/* For pseudo-dma (Sun floppy drives have no real DMA available to
- * them so we must eat the data fifo bytes directly ourselves) we have
- * three state variables.  doing_pdma tells our inline low-level
- * assembly floppy interrupt entry point whether it should sit and eat
- * bytes from the fifo or just transfer control up to the higher level
- * floppy interrupt c-code.  I tried very hard but I could not get the
- * pseudo-dma to work in c-code without getting many overruns and
- * underruns.  If non-zero, doing_pdma encodes the direction of
- * the transfer for debugging.  1=read 2=write
- */
-extern char *pdma_vaddr;
-extern unsigned long pdma_size;
-extern volatile int doing_pdma;
-
-/* This is software state */
-extern char *pdma_base;
-extern unsigned long pdma_areasize;
-
-/* Common routines to all controller types on the Sparc. */
-static inline void virtual_dma_init(void)
-{
-	/* nothing... */
-}
-
-static inline void sun_fd_disable_dma(void)
-{
-	doing_pdma = 0;
-	if (pdma_base) {
-		mmu_unlockarea(pdma_base, pdma_areasize);
-		pdma_base = NULL;
-	}
-}
-
-static inline void sun_fd_set_dma_mode(int mode)
-{
-	switch(mode) {
-	case DMA_MODE_READ:
-		doing_pdma = 1;
-		break;
-	case DMA_MODE_WRITE:
-		doing_pdma = 2;
-		break;
-	default:
-		printk("Unknown dma mode %d\n", mode);
-		panic("floppy: Giving up...");
-	}
-}
-
-static inline void sun_fd_set_dma_addr(char *buffer)
-{
-	pdma_vaddr = buffer;
-}
-
-static inline void sun_fd_set_dma_count(int length)
-{
-	pdma_size = length;
-}
-
-static inline void sun_fd_enable_dma(void)
-{
-	pdma_vaddr = mmu_lockarea(pdma_vaddr, pdma_size);
-	pdma_base = pdma_vaddr;
-	pdma_areasize = pdma_size;
-}
-
-/* Our low-level entry point in arch/sparc/kernel/entry.S */
-extern int sparc_floppy_request_irq(int irq, unsigned long flags,
-				    irq_handler_t irq_handler);
-
-static int sun_fd_request_irq(void)
-{
-	static int once = 0;
-	int error;
-
-	if(!once) {
-		once = 1;
-		error = sparc_floppy_request_irq(FLOPPY_IRQ,
-						 IRQF_DISABLED,
-						 floppy_interrupt);
-		return ((error == 0) ? 0 : -1);
-	} else return 0;
-}
-
-static struct linux_prom_registers fd_regs[2];
-
-static int sun_floppy_init(void)
-{
-	char state[128];
-	int tnode, fd_node, num_regs;
-	struct resource r;
-
-	use_virtual_dma = 1;
-	
-	FLOPPY_IRQ = 11;
-	/* Forget it if we aren't on a machine that could possibly
-	 * ever have a floppy drive.
-	 */
-	if((sparc_cpu_model != sun4c && sparc_cpu_model != sun4m) ||
-	   ((idprom->id_machtype == (SM_SUN4C | SM_4C_SLC)) ||
-	    (idprom->id_machtype == (SM_SUN4C | SM_4C_ELC)))) {
-		/* We certainly don't have a floppy controller. */
-		goto no_sun_fdc;
-	}
-	/* Well, try to find one. */
-	tnode = prom_getchild(prom_root_node);
-	fd_node = prom_searchsiblings(tnode, "obio");
-	if(fd_node != 0) {
-		tnode = prom_getchild(fd_node);
-		fd_node = prom_searchsiblings(tnode, "SUNW,fdtwo");
-	} else {
-		fd_node = prom_searchsiblings(tnode, "fd");
-	}
-	if(fd_node == 0) {
-		goto no_sun_fdc;
-	}
-
-	/* The sun4m lets us know if the controller is actually usable. */
-	if(sparc_cpu_model == sun4m &&
-	   prom_getproperty(fd_node, "status", state, sizeof(state)) != -1) {
-		if(!strcmp(state, "disabled")) {
-			goto no_sun_fdc;
-		}
-	}
-	num_regs = prom_getproperty(fd_node, "reg", (char *) fd_regs, sizeof(fd_regs));
-	num_regs = (num_regs / sizeof(fd_regs[0]));
-	prom_apply_obio_ranges(fd_regs, num_regs);
-	memset(&r, 0, sizeof(r));
-	r.flags = fd_regs[0].which_io;
-	r.start = fd_regs[0].phys_addr;
-	sun_fdc = (struct sun_flpy_controller *)
-	    sbus_ioremap(&r, 0, fd_regs[0].reg_size, "floppy");
-
-	/* Last minute sanity check... */
-	if(sun_fdc->status_82072 == 0xff) {
-		sun_fdc = NULL;
-		goto no_sun_fdc;
-	}
-
-	sun_fdops.fd_inb = sun_82077_fd_inb;
-	sun_fdops.fd_outb = sun_82077_fd_outb;
-	fdc_status = &sun_fdc->status_82077;
-
-	if (sun_fdc->dor_82077 == 0x80) {
-		sun_fdc->dor_82077 = 0x02;
-		if (sun_fdc->dor_82077 == 0x80) {
-			sun_fdops.fd_inb = sun_82072_fd_inb;
-			sun_fdops.fd_outb = sun_82072_fd_outb;
-			fdc_status = &sun_fdc->status_82072;
-		}
-	}
-
-	/* Success... */
-	allowed_drive_mask = 0x01;
-	return (int) sun_fdc;
-
-no_sun_fdc:
-	return -1;
-}
-
-static int sparc_eject(void)
-{
-	set_dor(0x00, 0xff, 0x90);
-	udelay(500);
-	set_dor(0x00, 0x6f, 0x00);
-	udelay(500);
-	return 0;
-}
-
-#define fd_eject(drive) sparc_eject()
-
-#define EXTRA_FLOPPY_PARAMS
-
-#endif /* !(__ASM_SPARC_FLOPPY_H) */
diff --git a/include/asm-sparc/floppy_32.h b/include/asm-sparc/floppy_32.h
new file mode 100644
index 00000000000..acdd06eafe5
--- /dev/null
+++ b/include/asm-sparc/floppy_32.h
@@ -0,0 +1,388 @@
+/* asm-sparc/floppy.h: Sparc specific parts of the Floppy driver.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef __ASM_SPARC_FLOPPY_H
+#define __ASM_SPARC_FLOPPY_H
+
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/system.h>
+#include <asm/idprom.h>
+#include <asm/machines.h>
+#include <asm/oplib.h>
+#include <asm/auxio.h>
+#include <asm/irq.h>
+
+/* We don't need no stinkin' I/O port allocation crap. */
+#undef release_region
+#undef request_region
+#define release_region(X, Y)	do { } while(0)
+#define request_region(X, Y, Z)	(1)
+
+/* References:
+ * 1) Netbsd Sun floppy driver.
+ * 2) NCR 82077 controller manual
+ * 3) Intel 82077 controller manual
+ */
+struct sun_flpy_controller {
+	volatile unsigned char status_82072;  /* Main Status reg. */
+#define dcr_82072              status_82072   /* Digital Control reg. */
+#define status1_82077          status_82072   /* Auxiliary Status reg. 1 */
+
+	volatile unsigned char data_82072;    /* Data fifo. */
+#define status2_82077          data_82072     /* Auxiliary Status reg. 2 */
+
+	volatile unsigned char dor_82077;     /* Digital Output reg. */
+	volatile unsigned char tapectl_82077; /* What the? Tape control reg? */
+
+	volatile unsigned char status_82077;  /* Main Status Register. */
+#define drs_82077              status_82077   /* Digital Rate Select reg. */
+
+	volatile unsigned char data_82077;    /* Data fifo. */
+	volatile unsigned char ___unused;
+	volatile unsigned char dir_82077;     /* Digital Input reg. */
+#define dcr_82077              dir_82077      /* Config Control reg. */
+};
+
+/* You'll only ever find one controller on a SparcStation anyways. */
+static struct sun_flpy_controller *sun_fdc = NULL;
+extern volatile unsigned char *fdc_status;
+
+struct sun_floppy_ops {
+	unsigned char (*fd_inb)(int port);
+	void (*fd_outb)(unsigned char value, int port);
+};
+
+static struct sun_floppy_ops sun_fdops;
+
+#define fd_inb(port)              sun_fdops.fd_inb(port)
+#define fd_outb(value,port)       sun_fdops.fd_outb(value,port)
+#define fd_enable_dma()           sun_fd_enable_dma()
+#define fd_disable_dma()          sun_fd_disable_dma()
+#define fd_request_dma()          (0) /* nothing... */
+#define fd_free_dma()             /* nothing... */
+#define fd_clear_dma_ff()         /* nothing... */
+#define fd_set_dma_mode(mode)     sun_fd_set_dma_mode(mode)
+#define fd_set_dma_addr(addr)     sun_fd_set_dma_addr(addr)
+#define fd_set_dma_count(count)   sun_fd_set_dma_count(count)
+#define fd_enable_irq()           /* nothing... */
+#define fd_disable_irq()          /* nothing... */
+#define fd_cacheflush(addr, size) /* nothing... */
+#define fd_request_irq()          sun_fd_request_irq()
+#define fd_free_irq()             /* nothing... */
+#if 0  /* P3: added by Alain, these cause a MMU corruption. 19960524 XXX */
+#define fd_dma_mem_alloc(size)    ((unsigned long) vmalloc(size))
+#define fd_dma_mem_free(addr,size) (vfree((void *)(addr)))
+#endif
+
+/* XXX This isn't really correct. XXX */
+#define get_dma_residue(x)        (0)
+
+#define FLOPPY0_TYPE  4
+#define FLOPPY1_TYPE  0
+
+/* Super paranoid... */
+#undef HAVE_DISABLE_HLT
+
+/* Here is where we catch the floppy driver trying to initialize,
+ * therefore this is where we call the PROM device tree probing
+ * routine etc. on the Sparc.
+ */
+#define FDC1                      sun_floppy_init()
+
+#define N_FDC    1
+#define N_DRIVE  8
+
+/* No 64k boundary crossing problems on the Sparc. */
+#define CROSS_64KB(a,s) (0)
+
+/* Routines unique to each controller type on a Sun. */
+static void sun_set_dor(unsigned char value, int fdc_82077)
+{
+	if (sparc_cpu_model == sun4c) {
+		unsigned int bits = 0;
+		if (value & 0x10)
+			bits |= AUXIO_FLPY_DSEL;
+		if ((value & 0x80) == 0)
+			bits |= AUXIO_FLPY_EJCT;
+		set_auxio(bits, (~bits) & (AUXIO_FLPY_DSEL|AUXIO_FLPY_EJCT));
+	}
+	if (fdc_82077) {
+		sun_fdc->dor_82077 = value;
+	}
+}
+
+static unsigned char sun_read_dir(void)
+{
+	if (sparc_cpu_model == sun4c)
+		return (get_auxio() & AUXIO_FLPY_DCHG) ? 0x80 : 0;
+	else
+		return sun_fdc->dir_82077;
+}
+
+static unsigned char sun_82072_fd_inb(int port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to read unknown port %d\n", port);
+		panic("floppy: Port bolixed.");
+	case 4: /* FD_STATUS */
+		return sun_fdc->status_82072 & ~STATUS_DMA;
+	case 5: /* FD_DATA */
+		return sun_fdc->data_82072;
+	case 7: /* FD_DIR */
+		return sun_read_dir();
+	};
+	panic("sun_82072_fd_inb: How did I get here?");
+}
+
+static void sun_82072_fd_outb(unsigned char value, int port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to write to unknown port %d\n", port);
+		panic("floppy: Port bolixed.");
+	case 2: /* FD_DOR */
+		sun_set_dor(value, 0);
+		break;
+	case 5: /* FD_DATA */
+		sun_fdc->data_82072 = value;
+		break;
+	case 7: /* FD_DCR */
+		sun_fdc->dcr_82072 = value;
+		break;
+	case 4: /* FD_STATUS */
+		sun_fdc->status_82072 = value;
+		break;
+	};
+	return;
+}
+
+static unsigned char sun_82077_fd_inb(int port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to read unknown port %d\n", port);
+		panic("floppy: Port bolixed.");
+	case 0: /* FD_STATUS_0 */
+		return sun_fdc->status1_82077;
+	case 1: /* FD_STATUS_1 */
+		return sun_fdc->status2_82077;
+	case 2: /* FD_DOR */
+		return sun_fdc->dor_82077;
+	case 3: /* FD_TDR */
+		return sun_fdc->tapectl_82077;
+	case 4: /* FD_STATUS */
+		return sun_fdc->status_82077 & ~STATUS_DMA;
+	case 5: /* FD_DATA */
+		return sun_fdc->data_82077;
+	case 7: /* FD_DIR */
+		return sun_read_dir();
+	};
+	panic("sun_82077_fd_inb: How did I get here?");
+}
+
+static void sun_82077_fd_outb(unsigned char value, int port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to write to unknown port %d\n", port);
+		panic("floppy: Port bolixed.");
+	case 2: /* FD_DOR */
+		sun_set_dor(value, 1);
+		break;
+	case 5: /* FD_DATA */
+		sun_fdc->data_82077 = value;
+		break;
+	case 7: /* FD_DCR */
+		sun_fdc->dcr_82077 = value;
+		break;
+	case 4: /* FD_STATUS */
+		sun_fdc->status_82077 = value;
+		break;
+	case 3: /* FD_TDR */
+		sun_fdc->tapectl_82077 = value;
+		break;
+	};
+	return;
+}
+
+/* For pseudo-dma (Sun floppy drives have no real DMA available to
+ * them so we must eat the data fifo bytes directly ourselves) we have
+ * three state variables.  doing_pdma tells our inline low-level
+ * assembly floppy interrupt entry point whether it should sit and eat
+ * bytes from the fifo or just transfer control up to the higher level
+ * floppy interrupt c-code.  I tried very hard but I could not get the
+ * pseudo-dma to work in c-code without getting many overruns and
+ * underruns.  If non-zero, doing_pdma encodes the direction of
+ * the transfer for debugging.  1=read 2=write
+ */
+extern char *pdma_vaddr;
+extern unsigned long pdma_size;
+extern volatile int doing_pdma;
+
+/* This is software state */
+extern char *pdma_base;
+extern unsigned long pdma_areasize;
+
+/* Common routines to all controller types on the Sparc. */
+static inline void virtual_dma_init(void)
+{
+	/* nothing... */
+}
+
+static inline void sun_fd_disable_dma(void)
+{
+	doing_pdma = 0;
+	if (pdma_base) {
+		mmu_unlockarea(pdma_base, pdma_areasize);
+		pdma_base = NULL;
+	}
+}
+
+static inline void sun_fd_set_dma_mode(int mode)
+{
+	switch(mode) {
+	case DMA_MODE_READ:
+		doing_pdma = 1;
+		break;
+	case DMA_MODE_WRITE:
+		doing_pdma = 2;
+		break;
+	default:
+		printk("Unknown dma mode %d\n", mode);
+		panic("floppy: Giving up...");
+	}
+}
+
+static inline void sun_fd_set_dma_addr(char *buffer)
+{
+	pdma_vaddr = buffer;
+}
+
+static inline void sun_fd_set_dma_count(int length)
+{
+	pdma_size = length;
+}
+
+static inline void sun_fd_enable_dma(void)
+{
+	pdma_vaddr = mmu_lockarea(pdma_vaddr, pdma_size);
+	pdma_base = pdma_vaddr;
+	pdma_areasize = pdma_size;
+}
+
+/* Our low-level entry point in arch/sparc/kernel/entry.S */
+extern int sparc_floppy_request_irq(int irq, unsigned long flags,
+				    irq_handler_t irq_handler);
+
+static int sun_fd_request_irq(void)
+{
+	static int once = 0;
+	int error;
+
+	if(!once) {
+		once = 1;
+		error = sparc_floppy_request_irq(FLOPPY_IRQ,
+						 IRQF_DISABLED,
+						 floppy_interrupt);
+		return ((error == 0) ? 0 : -1);
+	} else return 0;
+}
+
+static struct linux_prom_registers fd_regs[2];
+
+static int sun_floppy_init(void)
+{
+	char state[128];
+	int tnode, fd_node, num_regs;
+	struct resource r;
+
+	use_virtual_dma = 1;
+
+	FLOPPY_IRQ = 11;
+	/* Forget it if we aren't on a machine that could possibly
+	 * ever have a floppy drive.
+	 */
+	if((sparc_cpu_model != sun4c && sparc_cpu_model != sun4m) ||
+	   ((idprom->id_machtype == (SM_SUN4C | SM_4C_SLC)) ||
+	    (idprom->id_machtype == (SM_SUN4C | SM_4C_ELC)))) {
+		/* We certainly don't have a floppy controller. */
+		goto no_sun_fdc;
+	}
+	/* Well, try to find one. */
+	tnode = prom_getchild(prom_root_node);
+	fd_node = prom_searchsiblings(tnode, "obio");
+	if(fd_node != 0) {
+		tnode = prom_getchild(fd_node);
+		fd_node = prom_searchsiblings(tnode, "SUNW,fdtwo");
+	} else {
+		fd_node = prom_searchsiblings(tnode, "fd");
+	}
+	if(fd_node == 0) {
+		goto no_sun_fdc;
+	}
+
+	/* The sun4m lets us know if the controller is actually usable. */
+	if(sparc_cpu_model == sun4m &&
+	   prom_getproperty(fd_node, "status", state, sizeof(state)) != -1) {
+		if(!strcmp(state, "disabled")) {
+			goto no_sun_fdc;
+		}
+	}
+	num_regs = prom_getproperty(fd_node, "reg", (char *) fd_regs, sizeof(fd_regs));
+	num_regs = (num_regs / sizeof(fd_regs[0]));
+	prom_apply_obio_ranges(fd_regs, num_regs);
+	memset(&r, 0, sizeof(r));
+	r.flags = fd_regs[0].which_io;
+	r.start = fd_regs[0].phys_addr;
+	sun_fdc = (struct sun_flpy_controller *)
+	    sbus_ioremap(&r, 0, fd_regs[0].reg_size, "floppy");
+
+	/* Last minute sanity check... */
+	if(sun_fdc->status_82072 == 0xff) {
+		sun_fdc = NULL;
+		goto no_sun_fdc;
+	}
+
+	sun_fdops.fd_inb = sun_82077_fd_inb;
+	sun_fdops.fd_outb = sun_82077_fd_outb;
+	fdc_status = &sun_fdc->status_82077;
+
+	if (sun_fdc->dor_82077 == 0x80) {
+		sun_fdc->dor_82077 = 0x02;
+		if (sun_fdc->dor_82077 == 0x80) {
+			sun_fdops.fd_inb = sun_82072_fd_inb;
+			sun_fdops.fd_outb = sun_82072_fd_outb;
+			fdc_status = &sun_fdc->status_82072;
+		}
+	}
+
+	/* Success... */
+	allowed_drive_mask = 0x01;
+	return (int) sun_fdc;
+
+no_sun_fdc:
+	return -1;
+}
+
+static int sparc_eject(void)
+{
+	set_dor(0x00, 0xff, 0x90);
+	udelay(500);
+	set_dor(0x00, 0x6f, 0x00);
+	udelay(500);
+	return 0;
+}
+
+#define fd_eject(drive) sparc_eject()
+
+#define EXTRA_FLOPPY_PARAMS
+
+#endif /* !(__ASM_SPARC_FLOPPY_H) */
diff --git a/include/asm-sparc/floppy_64.h b/include/asm-sparc/floppy_64.h
new file mode 100644
index 00000000000..c39db1060bc
--- /dev/null
+++ b/include/asm-sparc/floppy_64.h
@@ -0,0 +1,782 @@
+/* floppy.h: Sparc specific parts of the Floppy driver.
+ *
+ * Copyright (C) 1996, 2007 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ *
+ * Ultra/PCI support added: Sep 1997  Eddie C. Dost  (ecd@skynet.be)
+ */
+
+#ifndef __ASM_SPARC64_FLOPPY_H
+#define __ASM_SPARC64_FLOPPY_H
+
+#include <linux/init.h>
+#include <linux/pci.h>
+
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/system.h>
+#include <asm/idprom.h>
+#include <asm/oplib.h>
+#include <asm/auxio.h>
+#include <asm/sbus.h>
+#include <asm/irq.h>
+
+
+/*
+ * Define this to enable exchanging drive 0 and 1 if only drive 1 is
+ * probed on PCI machines.
+ */
+#undef PCI_FDC_SWAP_DRIVES
+
+
+/* References:
+ * 1) Netbsd Sun floppy driver.
+ * 2) NCR 82077 controller manual
+ * 3) Intel 82077 controller manual
+ */
+struct sun_flpy_controller {
+	volatile unsigned char status1_82077; /* Auxiliary Status reg. 1 */
+	volatile unsigned char status2_82077; /* Auxiliary Status reg. 2 */
+	volatile unsigned char dor_82077;     /* Digital Output reg. */
+	volatile unsigned char tapectl_82077; /* Tape Control reg */
+	volatile unsigned char status_82077;  /* Main Status Register. */
+#define drs_82077              status_82077   /* Digital Rate Select reg. */
+	volatile unsigned char data_82077;    /* Data fifo. */
+	volatile unsigned char ___unused;
+	volatile unsigned char dir_82077;     /* Digital Input reg. */
+#define dcr_82077              dir_82077      /* Config Control reg. */
+};
+
+/* You'll only ever find one controller on an Ultra anyways. */
+static struct sun_flpy_controller *sun_fdc = (struct sun_flpy_controller *)-1;
+unsigned long fdc_status;
+static struct sbus_dev *floppy_sdev = NULL;
+
+struct sun_floppy_ops {
+	unsigned char	(*fd_inb) (unsigned long port);
+	void		(*fd_outb) (unsigned char value, unsigned long port);
+	void		(*fd_enable_dma) (void);
+	void		(*fd_disable_dma) (void);
+	void		(*fd_set_dma_mode) (int);
+	void		(*fd_set_dma_addr) (char *);
+	void		(*fd_set_dma_count) (int);
+	unsigned int	(*get_dma_residue) (void);
+	int		(*fd_request_irq) (void);
+	void		(*fd_free_irq) (void);
+	int		(*fd_eject) (int);
+};
+
+static struct sun_floppy_ops sun_fdops;
+
+#define fd_inb(port)              sun_fdops.fd_inb(port)
+#define fd_outb(value,port)       sun_fdops.fd_outb(value,port)
+#define fd_enable_dma()           sun_fdops.fd_enable_dma()
+#define fd_disable_dma()          sun_fdops.fd_disable_dma()
+#define fd_request_dma()          (0) /* nothing... */
+#define fd_free_dma()             /* nothing... */
+#define fd_clear_dma_ff()         /* nothing... */
+#define fd_set_dma_mode(mode)     sun_fdops.fd_set_dma_mode(mode)
+#define fd_set_dma_addr(addr)     sun_fdops.fd_set_dma_addr(addr)
+#define fd_set_dma_count(count)   sun_fdops.fd_set_dma_count(count)
+#define get_dma_residue(x)        sun_fdops.get_dma_residue()
+#define fd_cacheflush(addr, size) /* nothing... */
+#define fd_request_irq()          sun_fdops.fd_request_irq()
+#define fd_free_irq()             sun_fdops.fd_free_irq()
+#define fd_eject(drive)           sun_fdops.fd_eject(drive)
+
+/* Super paranoid... */
+#undef HAVE_DISABLE_HLT
+
+static int sun_floppy_types[2] = { 0, 0 };
+
+/* Here is where we catch the floppy driver trying to initialize,
+ * therefore this is where we call the PROM device tree probing
+ * routine etc. on the Sparc.
+ */
+#define FLOPPY0_TYPE		sun_floppy_init()
+#define FLOPPY1_TYPE		sun_floppy_types[1]
+
+#define FDC1			((unsigned long)sun_fdc)
+
+#define N_FDC    1
+#define N_DRIVE  8
+
+/* No 64k boundary crossing problems on the Sparc. */
+#define CROSS_64KB(a,s) (0)
+
+static unsigned char sun_82077_fd_inb(unsigned long port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to read unknown port %lx\n", port);
+		panic("floppy: Port bolixed.");
+	case 4: /* FD_STATUS */
+		return sbus_readb(&sun_fdc->status_82077) & ~STATUS_DMA;
+	case 5: /* FD_DATA */
+		return sbus_readb(&sun_fdc->data_82077);
+	case 7: /* FD_DIR */
+		/* XXX: Is DCL on 0x80 in sun4m? */
+		return sbus_readb(&sun_fdc->dir_82077);
+	};
+	panic("sun_82072_fd_inb: How did I get here?");
+}
+
+static void sun_82077_fd_outb(unsigned char value, unsigned long port)
+{
+	udelay(5);
+	switch(port & 7) {
+	default:
+		printk("floppy: Asked to write to unknown port %lx\n", port);
+		panic("floppy: Port bolixed.");
+	case 2: /* FD_DOR */
+		/* Happily, the 82077 has a real DOR register. */
+		sbus_writeb(value, &sun_fdc->dor_82077);
+		break;
+	case 5: /* FD_DATA */
+		sbus_writeb(value, &sun_fdc->data_82077);
+		break;
+	case 7: /* FD_DCR */
+		sbus_writeb(value, &sun_fdc->dcr_82077);
+		break;
+	case 4: /* FD_STATUS */
+		sbus_writeb(value, &sun_fdc->status_82077);
+		break;
+	};
+	return;
+}
+
+/* For pseudo-dma (Sun floppy drives have no real DMA available to
+ * them so we must eat the data fifo bytes directly ourselves) we have
+ * three state variables.  doing_pdma tells our inline low-level
+ * assembly floppy interrupt entry point whether it should sit and eat
+ * bytes from the fifo or just transfer control up to the higher level
+ * floppy interrupt c-code.  I tried very hard but I could not get the
+ * pseudo-dma to work in c-code without getting many overruns and
+ * underruns.  If non-zero, doing_pdma encodes the direction of
+ * the transfer for debugging.  1=read 2=write
+ */
+unsigned char *pdma_vaddr;
+unsigned long pdma_size;
+volatile int doing_pdma = 0;
+
+/* This is software state */
+char *pdma_base = NULL;
+unsigned long pdma_areasize;
+
+/* Common routines to all controller types on the Sparc. */
+static void sun_fd_disable_dma(void)
+{
+	doing_pdma = 0;
+	if (pdma_base) {
+		mmu_unlockarea(pdma_base, pdma_areasize);
+		pdma_base = NULL;
+	}
+}
+
+static void sun_fd_set_dma_mode(int mode)
+{
+	switch(mode) {
+	case DMA_MODE_READ:
+		doing_pdma = 1;
+		break;
+	case DMA_MODE_WRITE:
+		doing_pdma = 2;
+		break;
+	default:
+		printk("Unknown dma mode %d\n", mode);
+		panic("floppy: Giving up...");
+	}
+}
+
+static void sun_fd_set_dma_addr(char *buffer)
+{
+	pdma_vaddr = buffer;
+}
+
+static void sun_fd_set_dma_count(int length)
+{
+	pdma_size = length;
+}
+
+static void sun_fd_enable_dma(void)
+{
+	pdma_vaddr = mmu_lockarea(pdma_vaddr, pdma_size);
+	pdma_base = pdma_vaddr;
+	pdma_areasize = pdma_size;
+}
+
+irqreturn_t sparc_floppy_irq(int irq, void *dev_cookie)
+{
+	if (likely(doing_pdma)) {
+		void __iomem *stat = (void __iomem *) fdc_status;
+		unsigned char *vaddr = pdma_vaddr;
+		unsigned long size = pdma_size;
+		u8 val;
+
+		while (size) {
+			val = readb(stat);
+			if (unlikely(!(val & 0x80))) {
+				pdma_vaddr = vaddr;
+				pdma_size = size;
+				return IRQ_HANDLED;
+			}
+			if (unlikely(!(val & 0x20))) {
+				pdma_vaddr = vaddr;
+				pdma_size = size;
+				doing_pdma = 0;
+				goto main_interrupt;
+			}
+			if (val & 0x40) {
+				/* read */
+				*vaddr++ = readb(stat + 1);
+			} else {
+				unsigned char data = *vaddr++;
+
+				/* write */
+				writeb(data, stat + 1);
+			}
+			size--;
+		}
+
+		pdma_vaddr = vaddr;
+		pdma_size = size;
+
+		/* Send Terminal Count pulse to floppy controller. */
+		val = readb(auxio_register);
+		val |= AUXIO_AUX1_FTCNT;
+		writeb(val, auxio_register);
+		val &= ~AUXIO_AUX1_FTCNT;
+		writeb(val, auxio_register);
+
+		doing_pdma = 0;
+	}
+
+main_interrupt:
+	return floppy_interrupt(irq, dev_cookie);
+}
+
+static int sun_fd_request_irq(void)
+{
+	static int once = 0;
+	int error;
+
+	if(!once) {
+		once = 1;
+
+		error = request_irq(FLOPPY_IRQ, sparc_floppy_irq,
+				    IRQF_DISABLED, "floppy", NULL);
+
+		return ((error == 0) ? 0 : -1);
+	}
+	return 0;
+}
+
+static void sun_fd_free_irq(void)
+{
+}
+
+static unsigned int sun_get_dma_residue(void)
+{
+	/* XXX This isn't really correct. XXX */
+	return 0;
+}
+
+static int sun_fd_eject(int drive)
+{
+	set_dor(0x00, 0xff, 0x90);
+	udelay(500);
+	set_dor(0x00, 0x6f, 0x00);
+	udelay(500);
+	return 0;
+}
+
+#ifdef CONFIG_PCI
+#include <asm/ebus.h>
+#include <asm/ns87303.h>
+
+static struct ebus_dma_info sun_pci_fd_ebus_dma;
+static struct pci_dev *sun_pci_ebus_dev;
+static int sun_pci_broken_drive = -1;
+
+struct sun_pci_dma_op {
+	unsigned int 	addr;
+	int		len;
+	int		direction;
+	char		*buf;
+};
+static struct sun_pci_dma_op sun_pci_dma_current = { -1U, 0, 0, NULL};
+static struct sun_pci_dma_op sun_pci_dma_pending = { -1U, 0, 0, NULL};
+
+extern irqreturn_t floppy_interrupt(int irq, void *dev_id);
+
+static unsigned char sun_pci_fd_inb(unsigned long port)
+{
+	udelay(5);
+	return inb(port);
+}
+
+static void sun_pci_fd_outb(unsigned char val, unsigned long port)
+{
+	udelay(5);
+	outb(val, port);
+}
+
+static void sun_pci_fd_broken_outb(unsigned char val, unsigned long port)
+{
+	udelay(5);
+	/*
+	 * XXX: Due to SUN's broken floppy connector on AX and AXi
+	 *      we need to turn on MOTOR_0 also, if the floppy is
+	 *      jumpered to DS1 (like most PC floppies are). I hope
+	 *      this does not hurt correct hardware like the AXmp.
+	 *      (Eddie, Sep 12 1998).
+	 */
+	if (port == ((unsigned long)sun_fdc) + 2) {
+		if (((val & 0x03) == sun_pci_broken_drive) && (val & 0x20)) {
+			val |= 0x10;
+		}
+	}
+	outb(val, port);
+}
+
+#ifdef PCI_FDC_SWAP_DRIVES
+static void sun_pci_fd_lde_broken_outb(unsigned char val, unsigned long port)
+{
+	udelay(5);
+	/*
+	 * XXX: Due to SUN's broken floppy connector on AX and AXi
+	 *      we need to turn on MOTOR_0 also, if the floppy is
+	 *      jumpered to DS1 (like most PC floppies are). I hope
+	 *      this does not hurt correct hardware like the AXmp.
+	 *      (Eddie, Sep 12 1998).
+	 */
+	if (port == ((unsigned long)sun_fdc) + 2) {
+		if (((val & 0x03) == sun_pci_broken_drive) && (val & 0x10)) {
+			val &= ~(0x03);
+			val |= 0x21;
+		}
+	}
+	outb(val, port);
+}
+#endif /* PCI_FDC_SWAP_DRIVES */
+
+static void sun_pci_fd_enable_dma(void)
+{
+	BUG_ON((NULL == sun_pci_dma_pending.buf) 	||
+	    (0	  == sun_pci_dma_pending.len) 	||
+	    (0	  == sun_pci_dma_pending.direction));
+
+	sun_pci_dma_current.buf = sun_pci_dma_pending.buf;
+	sun_pci_dma_current.len = sun_pci_dma_pending.len;
+	sun_pci_dma_current.direction = sun_pci_dma_pending.direction;
+
+	sun_pci_dma_pending.buf  = NULL;
+	sun_pci_dma_pending.len  = 0;
+	sun_pci_dma_pending.direction = 0;
+	sun_pci_dma_pending.addr = -1U;
+
+	sun_pci_dma_current.addr =
+		pci_map_single(sun_pci_ebus_dev,
+			       sun_pci_dma_current.buf,
+			       sun_pci_dma_current.len,
+			       sun_pci_dma_current.direction);
+
+	ebus_dma_enable(&sun_pci_fd_ebus_dma, 1);
+
+	if (ebus_dma_request(&sun_pci_fd_ebus_dma,
+			     sun_pci_dma_current.addr,
+			     sun_pci_dma_current.len))
+		BUG();
+}
+
+static void sun_pci_fd_disable_dma(void)
+{
+	ebus_dma_enable(&sun_pci_fd_ebus_dma, 0);
+	if (sun_pci_dma_current.addr != -1U)
+		pci_unmap_single(sun_pci_ebus_dev,
+				 sun_pci_dma_current.addr,
+				 sun_pci_dma_current.len,
+				 sun_pci_dma_current.direction);
+	sun_pci_dma_current.addr = -1U;
+}
+
+static void sun_pci_fd_set_dma_mode(int mode)
+{
+	if (mode == DMA_MODE_WRITE)
+		sun_pci_dma_pending.direction = PCI_DMA_TODEVICE;
+	else
+		sun_pci_dma_pending.direction = PCI_DMA_FROMDEVICE;
+
+	ebus_dma_prepare(&sun_pci_fd_ebus_dma, mode != DMA_MODE_WRITE);
+}
+
+static void sun_pci_fd_set_dma_count(int length)
+{
+	sun_pci_dma_pending.len = length;
+}
+
+static void sun_pci_fd_set_dma_addr(char *buffer)
+{
+	sun_pci_dma_pending.buf = buffer;
+}
+
+static unsigned int sun_pci_get_dma_residue(void)
+{
+	return ebus_dma_residue(&sun_pci_fd_ebus_dma);
+}
+
+static int sun_pci_fd_request_irq(void)
+{
+	return ebus_dma_irq_enable(&sun_pci_fd_ebus_dma, 1);
+}
+
+static void sun_pci_fd_free_irq(void)
+{
+	ebus_dma_irq_enable(&sun_pci_fd_ebus_dma, 0);
+}
+
+static int sun_pci_fd_eject(int drive)
+{
+	return -EINVAL;
+}
+
+void sun_pci_fd_dma_callback(struct ebus_dma_info *p, int event, void *cookie)
+{
+	floppy_interrupt(0, NULL);
+}
+
+/*
+ * Floppy probing, we'd like to use /dev/fd0 for a single Floppy on PCI,
+ * even if this is configured using DS1, thus looks like /dev/fd1 with
+ * the cabling used in Ultras.
+ */
+#define DOR	(port + 2)
+#define MSR	(port + 4)
+#define FIFO	(port + 5)
+
+static void sun_pci_fd_out_byte(unsigned long port, unsigned char val,
+			        unsigned long reg)
+{
+	unsigned char status;
+	int timeout = 1000;
+
+	while (!((status = inb(MSR)) & 0x80) && --timeout)
+		udelay(100);
+	outb(val, reg);
+}
+
+static unsigned char sun_pci_fd_sensei(unsigned long port)
+{
+	unsigned char result[2] = { 0x70, 0x00 };
+	unsigned char status;
+	int i = 0;
+
+	sun_pci_fd_out_byte(port, 0x08, FIFO);
+	do {
+		int timeout = 1000;
+
+		while (!((status = inb(MSR)) & 0x80) && --timeout)
+			udelay(100);
+
+		if (!timeout)
+			break;
+
+		if ((status & 0xf0) == 0xd0)
+			result[i++] = inb(FIFO);
+		else
+			break;
+	} while (i < 2);
+
+	return result[0];
+}
+
+static void sun_pci_fd_reset(unsigned long port)
+{
+	unsigned char mask = 0x00;
+	unsigned char status;
+	int timeout = 10000;
+
+	outb(0x80, MSR);
+	do {
+		status = sun_pci_fd_sensei(port);
+		if ((status & 0xc0) == 0xc0)
+			mask |= 1 << (status & 0x03);
+		else
+			udelay(100);
+	} while ((mask != 0x0f) && --timeout);
+}
+
+static int sun_pci_fd_test_drive(unsigned long port, int drive)
+{
+	unsigned char status, data;
+	int timeout = 1000;
+	int ready;
+
+	sun_pci_fd_reset(port);
+
+	data = (0x10 << drive) | 0x0c | drive;
+	sun_pci_fd_out_byte(port, data, DOR);
+
+	sun_pci_fd_out_byte(port, 0x07, FIFO);
+	sun_pci_fd_out_byte(port, drive & 0x03, FIFO);
+
+	do {
+		udelay(100);
+		status = sun_pci_fd_sensei(port);
+	} while (((status & 0xc0) == 0x80) && --timeout);
+
+	if (!timeout)
+		ready = 0;
+	else
+		ready = (status & 0x10) ? 0 : 1;
+
+	sun_pci_fd_reset(port);
+	return ready;
+}
+#undef FIFO
+#undef MSR
+#undef DOR
+
+#endif /* CONFIG_PCI */
+
+#ifdef CONFIG_PCI
+static int __init ebus_fdthree_p(struct linux_ebus_device *edev)
+{
+	if (!strcmp(edev->prom_node->name, "fdthree"))
+		return 1;
+	if (!strcmp(edev->prom_node->name, "floppy")) {
+		const char *compat;
+
+		compat = of_get_property(edev->prom_node,
+					 "compatible", NULL);
+		if (compat && !strcmp(compat, "fdthree"))
+			return 1;
+	}
+	return 0;
+}
+#endif
+
+static unsigned long __init sun_floppy_init(void)
+{
+	char state[128];
+	struct sbus_bus *bus;
+	struct sbus_dev *sdev = NULL;
+	static int initialized = 0;
+
+	if (initialized)
+		return sun_floppy_types[0];
+	initialized = 1;
+
+	for_all_sbusdev (sdev, bus) {
+		if (!strcmp(sdev->prom_name, "SUNW,fdtwo"))
+			break;
+	}
+	if(sdev) {
+		floppy_sdev = sdev;
+		FLOPPY_IRQ = sdev->irqs[0];
+	} else {
+#ifdef CONFIG_PCI
+		struct linux_ebus *ebus;
+		struct linux_ebus_device *edev = NULL;
+		unsigned long config = 0;
+		void __iomem *auxio_reg;
+		const char *state_prop;
+
+		for_each_ebus(ebus) {
+			for_each_ebusdev(edev, ebus) {
+				if (ebus_fdthree_p(edev))
+					goto ebus_done;
+			}
+		}
+	ebus_done:
+		if (!edev)
+			return 0;
+
+		state_prop = of_get_property(edev->prom_node, "status", NULL);
+		if (state_prop && !strncmp(state_prop, "disabled", 8))
+			return 0;
+
+		FLOPPY_IRQ = edev->irqs[0];
+
+		/* Make sure the high density bit is set, some systems
+		 * (most notably Ultra5/Ultra10) come up with it clear.
+		 */
+		auxio_reg = (void __iomem *) edev->resource[2].start;
+		writel(readl(auxio_reg)|0x2, auxio_reg);
+
+		sun_pci_ebus_dev = ebus->self;
+
+		spin_lock_init(&sun_pci_fd_ebus_dma.lock);
+
+		/* XXX ioremap */
+		sun_pci_fd_ebus_dma.regs = (void __iomem *)
+			edev->resource[1].start;
+		if (!sun_pci_fd_ebus_dma.regs)
+			return 0;
+
+		sun_pci_fd_ebus_dma.flags = (EBUS_DMA_FLAG_USE_EBDMA_HANDLER |
+					     EBUS_DMA_FLAG_TCI_DISABLE);
+		sun_pci_fd_ebus_dma.callback = sun_pci_fd_dma_callback;
+		sun_pci_fd_ebus_dma.client_cookie = NULL;
+		sun_pci_fd_ebus_dma.irq = FLOPPY_IRQ;
+		strcpy(sun_pci_fd_ebus_dma.name, "floppy");
+		if (ebus_dma_register(&sun_pci_fd_ebus_dma))
+			return 0;
+
+		/* XXX ioremap */
+		sun_fdc = (struct sun_flpy_controller *)edev->resource[0].start;
+
+		sun_fdops.fd_inb = sun_pci_fd_inb;
+		sun_fdops.fd_outb = sun_pci_fd_outb;
+
+		can_use_virtual_dma = use_virtual_dma = 0;
+		sun_fdops.fd_enable_dma = sun_pci_fd_enable_dma;
+		sun_fdops.fd_disable_dma = sun_pci_fd_disable_dma;
+		sun_fdops.fd_set_dma_mode = sun_pci_fd_set_dma_mode;
+		sun_fdops.fd_set_dma_addr = sun_pci_fd_set_dma_addr;
+		sun_fdops.fd_set_dma_count = sun_pci_fd_set_dma_count;
+		sun_fdops.get_dma_residue = sun_pci_get_dma_residue;
+
+		sun_fdops.fd_request_irq = sun_pci_fd_request_irq;
+		sun_fdops.fd_free_irq = sun_pci_fd_free_irq;
+
+		sun_fdops.fd_eject = sun_pci_fd_eject;
+
+		fdc_status = (unsigned long) &sun_fdc->status_82077;
+
+		/*
+		 * XXX: Find out on which machines this is really needed.
+		 */
+		if (1) {
+			sun_pci_broken_drive = 1;
+			sun_fdops.fd_outb = sun_pci_fd_broken_outb;
+		}
+
+		allowed_drive_mask = 0;
+		if (sun_pci_fd_test_drive((unsigned long)sun_fdc, 0))
+			sun_floppy_types[0] = 4;
+		if (sun_pci_fd_test_drive((unsigned long)sun_fdc, 1))
+			sun_floppy_types[1] = 4;
+
+		/*
+		 * Find NS87303 SuperIO config registers (through ecpp).
+		 */
+		for_each_ebus(ebus) {
+			for_each_ebusdev(edev, ebus) {
+				if (!strcmp(edev->prom_node->name, "ecpp")) {
+					config = edev->resource[1].start;
+					goto config_done;
+				}
+			}
+		}
+	config_done:
+
+		/*
+		 * Sanity check, is this really the NS87303?
+		 */
+		switch (config & 0x3ff) {
+		case 0x02e:
+		case 0x15c:
+		case 0x26e:
+		case 0x398:
+			break;
+		default:
+			config = 0;
+		}
+
+		if (!config)
+			return sun_floppy_types[0];
+
+		/* Enable PC-AT mode. */
+		ns87303_modify(config, ASC, 0, 0xc0);
+
+#ifdef PCI_FDC_SWAP_DRIVES
+		/*
+		 * If only Floppy 1 is present, swap drives.
+		 */
+		if (!sun_floppy_types[0] && sun_floppy_types[1]) {
+			/*
+			 * Set the drive exchange bit in FCR on NS87303,
+			 * make sure other bits are sane before doing so.
+			 */
+			ns87303_modify(config, FER, FER_EDM, 0);
+			ns87303_modify(config, ASC, ASC_DRV2_SEL, 0);
+			ns87303_modify(config, FCR, 0, FCR_LDE);
+
+			config = sun_floppy_types[0];
+			sun_floppy_types[0] = sun_floppy_types[1];
+			sun_floppy_types[1] = config;
+
+			if (sun_pci_broken_drive != -1) {
+				sun_pci_broken_drive = 1 - sun_pci_broken_drive;
+				sun_fdops.fd_outb = sun_pci_fd_lde_broken_outb;
+			}
+		}
+#endif /* PCI_FDC_SWAP_DRIVES */
+
+		return sun_floppy_types[0];
+#else
+		return 0;
+#endif
+	}
+	prom_getproperty(sdev->prom_node, "status", state, sizeof(state));
+	if(!strncmp(state, "disabled", 8))
+		return 0;
+
+	/*
+	 * We cannot do sbus_ioremap here: it does request_region,
+	 * which the generic floppy driver tries to do once again.
+	 * But we must use the sdev resource values as they have
+	 * had parent ranges applied.
+	 */
+	sun_fdc = (struct sun_flpy_controller *)
+		(sdev->resource[0].start +
+		 ((sdev->resource[0].flags & 0x1ffUL) << 32UL));
+
+	/* Last minute sanity check... */
+	if(sbus_readb(&sun_fdc->status1_82077) == 0xff) {
+		sun_fdc = (struct sun_flpy_controller *)-1;
+		return 0;
+	}
+
+        sun_fdops.fd_inb = sun_82077_fd_inb;
+        sun_fdops.fd_outb = sun_82077_fd_outb;
+
+	can_use_virtual_dma = use_virtual_dma = 1;
+	sun_fdops.fd_enable_dma = sun_fd_enable_dma;
+	sun_fdops.fd_disable_dma = sun_fd_disable_dma;
+	sun_fdops.fd_set_dma_mode = sun_fd_set_dma_mode;
+	sun_fdops.fd_set_dma_addr = sun_fd_set_dma_addr;
+	sun_fdops.fd_set_dma_count = sun_fd_set_dma_count;
+	sun_fdops.get_dma_residue = sun_get_dma_residue;
+
+	sun_fdops.fd_request_irq = sun_fd_request_irq;
+	sun_fdops.fd_free_irq = sun_fd_free_irq;
+
+	sun_fdops.fd_eject = sun_fd_eject;
+
+        fdc_status = (unsigned long) &sun_fdc->status_82077;
+
+	/* Success... */
+	allowed_drive_mask = 0x01;
+	sun_floppy_types[0] = 4;
+	sun_floppy_types[1] = 0;
+
+	return sun_floppy_types[0];
+}
+
+#define EXTRA_FLOPPY_PARAMS
+
+static DEFINE_SPINLOCK(dma_spin_lock);
+
+#define claim_dma_lock() \
+({	unsigned long flags; \
+	spin_lock_irqsave(&dma_spin_lock, flags); \
+	flags; \
+})
+
+#define release_dma_lock(__flags) \
+	spin_unlock_irqrestore(&dma_spin_lock, __flags);
+
+#endif /* !(__ASM_SPARC64_FLOPPY_H) */
diff --git a/include/asm-sparc/fpumacro.h b/include/asm-sparc/fpumacro.h
new file mode 100644
index 00000000000..cc463fec806
--- /dev/null
+++ b/include/asm-sparc/fpumacro.h
@@ -0,0 +1,33 @@
+/* fpumacro.h: FPU related macros.
+ *
+ * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _SPARC64_FPUMACRO_H
+#define _SPARC64_FPUMACRO_H
+
+#include <asm/asi.h>
+#include <asm/visasm.h>
+
+struct fpustate {
+	u32	regs[64];
+};
+
+#define FPUSTATE (struct fpustate *)(current_thread_info()->fpregs)
+
+static inline unsigned long fprs_read(void)
+{
+	unsigned long retval;
+
+	__asm__ __volatile__("rd %%fprs, %0" : "=r" (retval));
+
+	return retval;
+}
+
+static inline void fprs_write(unsigned long val)
+{
+	__asm__ __volatile__("wr %0, 0x0, %%fprs" : : "r" (val));
+}
+
+#endif /* !(_SPARC64_FPUMACRO_H) */
diff --git a/include/asm-sparc/futex.h b/include/asm-sparc/futex.h
index 6a332a9f099..c6a9f038c53 100644
--- a/include/asm-sparc/futex.h
+++ b/include/asm-sparc/futex.h
@@ -1,6 +1,8 @@
-#ifndef _ASM_FUTEX_H
-#define _ASM_FUTEX_H
-
-#include <asm-generic/futex.h>
-
+#ifndef ___ASM_SPARC_FUTEX_H
+#define ___ASM_SPARC_FUTEX_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/futex_64.h>
+#else
+#include <asm-sparc/futex_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/futex_32.h b/include/asm-sparc/futex_32.h
new file mode 100644
index 00000000000..6a332a9f099
--- /dev/null
+++ b/include/asm-sparc/futex_32.h
@@ -0,0 +1,6 @@
+#ifndef _ASM_FUTEX_H
+#define _ASM_FUTEX_H
+
+#include <asm-generic/futex.h>
+
+#endif
diff --git a/include/asm-sparc/futex_64.h b/include/asm-sparc/futex_64.h
new file mode 100644
index 00000000000..d8378935ae9
--- /dev/null
+++ b/include/asm-sparc/futex_64.h
@@ -0,0 +1,110 @@
+#ifndef _SPARC64_FUTEX_H
+#define _SPARC64_FUTEX_H
+
+#include <linux/futex.h>
+#include <linux/uaccess.h>
+#include <asm/errno.h>
+#include <asm/system.h>
+
+#define __futex_cas_op(insn, ret, oldval, uaddr, oparg)	\
+	__asm__ __volatile__(				\
+	"\n1:	lduwa	[%3] %%asi, %2\n"		\
+	"	" insn "\n"				\
+	"2:	casa	[%3] %%asi, %2, %1\n"		\
+	"	cmp	%2, %1\n"			\
+	"	bne,pn	%%icc, 1b\n"			\
+	"	 mov	0, %0\n"			\
+	"3:\n"						\
+	"	.section .fixup,#alloc,#execinstr\n"	\
+	"	.align	4\n"				\
+	"4:	sethi	%%hi(3b), %0\n"			\
+	"	jmpl	%0 + %%lo(3b), %%g0\n"		\
+	"	 mov	%5, %0\n"			\
+	"	.previous\n"				\
+	"	.section __ex_table,\"a\"\n"		\
+	"	.align	4\n"				\
+	"	.word	1b, 4b\n"			\
+	"	.word	2b, 4b\n"			\
+	"	.previous\n"				\
+	: "=&r" (ret), "=&r" (oldval), "=&r" (tem)	\
+	: "r" (uaddr), "r" (oparg), "i" (-EFAULT)	\
+	: "memory")
+
+static inline int futex_atomic_op_inuser(int encoded_op, int __user *uaddr)
+{
+	int op = (encoded_op >> 28) & 7;
+	int cmp = (encoded_op >> 24) & 15;
+	int oparg = (encoded_op << 8) >> 20;
+	int cmparg = (encoded_op << 20) >> 20;
+	int oldval = 0, ret, tem;
+
+	if (unlikely(!access_ok(VERIFY_WRITE, uaddr, sizeof(int))))
+		return -EFAULT;
+	if (unlikely((((unsigned long) uaddr) & 0x3UL)))
+		return -EINVAL;
+
+	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
+		oparg = 1 << oparg;
+
+	pagefault_disable();
+
+	switch (op) {
+	case FUTEX_OP_SET:
+		__futex_cas_op("mov\t%4, %1", ret, oldval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ADD:
+		__futex_cas_op("add\t%2, %4, %1", ret, oldval, uaddr, oparg);
+		break;
+	case FUTEX_OP_OR:
+		__futex_cas_op("or\t%2, %4, %1", ret, oldval, uaddr, oparg);
+		break;
+	case FUTEX_OP_ANDN:
+		__futex_cas_op("and\t%2, %4, %1", ret, oldval, uaddr, oparg);
+		break;
+	case FUTEX_OP_XOR:
+		__futex_cas_op("xor\t%2, %4, %1", ret, oldval, uaddr, oparg);
+		break;
+	default:
+		ret = -ENOSYS;
+	}
+
+	pagefault_enable();
+
+	if (!ret) {
+		switch (cmp) {
+		case FUTEX_OP_CMP_EQ: ret = (oldval == cmparg); break;
+		case FUTEX_OP_CMP_NE: ret = (oldval != cmparg); break;
+		case FUTEX_OP_CMP_LT: ret = (oldval < cmparg); break;
+		case FUTEX_OP_CMP_GE: ret = (oldval >= cmparg); break;
+		case FUTEX_OP_CMP_LE: ret = (oldval <= cmparg); break;
+		case FUTEX_OP_CMP_GT: ret = (oldval > cmparg); break;
+		default: ret = -ENOSYS;
+		}
+	}
+	return ret;
+}
+
+static inline int
+futex_atomic_cmpxchg_inatomic(int __user *uaddr, int oldval, int newval)
+{
+	__asm__ __volatile__(
+	"\n1:	casa	[%3] %%asi, %2, %0\n"
+	"2:\n"
+	"	.section .fixup,#alloc,#execinstr\n"
+	"	.align	4\n"
+	"3:	sethi	%%hi(2b), %0\n"
+	"	jmpl	%0 + %%lo(2b), %%g0\n"
+	"	 mov	%4, %0\n"
+	"	.previous\n"
+	"	.section __ex_table,\"a\"\n"
+	"	.align	4\n"
+	"	.word	1b, 3b\n"
+	"	.previous\n"
+	: "=r" (newval)
+	: "0" (newval), "r" (oldval), "r" (uaddr), "i" (-EFAULT)
+	: "memory");
+
+	return newval;
+}
+
+#endif /* !(_SPARC64_FUTEX_H) */
diff --git a/include/asm-sparc/hardirq.h b/include/asm-sparc/hardirq.h
index 4f63ed8df55..15647877310 100644
--- a/include/asm-sparc/hardirq.h
+++ b/include/asm-sparc/hardirq.h
@@ -1,23 +1,8 @@
-/* hardirq.h: 32-bit Sparc hard IRQ support.
- *
- * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1998-2000 Anton Blanchard (anton@samba.org)
- */
-
-#ifndef __SPARC_HARDIRQ_H
-#define __SPARC_HARDIRQ_H
-
-#include <linux/threads.h>
-#include <linux/spinlock.h>
-#include <linux/cache.h>
-
-/* entry.S is sensitive to the offsets of these fields */ /* XXX P3 Is it? */
-typedef struct {
-	unsigned int __softirq_pending;
-} ____cacheline_aligned irq_cpustat_t;
-
-#include <linux/irq_cpustat.h>	/* Standard mappings for irq_cpustat_t above */
-
-#define HARDIRQ_BITS    8
-
-#endif /* __SPARC_HARDIRQ_H */
+#ifndef ___ASM_SPARC_HARDIRQ_H
+#define ___ASM_SPARC_HARDIRQ_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/hardirq_64.h>
+#else
+#include <asm-sparc/hardirq_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/hardirq_32.h b/include/asm-sparc/hardirq_32.h
new file mode 100644
index 00000000000..4f63ed8df55
--- /dev/null
+++ b/include/asm-sparc/hardirq_32.h
@@ -0,0 +1,23 @@
+/* hardirq.h: 32-bit Sparc hard IRQ support.
+ *
+ * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1998-2000 Anton Blanchard (anton@samba.org)
+ */
+
+#ifndef __SPARC_HARDIRQ_H
+#define __SPARC_HARDIRQ_H
+
+#include <linux/threads.h>
+#include <linux/spinlock.h>
+#include <linux/cache.h>
+
+/* entry.S is sensitive to the offsets of these fields */ /* XXX P3 Is it? */
+typedef struct {
+	unsigned int __softirq_pending;
+} ____cacheline_aligned irq_cpustat_t;
+
+#include <linux/irq_cpustat.h>	/* Standard mappings for irq_cpustat_t above */
+
+#define HARDIRQ_BITS    8
+
+#endif /* __SPARC_HARDIRQ_H */
diff --git a/include/asm-sparc/hardirq_64.h b/include/asm-sparc/hardirq_64.h
new file mode 100644
index 00000000000..7c29fd1a87a
--- /dev/null
+++ b/include/asm-sparc/hardirq_64.h
@@ -0,0 +1,19 @@
+/* hardirq.h: 64-bit Sparc hard IRQ support.
+ *
+ * Copyright (C) 1997, 1998, 2005 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef __SPARC64_HARDIRQ_H
+#define __SPARC64_HARDIRQ_H
+
+#include <asm/cpudata.h>
+
+#define __ARCH_IRQ_STAT
+#define local_softirq_pending() \
+	(local_cpu_data().__softirq_pending)
+
+void ack_bad_irq(unsigned int irq);
+
+#define HARDIRQ_BITS	8
+
+#endif /* !(__SPARC64_HARDIRQ_H) */
diff --git a/include/asm-sparc/head.h b/include/asm-sparc/head.h
index 7c35491a8b5..14652abdea3 100644
--- a/include/asm-sparc/head.h
+++ b/include/asm-sparc/head.h
@@ -1,102 +1,8 @@
-#ifndef __SPARC_HEAD_H
-#define __SPARC_HEAD_H
-
-#define KERNBASE        0xf0000000  /* First address the kernel will eventually be */
-#define LOAD_ADDR       0x4000      /* prom jumps to us here unless this is elf /boot */
-#define SUN4C_SEGSZ     (1 << 18)
-#define SRMMU_L1_KBASE_OFFSET ((KERNBASE>>24)<<2)  /* Used in boot remapping. */
-#define INTS_ENAB        0x01           /* entry.S uses this. */
-
-#define SUN4_PROM_VECTOR 0xFFE81000     /* SUN4 PROM needs to be hardwired */
-
-#define WRITE_PAUSE      nop; nop; nop; /* Have to do this after %wim/%psr chg */
-#define NOP_INSN         0x01000000     /* Used to patch sparc_save_state */
-
-/* Here are some trap goodies */
-
-/* Generic trap entry. */
-#define TRAP_ENTRY(type, label) \
-	rd %psr, %l0; b label; rd %wim, %l3; nop;
-
-/* Data/text faults. Defaults to sun4c version at boot time. */
-#define SPARC_TFAULT rd %psr, %l0; rd %wim, %l3; b sun4c_fault; mov 1, %l7;
-#define SPARC_DFAULT rd %psr, %l0; rd %wim, %l3; b sun4c_fault; mov 0, %l7;
-#define SRMMU_TFAULT rd %psr, %l0; rd %wim, %l3; b srmmu_fault; mov 1, %l7;
-#define SRMMU_DFAULT rd %psr, %l0; rd %wim, %l3; b srmmu_fault; mov 0, %l7;
-
-/* This is for traps we should NEVER get. */
-#define BAD_TRAP(num) \
-        rd %psr, %l0; mov num, %l7; b bad_trap_handler; rd %wim, %l3;
-
-/* This is for traps when we want just skip the instruction which caused it */
-#define SKIP_TRAP(type, name) \
-	jmpl %l2, %g0; rett %l2 + 4; nop; nop;
-
-/* Notice that for the system calls we pull a trick.  We load up a
- * different pointer to the system call vector table in %l7, but call
- * the same generic system call low-level entry point.  The trap table
- * entry sequences are also HyperSparc pipeline friendly ;-)
- */
-
-/* Software trap for Linux system calls. */
-#define LINUX_SYSCALL_TRAP \
-        sethi %hi(sys_call_table), %l7; \
-        or %l7, %lo(sys_call_table), %l7; \
-        b linux_sparc_syscall; \
-        rd %psr, %l0;
-
-#define BREAKPOINT_TRAP \
-	b breakpoint_trap; \
-	rd %psr,%l0; \
-	nop; \
-	nop;
-
-#ifdef CONFIG_KGDB
-#define KGDB_TRAP(num) \
-	b kgdb_trap_low; \
-	rd %psr,%l0; \
-	nop; \
-	nop;
+#ifndef ___ASM_SPARC_HEAD_H
+#define ___ASM_SPARC_HEAD_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/head_64.h>
 #else
-#define KGDB_TRAP(num) \
-	BAD_TRAP(num)
+#include <asm-sparc/head_32.h>
+#endif
 #endif
-
-/* The Get Condition Codes software trap for userland. */
-#define GETCC_TRAP \
-        b getcc_trap_handler; mov %psr, %l0; nop; nop;
-
-/* The Set Condition Codes software trap for userland. */
-#define SETCC_TRAP \
-        b setcc_trap_handler; mov %psr, %l0; nop; nop;
-
-/* The Get PSR software trap for userland. */
-#define GETPSR_TRAP \
-	mov %psr, %i0; jmp %l2; rett %l2 + 4; nop;
-
-/* This is for hard interrupts from level 1-14, 15 is non-maskable (nmi) and
- * gets handled with another macro.
- */
-#define TRAP_ENTRY_INTERRUPT(int_level) \
-        mov int_level, %l7; rd %psr, %l0; b real_irq_entry; rd %wim, %l3;
-
-/* NMI's (Non Maskable Interrupts) are special, you can't keep them
- * from coming in, and basically if you get one, the shows over. ;(
- * On the sun4c they are usually asynchronous memory errors, on the
- * the sun4m they could be either due to mem errors or a software
- * initiated interrupt from the prom/kern on an SMP box saying "I
- * command you to do CPU tricks, read your mailbox for more info."
- */
-#define NMI_TRAP \
-        rd %wim, %l3; b linux_trap_nmi_sun4c; mov %psr, %l0; nop;
-
-/* Window overflows/underflows are special and we need to try to be as
- * efficient as possible here....
- */
-#define WINDOW_SPILL \
-        rd %psr, %l0; rd %wim, %l3; b spill_window_entry; andcc %l0, PSR_PS, %g0;
-
-#define WINDOW_FILL \
-        rd %psr, %l0; rd %wim, %l3; b fill_window_entry; andcc %l0, PSR_PS, %g0;
-
-#endif /* __SPARC_HEAD_H */
diff --git a/include/asm-sparc/head_32.h b/include/asm-sparc/head_32.h
new file mode 100644
index 00000000000..7c35491a8b5
--- /dev/null
+++ b/include/asm-sparc/head_32.h
@@ -0,0 +1,102 @@
+#ifndef __SPARC_HEAD_H
+#define __SPARC_HEAD_H
+
+#define KERNBASE        0xf0000000  /* First address the kernel will eventually be */
+#define LOAD_ADDR       0x4000      /* prom jumps to us here unless this is elf /boot */
+#define SUN4C_SEGSZ     (1 << 18)
+#define SRMMU_L1_KBASE_OFFSET ((KERNBASE>>24)<<2)  /* Used in boot remapping. */
+#define INTS_ENAB        0x01           /* entry.S uses this. */
+
+#define SUN4_PROM_VECTOR 0xFFE81000     /* SUN4 PROM needs to be hardwired */
+
+#define WRITE_PAUSE      nop; nop; nop; /* Have to do this after %wim/%psr chg */
+#define NOP_INSN         0x01000000     /* Used to patch sparc_save_state */
+
+/* Here are some trap goodies */
+
+/* Generic trap entry. */
+#define TRAP_ENTRY(type, label) \
+	rd %psr, %l0; b label; rd %wim, %l3; nop;
+
+/* Data/text faults. Defaults to sun4c version at boot time. */
+#define SPARC_TFAULT rd %psr, %l0; rd %wim, %l3; b sun4c_fault; mov 1, %l7;
+#define SPARC_DFAULT rd %psr, %l0; rd %wim, %l3; b sun4c_fault; mov 0, %l7;
+#define SRMMU_TFAULT rd %psr, %l0; rd %wim, %l3; b srmmu_fault; mov 1, %l7;
+#define SRMMU_DFAULT rd %psr, %l0; rd %wim, %l3; b srmmu_fault; mov 0, %l7;
+
+/* This is for traps we should NEVER get. */
+#define BAD_TRAP(num) \
+        rd %psr, %l0; mov num, %l7; b bad_trap_handler; rd %wim, %l3;
+
+/* This is for traps when we want just skip the instruction which caused it */
+#define SKIP_TRAP(type, name) \
+	jmpl %l2, %g0; rett %l2 + 4; nop; nop;
+
+/* Notice that for the system calls we pull a trick.  We load up a
+ * different pointer to the system call vector table in %l7, but call
+ * the same generic system call low-level entry point.  The trap table
+ * entry sequences are also HyperSparc pipeline friendly ;-)
+ */
+
+/* Software trap for Linux system calls. */
+#define LINUX_SYSCALL_TRAP \
+        sethi %hi(sys_call_table), %l7; \
+        or %l7, %lo(sys_call_table), %l7; \
+        b linux_sparc_syscall; \
+        rd %psr, %l0;
+
+#define BREAKPOINT_TRAP \
+	b breakpoint_trap; \
+	rd %psr,%l0; \
+	nop; \
+	nop;
+
+#ifdef CONFIG_KGDB
+#define KGDB_TRAP(num) \
+	b kgdb_trap_low; \
+	rd %psr,%l0; \
+	nop; \
+	nop;
+#else
+#define KGDB_TRAP(num) \
+	BAD_TRAP(num)
+#endif
+
+/* The Get Condition Codes software trap for userland. */
+#define GETCC_TRAP \
+        b getcc_trap_handler; mov %psr, %l0; nop; nop;
+
+/* The Set Condition Codes software trap for userland. */
+#define SETCC_TRAP \
+        b setcc_trap_handler; mov %psr, %l0; nop; nop;
+
+/* The Get PSR software trap for userland. */
+#define GETPSR_TRAP \
+	mov %psr, %i0; jmp %l2; rett %l2 + 4; nop;
+
+/* This is for hard interrupts from level 1-14, 15 is non-maskable (nmi) and
+ * gets handled with another macro.
+ */
+#define TRAP_ENTRY_INTERRUPT(int_level) \
+        mov int_level, %l7; rd %psr, %l0; b real_irq_entry; rd %wim, %l3;
+
+/* NMI's (Non Maskable Interrupts) are special, you can't keep them
+ * from coming in, and basically if you get one, the shows over. ;(
+ * On the sun4c they are usually asynchronous memory errors, on the
+ * the sun4m they could be either due to mem errors or a software
+ * initiated interrupt from the prom/kern on an SMP box saying "I
+ * command you to do CPU tricks, read your mailbox for more info."
+ */
+#define NMI_TRAP \
+        rd %wim, %l3; b linux_trap_nmi_sun4c; mov %psr, %l0; nop;
+
+/* Window overflows/underflows are special and we need to try to be as
+ * efficient as possible here....
+ */
+#define WINDOW_SPILL \
+        rd %psr, %l0; rd %wim, %l3; b spill_window_entry; andcc %l0, PSR_PS, %g0;
+
+#define WINDOW_FILL \
+        rd %psr, %l0; rd %wim, %l3; b fill_window_entry; andcc %l0, PSR_PS, %g0;
+
+#endif /* __SPARC_HEAD_H */
diff --git a/include/asm-sparc/head_64.h b/include/asm-sparc/head_64.h
new file mode 100644
index 00000000000..10e9dabc4c4
--- /dev/null
+++ b/include/asm-sparc/head_64.h
@@ -0,0 +1,76 @@
+#ifndef _SPARC64_HEAD_H
+#define _SPARC64_HEAD_H
+
+#include <asm/pstate.h>
+
+	/* wrpr	%g0, val, %gl */
+#define SET_GL(val)	\
+	.word	0xa1902000 | val
+
+	/* rdpr %gl, %gN */
+#define GET_GL_GLOBAL(N)	\
+	.word	0x81540000 | (N << 25)
+
+#define KERNBASE	0x400000
+
+#define	PTREGS_OFF	(STACK_BIAS + STACKFRAME_SZ)
+
+#define __CHEETAH_ID	0x003e0014
+#define __JALAPENO_ID	0x003e0016
+#define __SERRANO_ID	0x003e0022
+
+#define CHEETAH_MANUF		0x003e
+#define CHEETAH_IMPL		0x0014 /* Ultra-III   */
+#define CHEETAH_PLUS_IMPL	0x0015 /* Ultra-III+  */
+#define JALAPENO_IMPL		0x0016 /* Ultra-IIIi  */
+#define JAGUAR_IMPL		0x0018 /* Ultra-IV    */
+#define PANTHER_IMPL		0x0019 /* Ultra-IV+   */
+#define SERRANO_IMPL		0x0022 /* Ultra-IIIi+ */
+
+#define BRANCH_IF_SUN4V(tmp1,label)		\
+	sethi	%hi(is_sun4v), %tmp1;		\
+	lduw	[%tmp1 + %lo(is_sun4v)], %tmp1; \
+	brnz,pn	%tmp1, label;			\
+	 nop
+
+#define BRANCH_IF_CHEETAH_BASE(tmp1,tmp2,label)	\
+	rdpr	%ver, %tmp1;			\
+	sethi	%hi(__CHEETAH_ID), %tmp2;	\
+	srlx	%tmp1, 32, %tmp1;		\
+	or	%tmp2, %lo(__CHEETAH_ID), %tmp2;\
+	cmp	%tmp1, %tmp2;			\
+	be,pn	%icc, label;			\
+	 nop;
+
+#define BRANCH_IF_JALAPENO(tmp1,tmp2,label)	\
+	rdpr	%ver, %tmp1;			\
+	sethi	%hi(__JALAPENO_ID), %tmp2;	\
+	srlx	%tmp1, 32, %tmp1;		\
+	or	%tmp2, %lo(__JALAPENO_ID), %tmp2;\
+	cmp	%tmp1, %tmp2;			\
+	be,pn	%icc, label;			\
+	 nop;
+
+#define BRANCH_IF_CHEETAH_PLUS_OR_FOLLOWON(tmp1,tmp2,label)	\
+	rdpr	%ver, %tmp1;			\
+	srlx	%tmp1, (32 + 16), %tmp2;	\
+	cmp	%tmp2, CHEETAH_MANUF;		\
+	bne,pt	%xcc, 99f;			\
+	 sllx	%tmp1, 16, %tmp1;		\
+	srlx	%tmp1, (32 + 16), %tmp2;	\
+	cmp	%tmp2, CHEETAH_PLUS_IMPL;	\
+	bgeu,pt	%xcc, label;			\
+99:	 nop;
+
+#define BRANCH_IF_ANY_CHEETAH(tmp1,tmp2,label)	\
+	rdpr	%ver, %tmp1;			\
+	srlx	%tmp1, (32 + 16), %tmp2;	\
+	cmp	%tmp2, CHEETAH_MANUF;		\
+	bne,pt	%xcc, 99f;			\
+	 sllx	%tmp1, 16, %tmp1;		\
+	srlx	%tmp1, (32 + 16), %tmp2;	\
+	cmp	%tmp2, CHEETAH_IMPL;		\
+	bgeu,pt	%xcc, label;			\
+99:	 nop;
+
+#endif /* !(_SPARC64_HEAD_H) */
diff --git a/include/asm-sparc/hugetlb.h b/include/asm-sparc/hugetlb.h
new file mode 100644
index 00000000000..412af58926a
--- /dev/null
+++ b/include/asm-sparc/hugetlb.h
@@ -0,0 +1,84 @@
+#ifndef _ASM_SPARC64_HUGETLB_H
+#define _ASM_SPARC64_HUGETLB_H
+
+#include <asm/page.h>
+
+
+void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
+		     pte_t *ptep, pte_t pte);
+
+pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr,
+			      pte_t *ptep);
+
+void hugetlb_prefault_arch_hook(struct mm_struct *mm);
+
+static inline int is_hugepage_only_range(struct mm_struct *mm,
+					 unsigned long addr,
+					 unsigned long len) {
+	return 0;
+}
+
+/*
+ * If the arch doesn't supply something else, assume that hugepage
+ * size aligned regions are ok without further preparation.
+ */
+static inline int prepare_hugepage_range(unsigned long addr, unsigned long len)
+{
+	if (len & ~HPAGE_MASK)
+		return -EINVAL;
+	if (addr & ~HPAGE_MASK)
+		return -EINVAL;
+	return 0;
+}
+
+static inline void hugetlb_free_pgd_range(struct mmu_gather **tlb,
+					  unsigned long addr, unsigned long end,
+					  unsigned long floor,
+					  unsigned long ceiling)
+{
+	free_pgd_range(tlb, addr, end, floor, ceiling);
+}
+
+static inline void huge_ptep_clear_flush(struct vm_area_struct *vma,
+					 unsigned long addr, pte_t *ptep)
+{
+}
+
+static inline int huge_pte_none(pte_t pte)
+{
+	return pte_none(pte);
+}
+
+static inline pte_t huge_pte_wrprotect(pte_t pte)
+{
+	return pte_wrprotect(pte);
+}
+
+static inline void huge_ptep_set_wrprotect(struct mm_struct *mm,
+					   unsigned long addr, pte_t *ptep)
+{
+	ptep_set_wrprotect(mm, addr, ptep);
+}
+
+static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma,
+					     unsigned long addr, pte_t *ptep,
+					     pte_t pte, int dirty)
+{
+	return ptep_set_access_flags(vma, addr, ptep, pte, dirty);
+}
+
+static inline pte_t huge_ptep_get(pte_t *ptep)
+{
+	return *ptep;
+}
+
+static inline int arch_prepare_hugepage(struct page *page)
+{
+	return 0;
+}
+
+static inline void arch_release_hugepage(struct page *page)
+{
+}
+
+#endif /* _ASM_SPARC64_HUGETLB_H */
diff --git a/include/asm-sparc/hvtramp.h b/include/asm-sparc/hvtramp.h
new file mode 100644
index 00000000000..b2b9b947b3a
--- /dev/null
+++ b/include/asm-sparc/hvtramp.h
@@ -0,0 +1,37 @@
+#ifndef _SPARC64_HVTRAP_H
+#define _SPARC64_HVTRAP_H
+
+#ifndef __ASSEMBLY__
+
+#include <linux/types.h>
+
+struct hvtramp_mapping {
+	__u64		vaddr;
+	__u64		tte;
+};
+
+struct hvtramp_descr {
+	__u32			cpu;
+	__u32			num_mappings;
+	__u64			fault_info_va;
+	__u64			fault_info_pa;
+	__u64			thread_reg;
+	struct hvtramp_mapping	maps[1];
+};
+
+extern void hv_cpu_startup(unsigned long hvdescr_pa);
+
+#endif
+
+#define HVTRAMP_DESCR_CPU		0x00
+#define HVTRAMP_DESCR_NUM_MAPPINGS	0x04
+#define HVTRAMP_DESCR_FAULT_INFO_VA	0x08
+#define HVTRAMP_DESCR_FAULT_INFO_PA	0x10
+#define HVTRAMP_DESCR_THREAD_REG	0x18
+#define HVTRAMP_DESCR_MAPS		0x20
+
+#define HVTRAMP_MAPPING_VADDR		0x00
+#define HVTRAMP_MAPPING_TTE		0x08
+#define HVTRAMP_MAPPING_SIZE		0x10
+
+#endif /* _SPARC64_HVTRAP_H */
diff --git a/include/asm-sparc/hypervisor.h b/include/asm-sparc/hypervisor.h
new file mode 100644
index 00000000000..109ae24ba24
--- /dev/null
+++ b/include/asm-sparc/hypervisor.h
@@ -0,0 +1,2949 @@
+#ifndef _SPARC64_HYPERVISOR_H
+#define _SPARC64_HYPERVISOR_H
+
+/* Sun4v hypervisor interfaces and defines.
+ *
+ * Hypervisor calls are made via traps to software traps number 0x80
+ * and above.  Registers %o0 to %o5 serve as argument, status, and
+ * return value registers.
+ *
+ * There are two kinds of these traps.  First there are the normal
+ * "fast traps" which use software trap 0x80 and encode the function
+ * to invoke by number in register %o5.  Argument and return value
+ * handling is as follows:
+ *
+ * -----------------------------------------------
+ * |  %o5  | function number |     undefined     |
+ * |  %o0  |   argument 0    |   return status   |
+ * |  %o1  |   argument 1    |   return value 1  |
+ * |  %o2  |   argument 2    |   return value 2  |
+ * |  %o3  |   argument 3    |   return value 3  |
+ * |  %o4  |   argument 4    |   return value 4  |
+ * -----------------------------------------------
+ *
+ * The second type are "hyper-fast traps" which encode the function
+ * number in the software trap number itself.  So these use trap
+ * numbers > 0x80.  The register usage for hyper-fast traps is as
+ * follows:
+ *
+ * -----------------------------------------------
+ * |  %o0  |   argument 0    |   return status   |
+ * |  %o1  |   argument 1    |   return value 1  |
+ * |  %o2  |   argument 2    |   return value 2  |
+ * |  %o3  |   argument 3    |   return value 3  |
+ * |  %o4  |   argument 4    |   return value 4  |
+ * -----------------------------------------------
+ *
+ * Registers providing explicit arguments to the hypervisor calls
+ * are volatile across the call.  Upon return their values are
+ * undefined unless explicitly specified as containing a particular
+ * return value by the specific call.  The return status is always
+ * returned in register %o0, zero indicates a successful execution of
+ * the hypervisor call and other values indicate an error status as
+ * defined below.  So, for example, if a hyper-fast trap takes
+ * arguments 0, 1, and 2, then %o0, %o1, and %o2 are volatile across
+ * the call and %o3, %o4, and %o5 would be preserved.
+ *
+ * If the hypervisor trap is invalid, or the fast trap function number
+ * is invalid, HV_EBADTRAP will be returned in %o0.  Also, all 64-bits
+ * of the argument and return values are significant.
+ */
+
+/* Trap numbers.  */
+#define HV_FAST_TRAP		0x80
+#define HV_MMU_MAP_ADDR_TRAP	0x83
+#define HV_MMU_UNMAP_ADDR_TRAP	0x84
+#define HV_TTRACE_ADDENTRY_TRAP	0x85
+#define HV_CORE_TRAP		0xff
+
+/* Error codes.  */
+#define HV_EOK				0  /* Successful return            */
+#define HV_ENOCPU			1  /* Invalid CPU id               */
+#define HV_ENORADDR			2  /* Invalid real address         */
+#define HV_ENOINTR			3  /* Invalid interrupt id         */
+#define HV_EBADPGSZ			4  /* Invalid pagesize encoding    */
+#define HV_EBADTSB			5  /* Invalid TSB description      */
+#define HV_EINVAL			6  /* Invalid argument             */
+#define HV_EBADTRAP			7  /* Invalid function number      */
+#define HV_EBADALIGN			8  /* Invalid address alignment    */
+#define HV_EWOULDBLOCK			9  /* Cannot complete w/o blocking */
+#define HV_ENOACCESS			10 /* No access to resource        */
+#define HV_EIO				11 /* I/O error                    */
+#define HV_ECPUERROR			12 /* CPU in error state           */
+#define HV_ENOTSUPPORTED		13 /* Function not supported       */
+#define HV_ENOMAP			14 /* No mapping found             */
+#define HV_ETOOMANY			15 /* Too many items specified     */
+#define HV_ECHANNEL			16 /* Invalid LDC channel          */
+#define HV_EBUSY			17 /* Resource busy                */
+
+/* mach_exit()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_EXIT
+ * ARG0:	exit code
+ * ERRORS:	This service does not return.
+ *
+ * Stop all CPUs in the virtual domain and place them into the stopped
+ * state.  The 64-bit exit code may be passed to a service entity as
+ * the domain's exit status.  On systems without a service entity, the
+ * domain will undergo a reset, and the boot firmware will be
+ * reloaded.
+ *
+ * This function will never return to the guest that invokes it.
+ *
+ * Note: By convention an exit code of zero denotes a successful exit by
+ *       the guest code.  A non-zero exit code denotes a guest specific
+ *       error indication.
+ *
+ */
+#define HV_FAST_MACH_EXIT		0x00
+
+#ifndef __ASSEMBLY__
+extern void sun4v_mach_exit(unsigned long exit_code);
+#endif
+
+/* Domain services.  */
+
+/* mach_desc()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_DESC
+ * ARG0:	buffer
+ * ARG1:	length
+ * RET0:	status
+ * RET1:	length
+ * ERRORS:	HV_EBADALIGN	Buffer is badly aligned
+ *		HV_ENORADDR	Buffer is to an illegal real address.
+ *		HV_EINVAL	Buffer length is too small for complete
+ *				machine description.
+ *
+ * Copy the most current machine description into the buffer indicated
+ * by the real address in ARG0.  The buffer provided must be 16 byte
+ * aligned.  Upon success or HV_EINVAL, this service returns the
+ * actual size of the machine description in the RET1 return value.
+ *
+ * Note: A method of determining the appropriate buffer size for the
+ *       machine description is to first call this service with a buffer
+ *       length of 0 bytes.
+ */
+#define HV_FAST_MACH_DESC		0x01
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mach_desc(unsigned long buffer_pa,
+				     unsigned long buf_len,
+				     unsigned long *real_buf_len);
+#endif
+
+/* mach_sir()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_SIR
+ * ERRORS:	This service does not return.
+ *
+ * Perform a software initiated reset of the virtual machine domain.
+ * All CPUs are captured as soon as possible, all hardware devices are
+ * returned to the entry default state, and the domain is restarted at
+ * the SIR (trap type 0x04) real trap table (RTBA) entry point on one
+ * of the CPUs.  The single CPU restarted is selected as determined by
+ * platform specific policy.  Memory is preserved across this
+ * operation.
+ */
+#define HV_FAST_MACH_SIR		0x02
+
+#ifndef __ASSEMBLY__
+extern void sun4v_mach_sir(void);
+#endif
+
+/* mach_set_watchdog()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_SET_WATCHDOG
+ * ARG0:	timeout in milliseconds
+ * RET0:	status
+ * RET1:	time remaining in milliseconds
+ *
+ * A guest uses this API to set a watchdog timer.  Once the gues has set
+ * the timer, it must call the timer service again either to disable or
+ * postpone the expiration.  If the timer expires before being reset or
+ * disabled, then the hypervisor take a platform specific action leading
+ * to guest termination within a bounded time period.  The platform action
+ * may include recovery actions such as reporting the expiration to a
+ * Service Processor, and/or automatically restarting the gues.
+ *
+ * The 'timeout' parameter is specified in milliseconds, however the
+ * implementated granularity is given by the 'watchdog-resolution'
+ * property in the 'platform' node of the guest's machine description.
+ * The largest allowed timeout value is specified by the
+ * 'watchdog-max-timeout' property of the 'platform' node.
+ *
+ * If the 'timeout' argument is not zero, the watchdog timer is set to
+ * expire after a minimum of 'timeout' milliseconds.
+ *
+ * If the 'timeout' argument is zero, the watchdog timer is disabled.
+ *
+ * If the 'timeout' value exceeds the value of the 'max-watchdog-timeout'
+ * property, the hypervisor leaves the watchdog timer state unchanged,
+ * and returns a status of EINVAL.
+ *
+ * The 'time remaining' return value is valid regardless of whether the
+ * return status is EOK or EINVAL.  A non-zero return value indicates the
+ * number of milliseconds that were remaining until the timer was to expire.
+ * If less than one millisecond remains, the return value is '1'.  If the
+ * watchdog timer was disabled at the time of the call, the return value is
+ * zero.
+ *
+ * If the hypervisor cannot support the exact timeout value requested, but
+ * can support a larger timeout value, the hypervisor may round the actual
+ * timeout to a value larger than the requested timeout, consequently the
+ * 'time remaining' return value may be larger than the previously requested
+ * timeout value.
+ *
+ * Any guest OS debugger should be aware that the watchdog service may be in
+ * use.  Consequently, it is recommended that the watchdog service is
+ * disabled upon debugger entry (e.g. reaching a breakpoint), and then
+ * re-enabled upon returning to normal execution.  The API has been designed
+ * with this in mind, and the 'time remaining' result of the disable call may
+ * be used directly as the timeout argument of the re-enable call.
+ */
+#define HV_FAST_MACH_SET_WATCHDOG	0x05
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mach_set_watchdog(unsigned long timeout,
+					     unsigned long *orig_timeout);
+#endif
+
+/* CPU services.
+ *
+ * CPUs represent devices that can execute software threads.  A single
+ * chip that contains multiple cores or strands is represented as
+ * multiple CPUs with unique CPU identifiers.  CPUs are exported to
+ * OBP via the machine description (and to the OS via the OBP device
+ * tree).  CPUs are always in one of three states: stopped, running,
+ * or error.
+ *
+ * A CPU ID is a pre-assigned 16-bit value that uniquely identifies a
+ * CPU within a logical domain.  Operations that are to be performed
+ * on multiple CPUs specify them via a CPU list.  A CPU list is an
+ * array in real memory, of which each 16-bit word is a CPU ID.  CPU
+ * lists are passed through the API as two arguments.  The first is
+ * the number of entries (16-bit words) in the CPU list, and the
+ * second is the (real address) pointer to the CPU ID list.
+ */
+
+/* cpu_start()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_START
+ * ARG0:	CPU ID
+ * ARG1:	PC
+ * ARG2:	RTBA
+ * ARG3:	target ARG0
+ * RET0:	status
+ * ERRORS:	ENOCPU		Invalid CPU ID
+ *		EINVAL		Target CPU ID is not in the stopped state
+ *		ENORADDR	Invalid PC or RTBA real address
+ *		EBADALIGN	Unaligned PC or unaligned RTBA
+ *		EWOULDBLOCK	Starting resources are not available
+ *
+ * Start CPU with given CPU ID with PC in %pc and with a real trap
+ * base address value of RTBA.  The indicated CPU must be in the
+ * stopped state.  The supplied RTBA must be aligned on a 256 byte
+ * boundary.  On successful completion, the specified CPU will be in
+ * the running state and will be supplied with "target ARG0" in %o0
+ * and RTBA in %tba.
+ */
+#define HV_FAST_CPU_START		0x10
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_cpu_start(unsigned long cpuid,
+				     unsigned long pc,
+				     unsigned long rtba,
+				     unsigned long arg0);
+#endif
+
+/* cpu_stop()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_STOP
+ * ARG0:	CPU ID
+ * RET0:	status
+ * ERRORS:	ENOCPU		Invalid CPU ID
+ *		EINVAL		Target CPU ID is the current cpu
+ *		EINVAL		Target CPU ID is not in the running state
+ *		EWOULDBLOCK	Stopping resources are not available
+ *		ENOTSUPPORTED	Not supported on this platform
+ *
+ * The specified CPU is stopped.  The indicated CPU must be in the
+ * running state.  On completion, it will be in the stopped state.  It
+ * is not legal to stop the current CPU.
+ *
+ * Note: As this service cannot be used to stop the current cpu, this service
+ *       may not be used to stop the last running CPU in a domain.  To stop
+ *       and exit a running domain, a guest must use the mach_exit() service.
+ */
+#define HV_FAST_CPU_STOP		0x11
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_cpu_stop(unsigned long cpuid);
+#endif
+
+/* cpu_yield()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_YIELD
+ * RET0:	status
+ * ERRORS:	No possible error.
+ *
+ * Suspend execution on the current CPU.  Execution will resume when
+ * an interrupt (device, %stick_compare, or cross-call) is targeted to
+ * the CPU.  On some CPUs, this API may be used by the hypervisor to
+ * save power by disabling hardware strands.
+ */
+#define HV_FAST_CPU_YIELD		0x12
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_cpu_yield(void);
+#endif
+
+/* cpu_qconf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_QCONF
+ * ARG0:	queue
+ * ARG1:	base real address
+ * ARG2:	number of entries
+ * RET0:	status
+ * ERRORS:	ENORADDR	Invalid base real address
+ *		EINVAL		Invalid queue or number of entries is less
+ *				than 2 or too large.
+ *		EBADALIGN	Base real address is not correctly aligned
+ *				for size.
+ *
+ * Configure the given queue to be placed at the given base real
+ * address, with the given number of entries.  The number of entries
+ * must be a power of 2.  The base real address must be aligned
+ * exactly to match the queue size.  Each queue entry is 64 bytes
+ * long, so for example a 32 entry queue must be aligned on a 2048
+ * byte real address boundary.
+ *
+ * The specified queue is unconfigured if the number of entries is given
+ * as zero.
+ *
+ * For the current version of this API service, the argument queue is defined
+ * as follows:
+ *
+ *	queue		description
+ *	-----		-------------------------
+ *	0x3c		cpu mondo queue
+ *	0x3d		device mondo queue
+ *	0x3e		resumable error queue
+ *	0x3f		non-resumable error queue
+ *
+ * Note: The maximum number of entries for each queue for a specific cpu may
+ *       be determined from the machine description.
+ */
+#define HV_FAST_CPU_QCONF		0x14
+#define  HV_CPU_QUEUE_CPU_MONDO		 0x3c
+#define  HV_CPU_QUEUE_DEVICE_MONDO	 0x3d
+#define  HV_CPU_QUEUE_RES_ERROR		 0x3e
+#define  HV_CPU_QUEUE_NONRES_ERROR	 0x3f
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_cpu_qconf(unsigned long type,
+				     unsigned long queue_paddr,
+				     unsigned long num_queue_entries);
+#endif
+
+/* cpu_qinfo()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_QINFO
+ * ARG0:	queue
+ * RET0:	status
+ * RET1:	base real address
+ * RET1:	number of entries
+ * ERRORS:	EINVAL		Invalid queue
+ *
+ * Return the configuration info for the given queue.  The base real
+ * address and number of entries of the defined queue are returned.
+ * The queue argument values are the same as for cpu_qconf() above.
+ *
+ * If the specified queue is a valid queue number, but no queue has
+ * been defined, the number of entries will be set to zero and the
+ * base real address returned is undefined.
+ */
+#define HV_FAST_CPU_QINFO		0x15
+
+/* cpu_mondo_send()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_MONDO_SEND
+ * ARG0-1:	CPU list
+ * ARG2:	data real address
+ * RET0:	status
+ * ERRORS:	EBADALIGN	Mondo data is not 64-byte aligned or CPU list
+ *				is not 2-byte aligned.
+ *		ENORADDR	Invalid data mondo address, or invalid cpu list
+ *				address.
+ *		ENOCPU		Invalid cpu in CPU list
+ *		EWOULDBLOCK	Some or all of the listed CPUs did not receive
+ *				the mondo
+ *		ECPUERROR	One or more of the listed CPUs are in error
+ *				state, use HV_FAST_CPU_STATE to see which ones
+ *		EINVAL		CPU list includes caller's CPU ID
+ *
+ * Send a mondo interrupt to the CPUs in the given CPU list with the
+ * 64-bytes at the given data real address.  The data must be 64-byte
+ * aligned.  The mondo data will be delivered to the cpu_mondo queues
+ * of the recipient CPUs.
+ *
+ * In all cases, error or not, the CPUs in the CPU list to which the
+ * mondo has been successfully delivered will be indicated by having
+ * their entry in CPU list updated with the value 0xffff.
+ */
+#define HV_FAST_CPU_MONDO_SEND		0x42
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_cpu_mondo_send(unsigned long cpu_count, unsigned long cpu_list_pa, unsigned long mondo_block_pa);
+#endif
+
+/* cpu_myid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_MYID
+ * RET0:	status
+ * RET1:	CPU ID
+ * ERRORS:	No errors defined.
+ *
+ * Return the hypervisor ID handle for the current CPU.  Use by a
+ * virtual CPU to discover it's own identity.
+ */
+#define HV_FAST_CPU_MYID		0x16
+
+/* cpu_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_STATE
+ * ARG0:	CPU ID
+ * RET0:	status
+ * RET1:	state
+ * ERRORS:	ENOCPU		Invalid CPU ID
+ *
+ * Retrieve the current state of the CPU with the given CPU ID.
+ */
+#define HV_FAST_CPU_STATE		0x17
+#define  HV_CPU_STATE_STOPPED		 0x01
+#define  HV_CPU_STATE_RUNNING		 0x02
+#define  HV_CPU_STATE_ERROR		 0x03
+
+#ifndef __ASSEMBLY__
+extern long sun4v_cpu_state(unsigned long cpuid);
+#endif
+
+/* cpu_set_rtba()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_SET_RTBA
+ * ARG0:	RTBA
+ * RET0:	status
+ * RET1:	previous RTBA
+ * ERRORS:	ENORADDR	Invalid RTBA real address
+ *		EBADALIGN	RTBA is incorrectly aligned for a trap table
+ *
+ * Set the real trap base address of the local cpu to the given RTBA.
+ * The supplied RTBA must be aligned on a 256 byte boundary.  Upon
+ * success the previous value of the RTBA is returned in RET1.
+ *
+ * Note: This service does not affect %tba
+ */
+#define HV_FAST_CPU_SET_RTBA		0x18
+
+/* cpu_set_rtba()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CPU_GET_RTBA
+ * RET0:	status
+ * RET1:	previous RTBA
+ * ERRORS:	No possible error.
+ *
+ * Returns the current value of RTBA in RET1.
+ */
+#define HV_FAST_CPU_GET_RTBA		0x19
+
+/* MMU services.
+ *
+ * Layout of a TSB description for mmu_tsb_ctx{,non}0() calls.
+ */
+#ifndef __ASSEMBLY__
+struct hv_tsb_descr {
+	unsigned short		pgsz_idx;
+	unsigned short		assoc;
+	unsigned int		num_ttes;	/* in TTEs */
+	unsigned int		ctx_idx;
+	unsigned int		pgsz_mask;
+	unsigned long		tsb_base;
+	unsigned long		resv;
+};
+#endif
+#define HV_TSB_DESCR_PGSZ_IDX_OFFSET	0x00
+#define HV_TSB_DESCR_ASSOC_OFFSET	0x02
+#define HV_TSB_DESCR_NUM_TTES_OFFSET	0x04
+#define HV_TSB_DESCR_CTX_IDX_OFFSET	0x08
+#define HV_TSB_DESCR_PGSZ_MASK_OFFSET	0x0c
+#define HV_TSB_DESCR_TSB_BASE_OFFSET	0x10
+#define HV_TSB_DESCR_RESV_OFFSET	0x18
+
+/* Page size bitmask.  */
+#define HV_PGSZ_MASK_8K			(1 << 0)
+#define HV_PGSZ_MASK_64K		(1 << 1)
+#define HV_PGSZ_MASK_512K		(1 << 2)
+#define HV_PGSZ_MASK_4MB		(1 << 3)
+#define HV_PGSZ_MASK_32MB		(1 << 4)
+#define HV_PGSZ_MASK_256MB		(1 << 5)
+#define HV_PGSZ_MASK_2GB		(1 << 6)
+#define HV_PGSZ_MASK_16GB		(1 << 7)
+
+/* Page size index.  The value given in the TSB descriptor must correspond
+ * to the smallest page size specified in the pgsz_mask page size bitmask.
+ */
+#define HV_PGSZ_IDX_8K			0
+#define HV_PGSZ_IDX_64K			1
+#define HV_PGSZ_IDX_512K		2
+#define HV_PGSZ_IDX_4MB			3
+#define HV_PGSZ_IDX_32MB		4
+#define HV_PGSZ_IDX_256MB		5
+#define HV_PGSZ_IDX_2GB			6
+#define HV_PGSZ_IDX_16GB		7
+
+/* MMU fault status area.
+ *
+ * MMU related faults have their status and fault address information
+ * placed into a memory region made available by privileged code.  Each
+ * virtual processor must make a mmu_fault_area_conf() call to tell the
+ * hypervisor where that processor's fault status should be stored.
+ *
+ * The fault status block is a multiple of 64-bytes and must be aligned
+ * on a 64-byte boundary.
+ */
+#ifndef __ASSEMBLY__
+struct hv_fault_status {
+	unsigned long		i_fault_type;
+	unsigned long		i_fault_addr;
+	unsigned long		i_fault_ctx;
+	unsigned long		i_reserved[5];
+	unsigned long		d_fault_type;
+	unsigned long		d_fault_addr;
+	unsigned long		d_fault_ctx;
+	unsigned long		d_reserved[5];
+};
+#endif
+#define HV_FAULT_I_TYPE_OFFSET	0x00
+#define HV_FAULT_I_ADDR_OFFSET	0x08
+#define HV_FAULT_I_CTX_OFFSET	0x10
+#define HV_FAULT_D_TYPE_OFFSET	0x40
+#define HV_FAULT_D_ADDR_OFFSET	0x48
+#define HV_FAULT_D_CTX_OFFSET	0x50
+
+#define HV_FAULT_TYPE_FAST_MISS	1
+#define HV_FAULT_TYPE_FAST_PROT	2
+#define HV_FAULT_TYPE_MMU_MISS	3
+#define HV_FAULT_TYPE_INV_RA	4
+#define HV_FAULT_TYPE_PRIV_VIOL	5
+#define HV_FAULT_TYPE_PROT_VIOL	6
+#define HV_FAULT_TYPE_NFO	7
+#define HV_FAULT_TYPE_NFO_SEFF	8
+#define HV_FAULT_TYPE_INV_VA	9
+#define HV_FAULT_TYPE_INV_ASI	10
+#define HV_FAULT_TYPE_NC_ATOMIC	11
+#define HV_FAULT_TYPE_PRIV_ACT	12
+#define HV_FAULT_TYPE_RESV1	13
+#define HV_FAULT_TYPE_UNALIGNED	14
+#define HV_FAULT_TYPE_INV_PGSZ	15
+/* Values 16 --> -2 are reserved.  */
+#define HV_FAULT_TYPE_MULTIPLE	-1
+
+/* Flags argument for mmu_{map,unmap}_addr(), mmu_demap_{page,context,all}(),
+ * and mmu_{map,unmap}_perm_addr().
+ */
+#define HV_MMU_DMMU			0x01
+#define HV_MMU_IMMU			0x02
+#define HV_MMU_ALL			(HV_MMU_DMMU | HV_MMU_IMMU)
+
+/* mmu_map_addr()
+ * TRAP:	HV_MMU_MAP_ADDR_TRAP
+ * ARG0:	virtual address
+ * ARG1:	mmu context
+ * ARG2:	TTE
+ * ARG3:	flags (HV_MMU_{IMMU,DMMU})
+ * ERRORS:	EINVAL		Invalid virtual address, mmu context, or flags
+ *		EBADPGSZ	Invalid page size value
+ *		ENORADDR	Invalid real address in TTE
+ *
+ * Create a non-permanent mapping using the given TTE, virtual
+ * address, and mmu context.  The flags argument determines which
+ * (data, or instruction, or both) TLB the mapping gets loaded into.
+ *
+ * The behavior is undefined if the valid bit is clear in the TTE.
+ *
+ * Note: This API call is for privileged code to specify temporary translation
+ *       mappings without the need to create and manage a TSB.
+ */
+
+/* mmu_unmap_addr()
+ * TRAP:	HV_MMU_UNMAP_ADDR_TRAP
+ * ARG0:	virtual address
+ * ARG1:	mmu context
+ * ARG2:	flags (HV_MMU_{IMMU,DMMU})
+ * ERRORS:	EINVAL		Invalid virtual address, mmu context, or flags
+ *
+ * Demaps the given virtual address in the given mmu context on this
+ * CPU.  This function is intended to be used to demap pages mapped
+ * with mmu_map_addr.  This service is equivalent to invoking
+ * mmu_demap_page() with only the current CPU in the CPU list. The
+ * flags argument determines which (data, or instruction, or both) TLB
+ * the mapping gets unmapped from.
+ *
+ * Attempting to perform an unmap operation for a previously defined
+ * permanent mapping will have undefined results.
+ */
+
+/* mmu_tsb_ctx0()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_TSB_CTX0
+ * ARG0:	number of TSB descriptions
+ * ARG1:	TSB descriptions pointer
+ * RET0:	status
+ * ERRORS:	ENORADDR		Invalid TSB descriptions pointer or
+ *					TSB base within a descriptor
+ *		EBADALIGN		TSB descriptions pointer is not aligned
+ *					to an 8-byte boundary, or TSB base
+ *					within a descriptor is not aligned for
+ *					the given TSB size
+ *		EBADPGSZ		Invalid page size in a TSB descriptor
+ *		EBADTSB			Invalid associativity or size in a TSB
+ *					descriptor
+ *		EINVAL			Invalid number of TSB descriptions, or
+ *					invalid context index in a TSB
+ *					descriptor, or index page size not
+ *					equal to smallest page size in page
+ *					size bitmask field.
+ *
+ * Configures the TSBs for the current CPU for virtual addresses with
+ * context zero.  The TSB descriptions pointer is a pointer to an
+ * array of the given number of TSB descriptions.
+ *
+ * Note: The maximum number of TSBs available to a virtual CPU is given by the
+ *       mmu-max-#tsbs property of the cpu's corresponding "cpu" node in the
+ *       machine description.
+ */
+#define HV_FAST_MMU_TSB_CTX0		0x20
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mmu_tsb_ctx0(unsigned long num_descriptions,
+					unsigned long tsb_desc_ra);
+#endif
+
+/* mmu_tsb_ctxnon0()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_TSB_CTXNON0
+ * ARG0:	number of TSB descriptions
+ * ARG1:	TSB descriptions pointer
+ * RET0:	status
+ * ERRORS:	Same as for mmu_tsb_ctx0() above.
+ *
+ * Configures the TSBs for the current CPU for virtual addresses with
+ * non-zero contexts.  The TSB descriptions pointer is a pointer to an
+ * array of the given number of TSB descriptions.
+ *
+ * Note: A maximum of 16 TSBs may be specified in the TSB description list.
+ */
+#define HV_FAST_MMU_TSB_CTXNON0		0x21
+
+/* mmu_demap_page()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_DEMAP_PAGE
+ * ARG0:	reserved, must be zero
+ * ARG1:	reserved, must be zero
+ * ARG2:	virtual address
+ * ARG3:	mmu context
+ * ARG4:	flags (HV_MMU_{IMMU,DMMU})
+ * RET0:	status
+ * ERRORS:	EINVAL			Invalid virutal address, context, or
+ *					flags value
+ *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
+ *
+ * Demaps any page mapping of the given virtual address in the given
+ * mmu context for the current virtual CPU.  Any virtually tagged
+ * caches are guaranteed to be kept consistent.  The flags argument
+ * determines which TLB (instruction, or data, or both) participate in
+ * the operation.
+ *
+ * ARG0 and ARG1 are both reserved and must be set to zero.
+ */
+#define HV_FAST_MMU_DEMAP_PAGE		0x22
+
+/* mmu_demap_ctx()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_DEMAP_CTX
+ * ARG0:	reserved, must be zero
+ * ARG1:	reserved, must be zero
+ * ARG2:	mmu context
+ * ARG3:	flags (HV_MMU_{IMMU,DMMU})
+ * RET0:	status
+ * ERRORS:	EINVAL			Invalid context or flags value
+ *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
+ *
+ * Demaps all non-permanent virtual page mappings previously specified
+ * for the given context for the current virtual CPU.  Any virtual
+ * tagged caches are guaranteed to be kept consistent.  The flags
+ * argument determines which TLB (instruction, or data, or both)
+ * participate in the operation.
+ *
+ * ARG0 and ARG1 are both reserved and must be set to zero.
+ */
+#define HV_FAST_MMU_DEMAP_CTX		0x23
+
+/* mmu_demap_all()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_DEMAP_ALL
+ * ARG0:	reserved, must be zero
+ * ARG1:	reserved, must be zero
+ * ARG2:	flags (HV_MMU_{IMMU,DMMU})
+ * RET0:	status
+ * ERRORS:	EINVAL			Invalid flags value
+ *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
+ *
+ * Demaps all non-permanent virtual page mappings previously specified
+ * for the current virtual CPU.  Any virtual tagged caches are
+ * guaranteed to be kept consistent.  The flags argument determines
+ * which TLB (instruction, or data, or both) participate in the
+ * operation.
+ *
+ * ARG0 and ARG1 are both reserved and must be set to zero.
+ */
+#define HV_FAST_MMU_DEMAP_ALL		0x24
+
+#ifndef __ASSEMBLY__
+extern void sun4v_mmu_demap_all(void);
+#endif
+
+/* mmu_map_perm_addr()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_MAP_PERM_ADDR
+ * ARG0:	virtual address
+ * ARG1:	reserved, must be zero
+ * ARG2:	TTE
+ * ARG3:	flags (HV_MMU_{IMMU,DMMU})
+ * RET0:	status
+ * ERRORS:	EINVAL			Invalid virutal address or flags value
+ *		EBADPGSZ		Invalid page size value
+ *		ENORADDR		Invalid real address in TTE
+ *		ETOOMANY		Too many mappings (max of 8 reached)
+ *
+ * Create a permanent mapping using the given TTE and virtual address
+ * for context 0 on the calling virtual CPU.  A maximum of 8 such
+ * permanent mappings may be specified by privileged code.  Mappings
+ * may be removed with mmu_unmap_perm_addr().
+ *
+ * The behavior is undefined if a TTE with the valid bit clear is given.
+ *
+ * Note: This call is used to specify address space mappings for which
+ *       privileged code does not expect to receive misses.  For example,
+ *       this mechanism can be used to map kernel nucleus code and data.
+ */
+#define HV_FAST_MMU_MAP_PERM_ADDR	0x25
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mmu_map_perm_addr(unsigned long vaddr,
+					     unsigned long set_to_zero,
+					     unsigned long tte,
+					     unsigned long flags);
+#endif
+
+/* mmu_fault_area_conf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_FAULT_AREA_CONF
+ * ARG0:	real address
+ * RET0:	status
+ * RET1:	previous mmu fault area real address
+ * ERRORS:	ENORADDR		Invalid real address
+ *		EBADALIGN		Invalid alignment for fault area
+ *
+ * Configure the MMU fault status area for the calling CPU.  A 64-byte
+ * aligned real address specifies where MMU fault status information
+ * is placed.  The return value is the previously specified area, or 0
+ * for the first invocation.  Specifying a fault area at real address
+ * 0 is not allowed.
+ */
+#define HV_FAST_MMU_FAULT_AREA_CONF	0x26
+
+/* mmu_enable()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_ENABLE
+ * ARG0:	enable flag
+ * ARG1:	return target address
+ * RET0:	status
+ * ERRORS:	ENORADDR		Invalid real address when disabling
+ *					translation.
+ *		EBADALIGN		The return target address is not
+ *					aligned to an instruction.
+ *		EINVAL			The enable flag request the current
+ *					operating mode (e.g. disable if already
+ *					disabled)
+ *
+ * Enable or disable virtual address translation for the calling CPU
+ * within the virtual machine domain.  If the enable flag is zero,
+ * translation is disabled, any non-zero value will enable
+ * translation.
+ *
+ * When this function returns, the newly selected translation mode
+ * will be active.  If the mmu is being enabled, then the return
+ * target address is a virtual address else it is a real address.
+ *
+ * Upon successful completion, control will be returned to the given
+ * return target address (ie. the cpu will jump to that address).  On
+ * failure, the previous mmu mode remains and the trap simply returns
+ * as normal with the appropriate error code in RET0.
+ */
+#define HV_FAST_MMU_ENABLE		0x27
+
+/* mmu_unmap_perm_addr()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_UNMAP_PERM_ADDR
+ * ARG0:	virtual address
+ * ARG1:	reserved, must be zero
+ * ARG2:	flags (HV_MMU_{IMMU,DMMU})
+ * RET0:	status
+ * ERRORS:	EINVAL			Invalid virutal address or flags value
+ *		ENOMAP			Specified mapping was not found
+ *
+ * Demaps any permanent page mapping (established via
+ * mmu_map_perm_addr()) at the given virtual address for context 0 on
+ * the current virtual CPU.  Any virtual tagged caches are guaranteed
+ * to be kept consistent.
+ */
+#define HV_FAST_MMU_UNMAP_PERM_ADDR	0x28
+
+/* mmu_tsb_ctx0_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_TSB_CTX0_INFO
+ * ARG0:	max TSBs
+ * ARG1:	buffer pointer
+ * RET0:	status
+ * RET1:	number of TSBs
+ * ERRORS:	EINVAL			Supplied buffer is too small
+ *		EBADALIGN		The buffer pointer is badly aligned
+ *		ENORADDR		Invalid real address for buffer pointer
+ *
+ * Return the TSB configuration as previous defined by mmu_tsb_ctx0()
+ * into the provided buffer.  The size of the buffer is given in ARG1
+ * in terms of the number of TSB description entries.
+ *
+ * Upon return, RET1 always contains the number of TSB descriptions
+ * previously configured.  If zero TSBs were configured, EOK is
+ * returned with RET1 containing 0.
+ */
+#define HV_FAST_MMU_TSB_CTX0_INFO	0x29
+
+/* mmu_tsb_ctxnon0_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_TSB_CTXNON0_INFO
+ * ARG0:	max TSBs
+ * ARG1:	buffer pointer
+ * RET0:	status
+ * RET1:	number of TSBs
+ * ERRORS:	EINVAL			Supplied buffer is too small
+ *		EBADALIGN		The buffer pointer is badly aligned
+ *		ENORADDR		Invalid real address for buffer pointer
+ *
+ * Return the TSB configuration as previous defined by
+ * mmu_tsb_ctxnon0() into the provided buffer.  The size of the buffer
+ * is given in ARG1 in terms of the number of TSB description entries.
+ *
+ * Upon return, RET1 always contains the number of TSB descriptions
+ * previously configured.  If zero TSBs were configured, EOK is
+ * returned with RET1 containing 0.
+ */
+#define HV_FAST_MMU_TSB_CTXNON0_INFO	0x2a
+
+/* mmu_fault_area_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMU_FAULT_AREA_INFO
+ * RET0:	status
+ * RET1:	fault area real address
+ * ERRORS:	No errors defined.
+ *
+ * Return the currently defined MMU fault status area for the current
+ * CPU.  The real address of the fault status area is returned in
+ * RET1, or 0 is returned in RET1 if no fault status area is defined.
+ *
+ * Note: mmu_fault_area_conf() may be called with the return value (RET1)
+ *       from this service if there is a need to save and restore the fault
+ *	 area for a cpu.
+ */
+#define HV_FAST_MMU_FAULT_AREA_INFO	0x2b
+
+/* Cache and Memory services. */
+
+/* mem_scrub()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MEM_SCRUB
+ * ARG0:	real address
+ * ARG1:	length
+ * RET0:	status
+ * RET1:	length scrubbed
+ * ERRORS:	ENORADDR	Invalid real address
+ *		EBADALIGN	Start address or length are not correctly
+ *				aligned
+ *		EINVAL		Length is zero
+ *
+ * Zero the memory contents in the range real address to real address
+ * plus length minus 1.  Also, valid ECC will be generated for that
+ * memory address range.  Scrubbing is started at the given real
+ * address, but may not scrub the entire given length.  The actual
+ * length scrubbed will be returned in RET1.
+ *
+ * The real address and length must be aligned on an 8K boundary, or
+ * contain the start address and length from a sun4v error report.
+ *
+ * Note: There are two uses for this function.  The first use is to block clear
+ *       and initialize memory and the second is to scrub an u ncorrectable
+ *       error reported via a resumable or non-resumable trap.  The second
+ *       use requires the arguments to be equal to the real address and length
+ *       provided in a sun4v memory error report.
+ */
+#define HV_FAST_MEM_SCRUB		0x31
+
+/* mem_sync()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MEM_SYNC
+ * ARG0:	real address
+ * ARG1:	length
+ * RET0:	status
+ * RET1:	length synced
+ * ERRORS:	ENORADDR	Invalid real address
+ *		EBADALIGN	Start address or length are not correctly
+ *				aligned
+ *		EINVAL		Length is zero
+ *
+ * Force the next access within the real address to real address plus
+ * length minus 1 to be fetches from main system memory.  Less than
+ * the given length may be synced, the actual amount synced is
+ * returned in RET1.  The real address and length must be aligned on
+ * an 8K boundary.
+ */
+#define HV_FAST_MEM_SYNC		0x32
+
+/* Time of day services.
+ *
+ * The hypervisor maintains the time of day on a per-domain basis.
+ * Changing the time of day in one domain does not affect the time of
+ * day on any other domain.
+ *
+ * Time is described by a single unsigned 64-bit word which is the
+ * number of seconds since the UNIX Epoch (00:00:00 UTC, January 1,
+ * 1970).
+ */
+
+/* tod_get()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TOD_GET
+ * RET0:	status
+ * RET1:	TOD
+ * ERRORS:	EWOULDBLOCK	TOD resource is temporarily unavailable
+ *		ENOTSUPPORTED	If TOD not supported on this platform
+ *
+ * Return the current time of day.  May block if TOD access is
+ * temporarily not possible.
+ */
+#define HV_FAST_TOD_GET			0x50
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_tod_get(unsigned long *time);
+#endif
+
+/* tod_set()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TOD_SET
+ * ARG0:	TOD
+ * RET0:	status
+ * ERRORS:	EWOULDBLOCK	TOD resource is temporarily unavailable
+ *		ENOTSUPPORTED	If TOD not supported on this platform
+ *
+ * The current time of day is set to the value specified in ARG0.  May
+ * block if TOD access is temporarily not possible.
+ */
+#define HV_FAST_TOD_SET			0x51
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_tod_set(unsigned long time);
+#endif
+
+/* Console services */
+
+/* con_getchar()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CONS_GETCHAR
+ * RET0:	status
+ * RET1:	character
+ * ERRORS:	EWOULDBLOCK	No character available.
+ *
+ * Returns a character from the console device.  If no character is
+ * available then an EWOULDBLOCK error is returned.  If a character is
+ * available, then the returned status is EOK and the character value
+ * is in RET1.
+ *
+ * A virtual BREAK is represented by the 64-bit value -1.
+ *
+ * A virtual HUP signal is represented by the 64-bit value -2.
+ */
+#define HV_FAST_CONS_GETCHAR		0x60
+
+/* con_putchar()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CONS_PUTCHAR
+ * ARG0:	character
+ * RET0:	status
+ * ERRORS:	EINVAL		Illegal character
+ *		EWOULDBLOCK	Output buffer currently full, would block
+ *
+ * Send a character to the console device.  Only character values
+ * between 0 and 255 may be used.  Values outside this range are
+ * invalid except for the 64-bit value -1 which is used to send a
+ * virtual BREAK.
+ */
+#define HV_FAST_CONS_PUTCHAR		0x61
+
+/* con_read()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CONS_READ
+ * ARG0:	buffer real address
+ * ARG1:	buffer size in bytes
+ * RET0:	status
+ * RET1:	bytes read or BREAK or HUP
+ * ERRORS:	EWOULDBLOCK	No character available.
+ *
+ * Reads characters into a buffer from the console device.  If no
+ * character is available then an EWOULDBLOCK error is returned.
+ * If a character is available, then the returned status is EOK
+ * and the number of bytes read into the given buffer is provided
+ * in RET1.
+ *
+ * A virtual BREAK is represented by the 64-bit RET1 value -1.
+ *
+ * A virtual HUP signal is represented by the 64-bit RET1 value -2.
+ *
+ * If BREAK or HUP are indicated, no bytes were read into buffer.
+ */
+#define HV_FAST_CONS_READ		0x62
+
+/* con_write()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_CONS_WRITE
+ * ARG0:	buffer real address
+ * ARG1:	buffer size in bytes
+ * RET0:	status
+ * RET1:	bytes written
+ * ERRORS:	EWOULDBLOCK	Output buffer currently full, would block
+ *
+ * Send a characters in buffer to the console device.  Breaks must be
+ * sent using con_putchar().
+ */
+#define HV_FAST_CONS_WRITE		0x63
+
+#ifndef __ASSEMBLY__
+extern long sun4v_con_getchar(long *status);
+extern long sun4v_con_putchar(long c);
+extern long sun4v_con_read(unsigned long buffer,
+			   unsigned long size,
+			   unsigned long *bytes_read);
+extern unsigned long sun4v_con_write(unsigned long buffer,
+				     unsigned long size,
+				     unsigned long *bytes_written);
+#endif
+
+/* mach_set_soft_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_SET_SOFT_STATE
+ * ARG0:	software state
+ * ARG1:	software state description pointer
+ * RET0:	status
+ * ERRORS:	EINVAL		software state not valid or software state
+ *				description is not NULL terminated
+ *		ENORADDR	software state description pointer is not a
+ *				valid real address
+ *		EBADALIGNED	software state description is not correctly
+ *				aligned
+ *
+ * This allows the guest to report it's soft state to the hypervisor.  There
+ * are two primary components to this state.  The first part states whether
+ * the guest software is running or not.  The second containts optional
+ * details specific to the software.
+ *
+ * The software state argument is defined below in HV_SOFT_STATE_*, and
+ * indicates whether the guest is operating normally or in a transitional
+ * state.
+ *
+ * The software state description argument is a real address of a data buffer
+ * of size 32-bytes aligned on a 32-byte boundary.  It is treated as a NULL
+ * terminated 7-bit ASCII string of up to 31 characters not including the
+ * NULL termination.
+ */
+#define HV_FAST_MACH_SET_SOFT_STATE	0x70
+#define  HV_SOFT_STATE_NORMAL		 0x01
+#define  HV_SOFT_STATE_TRANSITION	 0x02
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mach_set_soft_state(unsigned long soft_state,
+					       unsigned long msg_string_ra);
+#endif
+
+/* mach_get_soft_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MACH_GET_SOFT_STATE
+ * ARG0:	software state description pointer
+ * RET0:	status
+ * RET1:	software state
+ * ERRORS:	ENORADDR	software state description pointer is not a
+ *				valid real address
+ *		EBADALIGNED	software state description is not correctly
+ *				aligned
+ *
+ * Retrieve the current value of the guest's software state.  The rules
+ * for the software state pointer are the same as for mach_set_soft_state()
+ * above.
+ */
+#define HV_FAST_MACH_GET_SOFT_STATE	0x71
+
+/* svc_send()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SVC_SEND
+ * ARG0:	service ID
+ * ARG1:	buffer real address
+ * ARG2:	buffer size
+ * RET0:	STATUS
+ * RET1:	sent_bytes
+ *
+ * Be careful, all output registers are clobbered by this operation,
+ * so for example it is not possible to save away a value in %o4
+ * across the trap.
+ */
+#define HV_FAST_SVC_SEND		0x80
+
+/* svc_recv()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SVC_RECV
+ * ARG0:	service ID
+ * ARG1:	buffer real address
+ * ARG2:	buffer size
+ * RET0:	STATUS
+ * RET1:	recv_bytes
+ *
+ * Be careful, all output registers are clobbered by this operation,
+ * so for example it is not possible to save away a value in %o4
+ * across the trap.
+ */
+#define HV_FAST_SVC_RECV		0x81
+
+/* svc_getstatus()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SVC_GETSTATUS
+ * ARG0:	service ID
+ * RET0:	STATUS
+ * RET1:	status bits
+ */
+#define HV_FAST_SVC_GETSTATUS		0x82
+
+/* svc_setstatus()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SVC_SETSTATUS
+ * ARG0:	service ID
+ * ARG1:	bits to set
+ * RET0:	STATUS
+ */
+#define HV_FAST_SVC_SETSTATUS		0x83
+
+/* svc_clrstatus()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SVC_CLRSTATUS
+ * ARG0:	service ID
+ * ARG1:	bits to clear
+ * RET0:	STATUS
+ */
+#define HV_FAST_SVC_CLRSTATUS		0x84
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_svc_send(unsigned long svc_id,
+				    unsigned long buffer,
+				    unsigned long buffer_size,
+				    unsigned long *sent_bytes);
+extern unsigned long sun4v_svc_recv(unsigned long svc_id,
+				    unsigned long buffer,
+				    unsigned long buffer_size,
+				    unsigned long *recv_bytes);
+extern unsigned long sun4v_svc_getstatus(unsigned long svc_id,
+					 unsigned long *status_bits);
+extern unsigned long sun4v_svc_setstatus(unsigned long svc_id,
+					 unsigned long status_bits);
+extern unsigned long sun4v_svc_clrstatus(unsigned long svc_id,
+					 unsigned long status_bits);
+#endif
+
+/* Trap trace services.
+ *
+ * The hypervisor provides a trap tracing capability for privileged
+ * code running on each virtual CPU.  Privileged code provides a
+ * round-robin trap trace queue within which the hypervisor writes
+ * 64-byte entries detailing hyperprivileged traps taken n behalf of
+ * privileged code.  This is provided as a debugging capability for
+ * privileged code.
+ *
+ * The trap trace control structure is 64-bytes long and placed at the
+ * start (offset 0) of the trap trace buffer, and is described as
+ * follows:
+ */
+#ifndef __ASSEMBLY__
+struct hv_trap_trace_control {
+	unsigned long		head_offset;
+	unsigned long		tail_offset;
+	unsigned long		__reserved[0x30 / sizeof(unsigned long)];
+};
+#endif
+#define HV_TRAP_TRACE_CTRL_HEAD_OFFSET	0x00
+#define HV_TRAP_TRACE_CTRL_TAIL_OFFSET	0x08
+
+/* The head offset is the offset of the most recently completed entry
+ * in the trap-trace buffer.  The tail offset is the offset of the
+ * next entry to be written.  The control structure is owned and
+ * modified by the hypervisor.  A guest may not modify the control
+ * structure contents.  Attempts to do so will result in undefined
+ * behavior for the guest.
+ *
+ * Each trap trace buffer entry is layed out as follows:
+ */
+#ifndef __ASSEMBLY__
+struct hv_trap_trace_entry {
+	unsigned char	type;		/* Hypervisor or guest entry?	*/
+	unsigned char	hpstate;	/* Hyper-privileged state	*/
+	unsigned char	tl;		/* Trap level			*/
+	unsigned char	gl;		/* Global register level	*/
+	unsigned short	tt;		/* Trap type			*/
+	unsigned short	tag;		/* Extended trap identifier	*/
+	unsigned long	tstate;		/* Trap state			*/
+	unsigned long	tick;		/* Tick				*/
+	unsigned long	tpc;		/* Trap PC			*/
+	unsigned long	f1;		/* Entry specific		*/
+	unsigned long	f2;		/* Entry specific		*/
+	unsigned long	f3;		/* Entry specific		*/
+	unsigned long	f4;		/* Entry specific		*/
+};
+#endif
+#define HV_TRAP_TRACE_ENTRY_TYPE	0x00
+#define HV_TRAP_TRACE_ENTRY_HPSTATE	0x01
+#define HV_TRAP_TRACE_ENTRY_TL		0x02
+#define HV_TRAP_TRACE_ENTRY_GL		0x03
+#define HV_TRAP_TRACE_ENTRY_TT		0x04
+#define HV_TRAP_TRACE_ENTRY_TAG		0x06
+#define HV_TRAP_TRACE_ENTRY_TSTATE	0x08
+#define HV_TRAP_TRACE_ENTRY_TICK	0x10
+#define HV_TRAP_TRACE_ENTRY_TPC		0x18
+#define HV_TRAP_TRACE_ENTRY_F1		0x20
+#define HV_TRAP_TRACE_ENTRY_F2		0x28
+#define HV_TRAP_TRACE_ENTRY_F3		0x30
+#define HV_TRAP_TRACE_ENTRY_F4		0x38
+
+/* The type field is encoded as follows.  */
+#define HV_TRAP_TYPE_UNDEF		0x00 /* Entry content undefined     */
+#define HV_TRAP_TYPE_HV			0x01 /* Hypervisor trap entry       */
+#define HV_TRAP_TYPE_GUEST		0xff /* Added via ttrace_addentry() */
+
+/* ttrace_buf_conf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TTRACE_BUF_CONF
+ * ARG0:	real address
+ * ARG1:	number of entries
+ * RET0:	status
+ * RET1:	number of entries
+ * ERRORS:	ENORADDR	Invalid real address
+ *		EINVAL		Size is too small
+ *		EBADALIGN	Real address not aligned on 64-byte boundary
+ *
+ * Requests hypervisor trap tracing and declares a virtual CPU's trap
+ * trace buffer to the hypervisor.  The real address supplies the real
+ * base address of the trap trace queue and must be 64-byte aligned.
+ * Specifying a value of 0 for the number of entries disables trap
+ * tracing for the calling virtual CPU.  The buffer allocated must be
+ * sized for a power of two number of 64-byte trap trace entries plus
+ * an initial 64-byte control structure.
+ *
+ * This may be invoked any number of times so that a virtual CPU may
+ * relocate a trap trace buffer or create "snapshots" of information.
+ *
+ * If the real address is illegal or badly aligned, then trap tracing
+ * is disabled and an error is returned.
+ *
+ * Upon failure with EINVAL, this service call returns in RET1 the
+ * minimum number of buffer entries required.  Upon other failures
+ * RET1 is undefined.
+ */
+#define HV_FAST_TTRACE_BUF_CONF		0x90
+
+/* ttrace_buf_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TTRACE_BUF_INFO
+ * RET0:	status
+ * RET1:	real address
+ * RET2:	size
+ * ERRORS:	None defined.
+ *
+ * Returns the size and location of the previously declared trap-trace
+ * buffer.  In the event that no buffer was previously defined, or the
+ * buffer is disabled, this call will return a size of zero bytes.
+ */
+#define HV_FAST_TTRACE_BUF_INFO		0x91
+
+/* ttrace_enable()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TTRACE_ENABLE
+ * ARG0:	enable
+ * RET0:	status
+ * RET1:	previous enable state
+ * ERRORS:	EINVAL		No trap trace buffer currently defined
+ *
+ * Enable or disable trap tracing, and return the previous enabled
+ * state in RET1.  Future systems may define various flags for the
+ * enable argument (ARG0), for the moment a guest should pass
+ * "(uint64_t) -1" to enable, and "(uint64_t) 0" to disable all
+ * tracing - which will ensure future compatability.
+ */
+#define HV_FAST_TTRACE_ENABLE		0x92
+
+/* ttrace_freeze()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_TTRACE_FREEZE
+ * ARG0:	freeze
+ * RET0:	status
+ * RET1:	previous freeze state
+ * ERRORS:	EINVAL		No trap trace buffer currently defined
+ *
+ * Freeze or unfreeze trap tracing, returning the previous freeze
+ * state in RET1.  A guest should pass a non-zero value to freeze and
+ * a zero value to unfreeze all tracing.  The returned previous state
+ * is 0 for not frozen and 1 for frozen.
+ */
+#define HV_FAST_TTRACE_FREEZE		0x93
+
+/* ttrace_addentry()
+ * TRAP:	HV_TTRACE_ADDENTRY_TRAP
+ * ARG0:	tag (16-bits)
+ * ARG1:	data word 0
+ * ARG2:	data word 1
+ * ARG3:	data word 2
+ * ARG4:	data word 3
+ * RET0:	status
+ * ERRORS:	EINVAL		No trap trace buffer currently defined
+ *
+ * Add an entry to the trap trace buffer.  Upon return only ARG0/RET0
+ * is modified - none of the other registers holding arguments are
+ * volatile across this hypervisor service.
+ */
+
+/* Core dump services.
+ *
+ * Since the hypervisor viraulizes and thus obscures a lot of the
+ * physical machine layout and state, traditional OS crash dumps can
+ * be difficult to diagnose especially when the problem is a
+ * configuration error of some sort.
+ *
+ * The dump services provide an opaque buffer into which the
+ * hypervisor can place it's internal state in order to assist in
+ * debugging such situations.  The contents are opaque and extremely
+ * platform and hypervisor implementation specific.  The guest, during
+ * a core dump, requests that the hypervisor update any information in
+ * the dump buffer in preparation to being dumped as part of the
+ * domain's memory image.
+ */
+
+/* dump_buf_update()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_DUMP_BUF_UPDATE
+ * ARG0:	real address
+ * ARG1:	size
+ * RET0:	status
+ * RET1:	required size of dump buffer
+ * ERRORS:	ENORADDR	Invalid real address
+ *		EBADALIGN	Real address is not aligned on a 64-byte
+ *				boundary
+ *		EINVAL		Size is non-zero but less than minimum size
+ *				required
+ *		ENOTSUPPORTED	Operation not supported on current logical
+ *				domain
+ *
+ * Declare a domain dump buffer to the hypervisor.  The real address
+ * provided for the domain dump buffer must be 64-byte aligned.  The
+ * size specifies the size of the dump buffer and may be larger than
+ * the minimum size specified in the machine description.  The
+ * hypervisor will fill the dump buffer with opaque data.
+ *
+ * Note: A guest may elect to include dump buffer contents as part of a crash
+ *       dump to assist with debugging.  This function may be called any number
+ *       of times so that a guest may relocate a dump buffer, or create
+ *       "snapshots" of any dump-buffer information.  Each call to
+ *       dump_buf_update() atomically declares the new dump buffer to the
+ *       hypervisor.
+ *
+ * A specified size of 0 unconfigures the dump buffer.  If the real
+ * address is illegal or badly aligned, then any currently active dump
+ * buffer is disabled and an error is returned.
+ *
+ * In the event that the call fails with EINVAL, RET1 contains the
+ * minimum size requires by the hypervisor for a valid dump buffer.
+ */
+#define HV_FAST_DUMP_BUF_UPDATE		0x94
+
+/* dump_buf_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_DUMP_BUF_INFO
+ * RET0:	status
+ * RET1:	real address of current dump buffer
+ * RET2:	size of current dump buffer
+ * ERRORS:	No errors defined.
+ *
+ * Return the currently configures dump buffer description.  A
+ * returned size of 0 bytes indicates an undefined dump buffer.  In
+ * this case the return address in RET1 is undefined.
+ */
+#define HV_FAST_DUMP_BUF_INFO		0x95
+
+/* Device interrupt services.
+ *
+ * Device interrupts are allocated to system bus bridges by the hypervisor,
+ * and described to OBP in the machine description.  OBP then describes
+ * these interrupts to the OS via properties in the device tree.
+ *
+ * Terminology:
+ *
+ *	cpuid		Unique opaque value which represents a target cpu.
+ *
+ *	devhandle	Device handle.  It uniquely identifies a device, and
+ *			consistes of the lower 28-bits of the hi-cell of the
+ *			first entry of the device's "reg" property in the
+ *			OBP device tree.
+ *
+ *	devino		Device interrupt number.  Specifies the relative
+ *			interrupt number within the device.  The unique
+ *			combination of devhandle and devino are used to
+ *			identify a specific device interrupt.
+ *
+ *			Note: The devino value is the same as the values in the
+ *			      "interrupts" property or "interrupt-map" property
+ *			      in the OBP device tree for that device.
+ *
+ *	sysino		System interrupt number.  A 64-bit unsigned interger
+ *			representing a unique interrupt within a virtual
+ *			machine.
+ *
+ *	intr_state	A flag representing the interrupt state for a given
+ *			sysino.  The state values are defined below.
+ *
+ *	intr_enabled	A flag representing the 'enabled' state for a given
+ *			sysino.  The enable values are defined below.
+ */
+
+#define HV_INTR_STATE_IDLE		0 /* Nothing pending */
+#define HV_INTR_STATE_RECEIVED		1 /* Interrupt received by hardware */
+#define HV_INTR_STATE_DELIVERED		2 /* Interrupt delivered to queue */
+
+#define HV_INTR_DISABLED		0 /* sysino not enabled */
+#define HV_INTR_ENABLED			1 /* sysino enabled */
+
+/* intr_devino_to_sysino()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_DEVINO2SYSINO
+ * ARG0:	devhandle
+ * ARG1:	devino
+ * RET0:	status
+ * RET1:	sysino
+ * ERRORS:	EINVAL		Invalid devhandle/devino
+ *
+ * Converts a device specific interrupt number of the given
+ * devhandle/devino into a system specific ino (sysino).
+ */
+#define HV_FAST_INTR_DEVINO2SYSINO	0xa0
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_devino_to_sysino(unsigned long devhandle,
+					    unsigned long devino);
+#endif
+
+/* intr_getenabled()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_GETENABLED
+ * ARG0:	sysino
+ * RET0:	status
+ * RET1:	intr_enabled (HV_INTR_{DISABLED,ENABLED})
+ * ERRORS:	EINVAL		Invalid sysino
+ *
+ * Returns interrupt enabled state in RET1 for the interrupt defined
+ * by the given sysino.
+ */
+#define HV_FAST_INTR_GETENABLED		0xa1
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_getenabled(unsigned long sysino);
+#endif
+
+/* intr_setenabled()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_SETENABLED
+ * ARG0:	sysino
+ * ARG1:	intr_enabled (HV_INTR_{DISABLED,ENABLED})
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid sysino or intr_enabled value
+ *
+ * Set the 'enabled' state of the interrupt sysino.
+ */
+#define HV_FAST_INTR_SETENABLED		0xa2
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_setenabled(unsigned long sysino, unsigned long intr_enabled);
+#endif
+
+/* intr_getstate()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_GETSTATE
+ * ARG0:	sysino
+ * RET0:	status
+ * RET1:	intr_state (HV_INTR_STATE_*)
+ * ERRORS:	EINVAL		Invalid sysino
+ *
+ * Returns current state of the interrupt defined by the given sysino.
+ */
+#define HV_FAST_INTR_GETSTATE		0xa3
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_getstate(unsigned long sysino);
+#endif
+
+/* intr_setstate()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_SETSTATE
+ * ARG0:	sysino
+ * ARG1:	intr_state (HV_INTR_STATE_*)
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid sysino or intr_state value
+ *
+ * Sets the current state of the interrupt described by the given sysino
+ * value.
+ *
+ * Note: Setting the state to HV_INTR_STATE_IDLE clears any pending
+ *       interrupt for sysino.
+ */
+#define HV_FAST_INTR_SETSTATE		0xa4
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_setstate(unsigned long sysino, unsigned long intr_state);
+#endif
+
+/* intr_gettarget()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_GETTARGET
+ * ARG0:	sysino
+ * RET0:	status
+ * RET1:	cpuid
+ * ERRORS:	EINVAL		Invalid sysino
+ *
+ * Returns CPU that is the current target of the interrupt defined by
+ * the given sysino.  The CPU value returned is undefined if the target
+ * has not been set via intr_settarget().
+ */
+#define HV_FAST_INTR_GETTARGET		0xa5
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_gettarget(unsigned long sysino);
+#endif
+
+/* intr_settarget()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_INTR_SETTARGET
+ * ARG0:	sysino
+ * ARG1:	cpuid
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid sysino
+ *		ENOCPU		Invalid cpuid
+ *
+ * Set the target CPU for the interrupt defined by the given sysino.
+ */
+#define HV_FAST_INTR_SETTARGET		0xa6
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_intr_settarget(unsigned long sysino, unsigned long cpuid);
+#endif
+
+/* vintr_get_cookie()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_GET_COOKIE
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * RET0:	status
+ * RET1:	cookie
+ */
+#define HV_FAST_VINTR_GET_COOKIE	0xa7
+
+/* vintr_set_cookie()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_SET_COOKIE
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * ARG2:	cookie
+ * RET0:	status
+ */
+#define HV_FAST_VINTR_SET_COOKIE	0xa8
+
+/* vintr_get_valid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_GET_VALID
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * RET0:	status
+ * RET1:	valid state
+ */
+#define HV_FAST_VINTR_GET_VALID		0xa9
+
+/* vintr_set_valid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_SET_VALID
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * ARG2:	valid state
+ * RET0:	status
+ */
+#define HV_FAST_VINTR_SET_VALID		0xaa
+
+/* vintr_get_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_GET_STATE
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * RET0:	status
+ * RET1:	state
+ */
+#define HV_FAST_VINTR_GET_STATE		0xab
+
+/* vintr_set_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_SET_STATE
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * ARG2:	state
+ * RET0:	status
+ */
+#define HV_FAST_VINTR_SET_STATE		0xac
+
+/* vintr_get_target()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_GET_TARGET
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * RET0:	status
+ * RET1:	cpuid
+ */
+#define HV_FAST_VINTR_GET_TARGET	0xad
+
+/* vintr_set_target()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_VINTR_SET_TARGET
+ * ARG0:	device handle
+ * ARG1:	device ino
+ * ARG2:	cpuid
+ * RET0:	status
+ */
+#define HV_FAST_VINTR_SET_TARGET	0xae
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_vintr_get_cookie(unsigned long dev_handle,
+					    unsigned long dev_ino,
+					    unsigned long *cookie);
+extern unsigned long sun4v_vintr_set_cookie(unsigned long dev_handle,
+					    unsigned long dev_ino,
+					    unsigned long cookie);
+extern unsigned long sun4v_vintr_get_valid(unsigned long dev_handle,
+					   unsigned long dev_ino,
+					   unsigned long *valid);
+extern unsigned long sun4v_vintr_set_valid(unsigned long dev_handle,
+					   unsigned long dev_ino,
+					   unsigned long valid);
+extern unsigned long sun4v_vintr_get_state(unsigned long dev_handle,
+					   unsigned long dev_ino,
+					   unsigned long *state);
+extern unsigned long sun4v_vintr_set_state(unsigned long dev_handle,
+					   unsigned long dev_ino,
+					   unsigned long state);
+extern unsigned long sun4v_vintr_get_target(unsigned long dev_handle,
+					    unsigned long dev_ino,
+					    unsigned long *cpuid);
+extern unsigned long sun4v_vintr_set_target(unsigned long dev_handle,
+					    unsigned long dev_ino,
+					    unsigned long cpuid);
+#endif
+
+/* PCI IO services.
+ *
+ * See the terminology descriptions in the device interrupt services
+ * section above as those apply here too.  Here are terminology
+ * definitions specific to these PCI IO services:
+ *
+ *	tsbnum		TSB number.  Indentifies which io-tsb is used.
+ *			For this version of the specification, tsbnum
+ *			must be zero.
+ *
+ *	tsbindex	TSB index.  Identifies which entry in the TSB
+ *			is used.  The first entry is zero.
+ *
+ *	tsbid		A 64-bit aligned data structure which contains
+ *			a tsbnum and a tsbindex.  Bits 63:32 contain the
+ *			tsbnum and bits 31:00 contain the tsbindex.
+ *
+ *			Use the HV_PCI_TSBID() macro to construct such
+ * 			values.
+ *
+ *	io_attributes	IO attributes for IOMMU mappings.  One of more
+ *			of the attritbute bits are stores in a 64-bit
+ *			value.  The values are defined below.
+ *
+ *	r_addr		64-bit real address
+ *
+ *	pci_device	PCI device address.  A PCI device address identifies
+ *			a specific device on a specific PCI bus segment.
+ *			A PCI device address ia a 32-bit unsigned integer
+ *			with the following format:
+ *
+ *				00000000.bbbbbbbb.dddddfff.00000000
+ *
+ *			Use the HV_PCI_DEVICE_BUILD() macro to construct
+ *			such values.
+ *
+ *	pci_config_offset
+ *			PCI configureation space offset.  For conventional
+ *			PCI a value between 0 and 255.  For extended
+ *			configuration space, a value between 0 and 4095.
+ *
+ *			Note: For PCI configuration space accesses, the offset
+ *			      must be aligned to the access size.
+ *
+ *	error_flag	A return value which specifies if the action succeeded
+ *			or failed.  0 means no error, non-0 means some error
+ *			occurred while performing the service.
+ *
+ *	io_sync_direction
+ *			Direction definition for pci_dma_sync(), defined
+ *			below in HV_PCI_SYNC_*.
+ *
+ *	io_page_list	A list of io_page_addresses, an io_page_address is
+ *			a real address.
+ *
+ *	io_page_list_p	A pointer to an io_page_list.
+ *
+ *	"size based byte swap" - Some functions do size based byte swapping
+ *				 which allows sw to access pointers and
+ *				 counters in native form when the processor
+ *				 operates in a different endianness than the
+ *				 IO bus.  Size-based byte swapping converts a
+ *				 multi-byte field between big-endian and
+ *				 little-endian format.
+ */
+
+#define HV_PCI_MAP_ATTR_READ		0x01
+#define HV_PCI_MAP_ATTR_WRITE		0x02
+
+#define HV_PCI_DEVICE_BUILD(b,d,f)	\
+	((((b) & 0xff) << 16) | \
+	 (((d) & 0x1f) << 11) | \
+	 (((f) & 0x07) <<  8))
+
+#define HV_PCI_TSBID(__tsb_num, __tsb_index) \
+	((((u64)(__tsb_num)) << 32UL) | ((u64)(__tsb_index)))
+
+#define HV_PCI_SYNC_FOR_DEVICE		0x01
+#define HV_PCI_SYNC_FOR_CPU		0x02
+
+/* pci_iommu_map()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_IOMMU_MAP
+ * ARG0:	devhandle
+ * ARG1:	tsbid
+ * ARG2:	#ttes
+ * ARG3:	io_attributes
+ * ARG4:	io_page_list_p
+ * RET0:	status
+ * RET1:	#ttes mapped
+ * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex/io_attributes
+ *		EBADALIGN	Improperly aligned real address
+ *		ENORADDR	Invalid real address
+ *
+ * Create IOMMU mappings in the sun4v device defined by the given
+ * devhandle.  The mappings are created in the TSB defined by the
+ * tsbnum component of the given tsbid.  The first mapping is created
+ * in the TSB i ndex defined by the tsbindex component of the given tsbid.
+ * The call creates up to #ttes mappings, the first one at tsbnum, tsbindex,
+ * the second at tsbnum, tsbindex + 1, etc.
+ *
+ * All mappings are created with the attributes defined by the io_attributes
+ * argument.  The page mapping addresses are described in the io_page_list
+ * defined by the given io_page_list_p, which is a pointer to the io_page_list.
+ * The first entry in the io_page_list is the address for the first iotte, the
+ * 2nd for the 2nd iotte, and so on.
+ *
+ * Each io_page_address in the io_page_list must be appropriately aligned.
+ * #ttes must be greater than zero.  For this version of the spec, the tsbnum
+ * component of the given tsbid must be zero.
+ *
+ * Returns the actual number of mappings creates, which may be less than
+ * or equal to the argument #ttes.  If the function returns a value which
+ * is less than the #ttes, the caller may continus to call the function with
+ * an updated tsbid, #ttes, io_page_list_p arguments until all pages are
+ * mapped.
+ *
+ * Note: This function does not imply an iotte cache flush.  The guest must
+ *       demap an entry before re-mapping it.
+ */
+#define HV_FAST_PCI_IOMMU_MAP		0xb0
+
+/* pci_iommu_demap()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_IOMMU_DEMAP
+ * ARG0:	devhandle
+ * ARG1:	tsbid
+ * ARG2:	#ttes
+ * RET0:	status
+ * RET1:	#ttes demapped
+ * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex
+ *
+ * Demap and flush IOMMU mappings in the device defined by the given
+ * devhandle.  Demaps up to #ttes entries in the TSB defined by the tsbnum
+ * component of the given tsbid, starting at the TSB index defined by the
+ * tsbindex component of the given tsbid.
+ *
+ * For this version of the spec, the tsbnum of the given tsbid must be zero.
+ * #ttes must be greater than zero.
+ *
+ * Returns the actual number of ttes demapped, which may be less than or equal
+ * to the argument #ttes.  If #ttes demapped is less than #ttes, the caller
+ * may continue to call this function with updated tsbid and #ttes arguments
+ * until all pages are demapped.
+ *
+ * Note: Entries do not have to be mapped to be demapped.  A demap of an
+ *       unmapped page will flush the entry from the tte cache.
+ */
+#define HV_FAST_PCI_IOMMU_DEMAP		0xb1
+
+/* pci_iommu_getmap()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_IOMMU_GETMAP
+ * ARG0:	devhandle
+ * ARG1:	tsbid
+ * RET0:	status
+ * RET1:	io_attributes
+ * RET2:	real address
+ * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex
+ *		ENOMAP		Mapping is not valid, no translation exists
+ *
+ * Read and return the mapping in the device described by the given devhandle
+ * and tsbid.  If successful, the io_attributes shall be returned in RET1
+ * and the page address of the mapping shall be returned in RET2.
+ *
+ * For this version of the spec, the tsbnum component of the given tsbid
+ * must be zero.
+ */
+#define HV_FAST_PCI_IOMMU_GETMAP	0xb2
+
+/* pci_iommu_getbypass()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_IOMMU_GETBYPASS
+ * ARG0:	devhandle
+ * ARG1:	real address
+ * ARG2:	io_attributes
+ * RET0:	status
+ * RET1:	io_addr
+ * ERRORS:	EINVAL		Invalid devhandle/io_attributes
+ *		ENORADDR	Invalid real address
+ *		ENOTSUPPORTED	Function not supported in this implementation.
+ *
+ * Create a "special" mapping in the device described by the given devhandle,
+ * for the given real address and attributes.  Return the IO address in RET1
+ * if successful.
+ */
+#define HV_FAST_PCI_IOMMU_GETBYPASS	0xb3
+
+/* pci_config_get()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_CONFIG_GET
+ * ARG0:	devhandle
+ * ARG1:	pci_device
+ * ARG2:	pci_config_offset
+ * ARG3:	size
+ * RET0:	status
+ * RET1:	error_flag
+ * RET2:	data
+ * ERRORS:	EINVAL		Invalid devhandle/pci_device/offset/size
+ *		EBADALIGN	pci_config_offset not size aligned
+ *		ENOACCESS	Access to this offset is not permitted
+ *
+ * Read PCI configuration space for the adapter described by the given
+ * devhandle.  Read size (1, 2, or 4) bytes of data from the given
+ * pci_device, at pci_config_offset from the beginning of the device's
+ * configuration space.  If there was no error, RET1 is set to zero and
+ * RET2 is set to the data read.  Insignificant bits in RET2 are not
+ * guarenteed to have any specific value and therefore must be ignored.
+ *
+ * The data returned in RET2 is size based byte swapped.
+ *
+ * If an error occurs during the read, set RET1 to a non-zero value.  The
+ * given pci_config_offset must be 'size' aligned.
+ */
+#define HV_FAST_PCI_CONFIG_GET		0xb4
+
+/* pci_config_put()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_CONFIG_PUT
+ * ARG0:	devhandle
+ * ARG1:	pci_device
+ * ARG2:	pci_config_offset
+ * ARG3:	size
+ * ARG4:	data
+ * RET0:	status
+ * RET1:	error_flag
+ * ERRORS:	EINVAL		Invalid devhandle/pci_device/offset/size
+ *		EBADALIGN	pci_config_offset not size aligned
+ *		ENOACCESS	Access to this offset is not permitted
+ *
+ * Write PCI configuration space for the adapter described by the given
+ * devhandle.  Write size (1, 2, or 4) bytes of data in a single operation,
+ * at pci_config_offset from the beginning of the device's configuration
+ * space.  The data argument contains the data to be written to configuration
+ * space.  Prior to writing, the data is size based byte swapped.
+ *
+ * If an error occurs during the write access, do not generate an error
+ * report, do set RET1 to a non-zero value.  Otherwise RET1 is zero.
+ * The given pci_config_offset must be 'size' aligned.
+ *
+ * This function is permitted to read from offset zero in the configuration
+ * space described by the given pci_device if necessary to ensure that the
+ * write access to config space completes.
+ */
+#define HV_FAST_PCI_CONFIG_PUT		0xb5
+
+/* pci_peek()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_PEEK
+ * ARG0:	devhandle
+ * ARG1:	real address
+ * ARG2:	size
+ * RET0:	status
+ * RET1:	error_flag
+ * RET2:	data
+ * ERRORS:	EINVAL		Invalid devhandle or size
+ *		EBADALIGN	Improperly aligned real address
+ *		ENORADDR	Bad real address
+ *		ENOACCESS	Guest access prohibited
+ *
+ * Attempt to read the IO address given by the given devhandle, real address,
+ * and size.  Size must be 1, 2, 4, or 8.  The read is performed as a single
+ * access operation using the given size.  If an error occurs when reading
+ * from the given location, do not generate an error report, but return a
+ * non-zero value in RET1.  If the read was successful, return zero in RET1
+ * and return the actual data read in RET2.  The data returned is size based
+ * byte swapped.
+ *
+ * Non-significant bits in RET2 are not guarenteed to have any specific value
+ * and therefore must be ignored.  If RET1 is returned as non-zero, the data
+ * value is not guarenteed to have any specific value and should be ignored.
+ *
+ * The caller must have permission to read from the given devhandle, real
+ * address, which must be an IO address.  The argument real address must be a
+ * size aligned address.
+ *
+ * The hypervisor implementation of this function must block access to any
+ * IO address that the guest does not have explicit permission to access.
+ */
+#define HV_FAST_PCI_PEEK		0xb6
+
+/* pci_poke()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_POKE
+ * ARG0:	devhandle
+ * ARG1:	real address
+ * ARG2:	size
+ * ARG3:	data
+ * ARG4:	pci_device
+ * RET0:	status
+ * RET1:	error_flag
+ * ERRORS:	EINVAL		Invalid devhandle, size, or pci_device
+ *		EBADALIGN	Improperly aligned real address
+ *		ENORADDR	Bad real address
+ *		ENOACCESS	Guest access prohibited
+ *		ENOTSUPPORTED	Function is not supported by implementation
+ *
+ * Attempt to write data to the IO address given by the given devhandle,
+ * real address, and size.  Size must be 1, 2, 4, or 8.  The write is
+ * performed as a single access operation using the given size. Prior to
+ * writing the data is size based swapped.
+ *
+ * If an error occurs when writing to the given location, do not generate an
+ * error report, but return a non-zero value in RET1.  If the write was
+ * successful, return zero in RET1.
+ *
+ * pci_device describes the configuration address of the device being
+ * written to.  The implementation may safely read from offset 0 with
+ * the configuration space of the device described by devhandle and
+ * pci_device in order to guarantee that the write portion of the operation
+ * completes
+ *
+ * Any error that occurs due to the read shall be reported using the normal
+ * error reporting mechanisms .. the read error is not suppressed.
+ *
+ * The caller must have permission to write to the given devhandle, real
+ * address, which must be an IO address.  The argument real address must be a
+ * size aligned address.  The caller must have permission to read from
+ * the given devhandle, pci_device cofiguration space offset 0.
+ *
+ * The hypervisor implementation of this function must block access to any
+ * IO address that the guest does not have explicit permission to access.
+ */
+#define HV_FAST_PCI_POKE		0xb7
+
+/* pci_dma_sync()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_DMA_SYNC
+ * ARG0:	devhandle
+ * ARG1:	real address
+ * ARG2:	size
+ * ARG3:	io_sync_direction
+ * RET0:	status
+ * RET1:	#synced
+ * ERRORS:	EINVAL		Invalid devhandle or io_sync_direction
+ *		ENORADDR	Bad real address
+ *
+ * Synchronize a memory region described by the given real address and size,
+ * for the device defined by the given devhandle using the direction(s)
+ * defined by the given io_sync_direction.  The argument size is the size of
+ * the memory region in bytes.
+ *
+ * Return the actual number of bytes synchronized in the return value #synced,
+ * which may be less than or equal to the argument size.  If the return
+ * value #synced is less than size, the caller must continue to call this
+ * function with updated real address and size arguments until the entire
+ * memory region is synchronized.
+ */
+#define HV_FAST_PCI_DMA_SYNC		0xb8
+
+/* PCI MSI services.  */
+
+#define HV_MSITYPE_MSI32		0x00
+#define HV_MSITYPE_MSI64		0x01
+
+#define HV_MSIQSTATE_IDLE		0x00
+#define HV_MSIQSTATE_ERROR		0x01
+
+#define HV_MSIQ_INVALID			0x00
+#define HV_MSIQ_VALID			0x01
+
+#define HV_MSISTATE_IDLE		0x00
+#define HV_MSISTATE_DELIVERED		0x01
+
+#define HV_MSIVALID_INVALID		0x00
+#define HV_MSIVALID_VALID		0x01
+
+#define HV_PCIE_MSGTYPE_PME_MSG		0x18
+#define HV_PCIE_MSGTYPE_PME_ACK_MSG	0x1b
+#define HV_PCIE_MSGTYPE_CORR_MSG	0x30
+#define HV_PCIE_MSGTYPE_NONFATAL_MSG	0x31
+#define HV_PCIE_MSGTYPE_FATAL_MSG	0x33
+
+#define HV_MSG_INVALID			0x00
+#define HV_MSG_VALID			0x01
+
+/* pci_msiq_conf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_CONF
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * ARG2:	real address
+ * ARG3:	number of entries
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle, msiqid or nentries
+ *		EBADALIGN	Improperly aligned real address
+ *		ENORADDR	Bad real address
+ *
+ * Configure the MSI queue given by the devhandle and msiqid arguments,
+ * and to be placed at the given real address and be of the given
+ * number of entries.  The real address must be aligned exactly to match
+ * the queue size.  Each queue entry is 64-bytes long, so f.e. a 32 entry
+ * queue must be aligned on a 2048 byte real address boundary.  The MSI-EQ
+ * Head and Tail are initialized so that the MSI-EQ is 'empty'.
+ *
+ * Implementation Note: Certain implementations have fixed sized queues.  In
+ *                      that case, number of entries must contain the correct
+ *                      value.
+ */
+#define HV_FAST_PCI_MSIQ_CONF		0xc0
+
+/* pci_msiq_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_INFO
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * RET0:	status
+ * RET1:	real address
+ * RET2:	number of entries
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid
+ *
+ * Return the configuration information for the MSI queue described
+ * by the given devhandle and msiqid.  The base address of the queue
+ * is returned in ARG1 and the number of entries is returned in ARG2.
+ * If the queue is unconfigured, the real address is undefined and the
+ * number of entries will be returned as zero.
+ */
+#define HV_FAST_PCI_MSIQ_INFO		0xc1
+
+/* pci_msiq_getvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_GETVALID
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * RET0:	status
+ * RET1:	msiqvalid	(HV_MSIQ_VALID or HV_MSIQ_INVALID)
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid
+ *
+ * Get the valid state of the MSI-EQ described by the given devhandle and
+ * msiqid.
+ */
+#define HV_FAST_PCI_MSIQ_GETVALID	0xc2
+
+/* pci_msiq_setvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_SETVALID
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * ARG2:	msiqvalid	(HV_MSIQ_VALID or HV_MSIQ_INVALID)
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqvalid
+ *				value or MSI EQ is uninitialized
+ *
+ * Set the valid state of the MSI-EQ described by the given devhandle and
+ * msiqid to the given msiqvalid.
+ */
+#define HV_FAST_PCI_MSIQ_SETVALID	0xc3
+
+/* pci_msiq_getstate()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_GETSTATE
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * RET0:	status
+ * RET1:	msiqstate	(HV_MSIQSTATE_IDLE or HV_MSIQSTATE_ERROR)
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid
+ *
+ * Get the state of the MSI-EQ described by the given devhandle and
+ * msiqid.
+ */
+#define HV_FAST_PCI_MSIQ_GETSTATE	0xc4
+
+/* pci_msiq_getvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_GETVALID
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * ARG2:	msiqstate	(HV_MSIQSTATE_IDLE or HV_MSIQSTATE_ERROR)
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqstate
+ *				value or MSI EQ is uninitialized
+ *
+ * Set the state of the MSI-EQ described by the given devhandle and
+ * msiqid to the given msiqvalid.
+ */
+#define HV_FAST_PCI_MSIQ_SETSTATE	0xc5
+
+/* pci_msiq_gethead()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_GETHEAD
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * RET0:	status
+ * RET1:	msiqhead
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid
+ *
+ * Get the current MSI EQ queue head for the MSI-EQ described by the
+ * given devhandle and msiqid.
+ */
+#define HV_FAST_PCI_MSIQ_GETHEAD	0xc6
+
+/* pci_msiq_sethead()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_SETHEAD
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * ARG2:	msiqhead
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqhead,
+ *				or MSI EQ is uninitialized
+ *
+ * Set the current MSI EQ queue head for the MSI-EQ described by the
+ * given devhandle and msiqid.
+ */
+#define HV_FAST_PCI_MSIQ_SETHEAD	0xc7
+
+/* pci_msiq_gettail()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSIQ_GETTAIL
+ * ARG0:	devhandle
+ * ARG1:	msiqid
+ * RET0:	status
+ * RET1:	msiqtail
+ * ERRORS:	EINVAL		Invalid devhandle or msiqid
+ *
+ * Get the current MSI EQ queue tail for the MSI-EQ described by the
+ * given devhandle and msiqid.
+ */
+#define HV_FAST_PCI_MSIQ_GETTAIL	0xc8
+
+/* pci_msi_getvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_GETVALID
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * RET0:	status
+ * RET1:	msivalidstate
+ * ERRORS:	EINVAL		Invalid devhandle or msinum
+ *
+ * Get the current valid/enabled state for the MSI defined by the
+ * given devhandle and msinum.
+ */
+#define HV_FAST_PCI_MSI_GETVALID	0xc9
+
+/* pci_msi_setvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_SETVALID
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * ARG2:	msivalidstate
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msinum or msivalidstate
+ *
+ * Set the current valid/enabled state for the MSI defined by the
+ * given devhandle and msinum.
+ */
+#define HV_FAST_PCI_MSI_SETVALID	0xca
+
+/* pci_msi_getmsiq()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_GETMSIQ
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * RET0:	status
+ * RET1:	msiqid
+ * ERRORS:	EINVAL		Invalid devhandle or msinum or MSI is unbound
+ *
+ * Get the MSI EQ that the MSI defined by the given devhandle and
+ * msinum is bound to.
+ */
+#define HV_FAST_PCI_MSI_GETMSIQ		0xcb
+
+/* pci_msi_setmsiq()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_SETMSIQ
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * ARG2:	msitype
+ * ARG3:	msiqid
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msinum or msiqid
+ *
+ * Set the MSI EQ that the MSI defined by the given devhandle and
+ * msinum is bound to.
+ */
+#define HV_FAST_PCI_MSI_SETMSIQ		0xcc
+
+/* pci_msi_getstate()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_GETSTATE
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * RET0:	status
+ * RET1:	msistate
+ * ERRORS:	EINVAL		Invalid devhandle or msinum
+ *
+ * Get the state of the MSI defined by the given devhandle and msinum.
+ * If not initialized, return HV_MSISTATE_IDLE.
+ */
+#define HV_FAST_PCI_MSI_GETSTATE	0xcd
+
+/* pci_msi_setstate()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSI_SETSTATE
+ * ARG0:	devhandle
+ * ARG1:	msinum
+ * ARG2:	msistate
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msinum or msistate
+ *
+ * Set the state of the MSI defined by the given devhandle and msinum.
+ */
+#define HV_FAST_PCI_MSI_SETSTATE	0xce
+
+/* pci_msg_getmsiq()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSG_GETMSIQ
+ * ARG0:	devhandle
+ * ARG1:	msgtype
+ * RET0:	status
+ * RET1:	msiqid
+ * ERRORS:	EINVAL		Invalid devhandle or msgtype
+ *
+ * Get the MSI EQ of the MSG defined by the given devhandle and msgtype.
+ */
+#define HV_FAST_PCI_MSG_GETMSIQ		0xd0
+
+/* pci_msg_setmsiq()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSG_SETMSIQ
+ * ARG0:	devhandle
+ * ARG1:	msgtype
+ * ARG2:	msiqid
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle, msgtype, or msiqid
+ *
+ * Set the MSI EQ of the MSG defined by the given devhandle and msgtype.
+ */
+#define HV_FAST_PCI_MSG_SETMSIQ		0xd1
+
+/* pci_msg_getvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSG_GETVALID
+ * ARG0:	devhandle
+ * ARG1:	msgtype
+ * RET0:	status
+ * RET1:	msgvalidstate
+ * ERRORS:	EINVAL		Invalid devhandle or msgtype
+ *
+ * Get the valid/enabled state of the MSG defined by the given
+ * devhandle and msgtype.
+ */
+#define HV_FAST_PCI_MSG_GETVALID	0xd2
+
+/* pci_msg_setvalid()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_PCI_MSG_SETVALID
+ * ARG0:	devhandle
+ * ARG1:	msgtype
+ * ARG2:	msgvalidstate
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid devhandle or msgtype or msgvalidstate
+ *
+ * Set the valid/enabled state of the MSG defined by the given
+ * devhandle and msgtype.
+ */
+#define HV_FAST_PCI_MSG_SETVALID	0xd3
+
+/* Logical Domain Channel services.  */
+
+#define LDC_CHANNEL_DOWN		0
+#define LDC_CHANNEL_UP			1
+#define LDC_CHANNEL_RESETTING		2
+
+/* ldc_tx_qconf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_TX_QCONF
+ * ARG0:	channel ID
+ * ARG1:	real address base of queue
+ * ARG2:	num entries in queue
+ * RET0:	status
+ *
+ * Configure transmit queue for the LDC endpoint specified by the
+ * given channel ID, to be placed at the given real address, and
+ * be of the given num entries.  Num entries must be a power of two.
+ * The real address base of the queue must be aligned on the queue
+ * size.  Each queue entry is 64-bytes, so for example, a 32 entry
+ * queue must be aligned on a 2048 byte real address boundary.
+ *
+ * Upon configuration of a valid transmit queue the head and tail
+ * pointers are set to a hypervisor specific identical value indicating
+ * that the queue initially is empty.
+ *
+ * The endpoint's transmit queue is un-configured if num entries is zero.
+ *
+ * The maximum number of entries for each queue for a specific cpu may be
+ * determined from the machine description.  A transmit queue may be
+ * specified even in the event that the LDC is down (peer endpoint has no
+ * receive queue specified).  Transmission will begin as soon as the peer
+ * endpoint defines a receive queue.
+ *
+ * It is recommended that a guest wait for a transmit queue to empty prior
+ * to reconfiguring it, or un-configuring it.  Re or un-configuring of a
+ * non-empty transmit queue behaves exactly as defined above, however it
+ * is undefined as to how many of the pending entries in the original queue
+ * will be delivered prior to the re-configuration taking effect.
+ * Furthermore, as the queue configuration causes a reset of the head and
+ * tail pointers there is no way for a guest to determine how many entries
+ * have been sent after the configuration operation.
+ */
+#define HV_FAST_LDC_TX_QCONF		0xe0
+
+/* ldc_tx_qinfo()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_TX_QINFO
+ * ARG0:	channel ID
+ * RET0:	status
+ * RET1:	real address base of queue
+ * RET2:	num entries in queue
+ *
+ * Return the configuration info for the transmit queue of LDC endpoint
+ * defined by the given channel ID.  The real address is the currently
+ * defined real address base of the defined queue, and num entries is the
+ * size of the queue in terms of number of entries.
+ *
+ * If the specified channel ID is a valid endpoint number, but no transmit
+ * queue has been defined this service will return success, but with num
+ * entries set to zero and the real address will have an undefined value.
+ */
+#define HV_FAST_LDC_TX_QINFO		0xe1
+
+/* ldc_tx_get_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_TX_GET_STATE
+ * ARG0:	channel ID
+ * RET0:	status
+ * RET1:	head offset
+ * RET2:	tail offset
+ * RET3:	channel state
+ *
+ * Return the transmit state, and the head and tail queue pointers, for
+ * the transmit queue of the LDC endpoint defined by the given channel ID.
+ * The head and tail values are the byte offset of the head and tail
+ * positions of the transmit queue for the specified endpoint.
+ */
+#define HV_FAST_LDC_TX_GET_STATE	0xe2
+
+/* ldc_tx_set_qtail()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_TX_SET_QTAIL
+ * ARG0:	channel ID
+ * ARG1:	tail offset
+ * RET0:	status
+ *
+ * Update the tail pointer for the transmit queue associated with the LDC
+ * endpoint defined by the given channel ID.  The tail offset specified
+ * must be aligned on a 64 byte boundary, and calculated so as to increase
+ * the number of pending entries on the transmit queue.  Any attempt to
+ * decrease the number of pending transmit queue entires is considered
+ * an invalid tail offset and will result in an EINVAL error.
+ *
+ * Since the tail of the transmit queue may not be moved backwards, the
+ * transmit queue may be flushed by configuring a new transmit queue,
+ * whereupon the hypervisor will configure the initial transmit head and
+ * tail pointers to be equal.
+ */
+#define HV_FAST_LDC_TX_SET_QTAIL	0xe3
+
+/* ldc_rx_qconf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_RX_QCONF
+ * ARG0:	channel ID
+ * ARG1:	real address base of queue
+ * ARG2:	num entries in queue
+ * RET0:	status
+ *
+ * Configure receive queue for the LDC endpoint specified by the
+ * given channel ID, to be placed at the given real address, and
+ * be of the given num entries.  Num entries must be a power of two.
+ * The real address base of the queue must be aligned on the queue
+ * size.  Each queue entry is 64-bytes, so for example, a 32 entry
+ * queue must be aligned on a 2048 byte real address boundary.
+ *
+ * The endpoint's transmit queue is un-configured if num entries is zero.
+ *
+ * If a valid receive queue is specified for a local endpoint the LDC is
+ * in the up state for the purpose of transmission to this endpoint.
+ *
+ * The maximum number of entries for each queue for a specific cpu may be
+ * determined from the machine description.
+ *
+ * As receive queue configuration causes a reset of the queue's head and
+ * tail pointers there is no way for a gues to determine how many entries
+ * have been received between a preceeding ldc_get_rx_state() API call
+ * and the completion of the configuration operation.  It should be noted
+ * that datagram delivery is not guarenteed via domain channels anyway,
+ * and therefore any higher protocol should be resilient to datagram
+ * loss if necessary.  However, to overcome this specific race potential
+ * it is recommended, for example, that a higher level protocol be employed
+ * to ensure either retransmission, or ensure that no datagrams are pending
+ * on the peer endpoint's transmit queue prior to the configuration process.
+ */
+#define HV_FAST_LDC_RX_QCONF		0xe4
+
+/* ldc_rx_qinfo()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_RX_QINFO
+ * ARG0:	channel ID
+ * RET0:	status
+ * RET1:	real address base of queue
+ * RET2:	num entries in queue
+ *
+ * Return the configuration info for the receive queue of LDC endpoint
+ * defined by the given channel ID.  The real address is the currently
+ * defined real address base of the defined queue, and num entries is the
+ * size of the queue in terms of number of entries.
+ *
+ * If the specified channel ID is a valid endpoint number, but no receive
+ * queue has been defined this service will return success, but with num
+ * entries set to zero and the real address will have an undefined value.
+ */
+#define HV_FAST_LDC_RX_QINFO		0xe5
+
+/* ldc_rx_get_state()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_RX_GET_STATE
+ * ARG0:	channel ID
+ * RET0:	status
+ * RET1:	head offset
+ * RET2:	tail offset
+ * RET3:	channel state
+ *
+ * Return the receive state, and the head and tail queue pointers, for
+ * the receive queue of the LDC endpoint defined by the given channel ID.
+ * The head and tail values are the byte offset of the head and tail
+ * positions of the receive queue for the specified endpoint.
+ */
+#define HV_FAST_LDC_RX_GET_STATE	0xe6
+
+/* ldc_rx_set_qhead()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_RX_SET_QHEAD
+ * ARG0:	channel ID
+ * ARG1:	head offset
+ * RET0:	status
+ *
+ * Update the head pointer for the receive queue associated with the LDC
+ * endpoint defined by the given channel ID.  The head offset specified
+ * must be aligned on a 64 byte boundary, and calculated so as to decrease
+ * the number of pending entries on the receive queue.  Any attempt to
+ * increase the number of pending receive queue entires is considered
+ * an invalid head offset and will result in an EINVAL error.
+ *
+ * The receive queue may be flushed by setting the head offset equal
+ * to the current tail offset.
+ */
+#define HV_FAST_LDC_RX_SET_QHEAD	0xe7
+
+/* LDC Map Table Entry.  Each slot is defined by a translation table
+ * entry, as specified by the LDC_MTE_* bits below, and a 64-bit
+ * hypervisor invalidation cookie.
+ */
+#define LDC_MTE_PADDR	0x0fffffffffffe000 /* pa[55:13]          */
+#define LDC_MTE_COPY_W	0x0000000000000400 /* copy write access  */
+#define LDC_MTE_COPY_R	0x0000000000000200 /* copy read access   */
+#define LDC_MTE_IOMMU_W	0x0000000000000100 /* IOMMU write access */
+#define LDC_MTE_IOMMU_R	0x0000000000000080 /* IOMMU read access  */
+#define LDC_MTE_EXEC	0x0000000000000040 /* execute            */
+#define LDC_MTE_WRITE	0x0000000000000020 /* read               */
+#define LDC_MTE_READ	0x0000000000000010 /* write              */
+#define LDC_MTE_SZALL	0x000000000000000f /* page size bits     */
+#define LDC_MTE_SZ16GB	0x0000000000000007 /* 16GB page          */
+#define LDC_MTE_SZ2GB	0x0000000000000006 /* 2GB page           */
+#define LDC_MTE_SZ256MB	0x0000000000000005 /* 256MB page         */
+#define LDC_MTE_SZ32MB	0x0000000000000004 /* 32MB page          */
+#define LDC_MTE_SZ4MB	0x0000000000000003 /* 4MB page           */
+#define LDC_MTE_SZ512K	0x0000000000000002 /* 512K page          */
+#define LDC_MTE_SZ64K	0x0000000000000001 /* 64K page           */
+#define LDC_MTE_SZ8K	0x0000000000000000 /* 8K page            */
+
+#ifndef __ASSEMBLY__
+struct ldc_mtable_entry {
+	unsigned long	mte;
+	unsigned long	cookie;
+};
+#endif
+
+/* ldc_set_map_table()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_SET_MAP_TABLE
+ * ARG0:	channel ID
+ * ARG1:	table real address
+ * ARG2:	num entries
+ * RET0:	status
+ *
+ * Register the MTE table at the given table real address, with the
+ * specified num entries, for the LDC indicated by the given channel
+ * ID.
+ */
+#define HV_FAST_LDC_SET_MAP_TABLE	0xea
+
+/* ldc_get_map_table()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_GET_MAP_TABLE
+ * ARG0:	channel ID
+ * RET0:	status
+ * RET1:	table real address
+ * RET2:	num entries
+ *
+ * Return the configuration of the current mapping table registered
+ * for the given channel ID.
+ */
+#define HV_FAST_LDC_GET_MAP_TABLE	0xeb
+
+#define LDC_COPY_IN	0
+#define LDC_COPY_OUT	1
+
+/* ldc_copy()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_COPY
+ * ARG0:	channel ID
+ * ARG1:	LDC_COPY_* direction code
+ * ARG2:	target real address
+ * ARG3:	local real address
+ * ARG4:	length in bytes
+ * RET0:	status
+ * RET1:	actual length in bytes
+ */
+#define HV_FAST_LDC_COPY		0xec
+
+#define LDC_MEM_READ	1
+#define LDC_MEM_WRITE	2
+#define LDC_MEM_EXEC	4
+
+/* ldc_mapin()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_MAPIN
+ * ARG0:	channel ID
+ * ARG1:	cookie
+ * RET0:	status
+ * RET1:	real address
+ * RET2:	LDC_MEM_* permissions
+ */
+#define HV_FAST_LDC_MAPIN		0xed
+
+/* ldc_unmap()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_UNMAP
+ * ARG0:	real address
+ * RET0:	status
+ */
+#define HV_FAST_LDC_UNMAP		0xee
+
+/* ldc_revoke()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_LDC_REVOKE
+ * ARG0:	channel ID
+ * ARG1:	cookie
+ * ARG2:	ldc_mtable_entry cookie
+ * RET0:	status
+ */
+#define HV_FAST_LDC_REVOKE		0xef
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_ldc_tx_qconf(unsigned long channel,
+					unsigned long ra,
+					unsigned long num_entries);
+extern unsigned long sun4v_ldc_tx_qinfo(unsigned long channel,
+					unsigned long *ra,
+					unsigned long *num_entries);
+extern unsigned long sun4v_ldc_tx_get_state(unsigned long channel,
+					    unsigned long *head_off,
+					    unsigned long *tail_off,
+					    unsigned long *chan_state);
+extern unsigned long sun4v_ldc_tx_set_qtail(unsigned long channel,
+					    unsigned long tail_off);
+extern unsigned long sun4v_ldc_rx_qconf(unsigned long channel,
+					unsigned long ra,
+					unsigned long num_entries);
+extern unsigned long sun4v_ldc_rx_qinfo(unsigned long channel,
+					unsigned long *ra,
+					unsigned long *num_entries);
+extern unsigned long sun4v_ldc_rx_get_state(unsigned long channel,
+					    unsigned long *head_off,
+					    unsigned long *tail_off,
+					    unsigned long *chan_state);
+extern unsigned long sun4v_ldc_rx_set_qhead(unsigned long channel,
+					    unsigned long head_off);
+extern unsigned long sun4v_ldc_set_map_table(unsigned long channel,
+					     unsigned long ra,
+					     unsigned long num_entries);
+extern unsigned long sun4v_ldc_get_map_table(unsigned long channel,
+					     unsigned long *ra,
+					     unsigned long *num_entries);
+extern unsigned long sun4v_ldc_copy(unsigned long channel,
+				    unsigned long dir_code,
+				    unsigned long tgt_raddr,
+				    unsigned long lcl_raddr,
+				    unsigned long len,
+				    unsigned long *actual_len);
+extern unsigned long sun4v_ldc_mapin(unsigned long channel,
+				     unsigned long cookie,
+				     unsigned long *ra,
+				     unsigned long *perm);
+extern unsigned long sun4v_ldc_unmap(unsigned long ra);
+extern unsigned long sun4v_ldc_revoke(unsigned long channel,
+				      unsigned long cookie,
+				      unsigned long mte_cookie);
+#endif
+
+/* Performance counter services.  */
+
+#define HV_PERF_JBUS_PERF_CTRL_REG	0x00
+#define HV_PERF_JBUS_PERF_CNT_REG	0x01
+#define HV_PERF_DRAM_PERF_CTRL_REG_0	0x02
+#define HV_PERF_DRAM_PERF_CNT_REG_0	0x03
+#define HV_PERF_DRAM_PERF_CTRL_REG_1	0x04
+#define HV_PERF_DRAM_PERF_CNT_REG_1	0x05
+#define HV_PERF_DRAM_PERF_CTRL_REG_2	0x06
+#define HV_PERF_DRAM_PERF_CNT_REG_2	0x07
+#define HV_PERF_DRAM_PERF_CTRL_REG_3	0x08
+#define HV_PERF_DRAM_PERF_CNT_REG_3	0x09
+
+/* get_perfreg()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_GET_PERFREG
+ * ARG0:	performance reg number
+ * RET0:	status
+ * RET1:	performance reg value
+ * ERRORS:	EINVAL		Invalid performance register number
+ *		ENOACCESS	No access allowed to performance counters
+ *
+ * Read the value of the given DRAM/JBUS performance counter/control register.
+ */
+#define HV_FAST_GET_PERFREG		0x100
+
+/* set_perfreg()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_SET_PERFREG
+ * ARG0:	performance reg number
+ * ARG1:	performance reg value
+ * RET0:	status
+ * ERRORS:	EINVAL		Invalid performance register number
+ *		ENOACCESS	No access allowed to performance counters
+ *
+ * Write the given performance reg value to the given DRAM/JBUS
+ * performance counter/control register.
+ */
+#define HV_FAST_SET_PERFREG		0x101
+
+/* MMU statistics services.
+ *
+ * The hypervisor maintains MMU statistics and privileged code provides
+ * a buffer where these statistics can be collected.  It is continually
+ * updated once configured.  The layout is as follows:
+ */
+#ifndef __ASSEMBLY__
+struct hv_mmu_statistics {
+	unsigned long immu_tsb_hits_ctx0_8k_tte;
+	unsigned long immu_tsb_ticks_ctx0_8k_tte;
+	unsigned long immu_tsb_hits_ctx0_64k_tte;
+	unsigned long immu_tsb_ticks_ctx0_64k_tte;
+	unsigned long __reserved1[2];
+	unsigned long immu_tsb_hits_ctx0_4mb_tte;
+	unsigned long immu_tsb_ticks_ctx0_4mb_tte;
+	unsigned long __reserved2[2];
+	unsigned long immu_tsb_hits_ctx0_256mb_tte;
+	unsigned long immu_tsb_ticks_ctx0_256mb_tte;
+	unsigned long __reserved3[4];
+	unsigned long immu_tsb_hits_ctxnon0_8k_tte;
+	unsigned long immu_tsb_ticks_ctxnon0_8k_tte;
+	unsigned long immu_tsb_hits_ctxnon0_64k_tte;
+	unsigned long immu_tsb_ticks_ctxnon0_64k_tte;
+	unsigned long __reserved4[2];
+	unsigned long immu_tsb_hits_ctxnon0_4mb_tte;
+	unsigned long immu_tsb_ticks_ctxnon0_4mb_tte;
+	unsigned long __reserved5[2];
+	unsigned long immu_tsb_hits_ctxnon0_256mb_tte;
+	unsigned long immu_tsb_ticks_ctxnon0_256mb_tte;
+	unsigned long __reserved6[4];
+	unsigned long dmmu_tsb_hits_ctx0_8k_tte;
+	unsigned long dmmu_tsb_ticks_ctx0_8k_tte;
+	unsigned long dmmu_tsb_hits_ctx0_64k_tte;
+	unsigned long dmmu_tsb_ticks_ctx0_64k_tte;
+	unsigned long __reserved7[2];
+	unsigned long dmmu_tsb_hits_ctx0_4mb_tte;
+	unsigned long dmmu_tsb_ticks_ctx0_4mb_tte;
+	unsigned long __reserved8[2];
+	unsigned long dmmu_tsb_hits_ctx0_256mb_tte;
+	unsigned long dmmu_tsb_ticks_ctx0_256mb_tte;
+	unsigned long __reserved9[4];
+	unsigned long dmmu_tsb_hits_ctxnon0_8k_tte;
+	unsigned long dmmu_tsb_ticks_ctxnon0_8k_tte;
+	unsigned long dmmu_tsb_hits_ctxnon0_64k_tte;
+	unsigned long dmmu_tsb_ticks_ctxnon0_64k_tte;
+	unsigned long __reserved10[2];
+	unsigned long dmmu_tsb_hits_ctxnon0_4mb_tte;
+	unsigned long dmmu_tsb_ticks_ctxnon0_4mb_tte;
+	unsigned long __reserved11[2];
+	unsigned long dmmu_tsb_hits_ctxnon0_256mb_tte;
+	unsigned long dmmu_tsb_ticks_ctxnon0_256mb_tte;
+	unsigned long __reserved12[4];
+};
+#endif
+
+/* mmustat_conf()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMUSTAT_CONF
+ * ARG0:	real address
+ * RET0:	status
+ * RET1:	real address
+ * ERRORS:	ENORADDR	Invalid real address
+ *		EBADALIGN	Real address not aligned on 64-byte boundary
+ *		EBADTRAP	API not supported on this processor
+ *
+ * Enable MMU statistic gathering using the buffer at the given real
+ * address on the current virtual CPU.  The new buffer real address
+ * is given in ARG1, and the previously specified buffer real address
+ * is returned in RET1, or is returned as zero for the first invocation.
+ *
+ * If the passed in real address argument is zero, this will disable
+ * MMU statistic collection on the current virtual CPU.  If an error is
+ * returned then no statistics are collected.
+ *
+ * The buffer contents should be initialized to all zeros before being
+ * given to the hypervisor or else the statistics will be meaningless.
+ */
+#define HV_FAST_MMUSTAT_CONF		0x102
+
+/* mmustat_info()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_MMUSTAT_INFO
+ * RET0:	status
+ * RET1:	real address
+ * ERRORS:	EBADTRAP	API not supported on this processor
+ *
+ * Return the current state and real address of the currently configured
+ * MMU statistics buffer on the current virtual CPU.
+ */
+#define HV_FAST_MMUSTAT_INFO		0x103
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_mmustat_conf(unsigned long ra, unsigned long *orig_ra);
+extern unsigned long sun4v_mmustat_info(unsigned long *ra);
+#endif
+
+/* NCS crypto services  */
+
+/* ncs_request() sub-function numbers */
+#define HV_NCS_QCONF			0x01
+#define HV_NCS_QTAIL_UPDATE		0x02
+
+#ifndef __ASSEMBLY__
+struct hv_ncs_queue_entry {
+	/* MAU Control Register */
+	unsigned long	mau_control;
+#define MAU_CONTROL_INV_PARITY	0x0000000000002000
+#define MAU_CONTROL_STRAND	0x0000000000001800
+#define MAU_CONTROL_BUSY	0x0000000000000400
+#define MAU_CONTROL_INT		0x0000000000000200
+#define MAU_CONTROL_OP		0x00000000000001c0
+#define MAU_CONTROL_OP_SHIFT	6
+#define MAU_OP_LOAD_MA_MEMORY	0x0
+#define MAU_OP_STORE_MA_MEMORY	0x1
+#define MAU_OP_MODULAR_MULT	0x2
+#define MAU_OP_MODULAR_REDUCE	0x3
+#define MAU_OP_MODULAR_EXP_LOOP	0x4
+#define MAU_CONTROL_LEN		0x000000000000003f
+#define MAU_CONTROL_LEN_SHIFT	0
+
+	/* Real address of bytes to load or store bytes
+	 * into/out-of the MAU.
+	 */
+	unsigned long	mau_mpa;
+
+	/* Modular Arithmetic MA Offset Register.  */
+	unsigned long	mau_ma;
+
+	/* Modular Arithmetic N Prime Register.  */
+	unsigned long	mau_np;
+};
+
+struct hv_ncs_qconf_arg {
+	unsigned long	mid;      /* MAU ID, 1 per core on Niagara */
+	unsigned long	base;     /* Real address base of queue */
+	unsigned long	end;	  /* Real address end of queue */
+	unsigned long	num_ents; /* Number of entries in queue */
+};
+
+struct hv_ncs_qtail_update_arg {
+	unsigned long	mid;      /* MAU ID, 1 per core on Niagara */
+	unsigned long	tail;     /* New tail index to use */
+	unsigned long	syncflag; /* only SYNCFLAG_SYNC is implemented */
+#define HV_NCS_SYNCFLAG_SYNC	0x00
+#define HV_NCS_SYNCFLAG_ASYNC	0x01
+};
+#endif
+
+/* ncs_request()
+ * TRAP:	HV_FAST_TRAP
+ * FUNCTION:	HV_FAST_NCS_REQUEST
+ * ARG0:	NCS sub-function
+ * ARG1:	sub-function argument real address
+ * ARG2:	size in bytes of sub-function argument
+ * RET0:	status
+ *
+ * The MAU chip of the Niagara processor is not directly accessible
+ * to privileged code, instead it is programmed indirectly via this
+ * hypervisor API.
+ *
+ * The interfaces defines a queue of MAU operations to perform.
+ * Privileged code registers a queue with the hypervisor by invoking
+ * this HVAPI with the HV_NCS_QCONF sub-function, which defines the
+ * base, end, and number of entries of the queue.  Each queue entry
+ * contains a MAU register struct block.
+ *
+ * The privileged code then proceeds to add entries to the queue and
+ * then invoke the HV_NCS_QTAIL_UPDATE sub-function.  Since only
+ * synchronous operations are supported by the current hypervisor,
+ * HV_NCS_QTAIL_UPDATE will run all the pending queue entries to
+ * completion and return HV_EOK, or return an error code.
+ *
+ * The real address of the sub-function argument must be aligned on at
+ * least an 8-byte boundary.
+ *
+ * The tail argument of HV_NCS_QTAIL_UPDATE is an index, not a byte
+ * offset, into the queue and must be less than or equal the 'num_ents'
+ * argument given in the HV_NCS_QCONF call.
+ */
+#define HV_FAST_NCS_REQUEST		0x110
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_ncs_request(unsigned long request,
+				       unsigned long arg_ra,
+				       unsigned long arg_size);
+#endif
+
+#define HV_FAST_FIRE_GET_PERFREG	0x120
+#define HV_FAST_FIRE_SET_PERFREG	0x121
+
+/* Function numbers for HV_CORE_TRAP.  */
+#define HV_CORE_SET_VER			0x00
+#define HV_CORE_PUTCHAR			0x01
+#define HV_CORE_EXIT			0x02
+#define HV_CORE_GET_VER			0x03
+
+/* Hypervisor API groups for use with HV_CORE_SET_VER and
+ * HV_CORE_GET_VER.
+ */
+#define HV_GRP_SUN4V			0x0000
+#define HV_GRP_CORE			0x0001
+#define HV_GRP_INTR			0x0002
+#define HV_GRP_SOFT_STATE		0x0003
+#define HV_GRP_PCI			0x0100
+#define HV_GRP_LDOM			0x0101
+#define HV_GRP_SVC_CHAN			0x0102
+#define HV_GRP_NCS			0x0103
+#define HV_GRP_RNG			0x0104
+#define HV_GRP_NIAG_PERF		0x0200
+#define HV_GRP_FIRE_PERF		0x0201
+#define HV_GRP_N2_CPU			0x0202
+#define HV_GRP_NIU			0x0204
+#define HV_GRP_VF_CPU			0x0205
+#define HV_GRP_DIAG			0x0300
+
+#ifndef __ASSEMBLY__
+extern unsigned long sun4v_get_version(unsigned long group,
+				       unsigned long *major,
+				       unsigned long *minor);
+extern unsigned long sun4v_set_version(unsigned long group,
+				       unsigned long major,
+				       unsigned long minor,
+				       unsigned long *actual_minor);
+
+extern int sun4v_hvapi_register(unsigned long group, unsigned long major,
+				unsigned long *minor);
+extern void sun4v_hvapi_unregister(unsigned long group);
+extern int sun4v_hvapi_get(unsigned long group,
+			   unsigned long *major,
+			   unsigned long *minor);
+extern void sun4v_hvapi_init(void);
+#endif
+
+#endif /* !(_SPARC64_HYPERVISOR_H) */
diff --git a/include/asm-sparc/ide.h b/include/asm-sparc/ide.h
index afd1736ed48..a3c7f5ff270 100644
--- a/include/asm-sparc/ide.h
+++ b/include/asm-sparc/ide.h
@@ -1,95 +1,8 @@
-/* ide.h: SPARC PCI specific IDE glue.
- *
- * Copyright (C) 1997  David S. Miller (davem@davemloft.net)
- * Copyright (C) 1998  Eddie C. Dost   (ecd@skynet.be)
- * Adaptation from sparc64 version to sparc by Pete Zaitcev.
- */
-
-#ifndef _SPARC_IDE_H
-#define _SPARC_IDE_H
-
-#ifdef __KERNEL__
-
-#include <asm/pgtable.h>
-#include <asm/io.h>
-#include <asm/psr.h>
-
-#undef  MAX_HWIFS
-#define MAX_HWIFS	2
-
-#define __ide_insl(data_reg, buffer, wcount) \
-	__ide_insw(data_reg, buffer, (wcount)<<1)
-#define __ide_outsl(data_reg, buffer, wcount) \
-	__ide_outsw(data_reg, buffer, (wcount)<<1)
-
-/* On sparc, I/O ports and MMIO registers are accessed identically.  */
-#define __ide_mm_insw	__ide_insw
-#define __ide_mm_insl	__ide_insl
-#define __ide_mm_outsw	__ide_outsw
-#define __ide_mm_outsl	__ide_outsl
-
-static inline void __ide_insw(unsigned long port,
-				  void *dst,
-				  unsigned long count)
-{
-	volatile unsigned short *data_port;
-	/* unsigned long end = (unsigned long)dst + (count << 1); */ /* P3 */
-	u16 *ps = dst;
-	u32 *pi;
-
-	data_port = (volatile unsigned short *)port;
-
-	if(((unsigned long)ps) & 0x2) {
-		*ps++ = *data_port;
-		count--;
-	}
-	pi = (u32 *)ps;
-	while(count >= 2) {
-		u32 w;
-
-		w  = (*data_port) << 16;
-		w |= (*data_port);
-		*pi++ = w;
-		count -= 2;
-	}
-	ps = (u16 *)pi;
-	if(count)
-		*ps++ = *data_port;
-
-	/* __flush_dcache_range((unsigned long)dst, end); */ /* P3 see hme */
-}
-
-static inline void __ide_outsw(unsigned long port,
-				   const void *src,
-				   unsigned long count)
-{
-	volatile unsigned short *data_port;
-	/* unsigned long end = (unsigned long)src + (count << 1); */
-	const u16 *ps = src;
-	const u32 *pi;
-
-	data_port = (volatile unsigned short *)port;
-
-	if(((unsigned long)src) & 0x2) {
-		*data_port = *ps++;
-		count--;
-	}
-	pi = (const u32 *)ps;
-	while(count >= 2) {
-		u32 w;
-
-		w = *pi++;
-		*data_port = (w >> 16);
-		*data_port = w;
-		count -= 2;
-	}
-	ps = (const u16 *)pi;
-	if(count)
-		*data_port = *ps;
-
-	/* __flush_dcache_range((unsigned long)src, end); */ /* P3 see hme */
-}
-
-#endif /* __KERNEL__ */
-
-#endif /* _SPARC_IDE_H */
+#ifndef ___ASM_SPARC_IDE_H
+#define ___ASM_SPARC_IDE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/ide_64.h>
+#else
+#include <asm-sparc/ide_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/ide_32.h b/include/asm-sparc/ide_32.h
new file mode 100644
index 00000000000..afd1736ed48
--- /dev/null
+++ b/include/asm-sparc/ide_32.h
@@ -0,0 +1,95 @@
+/* ide.h: SPARC PCI specific IDE glue.
+ *
+ * Copyright (C) 1997  David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1998  Eddie C. Dost   (ecd@skynet.be)
+ * Adaptation from sparc64 version to sparc by Pete Zaitcev.
+ */
+
+#ifndef _SPARC_IDE_H
+#define _SPARC_IDE_H
+
+#ifdef __KERNEL__
+
+#include <asm/pgtable.h>
+#include <asm/io.h>
+#include <asm/psr.h>
+
+#undef  MAX_HWIFS
+#define MAX_HWIFS	2
+
+#define __ide_insl(data_reg, buffer, wcount) \
+	__ide_insw(data_reg, buffer, (wcount)<<1)
+#define __ide_outsl(data_reg, buffer, wcount) \
+	__ide_outsw(data_reg, buffer, (wcount)<<1)
+
+/* On sparc, I/O ports and MMIO registers are accessed identically.  */
+#define __ide_mm_insw	__ide_insw
+#define __ide_mm_insl	__ide_insl
+#define __ide_mm_outsw	__ide_outsw
+#define __ide_mm_outsl	__ide_outsl
+
+static inline void __ide_insw(unsigned long port,
+				  void *dst,
+				  unsigned long count)
+{
+	volatile unsigned short *data_port;
+	/* unsigned long end = (unsigned long)dst + (count << 1); */ /* P3 */
+	u16 *ps = dst;
+	u32 *pi;
+
+	data_port = (volatile unsigned short *)port;
+
+	if(((unsigned long)ps) & 0x2) {
+		*ps++ = *data_port;
+		count--;
+	}
+	pi = (u32 *)ps;
+	while(count >= 2) {
+		u32 w;
+
+		w  = (*data_port) << 16;
+		w |= (*data_port);
+		*pi++ = w;
+		count -= 2;
+	}
+	ps = (u16 *)pi;
+	if(count)
+		*ps++ = *data_port;
+
+	/* __flush_dcache_range((unsigned long)dst, end); */ /* P3 see hme */
+}
+
+static inline void __ide_outsw(unsigned long port,
+				   const void *src,
+				   unsigned long count)
+{
+	volatile unsigned short *data_port;
+	/* unsigned long end = (unsigned long)src + (count << 1); */
+	const u16 *ps = src;
+	const u32 *pi;
+
+	data_port = (volatile unsigned short *)port;
+
+	if(((unsigned long)src) & 0x2) {
+		*data_port = *ps++;
+		count--;
+	}
+	pi = (const u32 *)ps;
+	while(count >= 2) {
+		u32 w;
+
+		w = *pi++;
+		*data_port = (w >> 16);
+		*data_port = w;
+		count -= 2;
+	}
+	ps = (const u16 *)pi;
+	if(count)
+		*data_port = *ps;
+
+	/* __flush_dcache_range((unsigned long)src, end); */ /* P3 see hme */
+}
+
+#endif /* __KERNEL__ */
+
+#endif /* _SPARC_IDE_H */
diff --git a/include/asm-sparc/ide_64.h b/include/asm-sparc/ide_64.h
new file mode 100644
index 00000000000..1282676da1c
--- /dev/null
+++ b/include/asm-sparc/ide_64.h
@@ -0,0 +1,118 @@
+/*
+ * ide.h: Ultra/PCI specific IDE glue.
+ *
+ * Copyright (C) 1997  David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1998  Eddie C. Dost   (ecd@skynet.be)
+ */
+
+#ifndef _SPARC64_IDE_H
+#define _SPARC64_IDE_H
+
+#ifdef __KERNEL__
+
+#include <asm/pgalloc.h>
+#include <asm/io.h>
+#include <asm/spitfire.h>
+#include <asm/cacheflush.h>
+#include <asm/page.h>
+
+#ifndef MAX_HWIFS
+# ifdef CONFIG_BLK_DEV_IDEPCI
+#define MAX_HWIFS	10
+# else
+#define MAX_HWIFS	2
+# endif
+#endif
+
+#define __ide_insl(data_reg, buffer, wcount) \
+	__ide_insw(data_reg, buffer, (wcount)<<1)
+#define __ide_outsl(data_reg, buffer, wcount) \
+	__ide_outsw(data_reg, buffer, (wcount)<<1)
+
+/* On sparc64, I/O ports and MMIO registers are accessed identically.  */
+#define __ide_mm_insw	__ide_insw
+#define __ide_mm_insl	__ide_insl
+#define __ide_mm_outsw	__ide_outsw
+#define __ide_mm_outsl	__ide_outsl
+
+static inline unsigned int inw_be(void __iomem *addr)
+{
+	unsigned int ret;
+
+	__asm__ __volatile__("lduha [%1] %2, %0"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline void __ide_insw(void __iomem *port, void *dst, u32 count)
+{
+#ifdef DCACHE_ALIASING_POSSIBLE
+	unsigned long end = (unsigned long)dst + (count << 1);
+#endif
+	u16 *ps = dst;
+	u32 *pi;
+
+	if(((u64)ps) & 0x2) {
+		*ps++ = inw_be(port);
+		count--;
+	}
+	pi = (u32 *)ps;
+	while(count >= 2) {
+		u32 w;
+
+		w  = inw_be(port) << 16;
+		w |= inw_be(port);
+		*pi++ = w;
+		count -= 2;
+	}
+	ps = (u16 *)pi;
+	if(count)
+		*ps++ = inw_be(port);
+
+#ifdef DCACHE_ALIASING_POSSIBLE
+	__flush_dcache_range((unsigned long)dst, end);
+#endif
+}
+
+static inline void outw_be(unsigned short w, void __iomem *addr)
+{
+	__asm__ __volatile__("stha %0, [%1] %2"
+			     : /* no outputs */
+			     : "r" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void __ide_outsw(void __iomem *port, void *src, u32 count)
+{
+#ifdef DCACHE_ALIASING_POSSIBLE
+	unsigned long end = (unsigned long)src + (count << 1);
+#endif
+	const u16 *ps = src;
+	const u32 *pi;
+
+	if(((u64)src) & 0x2) {
+		outw_be(*ps++, port);
+		count--;
+	}
+	pi = (const u32 *)ps;
+	while(count >= 2) {
+		u32 w;
+
+		w = *pi++;
+		outw_be((w >> 16), port);
+		outw_be(w, port);
+		count -= 2;
+	}
+	ps = (const u16 *)pi;
+	if(count)
+		outw_be(*ps, port);
+
+#ifdef DCACHE_ALIASING_POSSIBLE
+	__flush_dcache_range((unsigned long)src, end);
+#endif
+}
+
+#endif /* __KERNEL__ */
+
+#endif /* _SPARC64_IDE_H */
diff --git a/include/asm-sparc/idprom.h b/include/asm-sparc/idprom.h
index 41adb417a4e..6976aa2439c 100644
--- a/include/asm-sparc/idprom.h
+++ b/include/asm-sparc/idprom.h
@@ -1,7 +1,7 @@
 /*
  * idprom.h: Macros and defines for idprom routines
  *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
  */
 
 #ifndef _SPARC_IDPROM_H
diff --git a/include/asm-sparc/intr_queue.h b/include/asm-sparc/intr_queue.h
new file mode 100644
index 00000000000..206077dedc2
--- /dev/null
+++ b/include/asm-sparc/intr_queue.h
@@ -0,0 +1,15 @@
+#ifndef _SPARC64_INTR_QUEUE_H
+#define _SPARC64_INTR_QUEUE_H
+
+/* Sun4v interrupt queue registers, accessed via ASI_QUEUE.  */
+
+#define INTRQ_CPU_MONDO_HEAD	  0x3c0 /* CPU mondo head	          */
+#define INTRQ_CPU_MONDO_TAIL	  0x3c8 /* CPU mondo tail	          */
+#define INTRQ_DEVICE_MONDO_HEAD	  0x3d0 /* Device mondo head	          */
+#define INTRQ_DEVICE_MONDO_TAIL	  0x3d8 /* Device mondo tail	          */
+#define INTRQ_RESUM_MONDO_HEAD	  0x3e0 /* Resumable error mondo head     */
+#define INTRQ_RESUM_MONDO_TAIL	  0x3e8 /* Resumable error mondo tail     */
+#define INTRQ_NONRESUM_MONDO_HEAD 0x3f0 /* Non-resumable error mondo head */
+#define INTRQ_NONRESUM_MONDO_TAIL 0x3f8 /* Non-resumable error mondo head */
+
+#endif /* !(_SPARC64_INTR_QUEUE_H) */
diff --git a/include/asm-sparc/io.h b/include/asm-sparc/io.h
index 3a3e7bdb06b..fc9024d3dfc 100644
--- a/include/asm-sparc/io.h
+++ b/include/asm-sparc/io.h
@@ -1,325 +1,8 @@
-#ifndef __SPARC_IO_H
-#define __SPARC_IO_H
-
-#include <linux/kernel.h>
-#include <linux/types.h>
-#include <linux/ioport.h>  /* struct resource */
-
-#include <asm/page.h>      /* IO address mapping routines need this */
-#include <asm/system.h>
-
-#define page_to_phys(page)	(((page) - mem_map) << PAGE_SHIFT)
-
-static inline u32 flip_dword (u32 l)
-{
-	return ((l&0xff)<<24) | (((l>>8)&0xff)<<16) | (((l>>16)&0xff)<<8)| ((l>>24)&0xff);
-}
-
-static inline u16 flip_word (u16 w)
-{
-	return ((w&0xff) << 8) | ((w>>8)&0xff);
-}
-
-#define mmiowb()
-
-/*
- * Memory mapped I/O to PCI
- */
-
-static inline u8 __raw_readb(const volatile void __iomem *addr)
-{
-	return *(__force volatile u8 *)addr;
-}
-
-static inline u16 __raw_readw(const volatile void __iomem *addr)
-{
-	return *(__force volatile u16 *)addr;
-}
-
-static inline u32 __raw_readl(const volatile void __iomem *addr)
-{
-	return *(__force volatile u32 *)addr;
-}
-
-static inline void __raw_writeb(u8 b, volatile void __iomem *addr)
-{
-	*(__force volatile u8 *)addr = b;
-}
-
-static inline void __raw_writew(u16 w, volatile void __iomem *addr)
-{
-	*(__force volatile u16 *)addr = w;
-}
-
-static inline void __raw_writel(u32 l, volatile void __iomem *addr)
-{
-	*(__force volatile u32 *)addr = l;
-}
-
-static inline u8 __readb(const volatile void __iomem *addr)
-{
-	return *(__force volatile u8 *)addr;
-}
-
-static inline u16 __readw(const volatile void __iomem *addr)
-{
-	return flip_word(*(__force volatile u16 *)addr);
-}
-
-static inline u32 __readl(const volatile void __iomem *addr)
-{
-	return flip_dword(*(__force volatile u32 *)addr);
-}
-
-static inline void __writeb(u8 b, volatile void __iomem *addr)
-{
-	*(__force volatile u8 *)addr = b;
-}
-
-static inline void __writew(u16 w, volatile void __iomem *addr)
-{
-	*(__force volatile u16 *)addr = flip_word(w);
-}
-
-static inline void __writel(u32 l, volatile void __iomem *addr)
-{
-	*(__force volatile u32 *)addr = flip_dword(l);
-}
-
-#define readb(__addr)		__readb(__addr)
-#define readw(__addr)		__readw(__addr)
-#define readl(__addr)		__readl(__addr)
-#define readb_relaxed(__addr)	readb(__addr)
-#define readw_relaxed(__addr)	readw(__addr)
-#define readl_relaxed(__addr)	readl(__addr)
-
-#define writeb(__b, __addr)	__writeb((__b),(__addr))
-#define writew(__w, __addr)	__writew((__w),(__addr))
-#define writel(__l, __addr)	__writel((__l),(__addr))
-
-/*
- * I/O space operations
- *
- * Arrangement on a Sun is somewhat complicated.
- *
- * First of all, we want to use standard Linux drivers
- * for keyboard, PC serial, etc. These drivers think
- * they access I/O space and use inb/outb.
- * On the other hand, EBus bridge accepts PCI *memory*
- * cycles and converts them into ISA *I/O* cycles.
- * Ergo, we want inb & outb to generate PCI memory cycles.
- *
- * If we want to issue PCI *I/O* cycles, we do this
- * with a low 64K fixed window in PCIC. This window gets
- * mapped somewhere into virtual kernel space and we
- * can use inb/outb again.
- */
-#define inb_local(__addr)	__readb((void __iomem *)(unsigned long)(__addr))
-#define inb(__addr)		__readb((void __iomem *)(unsigned long)(__addr))
-#define inw(__addr)		__readw((void __iomem *)(unsigned long)(__addr))
-#define inl(__addr)		__readl((void __iomem *)(unsigned long)(__addr))
-
-#define outb_local(__b, __addr)	__writeb(__b, (void __iomem *)(unsigned long)(__addr))
-#define outb(__b, __addr)	__writeb(__b, (void __iomem *)(unsigned long)(__addr))
-#define outw(__w, __addr)	__writew(__w, (void __iomem *)(unsigned long)(__addr))
-#define outl(__l, __addr)	__writel(__l, (void __iomem *)(unsigned long)(__addr))
-
-#define inb_p(__addr)		inb(__addr)
-#define outb_p(__b, __addr)	outb(__b, __addr)
-#define inw_p(__addr)		inw(__addr)
-#define outw_p(__w, __addr)	outw(__w, __addr)
-#define inl_p(__addr)		inl(__addr)
-#define outl_p(__l, __addr)	outl(__l, __addr)
-
-void outsb(unsigned long addr, const void *src, unsigned long cnt);
-void outsw(unsigned long addr, const void *src, unsigned long cnt);
-void outsl(unsigned long addr, const void *src, unsigned long cnt);
-void insb(unsigned long addr, void *dst, unsigned long count);
-void insw(unsigned long addr, void *dst, unsigned long count);
-void insl(unsigned long addr, void *dst, unsigned long count);
-
-#define IO_SPACE_LIMIT 0xffffffff
-
-/*
- * SBus accessors.
- *
- * SBus has only one, memory mapped, I/O space.
- * We do not need to flip bytes for SBus of course.
- */
-static inline u8 _sbus_readb(const volatile void __iomem *addr)
-{
-	return *(__force volatile u8 *)addr;
-}
-
-static inline u16 _sbus_readw(const volatile void __iomem *addr)
-{
-	return *(__force volatile u16 *)addr;
-}
-
-static inline u32 _sbus_readl(const volatile void __iomem *addr)
-{
-	return *(__force volatile u32 *)addr;
-}
-
-static inline void _sbus_writeb(u8 b, volatile void __iomem *addr)
-{
-	*(__force volatile u8 *)addr = b;
-}
-
-static inline void _sbus_writew(u16 w, volatile void __iomem *addr)
-{
-	*(__force volatile u16 *)addr = w;
-}
-
-static inline void _sbus_writel(u32 l, volatile void __iomem *addr)
-{
-	*(__force volatile u32 *)addr = l;
-}
-
-/*
- * The only reason for #define's is to hide casts to unsigned long.
- */
-#define sbus_readb(__addr)		_sbus_readb(__addr)
-#define sbus_readw(__addr)		_sbus_readw(__addr)
-#define sbus_readl(__addr)		_sbus_readl(__addr)
-#define sbus_writeb(__b, __addr)	_sbus_writeb(__b, __addr)
-#define sbus_writew(__w, __addr)	_sbus_writew(__w, __addr)
-#define sbus_writel(__l, __addr)	_sbus_writel(__l, __addr)
-
-static inline void sbus_memset_io(volatile void __iomem *__dst, int c, __kernel_size_t n)
-{
-	while(n--) {
-		sbus_writeb(c, __dst);
-		__dst++;
-	}
-}
-
-static inline void
-_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
-{
-	volatile void __iomem *d = dst;
-
-	while (n--) {
-		writeb(c, d);
-		d++;
-	}
-}
-
-#define memset_io(d,c,sz)	_memset_io(d,c,sz)
-
-static inline void
-_memcpy_fromio(void *dst, const volatile void __iomem *src, __kernel_size_t n)
-{
-	char *d = dst;
-
-	while (n--) {
-		char tmp = readb(src);
-		*d++ = tmp;
-		src++;
-	}
-}
-
-#define memcpy_fromio(d,s,sz)	_memcpy_fromio(d,s,sz)
-
-static inline void 
-_memcpy_toio(volatile void __iomem *dst, const void *src, __kernel_size_t n)
-{
-	const char *s = src;
-	volatile void __iomem *d = dst;
-
-	while (n--) {
-		char tmp = *s++;
-		writeb(tmp, d);
-		d++;
-	}
-}
-
-#define memcpy_toio(d,s,sz)	_memcpy_toio(d,s,sz)
-
-#ifdef __KERNEL__
-
-/*
- * Bus number may be embedded in the higher bits of the physical address.
- * This is why we have no bus number argument to ioremap().
- */
-extern void __iomem *ioremap(unsigned long offset, unsigned long size);
-#define ioremap_nocache(X,Y)	ioremap((X),(Y))
-extern void iounmap(volatile void __iomem *addr);
-
-#define ioread8(X)			readb(X)
-#define ioread16(X)			readw(X)
-#define ioread32(X)			readl(X)
-#define iowrite8(val,X)			writeb(val,X)
-#define iowrite16(val,X)		writew(val,X)
-#define iowrite32(val,X)		writel(val,X)
-
-static inline void ioread8_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insb((unsigned long __force)port, buf, count);
-}
-static inline void ioread16_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insw((unsigned long __force)port, buf, count);
-}
-
-static inline void ioread32_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insl((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite8_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsb((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite16_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsw((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite32_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsl((unsigned long __force)port, buf, count);
-}
-
-/* Create a virtual mapping cookie for an IO port range */
-extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
-extern void ioport_unmap(void __iomem *);
-
-/* Create a virtual mapping cookie for a PCI BAR (memory or IO) */
-struct pci_dev;
-extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
-extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
-
-/*
- * Bus number may be in res->flags... somewhere.
- */
-extern void __iomem *sbus_ioremap(struct resource *res, unsigned long offset,
-    unsigned long size, char *name);
-extern void sbus_iounmap(volatile void __iomem *vaddr, unsigned long size);
-
-
-/*
- * At the moment, we do not use CMOS_READ anywhere outside of rtc.c,
- * so rtc_port is static in it. This should not change unless a new
- * hardware pops up.
- */
-#define RTC_PORT(x)   (rtc_port + (x))
-#define RTC_ALWAYS_BCD  0
-
+#ifndef ___ASM_SPARC_IO_H
+#define ___ASM_SPARC_IO_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/io_64.h>
+#else
+#include <asm-sparc/io_32.h>
+#endif
 #endif
-
-#define __ARCH_HAS_NO_PAGE_ZERO_MAPPED		1
-
-/*
- * Convert a physical pointer to a virtual kernel pointer for /dev/mem
- * access
- */
-#define xlate_dev_mem_ptr(p)	__va(p)
-
-/*
- * Convert a virtual cached pointer to an uncached pointer
- */
-#define xlate_dev_kmem_ptr(p)	p
-
-#endif /* !(__SPARC_IO_H) */
diff --git a/include/asm-sparc/io_32.h b/include/asm-sparc/io_32.h
new file mode 100644
index 00000000000..c43af722ae8
--- /dev/null
+++ b/include/asm-sparc/io_32.h
@@ -0,0 +1,325 @@
+#ifndef __SPARC_IO_H
+#define __SPARC_IO_H
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/ioport.h>  /* struct resource */
+
+#include <asm/page.h>      /* IO address mapping routines need this */
+#include <asm/system.h>
+
+#define page_to_phys(page)	(((page) - mem_map) << PAGE_SHIFT)
+
+static inline u32 flip_dword (u32 l)
+{
+	return ((l&0xff)<<24) | (((l>>8)&0xff)<<16) | (((l>>16)&0xff)<<8)| ((l>>24)&0xff);
+}
+
+static inline u16 flip_word (u16 w)
+{
+	return ((w&0xff) << 8) | ((w>>8)&0xff);
+}
+
+#define mmiowb()
+
+/*
+ * Memory mapped I/O to PCI
+ */
+
+static inline u8 __raw_readb(const volatile void __iomem *addr)
+{
+	return *(__force volatile u8 *)addr;
+}
+
+static inline u16 __raw_readw(const volatile void __iomem *addr)
+{
+	return *(__force volatile u16 *)addr;
+}
+
+static inline u32 __raw_readl(const volatile void __iomem *addr)
+{
+	return *(__force volatile u32 *)addr;
+}
+
+static inline void __raw_writeb(u8 b, volatile void __iomem *addr)
+{
+	*(__force volatile u8 *)addr = b;
+}
+
+static inline void __raw_writew(u16 w, volatile void __iomem *addr)
+{
+	*(__force volatile u16 *)addr = w;
+}
+
+static inline void __raw_writel(u32 l, volatile void __iomem *addr)
+{
+	*(__force volatile u32 *)addr = l;
+}
+
+static inline u8 __readb(const volatile void __iomem *addr)
+{
+	return *(__force volatile u8 *)addr;
+}
+
+static inline u16 __readw(const volatile void __iomem *addr)
+{
+	return flip_word(*(__force volatile u16 *)addr);
+}
+
+static inline u32 __readl(const volatile void __iomem *addr)
+{
+	return flip_dword(*(__force volatile u32 *)addr);
+}
+
+static inline void __writeb(u8 b, volatile void __iomem *addr)
+{
+	*(__force volatile u8 *)addr = b;
+}
+
+static inline void __writew(u16 w, volatile void __iomem *addr)
+{
+	*(__force volatile u16 *)addr = flip_word(w);
+}
+
+static inline void __writel(u32 l, volatile void __iomem *addr)
+{
+	*(__force volatile u32 *)addr = flip_dword(l);
+}
+
+#define readb(__addr)		__readb(__addr)
+#define readw(__addr)		__readw(__addr)
+#define readl(__addr)		__readl(__addr)
+#define readb_relaxed(__addr)	readb(__addr)
+#define readw_relaxed(__addr)	readw(__addr)
+#define readl_relaxed(__addr)	readl(__addr)
+
+#define writeb(__b, __addr)	__writeb((__b),(__addr))
+#define writew(__w, __addr)	__writew((__w),(__addr))
+#define writel(__l, __addr)	__writel((__l),(__addr))
+
+/*
+ * I/O space operations
+ *
+ * Arrangement on a Sun is somewhat complicated.
+ *
+ * First of all, we want to use standard Linux drivers
+ * for keyboard, PC serial, etc. These drivers think
+ * they access I/O space and use inb/outb.
+ * On the other hand, EBus bridge accepts PCI *memory*
+ * cycles and converts them into ISA *I/O* cycles.
+ * Ergo, we want inb & outb to generate PCI memory cycles.
+ *
+ * If we want to issue PCI *I/O* cycles, we do this
+ * with a low 64K fixed window in PCIC. This window gets
+ * mapped somewhere into virtual kernel space and we
+ * can use inb/outb again.
+ */
+#define inb_local(__addr)	__readb((void __iomem *)(unsigned long)(__addr))
+#define inb(__addr)		__readb((void __iomem *)(unsigned long)(__addr))
+#define inw(__addr)		__readw((void __iomem *)(unsigned long)(__addr))
+#define inl(__addr)		__readl((void __iomem *)(unsigned long)(__addr))
+
+#define outb_local(__b, __addr)	__writeb(__b, (void __iomem *)(unsigned long)(__addr))
+#define outb(__b, __addr)	__writeb(__b, (void __iomem *)(unsigned long)(__addr))
+#define outw(__w, __addr)	__writew(__w, (void __iomem *)(unsigned long)(__addr))
+#define outl(__l, __addr)	__writel(__l, (void __iomem *)(unsigned long)(__addr))
+
+#define inb_p(__addr)		inb(__addr)
+#define outb_p(__b, __addr)	outb(__b, __addr)
+#define inw_p(__addr)		inw(__addr)
+#define outw_p(__w, __addr)	outw(__w, __addr)
+#define inl_p(__addr)		inl(__addr)
+#define outl_p(__l, __addr)	outl(__l, __addr)
+
+void outsb(unsigned long addr, const void *src, unsigned long cnt);
+void outsw(unsigned long addr, const void *src, unsigned long cnt);
+void outsl(unsigned long addr, const void *src, unsigned long cnt);
+void insb(unsigned long addr, void *dst, unsigned long count);
+void insw(unsigned long addr, void *dst, unsigned long count);
+void insl(unsigned long addr, void *dst, unsigned long count);
+
+#define IO_SPACE_LIMIT 0xffffffff
+
+/*
+ * SBus accessors.
+ *
+ * SBus has only one, memory mapped, I/O space.
+ * We do not need to flip bytes for SBus of course.
+ */
+static inline u8 _sbus_readb(const volatile void __iomem *addr)
+{
+	return *(__force volatile u8 *)addr;
+}
+
+static inline u16 _sbus_readw(const volatile void __iomem *addr)
+{
+	return *(__force volatile u16 *)addr;
+}
+
+static inline u32 _sbus_readl(const volatile void __iomem *addr)
+{
+	return *(__force volatile u32 *)addr;
+}
+
+static inline void _sbus_writeb(u8 b, volatile void __iomem *addr)
+{
+	*(__force volatile u8 *)addr = b;
+}
+
+static inline void _sbus_writew(u16 w, volatile void __iomem *addr)
+{
+	*(__force volatile u16 *)addr = w;
+}
+
+static inline void _sbus_writel(u32 l, volatile void __iomem *addr)
+{
+	*(__force volatile u32 *)addr = l;
+}
+
+/*
+ * The only reason for #define's is to hide casts to unsigned long.
+ */
+#define sbus_readb(__addr)		_sbus_readb(__addr)
+#define sbus_readw(__addr)		_sbus_readw(__addr)
+#define sbus_readl(__addr)		_sbus_readl(__addr)
+#define sbus_writeb(__b, __addr)	_sbus_writeb(__b, __addr)
+#define sbus_writew(__w, __addr)	_sbus_writew(__w, __addr)
+#define sbus_writel(__l, __addr)	_sbus_writel(__l, __addr)
+
+static inline void sbus_memset_io(volatile void __iomem *__dst, int c, __kernel_size_t n)
+{
+	while(n--) {
+		sbus_writeb(c, __dst);
+		__dst++;
+	}
+}
+
+static inline void
+_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
+{
+	volatile void __iomem *d = dst;
+
+	while (n--) {
+		writeb(c, d);
+		d++;
+	}
+}
+
+#define memset_io(d,c,sz)	_memset_io(d,c,sz)
+
+static inline void
+_memcpy_fromio(void *dst, const volatile void __iomem *src, __kernel_size_t n)
+{
+	char *d = dst;
+
+	while (n--) {
+		char tmp = readb(src);
+		*d++ = tmp;
+		src++;
+	}
+}
+
+#define memcpy_fromio(d,s,sz)	_memcpy_fromio(d,s,sz)
+
+static inline void
+_memcpy_toio(volatile void __iomem *dst, const void *src, __kernel_size_t n)
+{
+	const char *s = src;
+	volatile void __iomem *d = dst;
+
+	while (n--) {
+		char tmp = *s++;
+		writeb(tmp, d);
+		d++;
+	}
+}
+
+#define memcpy_toio(d,s,sz)	_memcpy_toio(d,s,sz)
+
+#ifdef __KERNEL__
+
+/*
+ * Bus number may be embedded in the higher bits of the physical address.
+ * This is why we have no bus number argument to ioremap().
+ */
+extern void __iomem *ioremap(unsigned long offset, unsigned long size);
+#define ioremap_nocache(X,Y)	ioremap((X),(Y))
+extern void iounmap(volatile void __iomem *addr);
+
+#define ioread8(X)			readb(X)
+#define ioread16(X)			readw(X)
+#define ioread32(X)			readl(X)
+#define iowrite8(val,X)			writeb(val,X)
+#define iowrite16(val,X)		writew(val,X)
+#define iowrite32(val,X)		writel(val,X)
+
+static inline void ioread8_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insb((unsigned long __force)port, buf, count);
+}
+static inline void ioread16_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insw((unsigned long __force)port, buf, count);
+}
+
+static inline void ioread32_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insl((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite8_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsb((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite16_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsw((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite32_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsl((unsigned long __force)port, buf, count);
+}
+
+/* Create a virtual mapping cookie for an IO port range */
+extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
+extern void ioport_unmap(void __iomem *);
+
+/* Create a virtual mapping cookie for a PCI BAR (memory or IO) */
+struct pci_dev;
+extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
+extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
+
+/*
+ * Bus number may be in res->flags... somewhere.
+ */
+extern void __iomem *sbus_ioremap(struct resource *res, unsigned long offset,
+    unsigned long size, char *name);
+extern void sbus_iounmap(volatile void __iomem *vaddr, unsigned long size);
+
+
+/*
+ * At the moment, we do not use CMOS_READ anywhere outside of rtc.c,
+ * so rtc_port is static in it. This should not change unless a new
+ * hardware pops up.
+ */
+#define RTC_PORT(x)   (rtc_port + (x))
+#define RTC_ALWAYS_BCD  0
+
+#endif
+
+#define __ARCH_HAS_NO_PAGE_ZERO_MAPPED		1
+
+/*
+ * Convert a physical pointer to a virtual kernel pointer for /dev/mem
+ * access
+ */
+#define xlate_dev_mem_ptr(p)	__va(p)
+
+/*
+ * Convert a virtual cached pointer to an uncached pointer
+ */
+#define xlate_dev_kmem_ptr(p)	p
+
+#endif /* !(__SPARC_IO_H) */
diff --git a/include/asm-sparc/io_64.h b/include/asm-sparc/io_64.h
new file mode 100644
index 00000000000..f4907414b39
--- /dev/null
+++ b/include/asm-sparc/io_64.h
@@ -0,0 +1,511 @@
+#ifndef __SPARC64_IO_H
+#define __SPARC64_IO_H
+
+#include <linux/kernel.h>
+#include <linux/compiler.h>
+#include <linux/types.h>
+
+#include <asm/page.h>      /* IO address mapping routines need this */
+#include <asm/system.h>
+#include <asm/asi.h>
+
+/* PC crapola... */
+#define __SLOW_DOWN_IO	do { } while (0)
+#define SLOW_DOWN_IO	do { } while (0)
+
+/* BIO layer definitions. */
+extern unsigned long kern_base, kern_size;
+#define page_to_phys(page)	(page_to_pfn(page) << PAGE_SHIFT)
+#define BIO_VMERGE_BOUNDARY	8192
+
+static inline u8 _inb(unsigned long addr)
+{
+	u8 ret;
+
+	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_inb */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u16 _inw(unsigned long addr)
+{
+	u16 ret;
+
+	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_inw */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u32 _inl(unsigned long addr)
+{
+	u32 ret;
+
+	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_inl */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline void _outb(u8 b, unsigned long addr)
+{
+	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_outb */"
+			     : /* no outputs */
+			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+static inline void _outw(u16 w, unsigned long addr)
+{
+	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_outw */"
+			     : /* no outputs */
+			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+static inline void _outl(u32 l, unsigned long addr)
+{
+	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_outl */"
+			     : /* no outputs */
+			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+#define inb(__addr)		(_inb((unsigned long)(__addr)))
+#define inw(__addr)		(_inw((unsigned long)(__addr)))
+#define inl(__addr)		(_inl((unsigned long)(__addr)))
+#define outb(__b, __addr)	(_outb((u8)(__b), (unsigned long)(__addr)))
+#define outw(__w, __addr)	(_outw((u16)(__w), (unsigned long)(__addr)))
+#define outl(__l, __addr)	(_outl((u32)(__l), (unsigned long)(__addr)))
+
+#define inb_p(__addr) 		inb(__addr)
+#define outb_p(__b, __addr)	outb(__b, __addr)
+#define inw_p(__addr)		inw(__addr)
+#define outw_p(__w, __addr)	outw(__w, __addr)
+#define inl_p(__addr)		inl(__addr)
+#define outl_p(__l, __addr)	outl(__l, __addr)
+
+extern void outsb(unsigned long, const void *, unsigned long);
+extern void outsw(unsigned long, const void *, unsigned long);
+extern void outsl(unsigned long, const void *, unsigned long);
+extern void insb(unsigned long, void *, unsigned long);
+extern void insw(unsigned long, void *, unsigned long);
+extern void insl(unsigned long, void *, unsigned long);
+
+static inline void ioread8_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insb((unsigned long __force)port, buf, count);
+}
+static inline void ioread16_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insw((unsigned long __force)port, buf, count);
+}
+
+static inline void ioread32_rep(void __iomem *port, void *buf, unsigned long count)
+{
+	insl((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite8_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsb((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite16_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsw((unsigned long __force)port, buf, count);
+}
+
+static inline void iowrite32_rep(void __iomem *port, const void *buf, unsigned long count)
+{
+	outsl((unsigned long __force)port, buf, count);
+}
+
+/* Memory functions, same as I/O accesses on Ultra. */
+static inline u8 _readb(const volatile void __iomem *addr)
+{	u8 ret;
+
+	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_readb */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+	return ret;
+}
+
+static inline u16 _readw(const volatile void __iomem *addr)
+{	u16 ret;
+
+	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_readw */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u32 _readl(const volatile void __iomem *addr)
+{	u32 ret;
+
+	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_readl */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u64 _readq(const volatile void __iomem *addr)
+{	u64 ret;
+
+	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* pci_readq */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+
+	return ret;
+}
+
+static inline void _writeb(u8 b, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_writeb */"
+			     : /* no outputs */
+			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+static inline void _writew(u16 w, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_writew */"
+			     : /* no outputs */
+			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+static inline void _writel(u32 l, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_writel */"
+			     : /* no outputs */
+			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+static inline void _writeq(u64 q, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* pci_writeq */"
+			     : /* no outputs */
+			     : "Jr" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
+			     : "memory");
+}
+
+#define readb(__addr)		_readb(__addr)
+#define readw(__addr)		_readw(__addr)
+#define readl(__addr)		_readl(__addr)
+#define readq(__addr)		_readq(__addr)
+#define readb_relaxed(__addr)	_readb(__addr)
+#define readw_relaxed(__addr)	_readw(__addr)
+#define readl_relaxed(__addr)	_readl(__addr)
+#define readq_relaxed(__addr)	_readq(__addr)
+#define writeb(__b, __addr)	_writeb(__b, __addr)
+#define writew(__w, __addr)	_writew(__w, __addr)
+#define writel(__l, __addr)	_writel(__l, __addr)
+#define writeq(__q, __addr)	_writeq(__q, __addr)
+
+/* Now versions without byte-swapping. */
+static inline u8 _raw_readb(unsigned long addr)
+{
+	u8 ret;
+
+	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_raw_readb */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline u16 _raw_readw(unsigned long addr)
+{
+	u16 ret;
+
+	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_raw_readw */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline u32 _raw_readl(unsigned long addr)
+{
+	u32 ret;
+
+	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_raw_readl */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline u64 _raw_readq(unsigned long addr)
+{
+	u64 ret;
+
+	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* pci_raw_readq */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline void _raw_writeb(u8 b, unsigned long addr)
+{
+	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_raw_writeb */"
+			     : /* no outputs */
+			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _raw_writew(u16 w, unsigned long addr)
+{
+	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_raw_writew */"
+			     : /* no outputs */
+			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _raw_writel(u32 l, unsigned long addr)
+{
+	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_raw_writel */"
+			     : /* no outputs */
+			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _raw_writeq(u64 q, unsigned long addr)
+{
+	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* pci_raw_writeq */"
+			     : /* no outputs */
+			     : "Jr" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+#define __raw_readb(__addr)		(_raw_readb((unsigned long)(__addr)))
+#define __raw_readw(__addr)		(_raw_readw((unsigned long)(__addr)))
+#define __raw_readl(__addr)		(_raw_readl((unsigned long)(__addr)))
+#define __raw_readq(__addr)		(_raw_readq((unsigned long)(__addr)))
+#define __raw_writeb(__b, __addr)	(_raw_writeb((u8)(__b), (unsigned long)(__addr)))
+#define __raw_writew(__w, __addr)	(_raw_writew((u16)(__w), (unsigned long)(__addr)))
+#define __raw_writel(__l, __addr)	(_raw_writel((u32)(__l), (unsigned long)(__addr)))
+#define __raw_writeq(__q, __addr)	(_raw_writeq((u64)(__q), (unsigned long)(__addr)))
+
+/* Valid I/O Space regions are anywhere, because each PCI bus supported
+ * can live in an arbitrary area of the physical address range.
+ */
+#define IO_SPACE_LIMIT 0xffffffffffffffffUL
+
+/* Now, SBUS variants, only difference from PCI is that we do
+ * not use little-endian ASIs.
+ */
+static inline u8 _sbus_readb(const volatile void __iomem *addr)
+{
+	u8 ret;
+
+	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* sbus_readb */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u16 _sbus_readw(const volatile void __iomem *addr)
+{
+	u16 ret;
+
+	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* sbus_readw */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u32 _sbus_readl(const volatile void __iomem *addr)
+{
+	u32 ret;
+
+	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* sbus_readl */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+
+	return ret;
+}
+
+static inline u64 _sbus_readq(const volatile void __iomem *addr)
+{
+	u64 ret;
+
+	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* sbus_readq */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+
+	return ret;
+}
+
+static inline void _sbus_writeb(u8 b, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* sbus_writeb */"
+			     : /* no outputs */
+			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+}
+
+static inline void _sbus_writew(u16 w, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* sbus_writew */"
+			     : /* no outputs */
+			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+}
+
+static inline void _sbus_writel(u32 l, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* sbus_writel */"
+			     : /* no outputs */
+			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+}
+
+static inline void _sbus_writeq(u64 l, volatile void __iomem *addr)
+{
+	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* sbus_writeq */"
+			     : /* no outputs */
+			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
+			     : "memory");
+}
+
+#define sbus_readb(__addr)		_sbus_readb(__addr)
+#define sbus_readw(__addr)		_sbus_readw(__addr)
+#define sbus_readl(__addr)		_sbus_readl(__addr)
+#define sbus_readq(__addr)		_sbus_readq(__addr)
+#define sbus_writeb(__b, __addr)	_sbus_writeb(__b, __addr)
+#define sbus_writew(__w, __addr)	_sbus_writew(__w, __addr)
+#define sbus_writel(__l, __addr)	_sbus_writel(__l, __addr)
+#define sbus_writeq(__l, __addr)	_sbus_writeq(__l, __addr)
+
+static inline void _sbus_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
+{
+	while(n--) {
+		sbus_writeb(c, dst);
+		dst++;
+	}
+}
+
+#define sbus_memset_io(d,c,sz)	_sbus_memset_io(d,c,sz)
+
+static inline void
+_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
+{
+	volatile void __iomem *d = dst;
+
+	while (n--) {
+		writeb(c, d);
+		d++;
+	}
+}
+
+#define memset_io(d,c,sz)	_memset_io(d,c,sz)
+
+static inline void
+_memcpy_fromio(void *dst, const volatile void __iomem *src, __kernel_size_t n)
+{
+	char *d = dst;
+
+	while (n--) {
+		char tmp = readb(src);
+		*d++ = tmp;
+		src++;
+	}
+}
+
+#define memcpy_fromio(d,s,sz)	_memcpy_fromio(d,s,sz)
+
+static inline void
+_memcpy_toio(volatile void __iomem *dst, const void *src, __kernel_size_t n)
+{
+	const char *s = src;
+	volatile void __iomem *d = dst;
+
+	while (n--) {
+		char tmp = *s++;
+		writeb(tmp, d);
+		d++;
+	}
+}
+
+#define memcpy_toio(d,s,sz)	_memcpy_toio(d,s,sz)
+
+#define mmiowb()
+
+#ifdef __KERNEL__
+
+/* On sparc64 we have the whole physical IO address space accessible
+ * using physically addressed loads and stores, so this does nothing.
+ */
+static inline void __iomem *ioremap(unsigned long offset, unsigned long size)
+{
+	return (void __iomem *)offset;
+}
+
+#define ioremap_nocache(X,Y)		ioremap((X),(Y))
+
+static inline void iounmap(volatile void __iomem *addr)
+{
+}
+
+#define ioread8(X)			readb(X)
+#define ioread16(X)			readw(X)
+#define ioread32(X)			readl(X)
+#define iowrite8(val,X)			writeb(val,X)
+#define iowrite16(val,X)		writew(val,X)
+#define iowrite32(val,X)		writel(val,X)
+
+/* Create a virtual mapping cookie for an IO port range */
+extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
+extern void ioport_unmap(void __iomem *);
+
+/* Create a virtual mapping cookie for a PCI BAR (memory or IO) */
+struct pci_dev;
+extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
+extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
+
+/* Similarly for SBUS. */
+#define sbus_ioremap(__res, __offset, __size, __name) \
+({	unsigned long __ret; \
+	__ret  = (__res)->start + (((__res)->flags & 0x1ffUL) << 32UL); \
+	__ret += (unsigned long) (__offset); \
+	if (! request_region((__ret), (__size), (__name))) \
+		__ret = 0UL; \
+	(void __iomem *) __ret; \
+})
+
+#define sbus_iounmap(__addr, __size)	\
+	release_region((unsigned long)(__addr), (__size))
+
+/*
+ * Convert a physical pointer to a virtual kernel pointer for /dev/mem
+ * access
+ */
+#define xlate_dev_mem_ptr(p)	__va(p)
+
+/*
+ * Convert a virtual cached pointer to an uncached pointer
+ */
+#define xlate_dev_kmem_ptr(p)	p
+
+#endif
+
+#endif /* !(__SPARC64_IO_H) */
diff --git a/include/asm-sparc/ioctls.h b/include/asm-sparc/ioctls.h
index 3f4d0087b6a..1fe6855c5c1 100644
--- a/include/asm-sparc/ioctls.h
+++ b/include/asm-sparc/ioctls.h
@@ -22,7 +22,7 @@
 
 /* Note that all the ioctls that are not available in Linux have a 
  * double underscore on the front to: a) avoid some programs to
- * thing we support some ioctls under Linux (autoconfiguration stuff)
+ * think we support some ioctls under Linux (autoconfiguration stuff)
  */
 /* Little t */
 #define TIOCGETD	_IOR('t', 0, int)
@@ -110,7 +110,7 @@
 #define TIOCSERGETLSR   0x5459 /* Get line status register */
 #define TIOCSERGETMULTI 0x545A /* Get multiport config  */
 #define TIOCSERSETMULTI 0x545B /* Set multiport config */
-#define TIOCMIWAIT	0x545C /* Wait input */
+#define TIOCMIWAIT	0x545C /* Wait for change on serial input line(s) */
 #define TIOCGICOUNT	0x545D /* Read serial port inline interrupt counts */
 
 /* Kernel definitions */
diff --git a/include/asm-sparc/iommu.h b/include/asm-sparc/iommu.h
index 70c589c05a1..91b072b0d7a 100644
--- a/include/asm-sparc/iommu.h
+++ b/include/asm-sparc/iommu.h
@@ -1,121 +1,8 @@
-/* iommu.h: Definitions for the sun4m IOMMU.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-#ifndef _SPARC_IOMMU_H
-#define _SPARC_IOMMU_H
-
-#include <asm/page.h>
-#include <asm/bitext.h>
-
-/* The iommu handles all virtual to physical address translations
- * that occur between the SBUS and physical memory.  Access by
- * the cpu to IO registers and similar go over the mbus so are
- * translated by the on chip SRMMU.  The iommu and the srmmu do
- * not need to have the same translations at all, in fact most
- * of the time the translations they handle are a disjunct set.
- * Basically the iommu handles all dvma sbus activity.
- */
-
-/* The IOMMU registers occupy three pages in IO space. */
-struct iommu_regs {
-	/* First page */
-	volatile unsigned long control;    /* IOMMU control */
-	volatile unsigned long base;       /* Physical base of iopte page table */
-	volatile unsigned long _unused1[3];
-	volatile unsigned long tlbflush;   /* write only */
-	volatile unsigned long pageflush;  /* write only */
-	volatile unsigned long _unused2[1017];
-	/* Second page */
-	volatile unsigned long afsr;       /* Async-fault status register */
-	volatile unsigned long afar;       /* Async-fault physical address */
-	volatile unsigned long _unused3[2];
-	volatile unsigned long sbuscfg0;   /* SBUS configuration registers, per-slot */
-	volatile unsigned long sbuscfg1;
-	volatile unsigned long sbuscfg2;
-	volatile unsigned long sbuscfg3;
-	volatile unsigned long mfsr;       /* Memory-fault status register */
-	volatile unsigned long mfar;       /* Memory-fault physical address */
-	volatile unsigned long _unused4[1014];
-	/* Third page */
-	volatile unsigned long mid;        /* IOMMU module-id */
-};
-
-#define IOMMU_CTRL_IMPL     0xf0000000 /* Implementation */
-#define IOMMU_CTRL_VERS     0x0f000000 /* Version */
-#define IOMMU_CTRL_RNGE     0x0000001c /* Mapping RANGE */
-#define IOMMU_RNGE_16MB     0x00000000 /* 0xff000000 -> 0xffffffff */
-#define IOMMU_RNGE_32MB     0x00000004 /* 0xfe000000 -> 0xffffffff */
-#define IOMMU_RNGE_64MB     0x00000008 /* 0xfc000000 -> 0xffffffff */
-#define IOMMU_RNGE_128MB    0x0000000c /* 0xf8000000 -> 0xffffffff */
-#define IOMMU_RNGE_256MB    0x00000010 /* 0xf0000000 -> 0xffffffff */
-#define IOMMU_RNGE_512MB    0x00000014 /* 0xe0000000 -> 0xffffffff */
-#define IOMMU_RNGE_1GB      0x00000018 /* 0xc0000000 -> 0xffffffff */
-#define IOMMU_RNGE_2GB      0x0000001c /* 0x80000000 -> 0xffffffff */
-#define IOMMU_CTRL_ENAB     0x00000001 /* IOMMU Enable */
-
-#define IOMMU_AFSR_ERR      0x80000000 /* LE, TO, or BE asserted */
-#define IOMMU_AFSR_LE       0x40000000 /* SBUS reports error after transaction */
-#define IOMMU_AFSR_TO       0x20000000 /* Write access took more than 12.8 us. */
-#define IOMMU_AFSR_BE       0x10000000 /* Write access received error acknowledge */
-#define IOMMU_AFSR_SIZE     0x0e000000 /* Size of transaction causing error */
-#define IOMMU_AFSR_S        0x01000000 /* Sparc was in supervisor mode */
-#define IOMMU_AFSR_RESV     0x00f00000 /* Reserver, forced to 0x8 by hardware */
-#define IOMMU_AFSR_ME       0x00080000 /* Multiple errors occurred */
-#define IOMMU_AFSR_RD       0x00040000 /* A read operation was in progress */
-#define IOMMU_AFSR_FAV      0x00020000 /* IOMMU afar has valid contents */
-
-#define IOMMU_SBCFG_SAB30   0x00010000 /* Phys-address bit 30 when bypass enabled */
-#define IOMMU_SBCFG_BA16    0x00000004 /* Slave supports 16 byte bursts */
-#define IOMMU_SBCFG_BA8     0x00000002 /* Slave supports 8 byte bursts */
-#define IOMMU_SBCFG_BYPASS  0x00000001 /* Bypass IOMMU, treat all addresses
-					  produced by this device as pure
-					  physical. */
-
-#define IOMMU_MFSR_ERR      0x80000000 /* One or more of PERR1 or PERR0 */
-#define IOMMU_MFSR_S        0x01000000 /* Sparc was in supervisor mode */
-#define IOMMU_MFSR_CPU      0x00800000 /* CPU transaction caused parity error */
-#define IOMMU_MFSR_ME       0x00080000 /* Multiple parity errors occurred */
-#define IOMMU_MFSR_PERR     0x00006000 /* high bit indicates parity error occurred
-					  on the even word of the access, low bit
-					  indicated odd word caused the parity error */
-#define IOMMU_MFSR_BM       0x00001000 /* Error occurred while in boot mode */
-#define IOMMU_MFSR_C        0x00000800 /* Address causing error was marked cacheable */
-#define IOMMU_MFSR_RTYP     0x000000f0 /* Memory request transaction type */
-
-#define IOMMU_MID_SBAE      0x001f0000 /* SBus arbitration enable */
-#define IOMMU_MID_SE        0x00100000 /* Enables SCSI/ETHERNET arbitration */
-#define IOMMU_MID_SB3       0x00080000 /* Enable SBUS device 3 arbitration */
-#define IOMMU_MID_SB2       0x00040000 /* Enable SBUS device 2 arbitration */
-#define IOMMU_MID_SB1       0x00020000 /* Enable SBUS device 1 arbitration */
-#define IOMMU_MID_SB0       0x00010000 /* Enable SBUS device 0 arbitration */
-#define IOMMU_MID_MID       0x0000000f /* Module-id, hardcoded to 0x8 */
-
-/* The format of an iopte in the page tables */
-#define IOPTE_PAGE          0x07ffff00 /* Physical page number (PA[30:12]) */
-#define IOPTE_CACHE         0x00000080 /* Cached (in vme IOCACHE or Viking/MXCC) */
-#define IOPTE_WRITE         0x00000004 /* Writeable */
-#define IOPTE_VALID         0x00000002 /* IOPTE is valid */
-#define IOPTE_WAZ           0x00000001 /* Write as zeros */
-
-struct iommu_struct {
-	struct iommu_regs *regs;
-	iopte_t *page_table;
-	/* For convenience */
-	unsigned long start; /* First managed virtual address */
-	unsigned long end;   /* Last managed virtual address */
-
-	struct bit_map usemap;
-};
-
-static inline void iommu_invalidate(struct iommu_regs *regs)
-{
-	regs->tlbflush = 0;
-}
-
-static inline void iommu_invalidate_page(struct iommu_regs *regs, unsigned long ba)
-{
-	regs->pageflush = (ba & PAGE_MASK);
-}
-
-#endif /* !(_SPARC_IOMMU_H) */
+#ifndef ___ASM_SPARC_IOMMU_H
+#define ___ASM_SPARC_IOMMU_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/iommu_64.h>
+#else
+#include <asm-sparc/iommu_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/iommu_32.h b/include/asm-sparc/iommu_32.h
new file mode 100644
index 00000000000..70c589c05a1
--- /dev/null
+++ b/include/asm-sparc/iommu_32.h
@@ -0,0 +1,121 @@
+/* iommu.h: Definitions for the sun4m IOMMU.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ */
+#ifndef _SPARC_IOMMU_H
+#define _SPARC_IOMMU_H
+
+#include <asm/page.h>
+#include <asm/bitext.h>
+
+/* The iommu handles all virtual to physical address translations
+ * that occur between the SBUS and physical memory.  Access by
+ * the cpu to IO registers and similar go over the mbus so are
+ * translated by the on chip SRMMU.  The iommu and the srmmu do
+ * not need to have the same translations at all, in fact most
+ * of the time the translations they handle are a disjunct set.
+ * Basically the iommu handles all dvma sbus activity.
+ */
+
+/* The IOMMU registers occupy three pages in IO space. */
+struct iommu_regs {
+	/* First page */
+	volatile unsigned long control;    /* IOMMU control */
+	volatile unsigned long base;       /* Physical base of iopte page table */
+	volatile unsigned long _unused1[3];
+	volatile unsigned long tlbflush;   /* write only */
+	volatile unsigned long pageflush;  /* write only */
+	volatile unsigned long _unused2[1017];
+	/* Second page */
+	volatile unsigned long afsr;       /* Async-fault status register */
+	volatile unsigned long afar;       /* Async-fault physical address */
+	volatile unsigned long _unused3[2];
+	volatile unsigned long sbuscfg0;   /* SBUS configuration registers, per-slot */
+	volatile unsigned long sbuscfg1;
+	volatile unsigned long sbuscfg2;
+	volatile unsigned long sbuscfg3;
+	volatile unsigned long mfsr;       /* Memory-fault status register */
+	volatile unsigned long mfar;       /* Memory-fault physical address */
+	volatile unsigned long _unused4[1014];
+	/* Third page */
+	volatile unsigned long mid;        /* IOMMU module-id */
+};
+
+#define IOMMU_CTRL_IMPL     0xf0000000 /* Implementation */
+#define IOMMU_CTRL_VERS     0x0f000000 /* Version */
+#define IOMMU_CTRL_RNGE     0x0000001c /* Mapping RANGE */
+#define IOMMU_RNGE_16MB     0x00000000 /* 0xff000000 -> 0xffffffff */
+#define IOMMU_RNGE_32MB     0x00000004 /* 0xfe000000 -> 0xffffffff */
+#define IOMMU_RNGE_64MB     0x00000008 /* 0xfc000000 -> 0xffffffff */
+#define IOMMU_RNGE_128MB    0x0000000c /* 0xf8000000 -> 0xffffffff */
+#define IOMMU_RNGE_256MB    0x00000010 /* 0xf0000000 -> 0xffffffff */
+#define IOMMU_RNGE_512MB    0x00000014 /* 0xe0000000 -> 0xffffffff */
+#define IOMMU_RNGE_1GB      0x00000018 /* 0xc0000000 -> 0xffffffff */
+#define IOMMU_RNGE_2GB      0x0000001c /* 0x80000000 -> 0xffffffff */
+#define IOMMU_CTRL_ENAB     0x00000001 /* IOMMU Enable */
+
+#define IOMMU_AFSR_ERR      0x80000000 /* LE, TO, or BE asserted */
+#define IOMMU_AFSR_LE       0x40000000 /* SBUS reports error after transaction */
+#define IOMMU_AFSR_TO       0x20000000 /* Write access took more than 12.8 us. */
+#define IOMMU_AFSR_BE       0x10000000 /* Write access received error acknowledge */
+#define IOMMU_AFSR_SIZE     0x0e000000 /* Size of transaction causing error */
+#define IOMMU_AFSR_S        0x01000000 /* Sparc was in supervisor mode */
+#define IOMMU_AFSR_RESV     0x00f00000 /* Reserver, forced to 0x8 by hardware */
+#define IOMMU_AFSR_ME       0x00080000 /* Multiple errors occurred */
+#define IOMMU_AFSR_RD       0x00040000 /* A read operation was in progress */
+#define IOMMU_AFSR_FAV      0x00020000 /* IOMMU afar has valid contents */
+
+#define IOMMU_SBCFG_SAB30   0x00010000 /* Phys-address bit 30 when bypass enabled */
+#define IOMMU_SBCFG_BA16    0x00000004 /* Slave supports 16 byte bursts */
+#define IOMMU_SBCFG_BA8     0x00000002 /* Slave supports 8 byte bursts */
+#define IOMMU_SBCFG_BYPASS  0x00000001 /* Bypass IOMMU, treat all addresses
+					  produced by this device as pure
+					  physical. */
+
+#define IOMMU_MFSR_ERR      0x80000000 /* One or more of PERR1 or PERR0 */
+#define IOMMU_MFSR_S        0x01000000 /* Sparc was in supervisor mode */
+#define IOMMU_MFSR_CPU      0x00800000 /* CPU transaction caused parity error */
+#define IOMMU_MFSR_ME       0x00080000 /* Multiple parity errors occurred */
+#define IOMMU_MFSR_PERR     0x00006000 /* high bit indicates parity error occurred
+					  on the even word of the access, low bit
+					  indicated odd word caused the parity error */
+#define IOMMU_MFSR_BM       0x00001000 /* Error occurred while in boot mode */
+#define IOMMU_MFSR_C        0x00000800 /* Address causing error was marked cacheable */
+#define IOMMU_MFSR_RTYP     0x000000f0 /* Memory request transaction type */
+
+#define IOMMU_MID_SBAE      0x001f0000 /* SBus arbitration enable */
+#define IOMMU_MID_SE        0x00100000 /* Enables SCSI/ETHERNET arbitration */
+#define IOMMU_MID_SB3       0x00080000 /* Enable SBUS device 3 arbitration */
+#define IOMMU_MID_SB2       0x00040000 /* Enable SBUS device 2 arbitration */
+#define IOMMU_MID_SB1       0x00020000 /* Enable SBUS device 1 arbitration */
+#define IOMMU_MID_SB0       0x00010000 /* Enable SBUS device 0 arbitration */
+#define IOMMU_MID_MID       0x0000000f /* Module-id, hardcoded to 0x8 */
+
+/* The format of an iopte in the page tables */
+#define IOPTE_PAGE          0x07ffff00 /* Physical page number (PA[30:12]) */
+#define IOPTE_CACHE         0x00000080 /* Cached (in vme IOCACHE or Viking/MXCC) */
+#define IOPTE_WRITE         0x00000004 /* Writeable */
+#define IOPTE_VALID         0x00000002 /* IOPTE is valid */
+#define IOPTE_WAZ           0x00000001 /* Write as zeros */
+
+struct iommu_struct {
+	struct iommu_regs *regs;
+	iopte_t *page_table;
+	/* For convenience */
+	unsigned long start; /* First managed virtual address */
+	unsigned long end;   /* Last managed virtual address */
+
+	struct bit_map usemap;
+};
+
+static inline void iommu_invalidate(struct iommu_regs *regs)
+{
+	regs->tlbflush = 0;
+}
+
+static inline void iommu_invalidate_page(struct iommu_regs *regs, unsigned long ba)
+{
+	regs->pageflush = (ba & PAGE_MASK);
+}
+
+#endif /* !(_SPARC_IOMMU_H) */
diff --git a/include/asm-sparc/iommu_64.h b/include/asm-sparc/iommu_64.h
new file mode 100644
index 00000000000..d7b9afcba08
--- /dev/null
+++ b/include/asm-sparc/iommu_64.h
@@ -0,0 +1,62 @@
+/* iommu.h: Definitions for the sun5 IOMMU.
+ *
+ * Copyright (C) 1996, 1999, 2007 David S. Miller (davem@davemloft.net)
+ */
+#ifndef _SPARC64_IOMMU_H
+#define _SPARC64_IOMMU_H
+
+/* The format of an iopte in the page tables. */
+#define IOPTE_VALID   0x8000000000000000UL
+#define IOPTE_64K     0x2000000000000000UL
+#define IOPTE_STBUF   0x1000000000000000UL
+#define IOPTE_INTRA   0x0800000000000000UL
+#define IOPTE_CONTEXT 0x07ff800000000000UL
+#define IOPTE_PAGE    0x00007fffffffe000UL
+#define IOPTE_CACHE   0x0000000000000010UL
+#define IOPTE_WRITE   0x0000000000000002UL
+
+#define IOMMU_NUM_CTXS	4096
+
+struct iommu_arena {
+	unsigned long	*map;
+	unsigned int	hint;
+	unsigned int	limit;
+};
+
+struct iommu {
+	spinlock_t		lock;
+	struct iommu_arena	arena;
+	void			(*flush_all)(struct iommu *);
+	iopte_t			*page_table;
+	u32			page_table_map_base;
+	unsigned long		iommu_control;
+	unsigned long		iommu_tsbbase;
+	unsigned long		iommu_flush;
+	unsigned long		iommu_flushinv;
+	unsigned long		iommu_tags;
+	unsigned long		iommu_ctxflush;
+	unsigned long		write_complete_reg;
+	unsigned long		dummy_page;
+	unsigned long		dummy_page_pa;
+	unsigned long		ctx_lowest_free;
+	DECLARE_BITMAP(ctx_bitmap, IOMMU_NUM_CTXS);
+	u32			dma_addr_mask;
+};
+
+struct strbuf {
+	int			strbuf_enabled;
+	unsigned long		strbuf_control;
+	unsigned long		strbuf_pflush;
+	unsigned long		strbuf_fsync;
+	unsigned long		strbuf_ctxflush;
+	unsigned long		strbuf_ctxmatch_base;
+	unsigned long		strbuf_flushflag_pa;
+	volatile unsigned long *strbuf_flushflag;
+	volatile unsigned long	__flushflag_buf[(64+(64-1)) / sizeof(long)];
+};
+
+extern int iommu_table_init(struct iommu *iommu, int tsbsize,
+			    u32 dma_offset, u32 dma_addr_mask,
+			    int numa_node);
+
+#endif /* !(_SPARC64_IOMMU_H) */
diff --git a/include/asm-sparc/ipcbuf.h b/include/asm-sparc/ipcbuf.h
index 9bef02d04e4..037605d986e 100644
--- a/include/asm-sparc/ipcbuf.h
+++ b/include/asm-sparc/ipcbuf.h
@@ -1,31 +1,8 @@
-#ifndef _SPARC_IPCBUF_H
-#define _SPARC_IPCBUF_H
-
-/* 
- * The ipc64_perm structure for sparc architecture.
- * Note extra padding because this structure is passed back and forth
- * between kernel and user space.
- *
- * Pad space is left for:
- * - 32-bit mode
- * - 32-bit seq
- * - 2 miscellaneous 64-bit values (so that this structure matches
- *				    sparc64 ipc64_perm)
- */
-
-struct ipc64_perm
-{
-	__kernel_key_t		key;
-	__kernel_uid32_t	uid;
-	__kernel_gid32_t	gid;
-	__kernel_uid32_t	cuid;
-	__kernel_gid32_t	cgid;
-	unsigned short		__pad1;
-	__kernel_mode_t		mode;
-	unsigned short		__pad2;
-	unsigned short		seq;
-	unsigned long long	__unused1;
-	unsigned long long	__unused2;
-};
-
-#endif /* _SPARC_IPCBUF_H */
+#ifndef ___ASM_SPARC_IPCBUF_H
+#define ___ASM_SPARC_IPCBUF_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/ipcbuf_64.h>
+#else
+#include <asm-sparc/ipcbuf_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/ipcbuf_32.h b/include/asm-sparc/ipcbuf_32.h
new file mode 100644
index 00000000000..6387209518f
--- /dev/null
+++ b/include/asm-sparc/ipcbuf_32.h
@@ -0,0 +1,31 @@
+#ifndef _SPARC_IPCBUF_H
+#define _SPARC_IPCBUF_H
+
+/*
+ * The ipc64_perm structure for sparc architecture.
+ * Note extra padding because this structure is passed back and forth
+ * between kernel and user space.
+ *
+ * Pad space is left for:
+ * - 32-bit mode
+ * - 32-bit seq
+ * - 2 miscellaneous 64-bit values (so that this structure matches
+ *				    sparc64 ipc64_perm)
+ */
+
+struct ipc64_perm
+{
+	__kernel_key_t		key;
+	__kernel_uid32_t	uid;
+	__kernel_gid32_t	gid;
+	__kernel_uid32_t	cuid;
+	__kernel_gid32_t	cgid;
+	unsigned short		__pad1;
+	__kernel_mode_t		mode;
+	unsigned short		__pad2;
+	unsigned short		seq;
+	unsigned long long	__unused1;
+	unsigned long long	__unused2;
+};
+
+#endif /* _SPARC_IPCBUF_H */
diff --git a/include/asm-sparc/ipcbuf_64.h b/include/asm-sparc/ipcbuf_64.h
new file mode 100644
index 00000000000..a44b855b98d
--- /dev/null
+++ b/include/asm-sparc/ipcbuf_64.h
@@ -0,0 +1,28 @@
+#ifndef _SPARC64_IPCBUF_H
+#define _SPARC64_IPCBUF_H
+
+/*
+ * The ipc64_perm structure for sparc64 architecture.
+ * Note extra padding because this structure is passed back and forth
+ * between kernel and user space.
+ *
+ * Pad space is left for:
+ * - 32-bit seq
+ * - 2 miscellaneous 64-bit values
+ */
+
+struct ipc64_perm
+{
+	__kernel_key_t	key;
+	__kernel_uid_t	uid;
+	__kernel_gid_t	gid;
+	__kernel_uid_t	cuid;
+	__kernel_gid_t	cgid;
+	__kernel_mode_t	mode;
+	unsigned short	__pad1;
+	unsigned short	seq;
+	unsigned long	__unused1;
+	unsigned long	__unused2;
+};
+
+#endif /* _SPARC64_IPCBUF_H */
diff --git a/include/asm-sparc/irq.h b/include/asm-sparc/irq.h
index fe205cc444b..7af6bb4aa09 100644
--- a/include/asm-sparc/irq.h
+++ b/include/asm-sparc/irq.h
@@ -1,15 +1,8 @@
-/* irq.h: IRQ registers on the Sparc.
- *
- * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC_IRQ_H
-#define _SPARC_IRQ_H
-
-#include <linux/interrupt.h>
-
-#define NR_IRQS    16
-
-#define irq_canonicalize(irq)	(irq)
-
+#ifndef ___ASM_SPARC_IRQ_H
+#define ___ASM_SPARC_IRQ_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/irq_64.h>
+#else
+#include <asm-sparc/irq_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/irq_32.h b/include/asm-sparc/irq_32.h
new file mode 100644
index 00000000000..fe205cc444b
--- /dev/null
+++ b/include/asm-sparc/irq_32.h
@@ -0,0 +1,15 @@
+/* irq.h: IRQ registers on the Sparc.
+ *
+ * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC_IRQ_H
+#define _SPARC_IRQ_H
+
+#include <linux/interrupt.h>
+
+#define NR_IRQS    16
+
+#define irq_canonicalize(irq)	(irq)
+
+#endif
diff --git a/include/asm-sparc/irq_64.h b/include/asm-sparc/irq_64.h
new file mode 100644
index 00000000000..0bb9bf53174
--- /dev/null
+++ b/include/asm-sparc/irq_64.h
@@ -0,0 +1,93 @@
+/* irq.h: IRQ registers on the 64-bit Sparc.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1998 Jakub Jelinek (jj@ultra.linux.cz)
+ */
+
+#ifndef _SPARC64_IRQ_H
+#define _SPARC64_IRQ_H
+
+#include <linux/linkage.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/interrupt.h>
+#include <asm/pil.h>
+#include <asm/ptrace.h>
+
+/* IMAP/ICLR register defines */
+#define IMAP_VALID		0x80000000UL	/* IRQ Enabled		*/
+#define IMAP_TID_UPA		0x7c000000UL	/* UPA TargetID		*/
+#define IMAP_TID_JBUS		0x7c000000UL	/* JBUS TargetID	*/
+#define IMAP_TID_SHIFT		26
+#define IMAP_AID_SAFARI		0x7c000000UL	/* Safari AgentID	*/
+#define IMAP_AID_SHIFT		26
+#define IMAP_NID_SAFARI		0x03e00000UL	/* Safari NodeID	*/
+#define IMAP_NID_SHIFT		21
+#define IMAP_IGN		0x000007c0UL	/* IRQ Group Number	*/
+#define IMAP_INO		0x0000003fUL	/* IRQ Number		*/
+#define IMAP_INR		0x000007ffUL	/* Full interrupt number*/
+
+#define ICLR_IDLE		0x00000000UL	/* Idle state		*/
+#define ICLR_TRANSMIT		0x00000001UL	/* Transmit state	*/
+#define ICLR_PENDING		0x00000003UL	/* Pending state	*/
+
+/* The largest number of unique interrupt sources we support.
+ * If this needs to ever be larger than 255, you need to change
+ * the type of ino_bucket->virt_irq as appropriate.
+ *
+ * ino_bucket->virt_irq allocation is made during {sun4v_,}build_irq().
+ */
+#define NR_IRQS    255
+
+extern void irq_install_pre_handler(int virt_irq,
+				    void (*func)(unsigned int, void *, void *),
+				    void *arg1, void *arg2);
+#define irq_canonicalize(irq)	(irq)
+extern unsigned int build_irq(int inofixup, unsigned long iclr, unsigned long imap);
+extern unsigned int sun4v_build_irq(u32 devhandle, unsigned int devino);
+extern unsigned int sun4v_build_virq(u32 devhandle, unsigned int devino);
+extern unsigned int sun4v_build_msi(u32 devhandle, unsigned int *virt_irq_p,
+				    unsigned int msi_devino_start,
+				    unsigned int msi_devino_end);
+extern void sun4v_destroy_msi(unsigned int virt_irq);
+extern unsigned int sun4u_build_msi(u32 portid, unsigned int *virt_irq_p,
+				    unsigned int msi_devino_start,
+				    unsigned int msi_devino_end,
+				    unsigned long imap_base,
+				    unsigned long iclr_base);
+extern void sun4u_destroy_msi(unsigned int virt_irq);
+extern unsigned int sbus_build_irq(void *sbus, unsigned int ino);
+
+extern unsigned char virt_irq_alloc(unsigned int dev_handle,
+				    unsigned int dev_ino);
+#ifdef CONFIG_PCI_MSI
+extern void virt_irq_free(unsigned int virt_irq);
+#endif
+
+extern void __init init_IRQ(void);
+extern void fixup_irqs(void);
+
+static inline void set_softint(unsigned long bits)
+{
+	__asm__ __volatile__("wr	%0, 0x0, %%set_softint"
+			     : /* No outputs */
+			     : "r" (bits));
+}
+
+static inline void clear_softint(unsigned long bits)
+{
+	__asm__ __volatile__("wr	%0, 0x0, %%clear_softint"
+			     : /* No outputs */
+			     : "r" (bits));
+}
+
+static inline unsigned long get_softint(void)
+{
+	unsigned long retval;
+
+	__asm__ __volatile__("rd	%%softint, %0"
+			     : "=r" (retval));
+	return retval;
+}
+
+#endif
diff --git a/include/asm-sparc/irqflags.h b/include/asm-sparc/irqflags.h
index db398fb3282..c6402b187e2 100644
--- a/include/asm-sparc/irqflags.h
+++ b/include/asm-sparc/irqflags.h
@@ -1,39 +1,8 @@
-/*
- * include/asm-sparc/irqflags.h
- *
- * IRQ flags handling
- *
- * This file gets included from lowlevel asm headers too, to provide
- * wrapped versions of the local_irq_*() APIs, based on the
- * raw_local_irq_*() functions from the lowlevel headers.
- */
-#ifndef _ASM_IRQFLAGS_H
-#define _ASM_IRQFLAGS_H
-
-#ifndef __ASSEMBLY__
-
-extern void raw_local_irq_restore(unsigned long);
-extern unsigned long __raw_local_irq_save(void);
-extern void raw_local_irq_enable(void);
-
-static inline unsigned long getipl(void)
-{
-        unsigned long retval;
-
-        __asm__ __volatile__("rd        %%psr, %0" : "=r" (retval));
-        return retval;
-}
-
-#define raw_local_save_flags(flags) ((flags) = getipl())
-#define raw_local_irq_save(flags)   ((flags) = __raw_local_irq_save())
-#define raw_local_irq_disable()     ((void) __raw_local_irq_save())
-#define raw_irqs_disabled()         ((getipl() & PSR_PIL) != 0)
-
-static inline int raw_irqs_disabled_flags(unsigned long flags)
-{
-        return ((flags & PSR_PIL) != 0);
-}
-
-#endif /* (__ASSEMBLY__) */
-
-#endif /* !(_ASM_IRQFLAGS_H) */
+#ifndef ___ASM_SPARC_IRQFLAGS_H
+#define ___ASM_SPARC_IRQFLAGS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/irqflags_64.h>
+#else
+#include <asm-sparc/irqflags_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/irqflags_32.h b/include/asm-sparc/irqflags_32.h
new file mode 100644
index 00000000000..db398fb3282
--- /dev/null
+++ b/include/asm-sparc/irqflags_32.h
@@ -0,0 +1,39 @@
+/*
+ * include/asm-sparc/irqflags.h
+ *
+ * IRQ flags handling
+ *
+ * This file gets included from lowlevel asm headers too, to provide
+ * wrapped versions of the local_irq_*() APIs, based on the
+ * raw_local_irq_*() functions from the lowlevel headers.
+ */
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#ifndef __ASSEMBLY__
+
+extern void raw_local_irq_restore(unsigned long);
+extern unsigned long __raw_local_irq_save(void);
+extern void raw_local_irq_enable(void);
+
+static inline unsigned long getipl(void)
+{
+        unsigned long retval;
+
+        __asm__ __volatile__("rd        %%psr, %0" : "=r" (retval));
+        return retval;
+}
+
+#define raw_local_save_flags(flags) ((flags) = getipl())
+#define raw_local_irq_save(flags)   ((flags) = __raw_local_irq_save())
+#define raw_local_irq_disable()     ((void) __raw_local_irq_save())
+#define raw_irqs_disabled()         ((getipl() & PSR_PIL) != 0)
+
+static inline int raw_irqs_disabled_flags(unsigned long flags)
+{
+        return ((flags & PSR_PIL) != 0);
+}
+
+#endif /* (__ASSEMBLY__) */
+
+#endif /* !(_ASM_IRQFLAGS_H) */
diff --git a/include/asm-sparc/irqflags_64.h b/include/asm-sparc/irqflags_64.h
new file mode 100644
index 00000000000..024fc54d068
--- /dev/null
+++ b/include/asm-sparc/irqflags_64.h
@@ -0,0 +1,89 @@
+/*
+ * include/asm-sparc64/irqflags.h
+ *
+ * IRQ flags handling
+ *
+ * This file gets included from lowlevel asm headers too, to provide
+ * wrapped versions of the local_irq_*() APIs, based on the
+ * raw_local_irq_*() functions from the lowlevel headers.
+ */
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#ifndef __ASSEMBLY__
+
+static inline unsigned long __raw_local_save_flags(void)
+{
+	unsigned long flags;
+
+	__asm__ __volatile__(
+		"rdpr	%%pil, %0"
+		: "=r" (flags)
+	);
+
+	return flags;
+}
+
+#define raw_local_save_flags(flags) \
+		do { (flags) = __raw_local_save_flags(); } while (0)
+
+static inline void raw_local_irq_restore(unsigned long flags)
+{
+	__asm__ __volatile__(
+		"wrpr	%0, %%pil"
+		: /* no output */
+		: "r" (flags)
+		: "memory"
+	);
+}
+
+static inline void raw_local_irq_disable(void)
+{
+	__asm__ __volatile__(
+		"wrpr	15, %%pil"
+		: /* no outputs */
+		: /* no inputs */
+		: "memory"
+	);
+}
+
+static inline void raw_local_irq_enable(void)
+{
+	__asm__ __volatile__(
+		"wrpr	0, %%pil"
+		: /* no outputs */
+		: /* no inputs */
+		: "memory"
+	);
+}
+
+static inline int raw_irqs_disabled_flags(unsigned long flags)
+{
+	return (flags > 0);
+}
+
+static inline int raw_irqs_disabled(void)
+{
+	unsigned long flags = __raw_local_save_flags();
+
+	return raw_irqs_disabled_flags(flags);
+}
+
+/*
+ * For spinlocks, etc:
+ */
+static inline unsigned long __raw_local_irq_save(void)
+{
+	unsigned long flags = __raw_local_save_flags();
+
+	raw_local_irq_disable();
+
+	return flags;
+}
+
+#define raw_local_irq_save(flags) \
+		do { (flags) = __raw_local_irq_save(); } while (0)
+
+#endif /* (__ASSEMBLY__) */
+
+#endif /* !(_ASM_IRQFLAGS_H) */
diff --git a/include/asm-sparc/kdebug.h b/include/asm-sparc/kdebug.h
index f69fe7d84b3..fe07d00d053 100644
--- a/include/asm-sparc/kdebug.h
+++ b/include/asm-sparc/kdebug.h
@@ -1,73 +1,8 @@
-/*
- * kdebug.h:  Defines and definitions for debugging the Linux kernel
- *            under various kernel debuggers.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-#ifndef _SPARC_KDEBUG_H
-#define _SPARC_KDEBUG_H
-
-#include <asm/openprom.h>
-#include <asm/vaddrs.h>
-
-/* Breakpoints are enter through trap table entry 126.  So in sparc assembly
- * if you want to drop into the debugger you do:
- *
- * t DEBUG_BP_TRAP
- */
-
-#define DEBUG_BP_TRAP     126
-
-#ifndef __ASSEMBLY__
-/* The debug vector is passed in %o1 at boot time.  It is a pointer to
- * a structure in the debuggers address space.  Here is its format.
- */
-
-typedef unsigned int (*debugger_funct)(void);
-
-struct kernel_debug {
-	/* First the entry point into the debugger.  You jump here
-	 * to give control over to the debugger.
-	 */
-	unsigned long kdebug_entry;
-	unsigned long kdebug_trapme;   /* Figure out later... */
-	/* The following is the number of pages that the debugger has
-	 * taken from to total pool.
-	 */
-	unsigned long *kdebug_stolen_pages;
-	/* Ok, after you remap yourself and/or change the trap table
-	 * from what you were left with at boot time you have to call
-	 * this synchronization function so the debugger can check out
-	 * what you have done.
-	 */
-	debugger_funct teach_debugger;
-}; /* I think that is it... */
-
-extern struct kernel_debug *linux_dbvec;
-
-/* Use this macro in C-code to enter the debugger. */
-static inline void sp_enter_debugger(void)
-{
-	__asm__ __volatile__("jmpl %0, %%o7\n\t"
-			     "nop\n\t" : :
-			     "r" (linux_dbvec) : "o7", "memory");
-}
-
-#define SP_ENTER_DEBUGGER do { \
-	     if((linux_dbvec!=0) && ((*(short *)linux_dbvec)!=-1)) \
-	       sp_enter_debugger(); \
-		       } while(0)
-
-enum die_val {
-	DIE_UNUSED,
-};
-
-#endif /* !(__ASSEMBLY__) */
-
-/* Some nice offset defines for assembler code. */
-#define KDEBUG_ENTRY_OFF    0x0
-#define KDEBUG_DUNNO_OFF    0x4
-#define KDEBUG_DUNNO2_OFF   0x8
-#define KDEBUG_TEACH_OFF    0xc
-
-#endif /* !(_SPARC_KDEBUG_H) */
+#ifndef ___ASM_SPARC_KDEBUG_H
+#define ___ASM_SPARC_KDEBUG_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/kdebug_64.h>
+#else
+#include <asm-sparc/kdebug_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/kdebug_32.h b/include/asm-sparc/kdebug_32.h
new file mode 100644
index 00000000000..f69fe7d84b3
--- /dev/null
+++ b/include/asm-sparc/kdebug_32.h
@@ -0,0 +1,73 @@
+/*
+ * kdebug.h:  Defines and definitions for debugging the Linux kernel
+ *            under various kernel debuggers.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+#ifndef _SPARC_KDEBUG_H
+#define _SPARC_KDEBUG_H
+
+#include <asm/openprom.h>
+#include <asm/vaddrs.h>
+
+/* Breakpoints are enter through trap table entry 126.  So in sparc assembly
+ * if you want to drop into the debugger you do:
+ *
+ * t DEBUG_BP_TRAP
+ */
+
+#define DEBUG_BP_TRAP     126
+
+#ifndef __ASSEMBLY__
+/* The debug vector is passed in %o1 at boot time.  It is a pointer to
+ * a structure in the debuggers address space.  Here is its format.
+ */
+
+typedef unsigned int (*debugger_funct)(void);
+
+struct kernel_debug {
+	/* First the entry point into the debugger.  You jump here
+	 * to give control over to the debugger.
+	 */
+	unsigned long kdebug_entry;
+	unsigned long kdebug_trapme;   /* Figure out later... */
+	/* The following is the number of pages that the debugger has
+	 * taken from to total pool.
+	 */
+	unsigned long *kdebug_stolen_pages;
+	/* Ok, after you remap yourself and/or change the trap table
+	 * from what you were left with at boot time you have to call
+	 * this synchronization function so the debugger can check out
+	 * what you have done.
+	 */
+	debugger_funct teach_debugger;
+}; /* I think that is it... */
+
+extern struct kernel_debug *linux_dbvec;
+
+/* Use this macro in C-code to enter the debugger. */
+static inline void sp_enter_debugger(void)
+{
+	__asm__ __volatile__("jmpl %0, %%o7\n\t"
+			     "nop\n\t" : :
+			     "r" (linux_dbvec) : "o7", "memory");
+}
+
+#define SP_ENTER_DEBUGGER do { \
+	     if((linux_dbvec!=0) && ((*(short *)linux_dbvec)!=-1)) \
+	       sp_enter_debugger(); \
+		       } while(0)
+
+enum die_val {
+	DIE_UNUSED,
+};
+
+#endif /* !(__ASSEMBLY__) */
+
+/* Some nice offset defines for assembler code. */
+#define KDEBUG_ENTRY_OFF    0x0
+#define KDEBUG_DUNNO_OFF    0x4
+#define KDEBUG_DUNNO2_OFF   0x8
+#define KDEBUG_TEACH_OFF    0xc
+
+#endif /* !(_SPARC_KDEBUG_H) */
diff --git a/include/asm-sparc/kdebug_64.h b/include/asm-sparc/kdebug_64.h
new file mode 100644
index 00000000000..f905b773235
--- /dev/null
+++ b/include/asm-sparc/kdebug_64.h
@@ -0,0 +1,19 @@
+#ifndef _SPARC64_KDEBUG_H
+#define _SPARC64_KDEBUG_H
+
+struct pt_regs;
+
+extern void bad_trap(struct pt_regs *, long);
+
+/* Grossly misnamed. */
+enum die_val {
+	DIE_OOPS = 1,
+	DIE_DEBUG,	/* ta 0x70 */
+	DIE_DEBUG_2,	/* ta 0x71 */
+	DIE_DIE,
+	DIE_TRAP,
+	DIE_TRAP_TL1,
+	DIE_CALL,
+};
+
+#endif
diff --git a/include/asm-sparc/kmap_types.h b/include/asm-sparc/kmap_types.h
index e215f710497..602f5e034f7 100644
--- a/include/asm-sparc/kmap_types.h
+++ b/include/asm-sparc/kmap_types.h
@@ -1,6 +1,10 @@
 #ifndef _ASM_KMAP_TYPES_H
 #define _ASM_KMAP_TYPES_H
 
+/* Dummy header just to define km_type.  None of this
+ * is actually used on sparc.  -DaveM
+ */
+
 enum km_type {
 	KM_BOUNCE_READ,
 	KM_SKB_SUNRPC_DATA,
diff --git a/include/asm-sparc/kprobes.h b/include/asm-sparc/kprobes.h
new file mode 100644
index 00000000000..5879d71afda
--- /dev/null
+++ b/include/asm-sparc/kprobes.h
@@ -0,0 +1,49 @@
+#ifndef _SPARC64_KPROBES_H
+#define _SPARC64_KPROBES_H
+
+#include <linux/types.h>
+#include <linux/percpu.h>
+
+typedef u32 kprobe_opcode_t;
+
+#define BREAKPOINT_INSTRUCTION   0x91d02070 /* ta 0x70 */
+#define BREAKPOINT_INSTRUCTION_2 0x91d02071 /* ta 0x71 */
+#define MAX_INSN_SIZE 2
+
+#define kretprobe_blacklist_size 0
+
+#define arch_remove_kprobe(p)	do {} while (0)
+
+#define flush_insn_slot(p)		\
+do { 	flushi(&(p)->ainsn.insn[0]);	\
+	flushi(&(p)->ainsn.insn[1]);	\
+} while (0)
+
+void kretprobe_trampoline(void);
+
+/* Architecture specific copy of original instruction*/
+struct arch_specific_insn {
+	/* copy of the original instruction */
+	kprobe_opcode_t insn[MAX_INSN_SIZE];
+};
+
+struct prev_kprobe {
+	struct kprobe *kp;
+	unsigned long status;
+	unsigned long orig_tnpc;
+	unsigned long orig_tstate_pil;
+};
+
+/* per-cpu kprobe control block */
+struct kprobe_ctlblk {
+	unsigned long kprobe_status;
+	unsigned long kprobe_orig_tnpc;
+	unsigned long kprobe_orig_tstate_pil;
+	struct pt_regs jprobe_saved_regs;
+	struct prev_kprobe prev_kprobe;
+};
+
+extern int kprobe_exceptions_notify(struct notifier_block *self,
+				    unsigned long val, void *data);
+extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
+#endif /* _SPARC64_KPROBES_H */
diff --git a/include/asm-sparc/ldc.h b/include/asm-sparc/ldc.h
new file mode 100644
index 00000000000..bdb524a7b81
--- /dev/null
+++ b/include/asm-sparc/ldc.h
@@ -0,0 +1,138 @@
+#ifndef _SPARC64_LDC_H
+#define _SPARC64_LDC_H
+
+#include <asm/hypervisor.h>
+
+extern int ldom_domaining_enabled;
+extern void ldom_set_var(const char *var, const char *value);
+extern void ldom_reboot(const char *boot_command);
+extern void ldom_power_off(void);
+
+/* The event handler will be evoked when link state changes
+ * or data becomes available on the receive side.
+ *
+ * For non-RAW links, if the LDC_EVENT_RESET event arrives the
+ * driver should reset all of it's internal state and reinvoke
+ * ldc_connect() to try and bring the link up again.
+ *
+ * For RAW links, ldc_connect() is not used.  Instead the driver
+ * just waits for the LDC_EVENT_UP event.
+ */
+struct ldc_channel_config {
+	void (*event)(void *arg, int event);
+
+	u32			mtu;
+	unsigned int		rx_irq;
+	unsigned int		tx_irq;
+	u8			mode;
+#define LDC_MODE_RAW		0x00
+#define LDC_MODE_UNRELIABLE	0x01
+#define LDC_MODE_RESERVED	0x02
+#define LDC_MODE_STREAM		0x03
+
+	u8			debug;
+#define LDC_DEBUG_HS		0x01
+#define LDC_DEBUG_STATE		0x02
+#define LDC_DEBUG_RX		0x04
+#define LDC_DEBUG_TX		0x08
+#define LDC_DEBUG_DATA		0x10
+};
+
+#define LDC_EVENT_RESET		0x01
+#define LDC_EVENT_UP		0x02
+#define LDC_EVENT_DATA_READY	0x04
+
+#define LDC_STATE_INVALID	0x00
+#define LDC_STATE_INIT		0x01
+#define LDC_STATE_BOUND		0x02
+#define LDC_STATE_READY		0x03
+#define LDC_STATE_CONNECTED	0x04
+
+struct ldc_channel;
+
+/* Allocate state for a channel.  */
+extern struct ldc_channel *ldc_alloc(unsigned long id,
+				     const struct ldc_channel_config *cfgp,
+				     void *event_arg);
+
+/* Shut down and free state for a channel.  */
+extern void ldc_free(struct ldc_channel *lp);
+
+/* Register TX and RX queues of the link with the hypervisor.  */
+extern int ldc_bind(struct ldc_channel *lp, const char *name);
+
+/* For non-RAW protocols we need to complete a handshake before
+ * communication can proceed.  ldc_connect() does that, if the
+ * handshake completes successfully, an LDC_EVENT_UP event will
+ * be sent up to the driver.
+ */
+extern int ldc_connect(struct ldc_channel *lp);
+extern int ldc_disconnect(struct ldc_channel *lp);
+
+extern int ldc_state(struct ldc_channel *lp);
+
+/* Read and write operations.  Only valid when the link is up.  */
+extern int ldc_write(struct ldc_channel *lp, const void *buf,
+		     unsigned int size);
+extern int ldc_read(struct ldc_channel *lp, void *buf, unsigned int size);
+
+#define LDC_MAP_SHADOW	0x01
+#define LDC_MAP_DIRECT	0x02
+#define LDC_MAP_IO	0x04
+#define LDC_MAP_R	0x08
+#define LDC_MAP_W	0x10
+#define LDC_MAP_X	0x20
+#define LDC_MAP_RW	(LDC_MAP_R | LDC_MAP_W)
+#define LDC_MAP_RWX	(LDC_MAP_R | LDC_MAP_W | LDC_MAP_X)
+#define LDC_MAP_ALL	0x03f
+
+struct ldc_trans_cookie {
+	u64			cookie_addr;
+	u64			cookie_size;
+};
+
+struct scatterlist;
+extern int ldc_map_sg(struct ldc_channel *lp,
+		      struct scatterlist *sg, int num_sg,
+		      struct ldc_trans_cookie *cookies, int ncookies,
+		      unsigned int map_perm);
+
+extern int ldc_map_single(struct ldc_channel *lp,
+			  void *buf, unsigned int len,
+			  struct ldc_trans_cookie *cookies, int ncookies,
+			  unsigned int map_perm);
+
+extern void ldc_unmap(struct ldc_channel *lp, struct ldc_trans_cookie *cookies,
+		      int ncookies);
+
+extern int ldc_copy(struct ldc_channel *lp, int copy_dir,
+		    void *buf, unsigned int len, unsigned long offset,
+		    struct ldc_trans_cookie *cookies, int ncookies);
+
+static inline int ldc_get_dring_entry(struct ldc_channel *lp,
+				      void *buf, unsigned int len,
+				      unsigned long offset,
+				      struct ldc_trans_cookie *cookies,
+				      int ncookies)
+{
+	return ldc_copy(lp, LDC_COPY_IN, buf, len, offset, cookies, ncookies);
+}
+
+static inline int ldc_put_dring_entry(struct ldc_channel *lp,
+				      void *buf, unsigned int len,
+				      unsigned long offset,
+				      struct ldc_trans_cookie *cookies,
+				      int ncookies)
+{
+	return ldc_copy(lp, LDC_COPY_OUT, buf, len, offset, cookies, ncookies);
+}
+
+extern void *ldc_alloc_exp_dring(struct ldc_channel *lp, unsigned int len,
+				 struct ldc_trans_cookie *cookies,
+				 int *ncookies, unsigned int map_perm);
+
+extern void ldc_free_exp_dring(struct ldc_channel *lp, void *buf,
+			       unsigned int len,
+			       struct ldc_trans_cookie *cookies, int ncookies);
+
+#endif /* _SPARC64_LDC_H */
diff --git a/include/asm-sparc/lmb.h b/include/asm-sparc/lmb.h
new file mode 100644
index 00000000000..6a352cbcf52
--- /dev/null
+++ b/include/asm-sparc/lmb.h
@@ -0,0 +1,10 @@
+#ifndef _SPARC64_LMB_H
+#define _SPARC64_LMB_H
+
+#include <asm/oplib.h>
+
+#define LMB_DBG(fmt...) prom_printf(fmt)
+
+#define LMB_REAL_LIMIT	0
+
+#endif /* !(_SPARC64_LMB_H) */
diff --git a/include/asm-sparc/lsu.h b/include/asm-sparc/lsu.h
new file mode 100644
index 00000000000..7190f8de90a
--- /dev/null
+++ b/include/asm-sparc/lsu.h
@@ -0,0 +1,19 @@
+#ifndef _SPARC64_LSU_H
+#define _SPARC64_LSU_H
+
+#include <linux/const.h>
+
+/* LSU Control Register */
+#define LSU_CONTROL_PM _AC(0x000001fe00000000,UL) /* Phys-watchpoint byte mask*/
+#define LSU_CONTROL_VM _AC(0x00000001fe000000,UL) /* Virt-watchpoint byte mask*/
+#define LSU_CONTROL_PR _AC(0x0000000001000000,UL) /* Phys-rd watchpoint enable*/
+#define LSU_CONTROL_PW _AC(0x0000000000800000,UL) /* Phys-wr watchpoint enable*/
+#define LSU_CONTROL_VR _AC(0x0000000000400000,UL) /* Virt-rd watchpoint enable*/
+#define LSU_CONTROL_VW _AC(0x0000000000200000,UL) /* Virt-wr watchpoint enable*/
+#define LSU_CONTROL_FM _AC(0x00000000000ffff0,UL) /* Parity mask enables.     */
+#define LSU_CONTROL_DM _AC(0x0000000000000008,UL) /* Data MMU enable.         */
+#define LSU_CONTROL_IM _AC(0x0000000000000004,UL) /* Instruction MMU enable.  */
+#define LSU_CONTROL_DC _AC(0x0000000000000002,UL) /* Data cache enable.       */
+#define LSU_CONTROL_IC _AC(0x0000000000000001,UL) /* Instruction cache enable.*/
+
+#endif /* !(_SPARC64_LSU_H) */
diff --git a/include/asm-sparc/machines.h b/include/asm-sparc/machines.h
index d6c6bf83620..c28c2f24879 100644
--- a/include/asm-sparc/machines.h
+++ b/include/asm-sparc/machines.h
@@ -17,8 +17,6 @@ struct Sun_Machine_Models {
  */
 #define NUM_SUN_MACHINES   15
 
-extern struct Sun_Machine_Models Sun_Machines[NUM_SUN_MACHINES];
-
 /* The machine type in the idprom area looks like this:
  *
  * ---------------
diff --git a/include/asm-sparc/mbus.h b/include/asm-sparc/mbus.h
index bb5ae614b16..69f07a022ee 100644
--- a/include/asm-sparc/mbus.h
+++ b/include/asm-sparc/mbus.h
@@ -43,8 +43,6 @@ extern unsigned int viking_rev, swift_rev, cypress_rev;
 #define HWBUG_SUPERSCALAR_BAD        0x00000080
 #define HWBUG_PACINIT_BITROT         0x00000100
 
-extern unsigned int hwbug_bitmask;
-
 /* First the module type values. To find out which you have, just load
  * the mmu control register from ASI_M_MMUREG alternate address space and
  * shift the value right 28 bits.
diff --git a/include/asm-sparc/mc146818rtc.h b/include/asm-sparc/mc146818rtc.h
index fa7eac92658..9ab65c21e9e 100644
--- a/include/asm-sparc/mc146818rtc.h
+++ b/include/asm-sparc/mc146818rtc.h
@@ -1,29 +1,8 @@
-/*
- * Machine dependent access functions for RTC registers.
- */
-#ifndef __ASM_SPARC_MC146818RTC_H
-#define __ASM_SPARC_MC146818RTC_H
-
-#include <asm/io.h>
-
-#ifndef RTC_PORT
-#define RTC_PORT(x)	(0x70 + (x))
-#define RTC_ALWAYS_BCD	1	/* RTC operates in binary mode */
+#ifndef ___ASM_SPARC_MC146818RTC_H
+#define ___ASM_SPARC_MC146818RTC_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/mc146818rtc_64.h>
+#else
+#include <asm-sparc/mc146818rtc_32.h>
+#endif
 #endif
-
-/*
- * The yet supported machines all access the RTC index register via
- * an ISA port access but the way to access the date register differs ...
- */
-#define CMOS_READ(addr) ({ \
-outb_p((addr),RTC_PORT(0)); \
-inb_p(RTC_PORT(1)); \
-})
-#define CMOS_WRITE(val, addr) ({ \
-outb_p((addr),RTC_PORT(0)); \
-outb_p((val),RTC_PORT(1)); \
-})
-
-#define RTC_IRQ 8
-
-#endif /* __ASM_SPARC_MC146818RTC_H */
diff --git a/include/asm-sparc/mc146818rtc_32.h b/include/asm-sparc/mc146818rtc_32.h
new file mode 100644
index 00000000000..fa7eac92658
--- /dev/null
+++ b/include/asm-sparc/mc146818rtc_32.h
@@ -0,0 +1,29 @@
+/*
+ * Machine dependent access functions for RTC registers.
+ */
+#ifndef __ASM_SPARC_MC146818RTC_H
+#define __ASM_SPARC_MC146818RTC_H
+
+#include <asm/io.h>
+
+#ifndef RTC_PORT
+#define RTC_PORT(x)	(0x70 + (x))
+#define RTC_ALWAYS_BCD	1	/* RTC operates in binary mode */
+#endif
+
+/*
+ * The yet supported machines all access the RTC index register via
+ * an ISA port access but the way to access the date register differs ...
+ */
+#define CMOS_READ(addr) ({ \
+outb_p((addr),RTC_PORT(0)); \
+inb_p(RTC_PORT(1)); \
+})
+#define CMOS_WRITE(val, addr) ({ \
+outb_p((addr),RTC_PORT(0)); \
+outb_p((val),RTC_PORT(1)); \
+})
+
+#define RTC_IRQ 8
+
+#endif /* __ASM_SPARC_MC146818RTC_H */
diff --git a/include/asm-sparc/mc146818rtc_64.h b/include/asm-sparc/mc146818rtc_64.h
new file mode 100644
index 00000000000..e9c0fcc25c6
--- /dev/null
+++ b/include/asm-sparc/mc146818rtc_64.h
@@ -0,0 +1,34 @@
+/*
+ * Machine dependent access functions for RTC registers.
+ */
+#ifndef __ASM_SPARC64_MC146818RTC_H
+#define __ASM_SPARC64_MC146818RTC_H
+
+#include <asm/io.h>
+
+#ifndef RTC_PORT
+#ifdef CONFIG_PCI
+extern unsigned long ds1287_regs;
+#else
+#define ds1287_regs (0UL)
+#endif
+#define RTC_PORT(x)	(ds1287_regs + (x))
+#define RTC_ALWAYS_BCD	0
+#endif
+
+/*
+ * The yet supported machines all access the RTC index register via
+ * an ISA port access but the way to access the date register differs ...
+ */
+#define CMOS_READ(addr) ({ \
+outb_p((addr),RTC_PORT(0)); \
+inb_p(RTC_PORT(1)); \
+})
+#define CMOS_WRITE(val, addr) ({ \
+outb_p((addr),RTC_PORT(0)); \
+outb_p((val),RTC_PORT(1)); \
+})
+
+#define RTC_IRQ 8
+
+#endif /* __ASM_SPARC64_MC146818RTC_H */
diff --git a/include/asm-sparc/mdesc.h b/include/asm-sparc/mdesc.h
new file mode 100644
index 00000000000..1acc7272e53
--- /dev/null
+++ b/include/asm-sparc/mdesc.h
@@ -0,0 +1,78 @@
+#ifndef _SPARC64_MDESC_H
+#define _SPARC64_MDESC_H
+
+#include <linux/types.h>
+#include <linux/cpumask.h>
+#include <asm/prom.h>
+
+struct mdesc_handle;
+
+/* Machine description operations are to be surrounded by grab and
+ * release calls.  The mdesc_handle returned from the grab is
+ * the first argument to all of the operational calls that work
+ * on mdescs.
+ */
+extern struct mdesc_handle *mdesc_grab(void);
+extern void mdesc_release(struct mdesc_handle *);
+
+#define MDESC_NODE_NULL		(~(u64)0)
+
+extern u64 mdesc_node_by_name(struct mdesc_handle *handle,
+			      u64 from_node, const char *name);
+#define mdesc_for_each_node_by_name(__hdl, __node, __name) \
+	for (__node = mdesc_node_by_name(__hdl, MDESC_NODE_NULL, __name); \
+	     (__node) != MDESC_NODE_NULL; \
+	     __node = mdesc_node_by_name(__hdl, __node, __name))
+
+/* Access to property values returned from mdesc_get_property() are
+ * only valid inside of a mdesc_grab()/mdesc_release() sequence.
+ * Once mdesc_release() is called, the memory backed up by these
+ * pointers may reference freed up memory.
+ *
+ * Therefore callers must make copies of any property values
+ * they need.
+ *
+ * These same rules apply to mdesc_node_name().
+ */
+extern const void *mdesc_get_property(struct mdesc_handle *handle,
+				      u64 node, const char *name, int *lenp);
+extern const char *mdesc_node_name(struct mdesc_handle *hp, u64 node);
+
+/* MD arc iteration, the standard sequence is:
+ *
+ *	unsigned long arc;
+ *	mdesc_for_each_arc(arc, handle, node, MDESC_ARC_TYPE_{FWD,BACK}) {
+ *		unsigned long target = mdesc_arc_target(handle, arc);
+ *		...
+ *	}
+ */
+
+#define MDESC_ARC_TYPE_FWD	"fwd"
+#define MDESC_ARC_TYPE_BACK	"back"
+
+extern u64 mdesc_next_arc(struct mdesc_handle *handle, u64 from,
+			  const char *arc_type);
+#define mdesc_for_each_arc(__arc, __hdl, __node, __type) \
+	for (__arc = mdesc_next_arc(__hdl, __node, __type); \
+	     (__arc) != MDESC_NODE_NULL; \
+	     __arc = mdesc_next_arc(__hdl, __arc, __type))
+
+extern u64 mdesc_arc_target(struct mdesc_handle *hp, u64 arc);
+
+extern void mdesc_update(void);
+
+struct mdesc_notifier_client {
+	void (*add)(struct mdesc_handle *handle, u64 node);
+	void (*remove)(struct mdesc_handle *handle, u64 node);
+
+	const char			*node_name;
+	struct mdesc_notifier_client	*next;
+};
+
+extern void mdesc_register_notifier(struct mdesc_notifier_client *client);
+
+extern void mdesc_fill_in_cpu_data(cpumask_t mask);
+
+extern void sun4v_mdesc_init(void);
+
+#endif
diff --git a/include/asm-sparc/mmu.h b/include/asm-sparc/mmu.h
index ccd36d26615..ee66bf6dcbd 100644
--- a/include/asm-sparc/mmu.h
+++ b/include/asm-sparc/mmu.h
@@ -1,7 +1,8 @@
-#ifndef __MMU_H
-#define __MMU_H
-
-/* Default "unsigned long" context */
-typedef unsigned long mm_context_t;
-
+#ifndef ___ASM_SPARC_MMU_H
+#define ___ASM_SPARC_MMU_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/mmu_64.h>
+#else
+#include <asm-sparc/mmu_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/mmu_32.h b/include/asm-sparc/mmu_32.h
new file mode 100644
index 00000000000..ccd36d26615
--- /dev/null
+++ b/include/asm-sparc/mmu_32.h
@@ -0,0 +1,7 @@
+#ifndef __MMU_H
+#define __MMU_H
+
+/* Default "unsigned long" context */
+typedef unsigned long mm_context_t;
+
+#endif
diff --git a/include/asm-sparc/mmu_64.h b/include/asm-sparc/mmu_64.h
new file mode 100644
index 00000000000..9067dc50053
--- /dev/null
+++ b/include/asm-sparc/mmu_64.h
@@ -0,0 +1,123 @@
+#ifndef __MMU_H
+#define __MMU_H
+
+#include <linux/const.h>
+#include <asm/page.h>
+#include <asm/hypervisor.h>
+
+#define CTX_NR_BITS		13
+
+#define TAG_CONTEXT_BITS	((_AC(1,UL) << CTX_NR_BITS) - _AC(1,UL))
+
+/* UltraSPARC-III+ and later have a feature whereby you can
+ * select what page size the various Data-TLB instances in the
+ * chip.  In order to gracefully support this, we put the version
+ * field in a spot outside of the areas of the context register
+ * where this parameter is specified.
+ */
+#define CTX_VERSION_SHIFT	22
+#define CTX_VERSION_MASK	((~0UL) << CTX_VERSION_SHIFT)
+
+#define CTX_PGSZ_8KB		_AC(0x0,UL)
+#define CTX_PGSZ_64KB		_AC(0x1,UL)
+#define CTX_PGSZ_512KB		_AC(0x2,UL)
+#define CTX_PGSZ_4MB		_AC(0x3,UL)
+#define CTX_PGSZ_BITS		_AC(0x7,UL)
+#define CTX_PGSZ0_NUC_SHIFT	61
+#define CTX_PGSZ1_NUC_SHIFT	58
+#define CTX_PGSZ0_SHIFT		16
+#define CTX_PGSZ1_SHIFT		19
+#define CTX_PGSZ_MASK		((CTX_PGSZ_BITS << CTX_PGSZ0_SHIFT) | \
+				 (CTX_PGSZ_BITS << CTX_PGSZ1_SHIFT))
+
+#if defined(CONFIG_SPARC64_PAGE_SIZE_8KB)
+#define CTX_PGSZ_BASE	CTX_PGSZ_8KB
+#elif defined(CONFIG_SPARC64_PAGE_SIZE_64KB)
+#define CTX_PGSZ_BASE	CTX_PGSZ_64KB
+#else
+#error No page size specified in kernel configuration
+#endif
+
+#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
+#define CTX_PGSZ_HUGE		CTX_PGSZ_4MB
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
+#define CTX_PGSZ_HUGE		CTX_PGSZ_512KB
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
+#define CTX_PGSZ_HUGE		CTX_PGSZ_64KB
+#endif
+
+#define CTX_PGSZ_KERN	CTX_PGSZ_4MB
+
+/* Thus, when running on UltraSPARC-III+ and later, we use the following
+ * PRIMARY_CONTEXT register values for the kernel context.
+ */
+#define CTX_CHEETAH_PLUS_NUC \
+	((CTX_PGSZ_KERN << CTX_PGSZ0_NUC_SHIFT) | \
+	 (CTX_PGSZ_BASE << CTX_PGSZ1_NUC_SHIFT))
+
+#define CTX_CHEETAH_PLUS_CTX0 \
+	((CTX_PGSZ_KERN << CTX_PGSZ0_SHIFT) | \
+	 (CTX_PGSZ_BASE << CTX_PGSZ1_SHIFT))
+
+/* If you want "the TLB context number" use CTX_NR_MASK.  If you
+ * want "the bits I program into the context registers" use
+ * CTX_HW_MASK.
+ */
+#define CTX_NR_MASK		TAG_CONTEXT_BITS
+#define CTX_HW_MASK		(CTX_NR_MASK | CTX_PGSZ_MASK)
+
+#define CTX_FIRST_VERSION	((_AC(1,UL) << CTX_VERSION_SHIFT) + _AC(1,UL))
+#define CTX_VALID(__ctx)	\
+	 (!(((__ctx.sparc64_ctx_val) ^ tlb_context_cache) & CTX_VERSION_MASK))
+#define CTX_HWBITS(__ctx)	((__ctx.sparc64_ctx_val) & CTX_HW_MASK)
+#define CTX_NRBITS(__ctx)	((__ctx.sparc64_ctx_val) & CTX_NR_MASK)
+
+#ifndef __ASSEMBLY__
+
+#define TSB_ENTRY_ALIGNMENT	16
+
+struct tsb {
+	unsigned long tag;
+	unsigned long pte;
+} __attribute__((aligned(TSB_ENTRY_ALIGNMENT)));
+
+extern void __tsb_insert(unsigned long ent, unsigned long tag, unsigned long pte);
+extern void tsb_flush(unsigned long ent, unsigned long tag);
+extern void tsb_init(struct tsb *tsb, unsigned long size);
+
+struct tsb_config {
+	struct tsb		*tsb;
+	unsigned long		tsb_rss_limit;
+	unsigned long		tsb_nentries;
+	unsigned long		tsb_reg_val;
+	unsigned long		tsb_map_vaddr;
+	unsigned long		tsb_map_pte;
+};
+
+#define MM_TSB_BASE	0
+
+#ifdef CONFIG_HUGETLB_PAGE
+#define MM_TSB_HUGE	1
+#define MM_NUM_TSBS	2
+#else
+#define MM_NUM_TSBS	1
+#endif
+
+typedef struct {
+	spinlock_t		lock;
+	unsigned long		sparc64_ctx_val;
+	unsigned long		huge_pte_count;
+	struct tsb_config	tsb_block[MM_NUM_TSBS];
+	struct hv_tsb_descr	tsb_descr[MM_NUM_TSBS];
+} mm_context_t;
+
+#endif /* !__ASSEMBLY__ */
+
+#define TSB_CONFIG_TSB		0x00
+#define TSB_CONFIG_RSS_LIMIT	0x08
+#define TSB_CONFIG_NENTRIES	0x10
+#define TSB_CONFIG_REG_VAL	0x18
+#define TSB_CONFIG_MAP_VADDR	0x20
+#define TSB_CONFIG_MAP_PTE	0x28
+
+#endif /* __MMU_H */
diff --git a/include/asm-sparc/mmu_context.h b/include/asm-sparc/mmu_context.h
index 671a997b9e6..e14efb9532f 100644
--- a/include/asm-sparc/mmu_context.h
+++ b/include/asm-sparc/mmu_context.h
@@ -1,42 +1,8 @@
-#ifndef __SPARC_MMU_CONTEXT_H
-#define __SPARC_MMU_CONTEXT_H
-
-#include <asm/btfixup.h>
-
-#ifndef __ASSEMBLY__
-
-#include <asm-generic/mm_hooks.h>
-
-static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
-{
-}
-
-/*
- * Initialize a new mmu context.  This is invoked when a new
- * address space instance (unique or shared) is instantiated.
- */
-#define init_new_context(tsk, mm) (((mm)->context = NO_CONTEXT), 0)
-
-/*
- * Destroy a dead context.  This occurs when mmput drops the
- * mm_users count to zero, the mmaps have been released, and
- * all the page tables have been flushed.  Our job is to destroy
- * any remaining processor-specific state.
- */
-BTFIXUPDEF_CALL(void, destroy_context, struct mm_struct *)
-
-#define destroy_context(mm) BTFIXUP_CALL(destroy_context)(mm)
-
-/* Switch the current MM context. */
-BTFIXUPDEF_CALL(void, switch_mm, struct mm_struct *, struct mm_struct *, struct task_struct *)
-
-#define switch_mm(old_mm, mm, tsk) BTFIXUP_CALL(switch_mm)(old_mm, mm, tsk)
-
-#define deactivate_mm(tsk,mm)	do { } while (0)
-
-/* Activate a new MM instance for the current task. */
-#define activate_mm(active_mm, mm) switch_mm((active_mm), (mm), NULL)
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC_MMU_CONTEXT_H) */
+#ifndef ___ASM_SPARC_MMU_CONTEXT_H
+#define ___ASM_SPARC_MMU_CONTEXT_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/mmu_context_64.h>
+#else
+#include <asm-sparc/mmu_context_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/mmu_context_32.h b/include/asm-sparc/mmu_context_32.h
new file mode 100644
index 00000000000..671a997b9e6
--- /dev/null
+++ b/include/asm-sparc/mmu_context_32.h
@@ -0,0 +1,42 @@
+#ifndef __SPARC_MMU_CONTEXT_H
+#define __SPARC_MMU_CONTEXT_H
+
+#include <asm/btfixup.h>
+
+#ifndef __ASSEMBLY__
+
+#include <asm-generic/mm_hooks.h>
+
+static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
+{
+}
+
+/*
+ * Initialize a new mmu context.  This is invoked when a new
+ * address space instance (unique or shared) is instantiated.
+ */
+#define init_new_context(tsk, mm) (((mm)->context = NO_CONTEXT), 0)
+
+/*
+ * Destroy a dead context.  This occurs when mmput drops the
+ * mm_users count to zero, the mmaps have been released, and
+ * all the page tables have been flushed.  Our job is to destroy
+ * any remaining processor-specific state.
+ */
+BTFIXUPDEF_CALL(void, destroy_context, struct mm_struct *)
+
+#define destroy_context(mm) BTFIXUP_CALL(destroy_context)(mm)
+
+/* Switch the current MM context. */
+BTFIXUPDEF_CALL(void, switch_mm, struct mm_struct *, struct mm_struct *, struct task_struct *)
+
+#define switch_mm(old_mm, mm, tsk) BTFIXUP_CALL(switch_mm)(old_mm, mm, tsk)
+
+#define deactivate_mm(tsk,mm)	do { } while (0)
+
+/* Activate a new MM instance for the current task. */
+#define activate_mm(active_mm, mm) switch_mm((active_mm), (mm), NULL)
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC_MMU_CONTEXT_H) */
diff --git a/include/asm-sparc/mmu_context_64.h b/include/asm-sparc/mmu_context_64.h
new file mode 100644
index 00000000000..5693ab48260
--- /dev/null
+++ b/include/asm-sparc/mmu_context_64.h
@@ -0,0 +1,155 @@
+#ifndef __SPARC64_MMU_CONTEXT_H
+#define __SPARC64_MMU_CONTEXT_H
+
+/* Derived heavily from Linus's Alpha/AXP ASN code... */
+
+#ifndef __ASSEMBLY__
+
+#include <linux/spinlock.h>
+#include <asm/system.h>
+#include <asm/spitfire.h>
+#include <asm-generic/mm_hooks.h>
+
+static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
+{
+}
+
+extern spinlock_t ctx_alloc_lock;
+extern unsigned long tlb_context_cache;
+extern unsigned long mmu_context_bmap[];
+
+extern void get_new_mmu_context(struct mm_struct *mm);
+#ifdef CONFIG_SMP
+extern void smp_new_mmu_context_version(void);
+#else
+#define smp_new_mmu_context_version() do { } while (0)
+#endif
+
+extern int init_new_context(struct task_struct *tsk, struct mm_struct *mm);
+extern void destroy_context(struct mm_struct *mm);
+
+extern void __tsb_context_switch(unsigned long pgd_pa,
+				 struct tsb_config *tsb_base,
+				 struct tsb_config *tsb_huge,
+				 unsigned long tsb_descr_pa);
+
+static inline void tsb_context_switch(struct mm_struct *mm)
+{
+	__tsb_context_switch(__pa(mm->pgd),
+			     &mm->context.tsb_block[0],
+#ifdef CONFIG_HUGETLB_PAGE
+			     (mm->context.tsb_block[1].tsb ?
+			      &mm->context.tsb_block[1] :
+			      NULL)
+#else
+			     NULL
+#endif
+			     , __pa(&mm->context.tsb_descr[0]));
+}
+
+extern void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long mm_rss);
+#ifdef CONFIG_SMP
+extern void smp_tsb_sync(struct mm_struct *mm);
+#else
+#define smp_tsb_sync(__mm) do { } while (0)
+#endif
+
+/* Set MMU context in the actual hardware. */
+#define load_secondary_context(__mm) \
+	__asm__ __volatile__( \
+	"\n661:	stxa		%0, [%1] %2\n" \
+	"	.section	.sun4v_1insn_patch, \"ax\"\n" \
+	"	.word		661b\n" \
+	"	stxa		%0, [%1] %3\n" \
+	"	.previous\n" \
+	"	flush		%%g6\n" \
+	: /* No outputs */ \
+	: "r" (CTX_HWBITS((__mm)->context)), \
+	  "r" (SECONDARY_CONTEXT), "i" (ASI_DMMU), "i" (ASI_MMU))
+
+extern void __flush_tlb_mm(unsigned long, unsigned long);
+
+/* Switch the current MM context.  Interrupts are disabled.  */
+static inline void switch_mm(struct mm_struct *old_mm, struct mm_struct *mm, struct task_struct *tsk)
+{
+	unsigned long ctx_valid, flags;
+	int cpu;
+
+	if (unlikely(mm == &init_mm))
+		return;
+
+	spin_lock_irqsave(&mm->context.lock, flags);
+	ctx_valid = CTX_VALID(mm->context);
+	if (!ctx_valid)
+		get_new_mmu_context(mm);
+
+	/* We have to be extremely careful here or else we will miss
+	 * a TSB grow if we switch back and forth between a kernel
+	 * thread and an address space which has it's TSB size increased
+	 * on another processor.
+	 *
+	 * It is possible to play some games in order to optimize the
+	 * switch, but the safest thing to do is to unconditionally
+	 * perform the secondary context load and the TSB context switch.
+	 *
+	 * For reference the bad case is, for address space "A":
+	 *
+	 *		CPU 0			CPU 1
+	 *	run address space A
+	 *	set cpu0's bits in cpu_vm_mask
+	 *	switch to kernel thread, borrow
+	 *	address space A via entry_lazy_tlb
+	 *					run address space A
+	 *					set cpu1's bit in cpu_vm_mask
+	 *					flush_tlb_pending()
+	 *					reset cpu_vm_mask to just cpu1
+	 *					TSB grow
+	 *	run address space A
+	 *	context was valid, so skip
+	 *	TSB context switch
+	 *
+	 * At that point cpu0 continues to use a stale TSB, the one from
+	 * before the TSB grow performed on cpu1.  cpu1 did not cross-call
+	 * cpu0 to update it's TSB because at that point the cpu_vm_mask
+	 * only had cpu1 set in it.
+	 */
+	load_secondary_context(mm);
+	tsb_context_switch(mm);
+
+	/* Any time a processor runs a context on an address space
+	 * for the first time, we must flush that context out of the
+	 * local TLB.
+	 */
+	cpu = smp_processor_id();
+	if (!ctx_valid || !cpu_isset(cpu, mm->cpu_vm_mask)) {
+		cpu_set(cpu, mm->cpu_vm_mask);
+		__flush_tlb_mm(CTX_HWBITS(mm->context),
+			       SECONDARY_CONTEXT);
+	}
+	spin_unlock_irqrestore(&mm->context.lock, flags);
+}
+
+#define deactivate_mm(tsk,mm)	do { } while (0)
+
+/* Activate a new MM instance for the current task. */
+static inline void activate_mm(struct mm_struct *active_mm, struct mm_struct *mm)
+{
+	unsigned long flags;
+	int cpu;
+
+	spin_lock_irqsave(&mm->context.lock, flags);
+	if (!CTX_VALID(mm->context))
+		get_new_mmu_context(mm);
+	cpu = smp_processor_id();
+	if (!cpu_isset(cpu, mm->cpu_vm_mask))
+		cpu_set(cpu, mm->cpu_vm_mask);
+
+	load_secondary_context(mm);
+	__flush_tlb_mm(CTX_HWBITS(mm->context), SECONDARY_CONTEXT);
+	tsb_context_switch(mm);
+	spin_unlock_irqrestore(&mm->context.lock, flags);
+}
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC64_MMU_CONTEXT_H) */
diff --git a/include/asm-sparc/mmzone.h b/include/asm-sparc/mmzone.h
new file mode 100644
index 00000000000..ebf5986c12e
--- /dev/null
+++ b/include/asm-sparc/mmzone.h
@@ -0,0 +1,17 @@
+#ifndef _SPARC64_MMZONE_H
+#define _SPARC64_MMZONE_H
+
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+
+extern struct pglist_data *node_data[];
+
+#define NODE_DATA(nid)		(node_data[nid])
+#define node_start_pfn(nid)	(NODE_DATA(nid)->node_start_pfn)
+#define node_end_pfn(nid)	(NODE_DATA(nid)->node_end_pfn)
+
+extern int numa_cpu_lookup_table[];
+extern cpumask_t numa_cpumask_lookup_table[];
+
+#endif /* CONFIG_NEED_MULTIPLE_NODES */
+
+#endif /* _SPARC64_MMZONE_H */
diff --git a/include/asm-sparc/module.h b/include/asm-sparc/module.h
index cbd9e67b0c0..516138fe681 100644
--- a/include/asm-sparc/module.h
+++ b/include/asm-sparc/module.h
@@ -1,7 +1,8 @@
-#ifndef _ASM_SPARC_MODULE_H
-#define _ASM_SPARC_MODULE_H
-struct mod_arch_specific { };
-#define Elf_Shdr Elf32_Shdr
-#define Elf_Sym Elf32_Sym
-#define Elf_Ehdr Elf32_Ehdr
-#endif /* _ASM_SPARC_MODULE_H */
+#ifndef ___ASM_SPARC_MODULE_H
+#define ___ASM_SPARC_MODULE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/module_64.h>
+#else
+#include <asm-sparc/module_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/module_32.h b/include/asm-sparc/module_32.h
new file mode 100644
index 00000000000..cbd9e67b0c0
--- /dev/null
+++ b/include/asm-sparc/module_32.h
@@ -0,0 +1,7 @@
+#ifndef _ASM_SPARC_MODULE_H
+#define _ASM_SPARC_MODULE_H
+struct mod_arch_specific { };
+#define Elf_Shdr Elf32_Shdr
+#define Elf_Sym Elf32_Sym
+#define Elf_Ehdr Elf32_Ehdr
+#endif /* _ASM_SPARC_MODULE_H */
diff --git a/include/asm-sparc/module_64.h b/include/asm-sparc/module_64.h
new file mode 100644
index 00000000000..3d77ba46578
--- /dev/null
+++ b/include/asm-sparc/module_64.h
@@ -0,0 +1,7 @@
+#ifndef _ASM_SPARC64_MODULE_H
+#define _ASM_SPARC64_MODULE_H
+struct mod_arch_specific { };
+#define Elf_Shdr Elf64_Shdr
+#define Elf_Sym Elf64_Sym
+#define Elf_Ehdr Elf64_Ehdr
+#endif /* _ASM_SPARC64_MODULE_H */
diff --git a/include/asm-sparc/mostek.h b/include/asm-sparc/mostek.h
index 29aad11b8f0..5b9f7fec7ee 100644
--- a/include/asm-sparc/mostek.h
+++ b/include/asm-sparc/mostek.h
@@ -1,173 +1,8 @@
-/*
- * mostek.h:  Describes the various Mostek time of day clock registers.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1996 Thomas K. Dyas (tdyas@eden.rutgers.edu)
- * Added intersil code 05/25/98 Chris Davis (cdavis@cois.on.ca)
- */
-
-#ifndef _SPARC_MOSTEK_H
-#define _SPARC_MOSTEK_H
-
-#include <asm/idprom.h>
-#include <asm/io.h>
-
-/*       M48T02 Register Map (adapted from Sun NVRAM/Hostid FAQ)
- *
- *                             Data
- * Address                                                 Function
- *        Bit 7 Bit 6 Bit 5 Bit 4Bit 3 Bit 2 Bit 1 Bit 0
- *   7ff  -     -     -     -    -     -     -     -       Year 00-99
- *   7fe  0     0     0     -    -     -     -     -      Month 01-12
- *   7fd  0     0     -     -    -     -     -     -       Date 01-31
- *   7fc  0     FT    0     0    0     -     -     -        Day 01-07
- *   7fb  KS    0     -     -    -     -     -     -      Hours 00-23
- *   7fa  0     -     -     -    -     -     -     -    Minutes 00-59
- *   7f9  ST    -     -     -    -     -     -     -    Seconds 00-59
- *   7f8  W     R     S     -    -     -     -     -    Control
- *
- *   * ST is STOP BIT
- *   * W is WRITE BIT
- *   * R is READ BIT
- *   * S is SIGN BIT
- *   * FT is FREQ TEST BIT
- *   * KS is KICK START BIT
- */
-
-/* The Mostek 48t02 real time clock and NVRAM chip. The registers
- * other than the control register are in binary coded decimal. Some
- * control bits also live outside the control register.
- */
-#define mostek_read(_addr)		readb(_addr)
-#define mostek_write(_addr,_val)	writeb(_val, _addr)
-#define MOSTEK_EEPROM		0x0000UL
-#define MOSTEK_IDPROM		0x07d8UL
-#define MOSTEK_CREG		0x07f8UL
-#define MOSTEK_SEC		0x07f9UL
-#define MOSTEK_MIN		0x07faUL
-#define MOSTEK_HOUR		0x07fbUL
-#define MOSTEK_DOW		0x07fcUL
-#define MOSTEK_DOM		0x07fdUL
-#define MOSTEK_MONTH		0x07feUL
-#define MOSTEK_YEAR		0x07ffUL
-
-struct mostek48t02 {
-	volatile char eeprom[2008];	/* This is the eeprom, don't touch! */
-	struct idprom idprom;		/* The idprom lives here. */
-	volatile unsigned char creg;	/* Control register */
-	volatile unsigned char sec;	/* Seconds (0-59) */
-	volatile unsigned char min;	/* Minutes (0-59) */
-	volatile unsigned char hour;	/* Hour (0-23) */
-	volatile unsigned char dow;	/* Day of the week (1-7) */
-	volatile unsigned char dom;	/* Day of the month (1-31) */
-	volatile unsigned char month;	/* Month of year (1-12) */
-	volatile unsigned char year;	/* Year (0-99) */
-};
-
-extern spinlock_t mostek_lock;
-extern void __iomem *mstk48t02_regs;
-
-/* Control register values. */
-#define	MSTK_CREG_WRITE	0x80	/* Must set this before placing values. */
-#define	MSTK_CREG_READ	0x40	/* Stop updates to allow a clean read. */
-#define	MSTK_CREG_SIGN	0x20	/* Slow/speed clock in calibration mode. */
-
-/* Control bits that live in the other registers. */
-#define	MSTK_STOP	0x80	/* Stop the clock oscillator. (sec) */
-#define	MSTK_KICK_START	0x80	/* Kick start the clock chip. (hour) */
-#define MSTK_FREQ_TEST	0x40	/* Frequency test mode. (day) */
-
-#define MSTK_YEAR_ZERO       1968   /* If year reg has zero, it is 1968. */
-#define MSTK_CVT_YEAR(yr)  ((yr) + MSTK_YEAR_ZERO)
-
-/* Masks that define how much space each value takes up. */
-#define	MSTK_SEC_MASK	0x7f
-#define	MSTK_MIN_MASK	0x7f
-#define	MSTK_HOUR_MASK	0x3f
-#define	MSTK_DOW_MASK	0x07
-#define	MSTK_DOM_MASK	0x3f
-#define	MSTK_MONTH_MASK	0x1f
-#define	MSTK_YEAR_MASK	0xffU
-
-/* Binary coded decimal conversion macros. */
-#define MSTK_REGVAL_TO_DECIMAL(x)  (((x) & 0x0F) + 0x0A * ((x) >> 0x04))
-#define MSTK_DECIMAL_TO_REGVAL(x)  ((((x) / 0x0A) << 0x04) + ((x) % 0x0A))
-
-/* Generic register set and get macros for internal use. */
-#define MSTK_GET(regs,var,mask) (MSTK_REGVAL_TO_DECIMAL(((struct mostek48t02 *)regs)->var & MSTK_ ## mask ## _MASK))
-#define MSTK_SET(regs,var,value,mask) do { ((struct mostek48t02 *)regs)->var &= ~(MSTK_ ## mask ## _MASK); ((struct mostek48t02 *)regs)->var |= MSTK_DECIMAL_TO_REGVAL(value) & (MSTK_ ## mask ## _MASK); } while (0)
-
-/* Macros to make register access easier on our fingers. These give you
- * the decimal value of the register requested if applicable. You pass
- * the a pointer to a 'struct mostek48t02'.
- */
-#define	MSTK_REG_CREG(regs)	(((struct mostek48t02 *)regs)->creg)
-#define	MSTK_REG_SEC(regs)	MSTK_GET(regs,sec,SEC)
-#define	MSTK_REG_MIN(regs)	MSTK_GET(regs,min,MIN)
-#define	MSTK_REG_HOUR(regs)	MSTK_GET(regs,hour,HOUR)
-#define	MSTK_REG_DOW(regs)	MSTK_GET(regs,dow,DOW)
-#define	MSTK_REG_DOM(regs)	MSTK_GET(regs,dom,DOM)
-#define	MSTK_REG_MONTH(regs)	MSTK_GET(regs,month,MONTH)
-#define	MSTK_REG_YEAR(regs)	MSTK_GET(regs,year,YEAR)
-
-#define	MSTK_SET_REG_SEC(regs,value)	MSTK_SET(regs,sec,value,SEC)
-#define	MSTK_SET_REG_MIN(regs,value)	MSTK_SET(regs,min,value,MIN)
-#define	MSTK_SET_REG_HOUR(regs,value)	MSTK_SET(regs,hour,value,HOUR)
-#define	MSTK_SET_REG_DOW(regs,value)	MSTK_SET(regs,dow,value,DOW)
-#define	MSTK_SET_REG_DOM(regs,value)	MSTK_SET(regs,dom,value,DOM)
-#define	MSTK_SET_REG_MONTH(regs,value)	MSTK_SET(regs,month,value,MONTH)
-#define	MSTK_SET_REG_YEAR(regs,value)	MSTK_SET(regs,year,value,YEAR)
-
-
-/* The Mostek 48t08 clock chip. Found on Sun4m's I think. It has the
- * same (basically) layout of the 48t02 chip except for the extra
- * NVRAM on board (8 KB against the 48t02's 2 KB).
- */
-struct mostek48t08 {
-	char offset[6*1024];         /* Magic things may be here, who knows? */
-	struct mostek48t02 regs;     /* Here is what we are interested in.   */
-};
-
-extern enum sparc_clock_type sp_clock_typ;
-
-#ifdef CONFIG_SUN4
-enum sparc_clock_type {	MSTK48T02, MSTK48T08, \
-INTERSIL, MSTK_INVALID };
+#ifndef ___ASM_SPARC_MOSTEK_H
+#define ___ASM_SPARC_MOSTEK_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/mostek_64.h>
 #else
-enum sparc_clock_type {	MSTK48T02, MSTK48T08, \
-MSTK_INVALID };
+#include <asm-sparc/mostek_32.h>
 #endif
-
-#ifdef CONFIG_SUN4
-/* intersil on a sun 4/260 code  data from harris doc */
-struct intersil_dt {
-        volatile unsigned char int_csec;
-        volatile unsigned char int_hour;
-        volatile unsigned char int_min;
-        volatile unsigned char int_sec;
-        volatile unsigned char int_month;
-        volatile unsigned char int_day;
-        volatile unsigned char int_year;
-        volatile unsigned char int_dow;
-};
-
-struct intersil {
-	struct intersil_dt clk;
-	struct intersil_dt cmp;
-	volatile unsigned char int_intr_reg;
-	volatile unsigned char int_cmd_reg;
-};
-
-#define INTERSIL_STOP        0x0
-#define INTERSIL_START       0x8
-#define INTERSIL_INTR_DISABLE   0x0
-#define INTERSIL_INTR_ENABLE   0x10
-#define INTERSIL_32K		0x0
-#define INTERSIL_NORMAL		0x0
-#define INTERSIL_24H		0x4 
-#define INTERSIL_INT_100HZ	0x2
-
-/* end of intersil info */
 #endif
-
-#endif /* !(_SPARC_MOSTEK_H) */
diff --git a/include/asm-sparc/mostek_32.h b/include/asm-sparc/mostek_32.h
new file mode 100644
index 00000000000..a99590c4c50
--- /dev/null
+++ b/include/asm-sparc/mostek_32.h
@@ -0,0 +1,171 @@
+/*
+ * mostek.h:  Describes the various Mostek time of day clock registers.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996 Thomas K. Dyas (tdyas@eden.rutgers.edu)
+ * Added intersil code 05/25/98 Chris Davis (cdavis@cois.on.ca)
+ */
+
+#ifndef _SPARC_MOSTEK_H
+#define _SPARC_MOSTEK_H
+
+#include <asm/idprom.h>
+#include <asm/io.h>
+
+/*       M48T02 Register Map (adapted from Sun NVRAM/Hostid FAQ)
+ *
+ *                             Data
+ * Address                                                 Function
+ *        Bit 7 Bit 6 Bit 5 Bit 4Bit 3 Bit 2 Bit 1 Bit 0
+ *   7ff  -     -     -     -    -     -     -     -       Year 00-99
+ *   7fe  0     0     0     -    -     -     -     -      Month 01-12
+ *   7fd  0     0     -     -    -     -     -     -       Date 01-31
+ *   7fc  0     FT    0     0    0     -     -     -        Day 01-07
+ *   7fb  KS    0     -     -    -     -     -     -      Hours 00-23
+ *   7fa  0     -     -     -    -     -     -     -    Minutes 00-59
+ *   7f9  ST    -     -     -    -     -     -     -    Seconds 00-59
+ *   7f8  W     R     S     -    -     -     -     -    Control
+ *
+ *   * ST is STOP BIT
+ *   * W is WRITE BIT
+ *   * R is READ BIT
+ *   * S is SIGN BIT
+ *   * FT is FREQ TEST BIT
+ *   * KS is KICK START BIT
+ */
+
+/* The Mostek 48t02 real time clock and NVRAM chip. The registers
+ * other than the control register are in binary coded decimal. Some
+ * control bits also live outside the control register.
+ */
+#define mostek_read(_addr)		readb(_addr)
+#define mostek_write(_addr,_val)	writeb(_val, _addr)
+#define MOSTEK_EEPROM		0x0000UL
+#define MOSTEK_IDPROM		0x07d8UL
+#define MOSTEK_CREG		0x07f8UL
+#define MOSTEK_SEC		0x07f9UL
+#define MOSTEK_MIN		0x07faUL
+#define MOSTEK_HOUR		0x07fbUL
+#define MOSTEK_DOW		0x07fcUL
+#define MOSTEK_DOM		0x07fdUL
+#define MOSTEK_MONTH		0x07feUL
+#define MOSTEK_YEAR		0x07ffUL
+
+struct mostek48t02 {
+	volatile char eeprom[2008];	/* This is the eeprom, don't touch! */
+	struct idprom idprom;		/* The idprom lives here. */
+	volatile unsigned char creg;	/* Control register */
+	volatile unsigned char sec;	/* Seconds (0-59) */
+	volatile unsigned char min;	/* Minutes (0-59) */
+	volatile unsigned char hour;	/* Hour (0-23) */
+	volatile unsigned char dow;	/* Day of the week (1-7) */
+	volatile unsigned char dom;	/* Day of the month (1-31) */
+	volatile unsigned char month;	/* Month of year (1-12) */
+	volatile unsigned char year;	/* Year (0-99) */
+};
+
+extern spinlock_t mostek_lock;
+extern void __iomem *mstk48t02_regs;
+
+/* Control register values. */
+#define	MSTK_CREG_WRITE	0x80	/* Must set this before placing values. */
+#define	MSTK_CREG_READ	0x40	/* Stop updates to allow a clean read. */
+#define	MSTK_CREG_SIGN	0x20	/* Slow/speed clock in calibration mode. */
+
+/* Control bits that live in the other registers. */
+#define	MSTK_STOP	0x80	/* Stop the clock oscillator. (sec) */
+#define	MSTK_KICK_START	0x80	/* Kick start the clock chip. (hour) */
+#define MSTK_FREQ_TEST	0x40	/* Frequency test mode. (day) */
+
+#define MSTK_YEAR_ZERO       1968   /* If year reg has zero, it is 1968. */
+#define MSTK_CVT_YEAR(yr)  ((yr) + MSTK_YEAR_ZERO)
+
+/* Masks that define how much space each value takes up. */
+#define	MSTK_SEC_MASK	0x7f
+#define	MSTK_MIN_MASK	0x7f
+#define	MSTK_HOUR_MASK	0x3f
+#define	MSTK_DOW_MASK	0x07
+#define	MSTK_DOM_MASK	0x3f
+#define	MSTK_MONTH_MASK	0x1f
+#define	MSTK_YEAR_MASK	0xffU
+
+/* Binary coded decimal conversion macros. */
+#define MSTK_REGVAL_TO_DECIMAL(x)  (((x) & 0x0F) + 0x0A * ((x) >> 0x04))
+#define MSTK_DECIMAL_TO_REGVAL(x)  ((((x) / 0x0A) << 0x04) + ((x) % 0x0A))
+
+/* Generic register set and get macros for internal use. */
+#define MSTK_GET(regs,var,mask) (MSTK_REGVAL_TO_DECIMAL(((struct mostek48t02 *)regs)->var & MSTK_ ## mask ## _MASK))
+#define MSTK_SET(regs,var,value,mask) do { ((struct mostek48t02 *)regs)->var &= ~(MSTK_ ## mask ## _MASK); ((struct mostek48t02 *)regs)->var |= MSTK_DECIMAL_TO_REGVAL(value) & (MSTK_ ## mask ## _MASK); } while (0)
+
+/* Macros to make register access easier on our fingers. These give you
+ * the decimal value of the register requested if applicable. You pass
+ * the a pointer to a 'struct mostek48t02'.
+ */
+#define	MSTK_REG_CREG(regs)	(((struct mostek48t02 *)regs)->creg)
+#define	MSTK_REG_SEC(regs)	MSTK_GET(regs,sec,SEC)
+#define	MSTK_REG_MIN(regs)	MSTK_GET(regs,min,MIN)
+#define	MSTK_REG_HOUR(regs)	MSTK_GET(regs,hour,HOUR)
+#define	MSTK_REG_DOW(regs)	MSTK_GET(regs,dow,DOW)
+#define	MSTK_REG_DOM(regs)	MSTK_GET(regs,dom,DOM)
+#define	MSTK_REG_MONTH(regs)	MSTK_GET(regs,month,MONTH)
+#define	MSTK_REG_YEAR(regs)	MSTK_GET(regs,year,YEAR)
+
+#define	MSTK_SET_REG_SEC(regs,value)	MSTK_SET(regs,sec,value,SEC)
+#define	MSTK_SET_REG_MIN(regs,value)	MSTK_SET(regs,min,value,MIN)
+#define	MSTK_SET_REG_HOUR(regs,value)	MSTK_SET(regs,hour,value,HOUR)
+#define	MSTK_SET_REG_DOW(regs,value)	MSTK_SET(regs,dow,value,DOW)
+#define	MSTK_SET_REG_DOM(regs,value)	MSTK_SET(regs,dom,value,DOM)
+#define	MSTK_SET_REG_MONTH(regs,value)	MSTK_SET(regs,month,value,MONTH)
+#define	MSTK_SET_REG_YEAR(regs,value)	MSTK_SET(regs,year,value,YEAR)
+
+
+/* The Mostek 48t08 clock chip. Found on Sun4m's I think. It has the
+ * same (basically) layout of the 48t02 chip except for the extra
+ * NVRAM on board (8 KB against the 48t02's 2 KB).
+ */
+struct mostek48t08 {
+	char offset[6*1024];         /* Magic things may be here, who knows? */
+	struct mostek48t02 regs;     /* Here is what we are interested in.   */
+};
+
+#ifdef CONFIG_SUN4
+enum sparc_clock_type {	MSTK48T02, MSTK48T08, \
+INTERSIL, MSTK_INVALID };
+#else
+enum sparc_clock_type {	MSTK48T02, MSTK48T08, \
+MSTK_INVALID };
+#endif
+
+#ifdef CONFIG_SUN4
+/* intersil on a sun 4/260 code  data from harris doc */
+struct intersil_dt {
+        volatile unsigned char int_csec;
+        volatile unsigned char int_hour;
+        volatile unsigned char int_min;
+        volatile unsigned char int_sec;
+        volatile unsigned char int_month;
+        volatile unsigned char int_day;
+        volatile unsigned char int_year;
+        volatile unsigned char int_dow;
+};
+
+struct intersil {
+	struct intersil_dt clk;
+	struct intersil_dt cmp;
+	volatile unsigned char int_intr_reg;
+	volatile unsigned char int_cmd_reg;
+};
+
+#define INTERSIL_STOP        0x0
+#define INTERSIL_START       0x8
+#define INTERSIL_INTR_DISABLE   0x0
+#define INTERSIL_INTR_ENABLE   0x10
+#define INTERSIL_32K		0x0
+#define INTERSIL_NORMAL		0x0
+#define INTERSIL_24H		0x4
+#define INTERSIL_INT_100HZ	0x2
+
+/* end of intersil info */
+#endif
+
+#endif /* !(_SPARC_MOSTEK_H) */
diff --git a/include/asm-sparc/mostek_64.h b/include/asm-sparc/mostek_64.h
new file mode 100644
index 00000000000..c5652de2ace
--- /dev/null
+++ b/include/asm-sparc/mostek_64.h
@@ -0,0 +1,143 @@
+/* mostek.h:  Describes the various Mostek time of day clock registers.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1996 Thomas K. Dyas (tdyas@eden.rutgers.edu)
+ */
+
+#ifndef _SPARC64_MOSTEK_H
+#define _SPARC64_MOSTEK_H
+
+#include <asm/idprom.h>
+
+/*       M48T02 Register Map (adapted from Sun NVRAM/Hostid FAQ)
+ *
+ *                             Data
+ * Address                                                 Function
+ *        Bit 7 Bit 6 Bit 5 Bit 4Bit 3 Bit 2 Bit 1 Bit 0
+ *   7ff  -     -     -     -    -     -     -     -       Year 00-99
+ *   7fe  0     0     0     -    -     -     -     -      Month 01-12
+ *   7fd  0     0     -     -    -     -     -     -       Date 01-31
+ *   7fc  0     FT    0     0    0     -     -     -        Day 01-07
+ *   7fb  KS    0     -     -    -     -     -     -      Hours 00-23
+ *   7fa  0     -     -     -    -     -     -     -    Minutes 00-59
+ *   7f9  ST    -     -     -    -     -     -     -    Seconds 00-59
+ *   7f8  W     R     S     -    -     -     -     -    Control
+ *
+ *   * ST is STOP BIT
+ *   * W is WRITE BIT
+ *   * R is READ BIT
+ *   * S is SIGN BIT
+ *   * FT is FREQ TEST BIT
+ *   * KS is KICK START BIT
+ */
+
+/* The Mostek 48t02 real time clock and NVRAM chip. The registers
+ * other than the control register are in binary coded decimal. Some
+ * control bits also live outside the control register.
+ *
+ * We now deal with physical addresses for I/O to the chip. -DaveM
+ */
+static inline u8 mostek_read(void __iomem *addr)
+{
+	u8 ret;
+
+	__asm__ __volatile__("lduba	[%1] %2, %0"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+	return ret;
+}
+
+static inline void mostek_write(void __iomem *addr, u8 val)
+{
+	__asm__ __volatile__("stba	%0, [%1] %2"
+			     : /* no outputs */
+			     : "r" (val), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+#define MOSTEK_EEPROM		0x0000UL
+#define MOSTEK_IDPROM		0x07d8UL
+#define MOSTEK_CREG		0x07f8UL
+#define MOSTEK_SEC		0x07f9UL
+#define MOSTEK_MIN		0x07faUL
+#define MOSTEK_HOUR		0x07fbUL
+#define MOSTEK_DOW		0x07fcUL
+#define MOSTEK_DOM		0x07fdUL
+#define MOSTEK_MONTH		0x07feUL
+#define MOSTEK_YEAR		0x07ffUL
+
+extern spinlock_t mostek_lock;
+extern void __iomem *mstk48t02_regs;
+
+/* Control register values. */
+#define	MSTK_CREG_WRITE	0x80	/* Must set this before placing values. */
+#define	MSTK_CREG_READ	0x40	/* Stop updates to allow a clean read. */
+#define	MSTK_CREG_SIGN	0x20	/* Slow/speed clock in calibration mode. */
+
+/* Control bits that live in the other registers. */
+#define	MSTK_STOP	0x80	/* Stop the clock oscillator. (sec) */
+#define	MSTK_KICK_START	0x80	/* Kick start the clock chip. (hour) */
+#define MSTK_FREQ_TEST	0x40	/* Frequency test mode. (day) */
+
+#define MSTK_YEAR_ZERO       1968   /* If year reg has zero, it is 1968. */
+#define MSTK_CVT_YEAR(yr)  ((yr) + MSTK_YEAR_ZERO)
+
+/* Masks that define how much space each value takes up. */
+#define	MSTK_SEC_MASK	0x7f
+#define	MSTK_MIN_MASK	0x7f
+#define	MSTK_HOUR_MASK	0x3f
+#define	MSTK_DOW_MASK	0x07
+#define	MSTK_DOM_MASK	0x3f
+#define	MSTK_MONTH_MASK	0x1f
+#define	MSTK_YEAR_MASK	0xffU
+
+/* Binary coded decimal conversion macros. */
+#define MSTK_REGVAL_TO_DECIMAL(x)  (((x) & 0x0F) + 0x0A * ((x) >> 0x04))
+#define MSTK_DECIMAL_TO_REGVAL(x)  ((((x) / 0x0A) << 0x04) + ((x) % 0x0A))
+
+/* Generic register set and get macros for internal use. */
+#define MSTK_GET(regs,name)	\
+	(MSTK_REGVAL_TO_DECIMAL(mostek_read(regs + MOSTEK_ ## name) & MSTK_ ## name ## _MASK))
+#define MSTK_SET(regs,name,value) \
+do {	u8 __val = mostek_read(regs + MOSTEK_ ## name); \
+	__val &= ~(MSTK_ ## name ## _MASK); \
+	__val |= (MSTK_DECIMAL_TO_REGVAL(value) & \
+		  (MSTK_ ## name ## _MASK)); \
+	mostek_write(regs + MOSTEK_ ## name, __val); \
+} while(0)
+
+/* Macros to make register access easier on our fingers. These give you
+ * the decimal value of the register requested if applicable. You pass
+ * the a pointer to a 'struct mostek48t02'.
+ */
+#define	MSTK_REG_CREG(regs)	(mostek_read((regs) + MOSTEK_CREG))
+#define	MSTK_REG_SEC(regs)	MSTK_GET(regs,SEC)
+#define	MSTK_REG_MIN(regs)	MSTK_GET(regs,MIN)
+#define	MSTK_REG_HOUR(regs)	MSTK_GET(regs,HOUR)
+#define	MSTK_REG_DOW(regs)	MSTK_GET(regs,DOW)
+#define	MSTK_REG_DOM(regs)	MSTK_GET(regs,DOM)
+#define	MSTK_REG_MONTH(regs)	MSTK_GET(regs,MONTH)
+#define	MSTK_REG_YEAR(regs)	MSTK_GET(regs,YEAR)
+
+#define	MSTK_SET_REG_SEC(regs,value)	MSTK_SET(regs,SEC,value)
+#define	MSTK_SET_REG_MIN(regs,value)	MSTK_SET(regs,MIN,value)
+#define	MSTK_SET_REG_HOUR(regs,value)	MSTK_SET(regs,HOUR,value)
+#define	MSTK_SET_REG_DOW(regs,value)	MSTK_SET(regs,DOW,value)
+#define	MSTK_SET_REG_DOM(regs,value)	MSTK_SET(regs,DOM,value)
+#define	MSTK_SET_REG_MONTH(regs,value)	MSTK_SET(regs,MONTH,value)
+#define	MSTK_SET_REG_YEAR(regs,value)	MSTK_SET(regs,YEAR,value)
+
+
+/* The Mostek 48t08 clock chip. Found on Sun4m's I think. It has the
+ * same (basically) layout of the 48t02 chip except for the extra
+ * NVRAM on board (8 KB against the 48t02's 2 KB).
+ */
+#define MOSTEK_48T08_OFFSET	0x0000UL	/* Lower NVRAM portions */
+#define MOSTEK_48T08_48T02	0x1800UL	/* Offset to 48T02 chip */
+
+/* SUN5 systems usually have 48t59 model clock chipsets.  But we keep the older
+ * clock chip definitions around just in case.
+ */
+#define MOSTEK_48T59_OFFSET	0x0000UL	/* Lower NVRAM portions */
+#define MOSTEK_48T59_48T02	0x1800UL	/* Offset to 48T02 chip */
+
+#endif /* !(_SPARC64_MOSTEK_H) */
diff --git a/include/asm-sparc/msgbuf.h b/include/asm-sparc/msgbuf.h
index 8cec9ad0b82..efc7cbe9788 100644
--- a/include/asm-sparc/msgbuf.h
+++ b/include/asm-sparc/msgbuf.h
@@ -1,7 +1,7 @@
-#ifndef _SPARC64_MSGBUF_H
-#define _SPARC64_MSGBUF_H
+#ifndef _SPARC_MSGBUF_H
+#define _SPARC_MSGBUF_H
 
-/* 
+/*
  * The msqid64_ds structure for sparc64 architecture.
  * Note extra padding because this structure is passed back and forth
  * between kernel and user space.
@@ -11,13 +11,20 @@
  * - 2 miscellaneous 32-bit values
  */
 
+#if defined(__sparc__) && defined(__arch64__)
+# define PADDING(x)
+#else
+# define PADDING(x) unsigned int x;
+#endif
+
+
 struct msqid64_ds {
 	struct ipc64_perm msg_perm;
-	unsigned int   __pad1;
+	PADDING(__pad1)
 	__kernel_time_t msg_stime;	/* last msgsnd time */
-	unsigned int   __pad2;
+	PADDING(__pad2)
 	__kernel_time_t msg_rtime;	/* last msgrcv time */
-	unsigned int   __pad3;
+	PADDING(__pad3)
 	__kernel_time_t msg_ctime;	/* last change time */
 	unsigned long  msg_cbytes;	/* current number of bytes on queue */
 	unsigned long  msg_qnum;	/* number of messages in queue */
@@ -27,5 +34,5 @@ struct msqid64_ds {
 	unsigned long  __unused1;
 	unsigned long  __unused2;
 };
-
-#endif /* _SPARC64_MSGBUF_H */
+#undef PADDING
+#endif /* _SPARC_MSGBUF_H */
diff --git a/include/asm-sparc/namei.h b/include/asm-sparc/namei.h
index 0646102fb02..eff944b8e32 100644
--- a/include/asm-sparc/namei.h
+++ b/include/asm-sparc/namei.h
@@ -1,13 +1,8 @@
-/*
- * linux/include/asm-sparc/namei.h
- *
- * Routines to handle famous /usr/gnemul/s*.
- * Included from linux/fs/namei.c
- */
-
-#ifndef __SPARC_NAMEI_H
-#define __SPARC_NAMEI_H
-
-#define __emul_prefix() NULL
-
-#endif /* __SPARC_NAMEI_H */
+#ifndef ___ASM_SPARC_NAMEI_H
+#define ___ASM_SPARC_NAMEI_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/namei_64.h>
+#else
+#include <asm-sparc/namei_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/namei_32.h b/include/asm-sparc/namei_32.h
new file mode 100644
index 00000000000..0646102fb02
--- /dev/null
+++ b/include/asm-sparc/namei_32.h
@@ -0,0 +1,13 @@
+/*
+ * linux/include/asm-sparc/namei.h
+ *
+ * Routines to handle famous /usr/gnemul/s*.
+ * Included from linux/fs/namei.c
+ */
+
+#ifndef __SPARC_NAMEI_H
+#define __SPARC_NAMEI_H
+
+#define __emul_prefix() NULL
+
+#endif /* __SPARC_NAMEI_H */
diff --git a/include/asm-sparc/namei_64.h b/include/asm-sparc/namei_64.h
new file mode 100644
index 00000000000..cbc1b4c0689
--- /dev/null
+++ b/include/asm-sparc/namei_64.h
@@ -0,0 +1,13 @@
+/*
+ * linux/include/asm-sparc64/namei.h
+ *
+ * Routines to handle famous /usr/gnemul/s*.
+ * Included from linux/fs/namei.c
+ */
+
+#ifndef __SPARC64_NAMEI_H
+#define __SPARC64_NAMEI_H
+
+#define __emul_prefix() NULL
+
+#endif /* __SPARC64_NAMEI_H */
diff --git a/include/asm-sparc/ns87303.h b/include/asm-sparc/ns87303.h
new file mode 100644
index 00000000000..686defe6aaa
--- /dev/null
+++ b/include/asm-sparc/ns87303.h
@@ -0,0 +1,118 @@
+/* ns87303.h: Configuration Register Description for the
+ *            National Semiconductor PC87303 (SuperIO).
+ *
+ * Copyright (C) 1997  Eddie C. Dost  (ecd@skynet.be)
+ */
+
+#ifndef _SPARC_NS87303_H
+#define _SPARC_NS87303_H 1
+
+/*
+ * Control Register Index Values
+ */
+#define FER	0x00
+#define FAR	0x01
+#define PTR	0x02
+#define FCR	0x03
+#define PCR	0x04
+#define KRR	0x05
+#define PMC	0x06
+#define TUP	0x07
+#define SID	0x08
+#define ASC	0x09
+#define CS0CF0	0x0a
+#define CS0CF1	0x0b
+#define CS1CF0	0x0c
+#define CS1CF1	0x0d
+
+/* Function Enable Register (FER) bits */
+#define FER_EDM		0x10	/* Encoded Drive and Motor pin information   */
+
+/* Function Address Register (FAR) bits */
+#define FAR_LPT_MASK	0x03
+#define FAR_LPTB	0x00
+#define FAR_LPTA	0x01
+#define FAR_LPTC	0x02
+
+/* Power and Test Register (PTR) bits */
+#define PTR_LPTB_IRQ7	0x08
+#define PTR_LEVEL_IRQ	0x80	/* When not ECP/EPP: Use level IRQ           */
+#define PTR_LPT_REG_DIR	0x80	/* When ECP/EPP: LPT CTR controlls direction */
+				/*               of the parallel port	     */
+
+/* Function Control Register (FCR) bits */
+#define FCR_LDE		0x10	/* Logical Drive Exchange                    */
+#define FCR_ZWS_ENA	0x20	/* Enable short host read/write in ECP/EPP   */
+
+/* Printer Control Register (PCR) bits */
+#define PCR_EPP_ENABLE	0x01
+#define PCR_EPP_IEEE	0x02	/* Enable EPP Version 1.9 (IEEE 1284)        */
+#define PCR_ECP_ENABLE	0x04
+#define PCR_ECP_CLK_ENA	0x08	/* If 0 ECP Clock is stopped on Power down   */
+#define PCR_IRQ_POLAR	0x20	/* If 0 IRQ is level high or negative pulse, */
+				/* if 1 polarity is inverted                 */
+#define PCR_IRQ_ODRAIN	0x40	/* If 1, IRQ is open drain                   */
+
+/* Tape UARTs and Parallel Port Config Register (TUP) bits */
+#define TUP_EPP_TIMO	0x02	/* Enable EPP timeout IRQ                    */
+
+/* Advanced SuperIO Config Register (ASC) bits */
+#define ASC_LPT_IRQ7	0x01	/* Always use IRQ7 for LPT                  */
+#define ASC_DRV2_SEL	0x02	/* Logical Drive Exchange controlled by TDR  */
+
+#define FER_RESERVED	0x00
+#define FAR_RESERVED	0x00
+#define PTR_RESERVED	0x73
+#define FCR_RESERVED	0xc4
+#define PCR_RESERVED	0x10
+#define KRR_RESERVED	0x00
+#define PMC_RESERVED	0x98
+#define TUP_RESERVED	0xfb
+#define SIP_RESERVED	0x00
+#define ASC_RESERVED	0x18
+#define CS0CF0_RESERVED	0x00
+#define CS0CF1_RESERVED	0x08
+#define CS1CF0_RESERVED	0x00
+#define CS1CF1_RESERVED	0x08
+
+#ifdef __KERNEL__
+
+#include <linux/spinlock.h>
+
+#include <asm/system.h>
+#include <asm/io.h>
+
+extern spinlock_t ns87303_lock;
+
+static inline int ns87303_modify(unsigned long port, unsigned int index,
+				     unsigned char clr, unsigned char set)
+{
+	static unsigned char reserved[] = {
+		FER_RESERVED, FAR_RESERVED, PTR_RESERVED, FCR_RESERVED,
+		PCR_RESERVED, KRR_RESERVED, PMC_RESERVED, TUP_RESERVED,
+		SIP_RESERVED, ASC_RESERVED, CS0CF0_RESERVED, CS0CF1_RESERVED,
+		CS1CF0_RESERVED, CS1CF1_RESERVED
+	};
+	unsigned long flags;
+	unsigned char value;
+
+	if (index > 0x0d)
+		return -EINVAL;
+
+	spin_lock_irqsave(&ns87303_lock, flags);
+
+	outb(index, port);
+	value = inb(port + 1);
+	value &= ~(reserved[index] | clr);
+	value |= set;
+	outb(value, port + 1);
+	outb(value, port + 1);
+
+	spin_unlock_irqrestore(&ns87303_lock, flags);
+
+	return 0;
+}
+
+#endif /* __KERNEL__ */
+
+#endif /* !(_SPARC_NS87303_H) */
diff --git a/include/asm-sparc/of_platform.h b/include/asm-sparc/of_platform.h
index 38334351c36..851eb84d737 100644
--- a/include/asm-sparc/of_platform.h
+++ b/include/asm-sparc/of_platform.h
@@ -1,24 +1,8 @@
-#ifndef _ASM_SPARC_OF_PLATFORM_H
-#define _ASM_SPARC_OF_PLATFORM_H
-/*
- *    Copyright (C) 2006 Benjamin Herrenschmidt, IBM Corp.
- *			 <benh@kernel.crashing.org>
- *    Modified for Sparc by merging parts of asm-sparc/of_device.h
- *		by Stephen Rothwell
- *
- *  This program is free software; you can redistribute it and/or
- *  modify it under the terms of the GNU General Public License
- *  as published by the Free Software Foundation; either version
- *  2 of the License, or (at your option) any later version.
- *
- */
-
-/* This is just here during the transition */
-#include <linux/of_platform.h>
-
-extern struct bus_type ebus_bus_type;
-extern struct bus_type sbus_bus_type;
-
-#define of_bus_type	of_platform_bus_type	/* for compatibility */
-
-#endif	/* _ASM_SPARC_OF_PLATFORM_H */
+#ifndef ___ASM_SPARC_OF_PLATFORM_H
+#define ___ASM_SPARC_OF_PLATFORM_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/of_platform_64.h>
+#else
+#include <asm-sparc/of_platform_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/of_platform_32.h b/include/asm-sparc/of_platform_32.h
new file mode 100644
index 00000000000..38334351c36
--- /dev/null
+++ b/include/asm-sparc/of_platform_32.h
@@ -0,0 +1,24 @@
+#ifndef _ASM_SPARC_OF_PLATFORM_H
+#define _ASM_SPARC_OF_PLATFORM_H
+/*
+ *    Copyright (C) 2006 Benjamin Herrenschmidt, IBM Corp.
+ *			 <benh@kernel.crashing.org>
+ *    Modified for Sparc by merging parts of asm-sparc/of_device.h
+ *		by Stephen Rothwell
+ *
+ *  This program is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU General Public License
+ *  as published by the Free Software Foundation; either version
+ *  2 of the License, or (at your option) any later version.
+ *
+ */
+
+/* This is just here during the transition */
+#include <linux/of_platform.h>
+
+extern struct bus_type ebus_bus_type;
+extern struct bus_type sbus_bus_type;
+
+#define of_bus_type	of_platform_bus_type	/* for compatibility */
+
+#endif	/* _ASM_SPARC_OF_PLATFORM_H */
diff --git a/include/asm-sparc/of_platform_64.h b/include/asm-sparc/of_platform_64.h
new file mode 100644
index 00000000000..78aa032b674
--- /dev/null
+++ b/include/asm-sparc/of_platform_64.h
@@ -0,0 +1,25 @@
+#ifndef _ASM_SPARC64_OF_PLATFORM_H
+#define _ASM_SPARC64_OF_PLATFORM_H
+/*
+ *    Copyright (C) 2006 Benjamin Herrenschmidt, IBM Corp.
+ *			 <benh@kernel.crashing.org>
+ *    Modified for Sparc by merging parts of asm-sparc/of_device.h
+ *		by Stephen Rothwell
+ *
+ *  This program is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU General Public License
+ *  as published by the Free Software Foundation; either version
+ *  2 of the License, or (at your option) any later version.
+ *
+ */
+
+/* This is just here during the transition */
+#include <linux/of_platform.h>
+
+extern struct bus_type isa_bus_type;
+extern struct bus_type ebus_bus_type;
+extern struct bus_type sbus_bus_type;
+
+#define of_bus_type	of_platform_bus_type	/* for compatibility */
+
+#endif	/* _ASM_SPARC64_OF_PLATFORM_H */
diff --git a/include/asm-sparc/openprom.h b/include/asm-sparc/openprom.h
index ed4b6bc2b10..8c349f06199 100644
--- a/include/asm-sparc/openprom.h
+++ b/include/asm-sparc/openprom.h
@@ -1,257 +1,8 @@
-#ifndef __SPARC_OPENPROM_H
-#define __SPARC_OPENPROM_H
-
-/* openprom.h:  Prom structures and defines for access to the OPENBOOT
- *              prom routines and data areas.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#include <asm/vaddrs.h>
-
-/* Empirical constants... */
-#define	LINUX_OPPROM_MAGIC      0x10010407
-
-#ifndef __ASSEMBLY__
-/* V0 prom device operations. */
-struct linux_dev_v0_funcs {
-	int (*v0_devopen)(char *device_str);
-	int (*v0_devclose)(int dev_desc);
-	int (*v0_rdblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
-	int (*v0_wrblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
-	int (*v0_wrnetdev)(int dev_desc, int num_bytes, char *buf);
-	int (*v0_rdnetdev)(int dev_desc, int num_bytes, char *buf);
-	int (*v0_rdchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
-	int (*v0_wrchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
-	int (*v0_seekdev)(int dev_desc, long logical_offst, int from);
-};
-
-/* V2 and later prom device operations. */
-struct linux_dev_v2_funcs {
-	int (*v2_inst2pkg)(int d);	/* Convert ihandle to phandle */
-	char * (*v2_dumb_mem_alloc)(char *va, unsigned sz);
-	void (*v2_dumb_mem_free)(char *va, unsigned sz);
-
-	/* To map devices into virtual I/O space. */
-	char * (*v2_dumb_mmap)(char *virta, int which_io, unsigned paddr, unsigned sz);
-	void (*v2_dumb_munmap)(char *virta, unsigned size);
-
-	int (*v2_dev_open)(char *devpath);
-	void (*v2_dev_close)(int d);
-	int (*v2_dev_read)(int d, char *buf, int nbytes);
-	int (*v2_dev_write)(int d, char *buf, int nbytes);
-	int (*v2_dev_seek)(int d, int hi, int lo);
-
-	/* Never issued (multistage load support) */
-	void (*v2_wheee2)(void);
-	void (*v2_wheee3)(void);
-};
-
-struct linux_mlist_v0 {
-	struct linux_mlist_v0 *theres_more;
-	char *start_adr;
-	unsigned num_bytes;
-};
-
-struct linux_mem_v0 {
-	struct linux_mlist_v0 **v0_totphys;
-	struct linux_mlist_v0 **v0_prommap;
-	struct linux_mlist_v0 **v0_available; /* What we can use */
-};
-
-/* Arguments sent to the kernel from the boot prompt. */
-struct linux_arguments_v0 {
-	char *argv[8];
-	char args[100];
-	char boot_dev[2];
-	int boot_dev_ctrl;
-	int boot_dev_unit;
-	int dev_partition;
-	char *kernel_file_name;
-	void *aieee1;           /* XXX */
-};
-
-/* V2 and up boot things. */
-struct linux_bootargs_v2 {
-	char **bootpath;
-	char **bootargs;
-	int *fd_stdin;
-	int *fd_stdout;
-};
-
-/* The top level PROM vector. */
-struct linux_romvec {
-	/* Version numbers. */
-	unsigned int pv_magic_cookie;
-	unsigned int pv_romvers;
-	unsigned int pv_plugin_revision;
-	unsigned int pv_printrev;
-
-	/* Version 0 memory descriptors. */
-	struct linux_mem_v0 pv_v0mem;
-
-	/* Node operations. */
-	struct linux_nodeops *pv_nodeops;
-
-	char **pv_bootstr;
-	struct linux_dev_v0_funcs pv_v0devops;
-
-	char *pv_stdin;
-	char *pv_stdout;
-#define	PROMDEV_KBD	0		/* input from keyboard */
-#define	PROMDEV_SCREEN	0		/* output to screen */
-#define	PROMDEV_TTYA	1		/* in/out to ttya */
-#define	PROMDEV_TTYB	2		/* in/out to ttyb */
-
-	/* Blocking getchar/putchar.  NOT REENTRANT! (grr) */
-	int (*pv_getchar)(void);
-	void (*pv_putchar)(int ch);
-
-	/* Non-blocking variants. */
-	int (*pv_nbgetchar)(void);
-	int (*pv_nbputchar)(int ch);
-
-	void (*pv_putstr)(char *str, int len);
-
-	/* Miscellany. */
-	void (*pv_reboot)(char *bootstr);
-	void (*pv_printf)(__const__ char *fmt, ...);
-	void (*pv_abort)(void);
-	__volatile__ int *pv_ticks;
-	void (*pv_halt)(void);
-	void (**pv_synchook)(void);
-
-	/* Evaluate a forth string, not different proto for V0 and V2->up. */
-	union {
-		void (*v0_eval)(int len, char *str);
-		void (*v2_eval)(char *str);
-	} pv_fortheval;
-
-	struct linux_arguments_v0 **pv_v0bootargs;
-
-	/* Get ether address. */
-	unsigned int (*pv_enaddr)(int d, char *enaddr);
-
-	struct linux_bootargs_v2 pv_v2bootargs;
-	struct linux_dev_v2_funcs pv_v2devops;
-
-	int filler[15];
-
-	/* This one is sun4c/sun4 only. */
-	void (*pv_setctxt)(int ctxt, char *va, int pmeg);
-
-	/* Prom version 3 Multiprocessor routines. This stuff is crazy.
-	 * No joke. Calling these when there is only one cpu probably
-	 * crashes the machine, have to test this. :-)
-	 */
-
-	/* v3_cpustart() will start the cpu 'whichcpu' in mmu-context
-	 * 'thiscontext' executing at address 'prog_counter'
-	 */
-	int (*v3_cpustart)(unsigned int whichcpu, int ctxtbl_ptr,
-			   int thiscontext, char *prog_counter);
-
-	/* v3_cpustop() will cause cpu 'whichcpu' to stop executing
-	 * until a resume cpu call is made.
-	 */
-	int (*v3_cpustop)(unsigned int whichcpu);
-
-	/* v3_cpuidle() will idle cpu 'whichcpu' until a stop or
-	 * resume cpu call is made.
-	 */
-	int (*v3_cpuidle)(unsigned int whichcpu);
-
-	/* v3_cpuresume() will resume processor 'whichcpu' executing
-	 * starting with whatever 'pc' and 'npc' were left at the
-	 * last 'idle' or 'stop' call.
-	 */
-	int (*v3_cpuresume)(unsigned int whichcpu);
-};
-
-/* Routines for traversing the prom device tree. */
-struct linux_nodeops {
-	int (*no_nextnode)(int node);
-	int (*no_child)(int node);
-	int (*no_proplen)(int node, char *name);
-	int (*no_getprop)(int node, char *name, char *val);
-	int (*no_setprop)(int node, char *name, char *val, int len);
-	char * (*no_nextprop)(int node, char *name);
-};
-
-/* More fun PROM structures for device probing. */
-#define PROMREG_MAX     16
-#define PROMVADDR_MAX   16
-#define PROMINTR_MAX    15
-
-struct linux_prom_registers {
-	unsigned int which_io;         /* is this in OBIO space? */
-	unsigned int phys_addr;        /* The physical address of this register */
-	unsigned int reg_size;         /* How many bytes does this register take up? */
-};
-
-struct linux_prom_irqs {
-	int pri;    /* IRQ priority */
-	int vector; /* This is foobar, what does it do? */
-};
-
-/* Element of the "ranges" vector */
-struct linux_prom_ranges {
-	unsigned int ot_child_space;
-	unsigned int ot_child_base;		/* Bus feels this */
-	unsigned int ot_parent_space;
-	unsigned int ot_parent_base;		/* CPU looks from here */
-	unsigned int or_size;
-};
-
-/* Ranges and reg properties are a bit different for PCI. */
-struct linux_prom_pci_registers {
-	/* 
-	 * We don't know what information this field contain.
-	 * We guess, PCI device function is in bits 15:8
-	 * So, ...
-	 */
-	unsigned int which_io;  /* Let it be which_io */
-
-	unsigned int phys_hi;
-	unsigned int phys_lo;
-
-	unsigned int size_hi;
-	unsigned int size_lo;
-};
-
-struct linux_prom_pci_ranges {
-	unsigned int child_phys_hi;	/* Only certain bits are encoded here. */
-	unsigned int child_phys_mid;
-	unsigned int child_phys_lo;
-
-	unsigned int parent_phys_hi;
-	unsigned int parent_phys_lo;
-
-	unsigned int size_hi;
-	unsigned int size_lo;
-};
-
-struct linux_prom_pci_assigned_addresses {
-	unsigned int which_io;
-
-	unsigned int phys_hi;
-	unsigned int phys_lo;
-
-	unsigned int size_hi;
-	unsigned int size_lo;
-};
-
-struct linux_prom_ebus_ranges {
-	unsigned int child_phys_hi;
-	unsigned int child_phys_lo;
-
-	unsigned int parent_phys_hi;
-	unsigned int parent_phys_mid;
-	unsigned int parent_phys_lo;
-
-	unsigned int size;
-};
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC_OPENPROM_H) */
+#ifndef ___ASM_SPARC_OPENPROM_H
+#define ___ASM_SPARC_OPENPROM_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/openprom_64.h>
+#else
+#include <asm-sparc/openprom_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/openprom_32.h b/include/asm-sparc/openprom_32.h
new file mode 100644
index 00000000000..8b1649f29ed
--- /dev/null
+++ b/include/asm-sparc/openprom_32.h
@@ -0,0 +1,255 @@
+#ifndef __SPARC_OPENPROM_H
+#define __SPARC_OPENPROM_H
+
+/* openprom.h:  Prom structures and defines for access to the OPENBOOT
+ *              prom routines and data areas.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+/* Empirical constants... */
+#define	LINUX_OPPROM_MAGIC      0x10010407
+
+#ifndef __ASSEMBLY__
+/* V0 prom device operations. */
+struct linux_dev_v0_funcs {
+	int (*v0_devopen)(char *device_str);
+	int (*v0_devclose)(int dev_desc);
+	int (*v0_rdblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
+	int (*v0_wrblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
+	int (*v0_wrnetdev)(int dev_desc, int num_bytes, char *buf);
+	int (*v0_rdnetdev)(int dev_desc, int num_bytes, char *buf);
+	int (*v0_rdchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
+	int (*v0_wrchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
+	int (*v0_seekdev)(int dev_desc, long logical_offst, int from);
+};
+
+/* V2 and later prom device operations. */
+struct linux_dev_v2_funcs {
+	int (*v2_inst2pkg)(int d);	/* Convert ihandle to phandle */
+	char * (*v2_dumb_mem_alloc)(char *va, unsigned sz);
+	void (*v2_dumb_mem_free)(char *va, unsigned sz);
+
+	/* To map devices into virtual I/O space. */
+	char * (*v2_dumb_mmap)(char *virta, int which_io, unsigned paddr, unsigned sz);
+	void (*v2_dumb_munmap)(char *virta, unsigned size);
+
+	int (*v2_dev_open)(char *devpath);
+	void (*v2_dev_close)(int d);
+	int (*v2_dev_read)(int d, char *buf, int nbytes);
+	int (*v2_dev_write)(int d, char *buf, int nbytes);
+	int (*v2_dev_seek)(int d, int hi, int lo);
+
+	/* Never issued (multistage load support) */
+	void (*v2_wheee2)(void);
+	void (*v2_wheee3)(void);
+};
+
+struct linux_mlist_v0 {
+	struct linux_mlist_v0 *theres_more;
+	char *start_adr;
+	unsigned num_bytes;
+};
+
+struct linux_mem_v0 {
+	struct linux_mlist_v0 **v0_totphys;
+	struct linux_mlist_v0 **v0_prommap;
+	struct linux_mlist_v0 **v0_available; /* What we can use */
+};
+
+/* Arguments sent to the kernel from the boot prompt. */
+struct linux_arguments_v0 {
+	char *argv[8];
+	char args[100];
+	char boot_dev[2];
+	int boot_dev_ctrl;
+	int boot_dev_unit;
+	int dev_partition;
+	char *kernel_file_name;
+	void *aieee1;           /* XXX */
+};
+
+/* V2 and up boot things. */
+struct linux_bootargs_v2 {
+	char **bootpath;
+	char **bootargs;
+	int *fd_stdin;
+	int *fd_stdout;
+};
+
+/* The top level PROM vector. */
+struct linux_romvec {
+	/* Version numbers. */
+	unsigned int pv_magic_cookie;
+	unsigned int pv_romvers;
+	unsigned int pv_plugin_revision;
+	unsigned int pv_printrev;
+
+	/* Version 0 memory descriptors. */
+	struct linux_mem_v0 pv_v0mem;
+
+	/* Node operations. */
+	struct linux_nodeops *pv_nodeops;
+
+	char **pv_bootstr;
+	struct linux_dev_v0_funcs pv_v0devops;
+
+	char *pv_stdin;
+	char *pv_stdout;
+#define	PROMDEV_KBD	0		/* input from keyboard */
+#define	PROMDEV_SCREEN	0		/* output to screen */
+#define	PROMDEV_TTYA	1		/* in/out to ttya */
+#define	PROMDEV_TTYB	2		/* in/out to ttyb */
+
+	/* Blocking getchar/putchar.  NOT REENTRANT! (grr) */
+	int (*pv_getchar)(void);
+	void (*pv_putchar)(int ch);
+
+	/* Non-blocking variants. */
+	int (*pv_nbgetchar)(void);
+	int (*pv_nbputchar)(int ch);
+
+	void (*pv_putstr)(char *str, int len);
+
+	/* Miscellany. */
+	void (*pv_reboot)(char *bootstr);
+	void (*pv_printf)(__const__ char *fmt, ...);
+	void (*pv_abort)(void);
+	__volatile__ int *pv_ticks;
+	void (*pv_halt)(void);
+	void (**pv_synchook)(void);
+
+	/* Evaluate a forth string, not different proto for V0 and V2->up. */
+	union {
+		void (*v0_eval)(int len, char *str);
+		void (*v2_eval)(char *str);
+	} pv_fortheval;
+
+	struct linux_arguments_v0 **pv_v0bootargs;
+
+	/* Get ether address. */
+	unsigned int (*pv_enaddr)(int d, char *enaddr);
+
+	struct linux_bootargs_v2 pv_v2bootargs;
+	struct linux_dev_v2_funcs pv_v2devops;
+
+	int filler[15];
+
+	/* This one is sun4c/sun4 only. */
+	void (*pv_setctxt)(int ctxt, char *va, int pmeg);
+
+	/* Prom version 3 Multiprocessor routines. This stuff is crazy.
+	 * No joke. Calling these when there is only one cpu probably
+	 * crashes the machine, have to test this. :-)
+	 */
+
+	/* v3_cpustart() will start the cpu 'whichcpu' in mmu-context
+	 * 'thiscontext' executing at address 'prog_counter'
+	 */
+	int (*v3_cpustart)(unsigned int whichcpu, int ctxtbl_ptr,
+			   int thiscontext, char *prog_counter);
+
+	/* v3_cpustop() will cause cpu 'whichcpu' to stop executing
+	 * until a resume cpu call is made.
+	 */
+	int (*v3_cpustop)(unsigned int whichcpu);
+
+	/* v3_cpuidle() will idle cpu 'whichcpu' until a stop or
+	 * resume cpu call is made.
+	 */
+	int (*v3_cpuidle)(unsigned int whichcpu);
+
+	/* v3_cpuresume() will resume processor 'whichcpu' executing
+	 * starting with whatever 'pc' and 'npc' were left at the
+	 * last 'idle' or 'stop' call.
+	 */
+	int (*v3_cpuresume)(unsigned int whichcpu);
+};
+
+/* Routines for traversing the prom device tree. */
+struct linux_nodeops {
+	int (*no_nextnode)(int node);
+	int (*no_child)(int node);
+	int (*no_proplen)(int node, char *name);
+	int (*no_getprop)(int node, char *name, char *val);
+	int (*no_setprop)(int node, char *name, char *val, int len);
+	char * (*no_nextprop)(int node, char *name);
+};
+
+/* More fun PROM structures for device probing. */
+#define PROMREG_MAX     16
+#define PROMVADDR_MAX   16
+#define PROMINTR_MAX    15
+
+struct linux_prom_registers {
+	unsigned int which_io;         /* is this in OBIO space? */
+	unsigned int phys_addr;        /* The physical address of this register */
+	unsigned int reg_size;         /* How many bytes does this register take up? */
+};
+
+struct linux_prom_irqs {
+	int pri;    /* IRQ priority */
+	int vector; /* This is foobar, what does it do? */
+};
+
+/* Element of the "ranges" vector */
+struct linux_prom_ranges {
+	unsigned int ot_child_space;
+	unsigned int ot_child_base;		/* Bus feels this */
+	unsigned int ot_parent_space;
+	unsigned int ot_parent_base;		/* CPU looks from here */
+	unsigned int or_size;
+};
+
+/* Ranges and reg properties are a bit different for PCI. */
+struct linux_prom_pci_registers {
+	/*
+	 * We don't know what information this field contain.
+	 * We guess, PCI device function is in bits 15:8
+	 * So, ...
+	 */
+	unsigned int which_io;  /* Let it be which_io */
+
+	unsigned int phys_hi;
+	unsigned int phys_lo;
+
+	unsigned int size_hi;
+	unsigned int size_lo;
+};
+
+struct linux_prom_pci_ranges {
+	unsigned int child_phys_hi;	/* Only certain bits are encoded here. */
+	unsigned int child_phys_mid;
+	unsigned int child_phys_lo;
+
+	unsigned int parent_phys_hi;
+	unsigned int parent_phys_lo;
+
+	unsigned int size_hi;
+	unsigned int size_lo;
+};
+
+struct linux_prom_pci_assigned_addresses {
+	unsigned int which_io;
+
+	unsigned int phys_hi;
+	unsigned int phys_lo;
+
+	unsigned int size_hi;
+	unsigned int size_lo;
+};
+
+struct linux_prom_ebus_ranges {
+	unsigned int child_phys_hi;
+	unsigned int child_phys_lo;
+
+	unsigned int parent_phys_hi;
+	unsigned int parent_phys_mid;
+	unsigned int parent_phys_lo;
+
+	unsigned int size;
+};
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC_OPENPROM_H) */
diff --git a/include/asm-sparc/openprom_64.h b/include/asm-sparc/openprom_64.h
new file mode 100644
index 00000000000..b69e4a8c917
--- /dev/null
+++ b/include/asm-sparc/openprom_64.h
@@ -0,0 +1,280 @@
+#ifndef __SPARC64_OPENPROM_H
+#define __SPARC64_OPENPROM_H
+
+/* openprom.h:  Prom structures and defines for access to the OPENBOOT
+ *              prom routines and data areas.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __ASSEMBLY__
+/* V0 prom device operations. */
+struct linux_dev_v0_funcs {
+	int (*v0_devopen)(char *device_str);
+	int (*v0_devclose)(int dev_desc);
+	int (*v0_rdblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
+	int (*v0_wrblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
+	int (*v0_wrnetdev)(int dev_desc, int num_bytes, char *buf);
+	int (*v0_rdnetdev)(int dev_desc, int num_bytes, char *buf);
+	int (*v0_rdchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
+	int (*v0_wrchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
+	int (*v0_seekdev)(int dev_desc, long logical_offst, int from);
+};
+
+/* V2 and later prom device operations. */
+struct linux_dev_v2_funcs {
+	int (*v2_inst2pkg)(int d);	/* Convert ihandle to phandle */
+	char * (*v2_dumb_mem_alloc)(char *va, unsigned sz);
+	void (*v2_dumb_mem_free)(char *va, unsigned sz);
+
+	/* To map devices into virtual I/O space. */
+	char * (*v2_dumb_mmap)(char *virta, int which_io, unsigned paddr, unsigned sz);
+	void (*v2_dumb_munmap)(char *virta, unsigned size);
+
+	int (*v2_dev_open)(char *devpath);
+	void (*v2_dev_close)(int d);
+	int (*v2_dev_read)(int d, char *buf, int nbytes);
+	int (*v2_dev_write)(int d, char *buf, int nbytes);
+	int (*v2_dev_seek)(int d, int hi, int lo);
+
+	/* Never issued (multistage load support) */
+	void (*v2_wheee2)(void);
+	void (*v2_wheee3)(void);
+};
+
+struct linux_mlist_v0 {
+	struct linux_mlist_v0 *theres_more;
+	unsigned start_adr;
+	unsigned num_bytes;
+};
+
+struct linux_mem_v0 {
+	struct linux_mlist_v0 **v0_totphys;
+	struct linux_mlist_v0 **v0_prommap;
+	struct linux_mlist_v0 **v0_available; /* What we can use */
+};
+
+/* Arguments sent to the kernel from the boot prompt. */
+struct linux_arguments_v0 {
+	char *argv[8];
+	char args[100];
+	char boot_dev[2];
+	int boot_dev_ctrl;
+	int boot_dev_unit;
+	int dev_partition;
+	char *kernel_file_name;
+	void *aieee1;           /* XXX */
+};
+
+/* V2 and up boot things. */
+struct linux_bootargs_v2 {
+	char **bootpath;
+	char **bootargs;
+	int *fd_stdin;
+	int *fd_stdout;
+};
+
+/* The top level PROM vector. */
+struct linux_romvec {
+	/* Version numbers. */
+	unsigned int pv_magic_cookie;
+	unsigned int pv_romvers;
+	unsigned int pv_plugin_revision;
+	unsigned int pv_printrev;
+
+	/* Version 0 memory descriptors. */
+	struct linux_mem_v0 pv_v0mem;
+
+	/* Node operations. */
+	struct linux_nodeops *pv_nodeops;
+
+	char **pv_bootstr;
+	struct linux_dev_v0_funcs pv_v0devops;
+
+	char *pv_stdin;
+	char *pv_stdout;
+#define	PROMDEV_KBD	0		/* input from keyboard */
+#define	PROMDEV_SCREEN	0		/* output to screen */
+#define	PROMDEV_TTYA	1		/* in/out to ttya */
+#define	PROMDEV_TTYB	2		/* in/out to ttyb */
+
+	/* Blocking getchar/putchar.  NOT REENTRANT! (grr) */
+	int (*pv_getchar)(void);
+	void (*pv_putchar)(int ch);
+
+	/* Non-blocking variants. */
+	int (*pv_nbgetchar)(void);
+	int (*pv_nbputchar)(int ch);
+
+	void (*pv_putstr)(char *str, int len);
+
+	/* Miscellany. */
+	void (*pv_reboot)(char *bootstr);
+	void (*pv_printf)(__const__ char *fmt, ...);
+	void (*pv_abort)(void);
+	__volatile__ int *pv_ticks;
+	void (*pv_halt)(void);
+	void (**pv_synchook)(void);
+
+	/* Evaluate a forth string, not different proto for V0 and V2->up. */
+	union {
+		void (*v0_eval)(int len, char *str);
+		void (*v2_eval)(char *str);
+	} pv_fortheval;
+
+	struct linux_arguments_v0 **pv_v0bootargs;
+
+	/* Get ether address. */
+	unsigned int (*pv_enaddr)(int d, char *enaddr);
+
+	struct linux_bootargs_v2 pv_v2bootargs;
+	struct linux_dev_v2_funcs pv_v2devops;
+
+	int filler[15];
+
+	/* This one is sun4c/sun4 only. */
+	void (*pv_setctxt)(int ctxt, char *va, int pmeg);
+
+	/* Prom version 3 Multiprocessor routines. This stuff is crazy.
+	 * No joke. Calling these when there is only one cpu probably
+	 * crashes the machine, have to test this. :-)
+	 */
+
+	/* v3_cpustart() will start the cpu 'whichcpu' in mmu-context
+	 * 'thiscontext' executing at address 'prog_counter'
+	 */
+	int (*v3_cpustart)(unsigned int whichcpu, int ctxtbl_ptr,
+			   int thiscontext, char *prog_counter);
+
+	/* v3_cpustop() will cause cpu 'whichcpu' to stop executing
+	 * until a resume cpu call is made.
+	 */
+	int (*v3_cpustop)(unsigned int whichcpu);
+
+	/* v3_cpuidle() will idle cpu 'whichcpu' until a stop or
+	 * resume cpu call is made.
+	 */
+	int (*v3_cpuidle)(unsigned int whichcpu);
+
+	/* v3_cpuresume() will resume processor 'whichcpu' executing
+	 * starting with whatever 'pc' and 'npc' were left at the
+	 * last 'idle' or 'stop' call.
+	 */
+	int (*v3_cpuresume)(unsigned int whichcpu);
+};
+
+/* Routines for traversing the prom device tree. */
+struct linux_nodeops {
+	int (*no_nextnode)(int node);
+	int (*no_child)(int node);
+	int (*no_proplen)(int node, char *name);
+	int (*no_getprop)(int node, char *name, char *val);
+	int (*no_setprop)(int node, char *name, char *val, int len);
+	char * (*no_nextprop)(int node, char *name);
+};
+
+/* More fun PROM structures for device probing. */
+#define PROMREG_MAX     24
+#define PROMVADDR_MAX   16
+#define PROMINTR_MAX    32
+
+struct linux_prom_registers {
+	unsigned which_io;	/* hi part of physical address			*/
+	unsigned phys_addr;	/* The physical address of this register	*/
+	int reg_size;		/* How many bytes does this register take up?	*/
+};
+
+struct linux_prom64_registers {
+	unsigned long phys_addr;
+	unsigned long reg_size;
+};
+
+struct linux_prom_irqs {
+	int pri;    /* IRQ priority */
+	int vector; /* This is foobar, what does it do? */
+};
+
+/* Element of the "ranges" vector */
+struct linux_prom_ranges {
+	unsigned int ot_child_space;
+	unsigned int ot_child_base;		/* Bus feels this */
+	unsigned int ot_parent_space;
+	unsigned int ot_parent_base;		/* CPU looks from here */
+	unsigned int or_size;
+};
+
+struct linux_prom64_ranges {
+	unsigned long ot_child_base;		/* Bus feels this */
+	unsigned long ot_parent_base;		/* CPU looks from here */
+	unsigned long or_size;
+};
+
+/* Ranges and reg properties are a bit different for PCI. */
+struct linux_prom_pci_registers {
+	unsigned int phys_hi;
+	unsigned int phys_mid;
+	unsigned int phys_lo;
+
+	unsigned int size_hi;
+	unsigned int size_lo;
+};
+
+struct linux_prom_pci_ranges {
+	unsigned int child_phys_hi;	/* Only certain bits are encoded here. */
+	unsigned int child_phys_mid;
+	unsigned int child_phys_lo;
+
+	unsigned int parent_phys_hi;
+	unsigned int parent_phys_lo;
+
+	unsigned int size_hi;
+	unsigned int size_lo;
+};
+
+struct linux_prom_pci_intmap {
+	unsigned int phys_hi;
+	unsigned int phys_mid;
+	unsigned int phys_lo;
+
+	unsigned int interrupt;
+
+	int          cnode;
+	unsigned int cinterrupt;
+};
+
+struct linux_prom_pci_intmask {
+	unsigned int phys_hi;
+	unsigned int phys_mid;
+	unsigned int phys_lo;
+	unsigned int interrupt;
+};
+
+struct linux_prom_ebus_ranges {
+	unsigned int child_phys_hi;
+	unsigned int child_phys_lo;
+
+	unsigned int parent_phys_hi;
+	unsigned int parent_phys_mid;
+	unsigned int parent_phys_lo;
+
+	unsigned int size;
+};
+
+struct linux_prom_ebus_intmap {
+	unsigned int phys_hi;
+	unsigned int phys_lo;
+
+	unsigned int interrupt;
+
+	int          cnode;
+	unsigned int cinterrupt;
+};
+
+struct linux_prom_ebus_intmask {
+	unsigned int phys_hi;
+	unsigned int phys_lo;
+	unsigned int interrupt;
+};
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC64_OPENPROM_H) */
diff --git a/include/asm-sparc/oplib.h b/include/asm-sparc/oplib.h
index 61c3ca6a8ac..e88d7c04a29 100644
--- a/include/asm-sparc/oplib.h
+++ b/include/asm-sparc/oplib.h
@@ -1,273 +1,8 @@
-/*
- * oplib.h:  Describes the interface and available routines in the
- *           Linux Prom library.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __SPARC_OPLIB_H
-#define __SPARC_OPLIB_H
-
-#include <asm/openprom.h>
-#include <linux/spinlock.h>
-#include <linux/compiler.h>
-
-/* The master romvec pointer... */
-extern struct linux_romvec *romvec;
-
-/* Enumeration to describe the prom major version we have detected. */
-enum prom_major_version {
-	PROM_V0,      /* Original sun4c V0 prom */
-	PROM_V2,      /* sun4c and early sun4m V2 prom */
-	PROM_V3,      /* sun4m and later, up to sun4d/sun4e machines V3 */
-	PROM_P1275,   /* IEEE compliant ISA based Sun PROM, only sun4u */
-        PROM_AP1000,  /* actually no prom at all */
-	PROM_SUN4,    /* Old sun4 proms are totally different, but we'll shoehorn it to make it fit */
-};
-
-extern enum prom_major_version prom_vers;
-/* Revision, and firmware revision. */
-extern unsigned int prom_rev, prom_prev;
-
-/* Root node of the prom device tree, this stays constant after
- * initialization is complete.
- */
-extern int prom_root_node;
-
-/* Pointer to prom structure containing the device tree traversal
- * and usage utility functions.  Only prom-lib should use these,
- * users use the interface defined by the library only!
- */
-extern struct linux_nodeops *prom_nodeops;
-
-/* The functions... */
-
-/* You must call prom_init() before using any of the library services,
- * preferably as early as possible.  Pass it the romvec pointer.
- */
-extern void prom_init(struct linux_romvec *rom_ptr);
-
-/* Boot argument acquisition, returns the boot command line string. */
-extern char *prom_getbootargs(void);
-
-/* Device utilities. */
-
-/* Map and unmap devices in IO space at virtual addresses. Note that the
- * virtual address you pass is a request and the prom may put your mappings
- * somewhere else, so check your return value as that is where your new
- * mappings really are!
- *
- * Another note, these are only available on V2 or higher proms!
- */
-extern char *prom_mapio(char *virt_hint, int io_space, unsigned int phys_addr, unsigned int num_bytes);
-extern void prom_unmapio(char *virt_addr, unsigned int num_bytes);
-
-/* Device operations. */
-
-/* Open the device described by the passed string.  Note, that the format
- * of the string is different on V0 vs. V2->higher proms.  The caller must
- * know what he/she is doing!  Returns the device descriptor, an int.
- */
-extern int prom_devopen(char *device_string);
-
-/* Close a previously opened device described by the passed integer
- * descriptor.
- */
-extern int prom_devclose(int device_handle);
-
-/* Do a seek operation on the device described by the passed integer
- * descriptor.
- */
-extern void prom_seek(int device_handle, unsigned int seek_hival,
-		      unsigned int seek_lowval);
-
-/* Miscellaneous routines, don't really fit in any category per se. */
-
-/* Reboot the machine with the command line passed. */
-extern void prom_reboot(char *boot_command);
-
-/* Evaluate the forth string passed. */
-extern void prom_feval(char *forth_string);
-
-/* Enter the prom, with possibility of continuation with the 'go'
- * command in newer proms.
- */
-extern void prom_cmdline(void);
-
-/* Enter the prom, with no chance of continuation for the stand-alone
- * which calls this.
- */
-extern void prom_halt(void) __attribute__ ((noreturn));
-
-/* Set the PROM 'sync' callback function to the passed function pointer.
- * When the user gives the 'sync' command at the prom prompt while the
- * kernel is still active, the prom will call this routine.
- *
- * XXX The arguments are different on V0 vs. V2->higher proms, grrr! XXX
- */
-typedef void (*sync_func_t)(void);
-extern void prom_setsync(sync_func_t func_ptr);
-
-/* Acquire the IDPROM of the root node in the prom device tree.  This
- * gets passed a buffer where you would like it stuffed.  The return value
- * is the format type of this idprom or 0xff on error.
- */
-extern unsigned char prom_get_idprom(char *idp_buffer, int idpbuf_size);
-
-/* Get the prom major version. */
-extern int prom_version(void);
-
-/* Get the prom plugin revision. */
-extern int prom_getrev(void);
-
-/* Get the prom firmware revision. */
-extern int prom_getprev(void);
-
-/* Character operations to/from the console.... */
-
-/* Non-blocking get character from console. */
-extern int prom_nbgetchar(void);
-
-/* Non-blocking put character to console. */
-extern int prom_nbputchar(char character);
-
-/* Blocking get character from console. */
-extern char prom_getchar(void);
-
-/* Blocking put character to console. */
-extern void prom_putchar(char character);
-
-/* Prom's internal routines, don't use in kernel/boot code. */
-extern void prom_printf(char *fmt, ...);
-extern void prom_write(const char *buf, unsigned int len);
-
-/* Multiprocessor operations... */
-
-/* Start the CPU with the given device tree node, context table, and context
- * at the passed program counter.
- */
-extern int prom_startcpu(int cpunode, struct linux_prom_registers *context_table,
-			 int context, char *program_counter);
-
-/* Stop the CPU with the passed device tree node. */
-extern int prom_stopcpu(int cpunode);
-
-/* Idle the CPU with the passed device tree node. */
-extern int prom_idlecpu(int cpunode);
-
-/* Re-Start the CPU with the passed device tree node. */
-extern int prom_restartcpu(int cpunode);
-
-/* PROM memory allocation facilities... */
-
-/* Allocated at possibly the given virtual address a chunk of the
- * indicated size.
- */
-extern char *prom_alloc(char *virt_hint, unsigned int size);
-
-/* Free a previously allocated chunk. */
-extern void prom_free(char *virt_addr, unsigned int size);
-
-/* Sun4/sun4c specific memory-management startup hook. */
-
-/* Map the passed segment in the given context at the passed
- * virtual address.
- */
-extern void prom_putsegment(int context, unsigned long virt_addr,
-			    int physical_segment);
-
-
-/* PROM device tree traversal functions... */
-
-#ifdef PROMLIB_INTERNAL
-
-/* Internal version of prom_getchild. */
-extern int __prom_getchild(int parent_node);
-
-/* Internal version of prom_getsibling. */
-extern int __prom_getsibling(int node);
-
+#ifndef ___ASM_SPARC_OPLIB_H
+#define ___ASM_SPARC_OPLIB_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/oplib_64.h>
+#else
+#include <asm-sparc/oplib_32.h>
+#endif
 #endif
-
-
-/* Get the child node of the given node, or zero if no child exists. */
-extern int prom_getchild(int parent_node);
-
-/* Get the next sibling node of the given node, or zero if no further
- * siblings exist.
- */
-extern int prom_getsibling(int node);
-
-/* Get the length, at the passed node, of the given property type.
- * Returns -1 on error (ie. no such property at this node).
- */
-extern int prom_getproplen(int thisnode, char *property);
-
-/* Fetch the requested property using the given buffer.  Returns
- * the number of bytes the prom put into your buffer or -1 on error.
- */
-extern int __must_check prom_getproperty(int thisnode, char *property,
-					 char *prop_buffer, int propbuf_size);
-
-/* Acquire an integer property. */
-extern int prom_getint(int node, char *property);
-
-/* Acquire an integer property, with a default value. */
-extern int prom_getintdefault(int node, char *property, int defval);
-
-/* Acquire a boolean property, 0=FALSE 1=TRUE. */
-extern int prom_getbool(int node, char *prop);
-
-/* Acquire a string property, null string on error. */
-extern void prom_getstring(int node, char *prop, char *buf, int bufsize);
-
-/* Does the passed node have the given "name"? YES=1 NO=0 */
-extern int prom_nodematch(int thisnode, char *name);
-
-/* Search all siblings starting at the passed node for "name" matching
- * the given string.  Returns the node on success, zero on failure.
- */
-extern int prom_searchsiblings(int node_start, char *name);
-
-/* Return the first property type, as a string, for the given node.
- * Returns a null string on error.
- */
-extern char *prom_firstprop(int node, char *buffer);
-
-/* Returns the next property after the passed property for the given
- * node.  Returns null string on failure.
- */
-extern char *prom_nextprop(int node, char *prev_property, char *buffer);
-
-/* Returns phandle of the path specified */
-extern int prom_finddevice(char *name);
-
-/* Returns 1 if the specified node has given property. */
-extern int prom_node_has_property(int node, char *property);
-
-/* Set the indicated property at the given node with the passed value.
- * Returns the number of bytes of your value that the prom took.
- */
-extern int prom_setprop(int node, char *prop_name, char *prop_value,
-			int value_size);
-			
-extern int prom_pathtoinode(char *path);
-extern int prom_inst2pkg(int);
-
-/* Dorking with Bus ranges... */
-
-/* Apply promlib probes OBIO ranges to registers. */
-extern void prom_apply_obio_ranges(struct linux_prom_registers *obioregs, int nregs);
-
-/* Apply ranges of any prom node (and optionally parent node as well) to registers. */
-extern void prom_apply_generic_ranges(int node, int parent, 
-				      struct linux_prom_registers *sbusregs, int nregs);
-
-/* CPU probing helpers.  */
-int cpu_find_by_instance(int instance, int *prom_node, int *mid);
-int cpu_find_by_mid(int mid, int *prom_node);
-int cpu_get_hwmid(int prom_node);
-
-extern spinlock_t prom_lock;
-
-#endif /* !(__SPARC_OPLIB_H) */
diff --git a/include/asm-sparc/oplib_32.h b/include/asm-sparc/oplib_32.h
new file mode 100644
index 00000000000..b2631da259e
--- /dev/null
+++ b/include/asm-sparc/oplib_32.h
@@ -0,0 +1,272 @@
+/*
+ * oplib.h:  Describes the interface and available routines in the
+ *           Linux Prom library.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __SPARC_OPLIB_H
+#define __SPARC_OPLIB_H
+
+#include <asm/openprom.h>
+#include <linux/spinlock.h>
+#include <linux/compiler.h>
+
+/* The master romvec pointer... */
+extern struct linux_romvec *romvec;
+
+/* Enumeration to describe the prom major version we have detected. */
+enum prom_major_version {
+	PROM_V0,      /* Original sun4c V0 prom */
+	PROM_V2,      /* sun4c and early sun4m V2 prom */
+	PROM_V3,      /* sun4m and later, up to sun4d/sun4e machines V3 */
+	PROM_P1275,   /* IEEE compliant ISA based Sun PROM, only sun4u */
+	PROM_SUN4,    /* Old sun4 proms are totally different, but we'll shoehorn it to make it fit */
+};
+
+extern enum prom_major_version prom_vers;
+/* Revision, and firmware revision. */
+extern unsigned int prom_rev, prom_prev;
+
+/* Root node of the prom device tree, this stays constant after
+ * initialization is complete.
+ */
+extern int prom_root_node;
+
+/* Pointer to prom structure containing the device tree traversal
+ * and usage utility functions.  Only prom-lib should use these,
+ * users use the interface defined by the library only!
+ */
+extern struct linux_nodeops *prom_nodeops;
+
+/* The functions... */
+
+/* You must call prom_init() before using any of the library services,
+ * preferably as early as possible.  Pass it the romvec pointer.
+ */
+extern void prom_init(struct linux_romvec *rom_ptr);
+
+/* Boot argument acquisition, returns the boot command line string. */
+extern char *prom_getbootargs(void);
+
+/* Device utilities. */
+
+/* Map and unmap devices in IO space at virtual addresses. Note that the
+ * virtual address you pass is a request and the prom may put your mappings
+ * somewhere else, so check your return value as that is where your new
+ * mappings really are!
+ *
+ * Another note, these are only available on V2 or higher proms!
+ */
+extern char *prom_mapio(char *virt_hint, int io_space, unsigned int phys_addr, unsigned int num_bytes);
+extern void prom_unmapio(char *virt_addr, unsigned int num_bytes);
+
+/* Device operations. */
+
+/* Open the device described by the passed string.  Note, that the format
+ * of the string is different on V0 vs. V2->higher proms.  The caller must
+ * know what he/she is doing!  Returns the device descriptor, an int.
+ */
+extern int prom_devopen(char *device_string);
+
+/* Close a previously opened device described by the passed integer
+ * descriptor.
+ */
+extern int prom_devclose(int device_handle);
+
+/* Do a seek operation on the device described by the passed integer
+ * descriptor.
+ */
+extern void prom_seek(int device_handle, unsigned int seek_hival,
+		      unsigned int seek_lowval);
+
+/* Miscellaneous routines, don't really fit in any category per se. */
+
+/* Reboot the machine with the command line passed. */
+extern void prom_reboot(char *boot_command);
+
+/* Evaluate the forth string passed. */
+extern void prom_feval(char *forth_string);
+
+/* Enter the prom, with possibility of continuation with the 'go'
+ * command in newer proms.
+ */
+extern void prom_cmdline(void);
+
+/* Enter the prom, with no chance of continuation for the stand-alone
+ * which calls this.
+ */
+extern void prom_halt(void) __attribute__ ((noreturn));
+
+/* Set the PROM 'sync' callback function to the passed function pointer.
+ * When the user gives the 'sync' command at the prom prompt while the
+ * kernel is still active, the prom will call this routine.
+ *
+ * XXX The arguments are different on V0 vs. V2->higher proms, grrr! XXX
+ */
+typedef void (*sync_func_t)(void);
+extern void prom_setsync(sync_func_t func_ptr);
+
+/* Acquire the IDPROM of the root node in the prom device tree.  This
+ * gets passed a buffer where you would like it stuffed.  The return value
+ * is the format type of this idprom or 0xff on error.
+ */
+extern unsigned char prom_get_idprom(char *idp_buffer, int idpbuf_size);
+
+/* Get the prom major version. */
+extern int prom_version(void);
+
+/* Get the prom plugin revision. */
+extern int prom_getrev(void);
+
+/* Get the prom firmware revision. */
+extern int prom_getprev(void);
+
+/* Character operations to/from the console.... */
+
+/* Non-blocking get character from console. */
+extern int prom_nbgetchar(void);
+
+/* Non-blocking put character to console. */
+extern int prom_nbputchar(char character);
+
+/* Blocking get character from console. */
+extern char prom_getchar(void);
+
+/* Blocking put character to console. */
+extern void prom_putchar(char character);
+
+/* Prom's internal routines, don't use in kernel/boot code. */
+extern void prom_printf(char *fmt, ...);
+extern void prom_write(const char *buf, unsigned int len);
+
+/* Multiprocessor operations... */
+
+/* Start the CPU with the given device tree node, context table, and context
+ * at the passed program counter.
+ */
+extern int prom_startcpu(int cpunode, struct linux_prom_registers *context_table,
+			 int context, char *program_counter);
+
+/* Stop the CPU with the passed device tree node. */
+extern int prom_stopcpu(int cpunode);
+
+/* Idle the CPU with the passed device tree node. */
+extern int prom_idlecpu(int cpunode);
+
+/* Re-Start the CPU with the passed device tree node. */
+extern int prom_restartcpu(int cpunode);
+
+/* PROM memory allocation facilities... */
+
+/* Allocated at possibly the given virtual address a chunk of the
+ * indicated size.
+ */
+extern char *prom_alloc(char *virt_hint, unsigned int size);
+
+/* Free a previously allocated chunk. */
+extern void prom_free(char *virt_addr, unsigned int size);
+
+/* Sun4/sun4c specific memory-management startup hook. */
+
+/* Map the passed segment in the given context at the passed
+ * virtual address.
+ */
+extern void prom_putsegment(int context, unsigned long virt_addr,
+			    int physical_segment);
+
+
+/* PROM device tree traversal functions... */
+
+#ifdef PROMLIB_INTERNAL
+
+/* Internal version of prom_getchild. */
+extern int __prom_getchild(int parent_node);
+
+/* Internal version of prom_getsibling. */
+extern int __prom_getsibling(int node);
+
+#endif
+
+
+/* Get the child node of the given node, or zero if no child exists. */
+extern int prom_getchild(int parent_node);
+
+/* Get the next sibling node of the given node, or zero if no further
+ * siblings exist.
+ */
+extern int prom_getsibling(int node);
+
+/* Get the length, at the passed node, of the given property type.
+ * Returns -1 on error (ie. no such property at this node).
+ */
+extern int prom_getproplen(int thisnode, char *property);
+
+/* Fetch the requested property using the given buffer.  Returns
+ * the number of bytes the prom put into your buffer or -1 on error.
+ */
+extern int __must_check prom_getproperty(int thisnode, char *property,
+					 char *prop_buffer, int propbuf_size);
+
+/* Acquire an integer property. */
+extern int prom_getint(int node, char *property);
+
+/* Acquire an integer property, with a default value. */
+extern int prom_getintdefault(int node, char *property, int defval);
+
+/* Acquire a boolean property, 0=FALSE 1=TRUE. */
+extern int prom_getbool(int node, char *prop);
+
+/* Acquire a string property, null string on error. */
+extern void prom_getstring(int node, char *prop, char *buf, int bufsize);
+
+/* Does the passed node have the given "name"? YES=1 NO=0 */
+extern int prom_nodematch(int thisnode, char *name);
+
+/* Search all siblings starting at the passed node for "name" matching
+ * the given string.  Returns the node on success, zero on failure.
+ */
+extern int prom_searchsiblings(int node_start, char *name);
+
+/* Return the first property type, as a string, for the given node.
+ * Returns a null string on error.
+ */
+extern char *prom_firstprop(int node, char *buffer);
+
+/* Returns the next property after the passed property for the given
+ * node.  Returns null string on failure.
+ */
+extern char *prom_nextprop(int node, char *prev_property, char *buffer);
+
+/* Returns phandle of the path specified */
+extern int prom_finddevice(char *name);
+
+/* Returns 1 if the specified node has given property. */
+extern int prom_node_has_property(int node, char *property);
+
+/* Set the indicated property at the given node with the passed value.
+ * Returns the number of bytes of your value that the prom took.
+ */
+extern int prom_setprop(int node, char *prop_name, char *prop_value,
+			int value_size);
+
+extern int prom_pathtoinode(char *path);
+extern int prom_inst2pkg(int);
+
+/* Dorking with Bus ranges... */
+
+/* Apply promlib probes OBIO ranges to registers. */
+extern void prom_apply_obio_ranges(struct linux_prom_registers *obioregs, int nregs);
+
+/* Apply ranges of any prom node (and optionally parent node as well) to registers. */
+extern void prom_apply_generic_ranges(int node, int parent,
+				      struct linux_prom_registers *sbusregs, int nregs);
+
+/* CPU probing helpers.  */
+int cpu_find_by_instance(int instance, int *prom_node, int *mid);
+int cpu_find_by_mid(int mid, int *prom_node);
+int cpu_get_hwmid(int prom_node);
+
+extern spinlock_t prom_lock;
+
+#endif /* !(__SPARC_OPLIB_H) */
diff --git a/include/asm-sparc/oplib_64.h b/include/asm-sparc/oplib_64.h
new file mode 100644
index 00000000000..6d2c2ca9803
--- /dev/null
+++ b/include/asm-sparc/oplib_64.h
@@ -0,0 +1,322 @@
+/* oplib.h:  Describes the interface and available routines in the
+ *           Linux Prom library.
+ *
+ * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1996 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+
+#ifndef __SPARC64_OPLIB_H
+#define __SPARC64_OPLIB_H
+
+#include <asm/openprom.h>
+
+/* OBP version string. */
+extern char prom_version[];
+
+/* Root node of the prom device tree, this stays constant after
+ * initialization is complete.
+ */
+extern int prom_root_node;
+
+/* PROM stdin and stdout */
+extern int prom_stdin, prom_stdout;
+
+/* /chosen node of the prom device tree, this stays constant after
+ * initialization is complete.
+ */
+extern int prom_chosen_node;
+
+/* Helper values and strings in arch/sparc64/kernel/head.S */
+extern const char prom_peer_name[];
+extern const char prom_compatible_name[];
+extern const char prom_root_compatible[];
+extern const char prom_cpu_compatible[];
+extern const char prom_finddev_name[];
+extern const char prom_chosen_path[];
+extern const char prom_cpu_path[];
+extern const char prom_getprop_name[];
+extern const char prom_mmu_name[];
+extern const char prom_callmethod_name[];
+extern const char prom_translate_name[];
+extern const char prom_map_name[];
+extern const char prom_unmap_name[];
+extern int prom_mmu_ihandle_cache;
+extern unsigned int prom_boot_mapped_pc;
+extern unsigned int prom_boot_mapping_mode;
+extern unsigned long prom_boot_mapping_phys_high, prom_boot_mapping_phys_low;
+
+struct linux_mlist_p1275 {
+	struct linux_mlist_p1275 *theres_more;
+	unsigned long start_adr;
+	unsigned long num_bytes;
+};
+
+struct linux_mem_p1275 {
+	struct linux_mlist_p1275 **p1275_totphys;
+	struct linux_mlist_p1275 **p1275_prommap;
+	struct linux_mlist_p1275 **p1275_available; /* What we can use */
+};
+
+/* The functions... */
+
+/* You must call prom_init() before using any of the library services,
+ * preferably as early as possible.  Pass it the romvec pointer.
+ */
+extern void prom_init(void *cif_handler, void *cif_stack);
+
+/* Boot argument acquisition, returns the boot command line string. */
+extern char *prom_getbootargs(void);
+
+/* Device utilities. */
+
+/* Device operations. */
+
+/* Open the device described by the passed string.  Note, that the format
+ * of the string is different on V0 vs. V2->higher proms.  The caller must
+ * know what he/she is doing!  Returns the device descriptor, an int.
+ */
+extern int prom_devopen(const char *device_string);
+
+/* Close a previously opened device described by the passed integer
+ * descriptor.
+ */
+extern int prom_devclose(int device_handle);
+
+/* Do a seek operation on the device described by the passed integer
+ * descriptor.
+ */
+extern void prom_seek(int device_handle, unsigned int seek_hival,
+		      unsigned int seek_lowval);
+
+/* Miscellaneous routines, don't really fit in any category per se. */
+
+/* Reboot the machine with the command line passed. */
+extern void prom_reboot(const char *boot_command);
+
+/* Evaluate the forth string passed. */
+extern void prom_feval(const char *forth_string);
+
+/* Enter the prom, with possibility of continuation with the 'go'
+ * command in newer proms.
+ */
+extern void prom_cmdline(void);
+
+/* Enter the prom, with no chance of continuation for the stand-alone
+ * which calls this.
+ */
+extern void prom_halt(void) __attribute__ ((noreturn));
+
+/* Halt and power-off the machine. */
+extern void prom_halt_power_off(void) __attribute__ ((noreturn));
+
+/* Set the PROM 'sync' callback function to the passed function pointer.
+ * When the user gives the 'sync' command at the prom prompt while the
+ * kernel is still active, the prom will call this routine.
+ *
+ */
+typedef int (*callback_func_t)(long *cmd);
+extern void prom_setcallback(callback_func_t func_ptr);
+
+/* Acquire the IDPROM of the root node in the prom device tree.  This
+ * gets passed a buffer where you would like it stuffed.  The return value
+ * is the format type of this idprom or 0xff on error.
+ */
+extern unsigned char prom_get_idprom(char *idp_buffer, int idpbuf_size);
+
+/* Character operations to/from the console.... */
+
+/* Non-blocking get character from console. */
+extern int prom_nbgetchar(void);
+
+/* Non-blocking put character to console. */
+extern int prom_nbputchar(char character);
+
+/* Blocking get character from console. */
+extern char prom_getchar(void);
+
+/* Blocking put character to console. */
+extern void prom_putchar(char character);
+
+/* Prom's internal routines, don't use in kernel/boot code. */
+extern void prom_printf(const char *fmt, ...);
+extern void prom_write(const char *buf, unsigned int len);
+
+/* Multiprocessor operations... */
+#ifdef CONFIG_SMP
+/* Start the CPU with the given device tree node at the passed program
+ * counter with the given arg passed in via register %o0.
+ */
+extern void prom_startcpu(int cpunode, unsigned long pc, unsigned long arg);
+
+/* Start the CPU with the given cpu ID at the passed program
+ * counter with the given arg passed in via register %o0.
+ */
+extern void prom_startcpu_cpuid(int cpuid, unsigned long pc, unsigned long arg);
+
+/* Stop the CPU with the given cpu ID.  */
+extern void prom_stopcpu_cpuid(int cpuid);
+
+/* Stop the current CPU. */
+extern void prom_stopself(void);
+
+/* Idle the current CPU. */
+extern void prom_idleself(void);
+
+/* Resume the CPU with the passed device tree node. */
+extern void prom_resumecpu(int cpunode);
+#endif
+
+/* Power management interfaces. */
+
+/* Put the current CPU to sleep. */
+extern void prom_sleepself(void);
+
+/* Put the entire system to sleep. */
+extern int prom_sleepsystem(void);
+
+/* Initiate a wakeup event. */
+extern int prom_wakeupsystem(void);
+
+/* MMU and memory related OBP interfaces. */
+
+/* Get unique string identifying SIMM at given physical address. */
+extern int prom_getunumber(int syndrome_code,
+			   unsigned long phys_addr,
+			   char *buf, int buflen);
+
+/* Retain physical memory to the caller across soft resets. */
+extern unsigned long prom_retain(const char *name,
+				 unsigned long pa_low, unsigned long pa_high,
+				 long size, long align);
+
+/* Load explicit I/D TLB entries into the calling processor. */
+extern long prom_itlb_load(unsigned long index,
+			   unsigned long tte_data,
+			   unsigned long vaddr);
+
+extern long prom_dtlb_load(unsigned long index,
+			   unsigned long tte_data,
+			   unsigned long vaddr);
+
+/* Map/Unmap client program address ranges.  First the format of
+ * the mapping mode argument.
+ */
+#define PROM_MAP_WRITE	0x0001 /* Writable */
+#define PROM_MAP_READ	0x0002 /* Readable - sw */
+#define PROM_MAP_EXEC	0x0004 /* Executable - sw */
+#define PROM_MAP_LOCKED	0x0010 /* Locked, use i/dtlb load calls for this instead */
+#define PROM_MAP_CACHED	0x0020 /* Cacheable in both L1 and L2 caches */
+#define PROM_MAP_SE	0x0040 /* Side-Effects */
+#define PROM_MAP_GLOB	0x0080 /* Global */
+#define PROM_MAP_IE	0x0100 /* Invert-Endianness */
+#define PROM_MAP_DEFAULT (PROM_MAP_WRITE | PROM_MAP_READ | PROM_MAP_EXEC | PROM_MAP_CACHED)
+
+extern int prom_map(int mode, unsigned long size,
+		    unsigned long vaddr, unsigned long paddr);
+extern void prom_unmap(unsigned long size, unsigned long vaddr);
+
+
+/* PROM device tree traversal functions... */
+
+#ifdef PROMLIB_INTERNAL
+
+/* Internal version of prom_getchild. */
+extern int __prom_getchild(int parent_node);
+
+/* Internal version of prom_getsibling. */
+extern int __prom_getsibling(int node);
+
+#endif
+
+/* Get the child node of the given node, or zero if no child exists. */
+extern int prom_getchild(int parent_node);
+
+/* Get the next sibling node of the given node, or zero if no further
+ * siblings exist.
+ */
+extern int prom_getsibling(int node);
+
+/* Get the length, at the passed node, of the given property type.
+ * Returns -1 on error (ie. no such property at this node).
+ */
+extern int prom_getproplen(int thisnode, const char *property);
+
+/* Fetch the requested property using the given buffer.  Returns
+ * the number of bytes the prom put into your buffer or -1 on error.
+ */
+extern int prom_getproperty(int thisnode, const char *property,
+			    char *prop_buffer, int propbuf_size);
+
+/* Acquire an integer property. */
+extern int prom_getint(int node, const char *property);
+
+/* Acquire an integer property, with a default value. */
+extern int prom_getintdefault(int node, const char *property, int defval);
+
+/* Acquire a boolean property, 0=FALSE 1=TRUE. */
+extern int prom_getbool(int node, const char *prop);
+
+/* Acquire a string property, null string on error. */
+extern void prom_getstring(int node, const char *prop, char *buf, int bufsize);
+
+/* Does the passed node have the given "name"? YES=1 NO=0 */
+extern int prom_nodematch(int thisnode, const char *name);
+
+/* Search all siblings starting at the passed node for "name" matching
+ * the given string.  Returns the node on success, zero on failure.
+ */
+extern int prom_searchsiblings(int node_start, const char *name);
+
+/* Return the first property type, as a string, for the given node.
+ * Returns a null string on error. Buffer should be at least 32B long.
+ */
+extern char *prom_firstprop(int node, char *buffer);
+
+/* Returns the next property after the passed property for the given
+ * node.  Returns null string on failure. Buffer should be at least 32B long.
+ */
+extern char *prom_nextprop(int node, const char *prev_property, char *buffer);
+
+/* Returns 1 if the specified node has given property. */
+extern int prom_node_has_property(int node, const char *property);
+
+/* Returns phandle of the path specified */
+extern int prom_finddevice(const char *name);
+
+/* Set the indicated property at the given node with the passed value.
+ * Returns the number of bytes of your value that the prom took.
+ */
+extern int prom_setprop(int node, const char *prop_name, char *prop_value,
+			int value_size);
+
+extern int prom_pathtoinode(const char *path);
+extern int prom_inst2pkg(int);
+extern int prom_service_exists(const char *service_name);
+extern void prom_sun4v_guest_soft_state(void);
+
+extern int prom_ihandle2path(int handle, char *buffer, int bufsize);
+
+/* Client interface level routines. */
+extern long p1275_cmd(const char *, long, ...);
+
+#if 0
+#define P1275_SIZE(x) ((((long)((x) / 32)) << 32) | (x))
+#else
+#define P1275_SIZE(x) x
+#endif
+
+/* We support at most 16 input and 1 output argument */
+#define P1275_ARG_NUMBER		0
+#define P1275_ARG_IN_STRING		1
+#define P1275_ARG_OUT_BUF		2
+#define P1275_ARG_OUT_32B		3
+#define P1275_ARG_IN_FUNCTION		4
+#define P1275_ARG_IN_BUF		5
+#define P1275_ARG_IN_64B		6
+
+#define P1275_IN(x) ((x) & 0xf)
+#define P1275_OUT(x) (((x) << 4) & 0xf0)
+#define P1275_INOUT(i,o) (P1275_IN(i)|P1275_OUT(o))
+#define P1275_ARG(n,x) ((x) << ((n)*3 + 8))
+
+#endif /* !(__SPARC64_OPLIB_H) */
diff --git a/include/asm-sparc/page.h b/include/asm-sparc/page.h
index 6aa9e4c910c..f32f49fcf75 100644
--- a/include/asm-sparc/page.h
+++ b/include/asm-sparc/page.h
@@ -1,165 +1,8 @@
-/*
- * page.h:  Various defines and such for MMU operations on the Sparc for
- *          the Linux kernel.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC_PAGE_H
-#define _SPARC_PAGE_H
-
-#ifdef CONFIG_SUN4
-#define PAGE_SHIFT   13
+#ifndef ___ASM_SPARC_PAGE_H
+#define ___ASM_SPARC_PAGE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/page_64.h>
 #else
-#define PAGE_SHIFT   12
+#include <asm-sparc/page_32.h>
 #endif
-#ifndef __ASSEMBLY__
-/* I have my suspicions... -DaveM */
-#define PAGE_SIZE    (1UL << PAGE_SHIFT)
-#else
-#define PAGE_SIZE    (1 << PAGE_SHIFT)
-#endif
-#define PAGE_MASK    (~(PAGE_SIZE-1))
-
-#include <asm/btfixup.h>
-
-#ifndef __ASSEMBLY__
-
-#define clear_page(page)	 memset((void *)(page), 0, PAGE_SIZE)
-#define copy_page(to,from) 	memcpy((void *)(to), (void *)(from), PAGE_SIZE)
-#define clear_user_page(addr, vaddr, page)	\
-	do { 	clear_page(addr);		\
-		sparc_flush_page_to_ram(page);	\
-	} while (0)
-#define copy_user_page(to, from, vaddr, page)	\
-	do {	copy_page(to, from);		\
-		sparc_flush_page_to_ram(page);	\
-	} while (0)
-
-/* The following structure is used to hold the physical
- * memory configuration of the machine.  This is filled in
- * prom_meminit() and is later used by mem_init() to set up
- * mem_map[].  We statically allocate SPARC_PHYS_BANKS+1 of
- * these structs, this is arbitrary.  The entry after the
- * last valid one has num_bytes==0.
- */
-struct sparc_phys_banks {
-  unsigned long base_addr;
-  unsigned long num_bytes;
-};
-
-#define SPARC_PHYS_BANKS 32
-
-extern struct sparc_phys_banks sp_banks[SPARC_PHYS_BANKS+1];
-
-/* Cache alias structure.  Entry is valid if context != -1. */
-struct cache_palias {
-	unsigned long vaddr;
-	int context;
-};
-
-extern struct cache_palias *sparc_aliases;
-
-/* passing structs on the Sparc slow us down tremendously... */
-
-/* #define STRICT_MM_TYPECHECKS */
-
-#ifdef STRICT_MM_TYPECHECKS
-/*
- * These are used to make use of C type-checking..
- */
-typedef struct { unsigned long pte; } pte_t;
-typedef struct { unsigned long iopte; } iopte_t;
-typedef struct { unsigned long pmdv[16]; } pmd_t;
-typedef struct { unsigned long pgd; } pgd_t;
-typedef struct { unsigned long ctxd; } ctxd_t;
-typedef struct { unsigned long pgprot; } pgprot_t;
-typedef struct { unsigned long iopgprot; } iopgprot_t;
-
-#define pte_val(x)	((x).pte)
-#define iopte_val(x)	((x).iopte)
-#define pmd_val(x)      ((x).pmdv[0])
-#define pgd_val(x)	((x).pgd)
-#define ctxd_val(x)	((x).ctxd)
-#define pgprot_val(x)	((x).pgprot)
-#define iopgprot_val(x)	((x).iopgprot)
-
-#define __pte(x)	((pte_t) { (x) } )
-#define __iopte(x)	((iopte_t) { (x) } )
-/* #define __pmd(x)        ((pmd_t) { (x) } ) */ /* XXX procedure with loop */
-#define __pgd(x)	((pgd_t) { (x) } )
-#define __ctxd(x)	((ctxd_t) { (x) } )
-#define __pgprot(x)	((pgprot_t) { (x) } )
-#define __iopgprot(x)	((iopgprot_t) { (x) } )
-
-#else
-/*
- * .. while these make it easier on the compiler
- */
-typedef unsigned long pte_t;
-typedef unsigned long iopte_t;
-typedef struct { unsigned long pmdv[16]; } pmd_t;
-typedef unsigned long pgd_t;
-typedef unsigned long ctxd_t;
-typedef unsigned long pgprot_t;
-typedef unsigned long iopgprot_t;
-
-#define pte_val(x)	(x)
-#define iopte_val(x)	(x)
-#define pmd_val(x)      ((x).pmdv[0])
-#define pgd_val(x)	(x)
-#define ctxd_val(x)	(x)
-#define pgprot_val(x)	(x)
-#define iopgprot_val(x)	(x)
-
-#define __pte(x)	(x)
-#define __iopte(x)	(x)
-/* #define __pmd(x)        (x) */ /* XXX later */
-#define __pgd(x)	(x)
-#define __ctxd(x)	(x)
-#define __pgprot(x)	(x)
-#define __iopgprot(x)	(x)
-
-#endif
-
-typedef struct page *pgtable_t;
-
-extern unsigned long sparc_unmapped_base;
-
-BTFIXUPDEF_SETHI(sparc_unmapped_base)
-
-#define TASK_UNMAPPED_BASE	BTFIXUP_SETHI(sparc_unmapped_base)
-
-#else /* !(__ASSEMBLY__) */
-
-#define __pgprot(x)	(x)
-
-#endif /* !(__ASSEMBLY__) */
-
-/* to align the pointer to the (next) page boundary */
-#define PAGE_ALIGN(addr)  (((addr)+PAGE_SIZE-1)&PAGE_MASK)
-
-#define PAGE_OFFSET	0xf0000000
-#ifndef __ASSEMBLY__
-extern unsigned long phys_base;
-extern unsigned long pfn_base;
 #endif
-#define __pa(x)			((unsigned long)(x) - PAGE_OFFSET + phys_base)
-#define __va(x)			((void *)((unsigned long) (x) - phys_base + PAGE_OFFSET))
-
-#define virt_to_phys		__pa
-#define phys_to_virt		__va
-
-#define ARCH_PFN_OFFSET		(pfn_base)
-#define virt_to_page(kaddr)	(mem_map + ((((unsigned long)(kaddr)-PAGE_OFFSET)>>PAGE_SHIFT)))
-
-#define pfn_valid(pfn)		(((pfn) >= (pfn_base)) && (((pfn)-(pfn_base)) < max_mapnr))
-#define virt_addr_valid(kaddr)	((((unsigned long)(kaddr)-PAGE_OFFSET)>>PAGE_SHIFT) < max_mapnr)
-
-#define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \
-				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
-
-#include <asm-generic/memory_model.h>
-#include <asm-generic/page.h>
-
-#endif /* _SPARC_PAGE_H */
diff --git a/include/asm-sparc/page_32.h b/include/asm-sparc/page_32.h
new file mode 100644
index 00000000000..14de518cc38
--- /dev/null
+++ b/include/asm-sparc/page_32.h
@@ -0,0 +1,163 @@
+/*
+ * page.h:  Various defines and such for MMU operations on the Sparc for
+ *          the Linux kernel.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _SPARC_PAGE_H
+#define _SPARC_PAGE_H
+
+#ifdef CONFIG_SUN4
+#define PAGE_SHIFT   13
+#else
+#define PAGE_SHIFT   12
+#endif
+#ifndef __ASSEMBLY__
+/* I have my suspicions... -DaveM */
+#define PAGE_SIZE    (1UL << PAGE_SHIFT)
+#else
+#define PAGE_SIZE    (1 << PAGE_SHIFT)
+#endif
+#define PAGE_MASK    (~(PAGE_SIZE-1))
+
+#include <asm/btfixup.h>
+
+#ifndef __ASSEMBLY__
+
+#define clear_page(page)	 memset((void *)(page), 0, PAGE_SIZE)
+#define copy_page(to,from) 	memcpy((void *)(to), (void *)(from), PAGE_SIZE)
+#define clear_user_page(addr, vaddr, page)	\
+	do { 	clear_page(addr);		\
+		sparc_flush_page_to_ram(page);	\
+	} while (0)
+#define copy_user_page(to, from, vaddr, page)	\
+	do {	copy_page(to, from);		\
+		sparc_flush_page_to_ram(page);	\
+	} while (0)
+
+/* The following structure is used to hold the physical
+ * memory configuration of the machine.  This is filled in
+ * prom_meminit() and is later used by mem_init() to set up
+ * mem_map[].  We statically allocate SPARC_PHYS_BANKS+1 of
+ * these structs, this is arbitrary.  The entry after the
+ * last valid one has num_bytes==0.
+ */
+struct sparc_phys_banks {
+  unsigned long base_addr;
+  unsigned long num_bytes;
+};
+
+#define SPARC_PHYS_BANKS 32
+
+extern struct sparc_phys_banks sp_banks[SPARC_PHYS_BANKS+1];
+
+/* Cache alias structure.  Entry is valid if context != -1. */
+struct cache_palias {
+	unsigned long vaddr;
+	int context;
+};
+
+/* passing structs on the Sparc slow us down tremendously... */
+
+/* #define STRICT_MM_TYPECHECKS */
+
+#ifdef STRICT_MM_TYPECHECKS
+/*
+ * These are used to make use of C type-checking..
+ */
+typedef struct { unsigned long pte; } pte_t;
+typedef struct { unsigned long iopte; } iopte_t;
+typedef struct { unsigned long pmdv[16]; } pmd_t;
+typedef struct { unsigned long pgd; } pgd_t;
+typedef struct { unsigned long ctxd; } ctxd_t;
+typedef struct { unsigned long pgprot; } pgprot_t;
+typedef struct { unsigned long iopgprot; } iopgprot_t;
+
+#define pte_val(x)	((x).pte)
+#define iopte_val(x)	((x).iopte)
+#define pmd_val(x)      ((x).pmdv[0])
+#define pgd_val(x)	((x).pgd)
+#define ctxd_val(x)	((x).ctxd)
+#define pgprot_val(x)	((x).pgprot)
+#define iopgprot_val(x)	((x).iopgprot)
+
+#define __pte(x)	((pte_t) { (x) } )
+#define __iopte(x)	((iopte_t) { (x) } )
+/* #define __pmd(x)        ((pmd_t) { (x) } ) */ /* XXX procedure with loop */
+#define __pgd(x)	((pgd_t) { (x) } )
+#define __ctxd(x)	((ctxd_t) { (x) } )
+#define __pgprot(x)	((pgprot_t) { (x) } )
+#define __iopgprot(x)	((iopgprot_t) { (x) } )
+
+#else
+/*
+ * .. while these make it easier on the compiler
+ */
+typedef unsigned long pte_t;
+typedef unsigned long iopte_t;
+typedef struct { unsigned long pmdv[16]; } pmd_t;
+typedef unsigned long pgd_t;
+typedef unsigned long ctxd_t;
+typedef unsigned long pgprot_t;
+typedef unsigned long iopgprot_t;
+
+#define pte_val(x)	(x)
+#define iopte_val(x)	(x)
+#define pmd_val(x)      ((x).pmdv[0])
+#define pgd_val(x)	(x)
+#define ctxd_val(x)	(x)
+#define pgprot_val(x)	(x)
+#define iopgprot_val(x)	(x)
+
+#define __pte(x)	(x)
+#define __iopte(x)	(x)
+/* #define __pmd(x)        (x) */ /* XXX later */
+#define __pgd(x)	(x)
+#define __ctxd(x)	(x)
+#define __pgprot(x)	(x)
+#define __iopgprot(x)	(x)
+
+#endif
+
+typedef struct page *pgtable_t;
+
+extern unsigned long sparc_unmapped_base;
+
+BTFIXUPDEF_SETHI(sparc_unmapped_base)
+
+#define TASK_UNMAPPED_BASE	BTFIXUP_SETHI(sparc_unmapped_base)
+
+#else /* !(__ASSEMBLY__) */
+
+#define __pgprot(x)	(x)
+
+#endif /* !(__ASSEMBLY__) */
+
+/* to align the pointer to the (next) page boundary */
+#define PAGE_ALIGN(addr)  (((addr)+PAGE_SIZE-1)&PAGE_MASK)
+
+#define PAGE_OFFSET	0xf0000000
+#ifndef __ASSEMBLY__
+extern unsigned long phys_base;
+extern unsigned long pfn_base;
+#endif
+#define __pa(x)			((unsigned long)(x) - PAGE_OFFSET + phys_base)
+#define __va(x)			((void *)((unsigned long) (x) - phys_base + PAGE_OFFSET))
+
+#define virt_to_phys		__pa
+#define phys_to_virt		__va
+
+#define ARCH_PFN_OFFSET		(pfn_base)
+#define virt_to_page(kaddr)	(mem_map + ((((unsigned long)(kaddr)-PAGE_OFFSET)>>PAGE_SHIFT)))
+
+#define pfn_valid(pfn)		(((pfn) >= (pfn_base)) && (((pfn)-(pfn_base)) < max_mapnr))
+#define virt_addr_valid(kaddr)	((((unsigned long)(kaddr)-PAGE_OFFSET)>>PAGE_SHIFT) < max_mapnr)
+
+#define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \
+				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+#include <asm-generic/memory_model.h>
+#include <asm-generic/page.h>
+
+#endif /* _SPARC_PAGE_H */
diff --git a/include/asm-sparc/page_64.h b/include/asm-sparc/page_64.h
new file mode 100644
index 00000000000..a8a2bba032c
--- /dev/null
+++ b/include/asm-sparc/page_64.h
@@ -0,0 +1,138 @@
+#ifndef _SPARC64_PAGE_H
+#define _SPARC64_PAGE_H
+
+#include <linux/const.h>
+
+#if defined(CONFIG_SPARC64_PAGE_SIZE_8KB)
+#define PAGE_SHIFT   13
+#elif defined(CONFIG_SPARC64_PAGE_SIZE_64KB)
+#define PAGE_SHIFT   16
+#else
+#error No page size specified in kernel configuration
+#endif
+
+#define PAGE_SIZE    (_AC(1,UL) << PAGE_SHIFT)
+#define PAGE_MASK    (~(PAGE_SIZE-1))
+
+/* Flushing for D-cache alias handling is only needed if
+ * the page size is smaller than 16K.
+ */
+#if PAGE_SHIFT < 14
+#define DCACHE_ALIASING_POSSIBLE
+#endif
+
+#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
+#define HPAGE_SHIFT		22
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
+#define HPAGE_SHIFT		19
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
+#define HPAGE_SHIFT		16
+#endif
+
+#ifdef CONFIG_HUGETLB_PAGE
+#define HPAGE_SIZE		(_AC(1,UL) << HPAGE_SHIFT)
+#define HPAGE_MASK		(~(HPAGE_SIZE - 1UL))
+#define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
+#define HAVE_ARCH_HUGETLB_UNMAPPED_AREA
+#endif
+
+#ifndef __ASSEMBLY__
+
+extern void _clear_page(void *page);
+#define clear_page(X)	_clear_page((void *)(X))
+struct page;
+extern void clear_user_page(void *addr, unsigned long vaddr, struct page *page);
+#define copy_page(X,Y)	memcpy((void *)(X), (void *)(Y), PAGE_SIZE)
+extern void copy_user_page(void *to, void *from, unsigned long vaddr, struct page *topage);
+
+/* Unlike sparc32, sparc64's parameter passing API is more
+ * sane in that structures which as small enough are passed
+ * in registers instead of on the stack.  Thus, setting
+ * STRICT_MM_TYPECHECKS does not generate worse code so
+ * let's enable it to get the type checking.
+ */
+
+#define STRICT_MM_TYPECHECKS
+
+#ifdef STRICT_MM_TYPECHECKS
+/* These are used to make use of C type-checking.. */
+typedef struct { unsigned long pte; } pte_t;
+typedef struct { unsigned long iopte; } iopte_t;
+typedef struct { unsigned int pmd; } pmd_t;
+typedef struct { unsigned int pgd; } pgd_t;
+typedef struct { unsigned long pgprot; } pgprot_t;
+
+#define pte_val(x)	((x).pte)
+#define iopte_val(x)	((x).iopte)
+#define pmd_val(x)      ((x).pmd)
+#define pgd_val(x)	((x).pgd)
+#define pgprot_val(x)	((x).pgprot)
+
+#define __pte(x)	((pte_t) { (x) } )
+#define __iopte(x)	((iopte_t) { (x) } )
+#define __pmd(x)        ((pmd_t) { (x) } )
+#define __pgd(x)	((pgd_t) { (x) } )
+#define __pgprot(x)	((pgprot_t) { (x) } )
+
+#else
+/* .. while these make it easier on the compiler */
+typedef unsigned long pte_t;
+typedef unsigned long iopte_t;
+typedef unsigned int pmd_t;
+typedef unsigned int pgd_t;
+typedef unsigned long pgprot_t;
+
+#define pte_val(x)	(x)
+#define iopte_val(x)	(x)
+#define pmd_val(x)      (x)
+#define pgd_val(x)	(x)
+#define pgprot_val(x)	(x)
+
+#define __pte(x)	(x)
+#define __iopte(x)	(x)
+#define __pmd(x)        (x)
+#define __pgd(x)	(x)
+#define __pgprot(x)	(x)
+
+#endif /* (STRICT_MM_TYPECHECKS) */
+
+typedef struct page *pgtable_t;
+
+#define TASK_UNMAPPED_BASE	(test_thread_flag(TIF_32BIT) ? \
+				 (_AC(0x0000000070000000,UL)) : \
+				 (_AC(0xfffff80000000000,UL) + (1UL << 32UL)))
+
+#include <asm-generic/memory_model.h>
+
+#endif /* !(__ASSEMBLY__) */
+
+/* to align the pointer to the (next) page boundary */
+#define PAGE_ALIGN(addr)	(((addr)+PAGE_SIZE-1)&PAGE_MASK)
+
+/* We used to stick this into a hard-coded global register (%g4)
+ * but that does not make sense anymore.
+ */
+#define PAGE_OFFSET		_AC(0xFFFFF80000000000,UL)
+
+#ifndef __ASSEMBLY__
+
+#define __pa(x)			((unsigned long)(x) - PAGE_OFFSET)
+#define __va(x)			((void *)((unsigned long) (x) + PAGE_OFFSET))
+
+#define pfn_to_kaddr(pfn)	__va((pfn) << PAGE_SHIFT)
+
+#define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr)>>PAGE_SHIFT)
+
+#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
+
+#define virt_to_phys __pa
+#define phys_to_virt __va
+
+#endif /* !(__ASSEMBLY__) */
+
+#define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \
+				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+#include <asm-generic/page.h>
+
+#endif /* _SPARC64_PAGE_H */
diff --git a/include/asm-sparc/parport.h b/include/asm-sparc/parport.h
new file mode 100644
index 00000000000..7818b2523b8
--- /dev/null
+++ b/include/asm-sparc/parport.h
@@ -0,0 +1,246 @@
+/* parport.h: sparc64 specific parport initialization and dma.
+ *
+ * Copyright (C) 1999  Eddie C. Dost  (ecd@skynet.be)
+ */
+
+#ifndef _ASM_SPARC64_PARPORT_H
+#define _ASM_SPARC64_PARPORT_H 1
+
+#include <asm/ebus.h>
+#include <asm/ns87303.h>
+#include <asm/of_device.h>
+#include <asm/prom.h>
+
+#define PARPORT_PC_MAX_PORTS	PARPORT_MAX
+
+/*
+ * While sparc64 doesn't have an ISA DMA API, we provide something that looks
+ * close enough to make parport_pc happy
+ */
+#define HAS_DMA
+
+static DEFINE_SPINLOCK(dma_spin_lock);
+
+#define claim_dma_lock() \
+({	unsigned long flags; \
+	spin_lock_irqsave(&dma_spin_lock, flags); \
+	flags; \
+})
+
+#define release_dma_lock(__flags) \
+	spin_unlock_irqrestore(&dma_spin_lock, __flags);
+
+static struct sparc_ebus_info {
+	struct ebus_dma_info info;
+	unsigned int addr;
+	unsigned int count;
+	int lock;
+
+	struct parport *port;
+} sparc_ebus_dmas[PARPORT_PC_MAX_PORTS];
+
+static DECLARE_BITMAP(dma_slot_map, PARPORT_PC_MAX_PORTS);
+
+static inline int request_dma(unsigned int dmanr, const char *device_id)
+{
+	if (dmanr >= PARPORT_PC_MAX_PORTS)
+		return -EINVAL;
+	if (xchg(&sparc_ebus_dmas[dmanr].lock, 1) != 0)
+		return -EBUSY;
+	return 0;
+}
+
+static inline void free_dma(unsigned int dmanr)
+{
+	if (dmanr >= PARPORT_PC_MAX_PORTS) {
+		printk(KERN_WARNING "Trying to free DMA%d\n", dmanr);
+		return;
+	}
+	if (xchg(&sparc_ebus_dmas[dmanr].lock, 0) == 0) {
+		printk(KERN_WARNING "Trying to free free DMA%d\n", dmanr);
+		return;
+	}
+}
+
+static inline void enable_dma(unsigned int dmanr)
+{
+	ebus_dma_enable(&sparc_ebus_dmas[dmanr].info, 1);
+
+	if (ebus_dma_request(&sparc_ebus_dmas[dmanr].info,
+			     sparc_ebus_dmas[dmanr].addr,
+			     sparc_ebus_dmas[dmanr].count))
+		BUG();
+}
+
+static inline void disable_dma(unsigned int dmanr)
+{
+	ebus_dma_enable(&sparc_ebus_dmas[dmanr].info, 0);
+}
+
+static inline void clear_dma_ff(unsigned int dmanr)
+{
+	/* nothing */
+}
+
+static inline void set_dma_mode(unsigned int dmanr, char mode)
+{
+	ebus_dma_prepare(&sparc_ebus_dmas[dmanr].info, (mode != DMA_MODE_WRITE));
+}
+
+static inline void set_dma_addr(unsigned int dmanr, unsigned int addr)
+{
+	sparc_ebus_dmas[dmanr].addr = addr;
+}
+
+static inline void set_dma_count(unsigned int dmanr, unsigned int count)
+{
+	sparc_ebus_dmas[dmanr].count = count;
+}
+
+static inline unsigned int get_dma_residue(unsigned int dmanr)
+{
+	return ebus_dma_residue(&sparc_ebus_dmas[dmanr].info);
+}
+
+static int __devinit ecpp_probe(struct of_device *op, const struct of_device_id *match)
+{
+	unsigned long base = op->resource[0].start;
+	unsigned long config = op->resource[1].start;
+	unsigned long d_base = op->resource[2].start;
+	unsigned long d_len;
+	struct device_node *parent;
+	struct parport *p;
+	int slot, err;
+
+	parent = op->node->parent;
+	if (!strcmp(parent->name, "dma")) {
+		p = parport_pc_probe_port(base, base + 0x400,
+					  op->irqs[0], PARPORT_DMA_NOFIFO,
+					  op->dev.parent->parent);
+		if (!p)
+			return -ENOMEM;
+		dev_set_drvdata(&op->dev, p);
+		return 0;
+	}
+
+	for (slot = 0; slot < PARPORT_PC_MAX_PORTS; slot++) {
+		if (!test_and_set_bit(slot, dma_slot_map))
+			break;
+	}
+	err = -ENODEV;
+	if (slot >= PARPORT_PC_MAX_PORTS)
+		goto out_err;
+
+	spin_lock_init(&sparc_ebus_dmas[slot].info.lock);
+
+	d_len = (op->resource[2].end - d_base) + 1UL;
+	sparc_ebus_dmas[slot].info.regs =
+		of_ioremap(&op->resource[2], 0, d_len, "ECPP DMA");
+
+	if (!sparc_ebus_dmas[slot].info.regs)
+		goto out_clear_map;
+
+	sparc_ebus_dmas[slot].info.flags = 0;
+	sparc_ebus_dmas[slot].info.callback = NULL;
+	sparc_ebus_dmas[slot].info.client_cookie = NULL;
+	sparc_ebus_dmas[slot].info.irq = 0xdeadbeef;
+	strcpy(sparc_ebus_dmas[slot].info.name, "parport");
+	if (ebus_dma_register(&sparc_ebus_dmas[slot].info))
+		goto out_unmap_regs;
+
+	ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 1);
+
+	/* Configure IRQ to Push Pull, Level Low */
+	/* Enable ECP, set bit 2 of the CTR first */
+	outb(0x04, base + 0x02);
+	ns87303_modify(config, PCR,
+		       PCR_EPP_ENABLE |
+		       PCR_IRQ_ODRAIN,
+		       PCR_ECP_ENABLE |
+		       PCR_ECP_CLK_ENA |
+		       PCR_IRQ_POLAR);
+
+	/* CTR bit 5 controls direction of port */
+	ns87303_modify(config, PTR,
+		       0, PTR_LPT_REG_DIR);
+
+	p = parport_pc_probe_port(base, base + 0x400,
+				  op->irqs[0],
+				  slot,
+				  op->dev.parent);
+	err = -ENOMEM;
+	if (!p)
+		goto out_disable_irq;
+
+	dev_set_drvdata(&op->dev, p);
+
+	return 0;
+
+out_disable_irq:
+	ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 0);
+	ebus_dma_unregister(&sparc_ebus_dmas[slot].info);
+
+out_unmap_regs:
+	of_iounmap(&op->resource[2], sparc_ebus_dmas[slot].info.regs, d_len);
+
+out_clear_map:
+	clear_bit(slot, dma_slot_map);
+
+out_err:
+	return err;
+}
+
+static int __devexit ecpp_remove(struct of_device *op)
+{
+	struct parport *p = dev_get_drvdata(&op->dev);
+	int slot = p->dma;
+
+	parport_pc_unregister_port(p);
+
+	if (slot != PARPORT_DMA_NOFIFO) {
+		unsigned long d_base = op->resource[2].start;
+		unsigned long d_len;
+
+		d_len = (op->resource[2].end - d_base) + 1UL;
+
+		ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 0);
+		ebus_dma_unregister(&sparc_ebus_dmas[slot].info);
+		of_iounmap(&op->resource[2],
+			   sparc_ebus_dmas[slot].info.regs,
+			   d_len);
+		clear_bit(slot, dma_slot_map);
+	}
+
+	return 0;
+}
+
+static struct of_device_id ecpp_match[] = {
+	{
+		.name = "ecpp",
+	},
+	{
+		.name = "parallel",
+		.compatible = "ecpp",
+	},
+	{
+		.name = "parallel",
+		.compatible = "ns87317-ecpp",
+	},
+	{},
+};
+
+static struct of_platform_driver ecpp_driver = {
+	.name			= "ecpp",
+	.match_table		= ecpp_match,
+	.probe			= ecpp_probe,
+	.remove			= __devexit_p(ecpp_remove),
+};
+
+static int parport_pc_find_nonpci_ports(int autoirq, int autodma)
+{
+	of_register_driver(&ecpp_driver, &of_bus_type);
+
+	return 0;
+}
+
+#endif /* !(_ASM_SPARC64_PARPORT_H */
diff --git a/include/asm-sparc/pci.h b/include/asm-sparc/pci.h
index b93b6c79e08..b807d52a480 100644
--- a/include/asm-sparc/pci.h
+++ b/include/asm-sparc/pci.h
@@ -1,170 +1,8 @@
-#ifndef __SPARC_PCI_H
-#define __SPARC_PCI_H
-
-#ifdef __KERNEL__
-
-/* Can be used to override the logic in pci_scan_bus for skipping
- * already-configured bus numbers - to be used for buggy BIOSes
- * or architectures with incomplete PCI setup by the loader.
- */
-#define pcibios_assign_all_busses()	0
-#define pcibios_scan_all_fns(a, b)	0
-
-#define PCIBIOS_MIN_IO		0UL
-#define PCIBIOS_MIN_MEM		0UL
-
-#define PCI_IRQ_NONE		0xffffffff
-
-static inline void pcibios_set_master(struct pci_dev *dev)
-{
-	/* No special bus mastering setup handling */
-}
-
-static inline void pcibios_penalize_isa_irq(int irq, int active)
-{
-	/* We don't do dynamic PCI IRQ allocation */
-}
-
-/* Dynamic DMA mapping stuff.
- */
-#define PCI_DMA_BUS_IS_PHYS	(0)
-
-#include <asm/scatterlist.h>
-
-struct pci_dev;
-
-/* Allocate and map kernel buffer using consistent mode DMA for a device.
- * hwdev should be valid struct pci_dev pointer for PCI devices.
- */
-extern void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size, dma_addr_t *dma_handle);
-
-/* Free and unmap a consistent DMA buffer.
- * cpu_addr is what was returned from pci_alloc_consistent,
- * size must be the same as what as passed into pci_alloc_consistent,
- * and likewise dma_addr must be the same as what *dma_addrp was set to.
- *
- * References to the memory and mappings assosciated with cpu_addr/dma_addr
- * past this call are illegal.
- */
-extern void pci_free_consistent(struct pci_dev *hwdev, size_t size, void *vaddr, dma_addr_t dma_handle);
-
-/* Map a single buffer of the indicated size for DMA in streaming mode.
- * The 32-bit bus address to use is returned.
- *
- * Once the device is given the dma address, the device owns this memory
- * until either pci_unmap_single or pci_dma_sync_single_for_cpu is performed.
- */
-extern dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr, size_t size, int direction);
-
-/* Unmap a single streaming mode DMA translation.  The dma_addr and size
- * must match what was provided for in a previous pci_map_single call.  All
- * other usages are undefined.
- *
- * After this call, reads by the cpu to the buffer are guaranteed to see
- * whatever the device wrote there.
- */
-extern void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t dma_addr, size_t size, int direction);
-
-/* pci_unmap_{single,page} is not a nop, thus... */
-#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
-	dma_addr_t ADDR_NAME;
-#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
-	__u32 LEN_NAME;
-#define pci_unmap_addr(PTR, ADDR_NAME)			\
-	((PTR)->ADDR_NAME)
-#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
-	(((PTR)->ADDR_NAME) = (VAL))
-#define pci_unmap_len(PTR, LEN_NAME)			\
-	((PTR)->LEN_NAME)
-#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
-	(((PTR)->LEN_NAME) = (VAL))
-
-/*
- * Same as above, only with pages instead of mapped addresses.
- */
-extern dma_addr_t pci_map_page(struct pci_dev *hwdev, struct page *page,
-			unsigned long offset, size_t size, int direction);
-extern void pci_unmap_page(struct pci_dev *hwdev,
-			dma_addr_t dma_address, size_t size, int direction);
-
-/* Map a set of buffers described by scatterlist in streaming
- * mode for DMA.  This is the scather-gather version of the
- * above pci_map_single interface.  Here the scatter gather list
- * elements are each tagged with the appropriate dma address
- * and length.  They are obtained via sg_dma_{address,length}(SG).
- *
- * NOTE: An implementation may be able to use a smaller number of
- *       DMA address/length pairs than there are SG table elements.
- *       (for example via virtual mapping capabilities)
- *       The routine returns the number of addr/length pairs actually
- *       used, at most nents.
- *
- * Device ownership issues as mentioned above for pci_map_single are
- * the same here.
- */
-extern int pci_map_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nents, int direction);
-
-/* Unmap a set of streaming mode DMA translations.
- * Again, cpu read rules concerning calls here are the same as for
- * pci_unmap_single() above.
- */
-extern void pci_unmap_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nhwents, int direction);
-
-/* Make physical memory consistent for a single
- * streaming mode DMA translation after a transfer.
- *
- * If you perform a pci_map_single() but wish to interrogate the
- * buffer using the cpu, yet do not wish to teardown the PCI dma
- * mapping, you must call this function before doing so.  At the
- * next point you give the PCI dma address back to the card, you
- * must first perform a pci_dma_sync_for_device, and then the device
- * again owns the buffer.
- */
-extern void pci_dma_sync_single_for_cpu(struct pci_dev *hwdev, dma_addr_t dma_handle, size_t size, int direction);
-extern void pci_dma_sync_single_for_device(struct pci_dev *hwdev, dma_addr_t dma_handle, size_t size, int direction);
-
-/* Make physical memory consistent for a set of streaming
- * mode DMA translations after a transfer.
- *
- * The same as pci_dma_sync_single_* but for a scatter-gather list,
- * same rules and usage.
- */
-extern void pci_dma_sync_sg_for_cpu(struct pci_dev *hwdev, struct scatterlist *sg, int nelems, int direction);
-extern void pci_dma_sync_sg_for_device(struct pci_dev *hwdev, struct scatterlist *sg, int nelems, int direction);
-
-/* Return whether the given PCI device DMA address mask can
- * be supported properly.  For example, if your device can
- * only drive the low 24-bits during PCI bus mastering, then
- * you would pass 0x00ffffff as the mask to this function.
- */
-static inline int pci_dma_supported(struct pci_dev *hwdev, u64 mask)
-{
-	return 1;
-}
-
-#ifdef CONFIG_PCI
-static inline void pci_dma_burst_advice(struct pci_dev *pdev,
-					enum pci_dma_burst_strategy *strat,
-					unsigned long *strategy_parameter)
-{
-	*strat = PCI_DMA_BURST_INFINITY;
-	*strategy_parameter = ~0UL;
-}
+#ifndef ___ASM_SPARC_PCI_H
+#define ___ASM_SPARC_PCI_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/pci_64.h>
+#else
+#include <asm-sparc/pci_32.h>
+#endif
 #endif
-
-#define PCI_DMA_ERROR_CODE      (~(dma_addr_t)0x0)
-
-static inline int pci_dma_mapping_error(dma_addr_t dma_addr)
-{
-        return (dma_addr == PCI_DMA_ERROR_CODE);
-}
-
-struct device_node;
-extern struct device_node *pci_device_to_OF_node(struct pci_dev *pdev);
-
-#endif /* __KERNEL__ */
-
-/* generic pci stuff */
-#include <asm-generic/pci.h>
-
-#endif /* __SPARC_PCI_H */
diff --git a/include/asm-sparc/pci_32.h b/include/asm-sparc/pci_32.h
new file mode 100644
index 00000000000..b93b6c79e08
--- /dev/null
+++ b/include/asm-sparc/pci_32.h
@@ -0,0 +1,170 @@
+#ifndef __SPARC_PCI_H
+#define __SPARC_PCI_H
+
+#ifdef __KERNEL__
+
+/* Can be used to override the logic in pci_scan_bus for skipping
+ * already-configured bus numbers - to be used for buggy BIOSes
+ * or architectures with incomplete PCI setup by the loader.
+ */
+#define pcibios_assign_all_busses()	0
+#define pcibios_scan_all_fns(a, b)	0
+
+#define PCIBIOS_MIN_IO		0UL
+#define PCIBIOS_MIN_MEM		0UL
+
+#define PCI_IRQ_NONE		0xffffffff
+
+static inline void pcibios_set_master(struct pci_dev *dev)
+{
+	/* No special bus mastering setup handling */
+}
+
+static inline void pcibios_penalize_isa_irq(int irq, int active)
+{
+	/* We don't do dynamic PCI IRQ allocation */
+}
+
+/* Dynamic DMA mapping stuff.
+ */
+#define PCI_DMA_BUS_IS_PHYS	(0)
+
+#include <asm/scatterlist.h>
+
+struct pci_dev;
+
+/* Allocate and map kernel buffer using consistent mode DMA for a device.
+ * hwdev should be valid struct pci_dev pointer for PCI devices.
+ */
+extern void *pci_alloc_consistent(struct pci_dev *hwdev, size_t size, dma_addr_t *dma_handle);
+
+/* Free and unmap a consistent DMA buffer.
+ * cpu_addr is what was returned from pci_alloc_consistent,
+ * size must be the same as what as passed into pci_alloc_consistent,
+ * and likewise dma_addr must be the same as what *dma_addrp was set to.
+ *
+ * References to the memory and mappings assosciated with cpu_addr/dma_addr
+ * past this call are illegal.
+ */
+extern void pci_free_consistent(struct pci_dev *hwdev, size_t size, void *vaddr, dma_addr_t dma_handle);
+
+/* Map a single buffer of the indicated size for DMA in streaming mode.
+ * The 32-bit bus address to use is returned.
+ *
+ * Once the device is given the dma address, the device owns this memory
+ * until either pci_unmap_single or pci_dma_sync_single_for_cpu is performed.
+ */
+extern dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr, size_t size, int direction);
+
+/* Unmap a single streaming mode DMA translation.  The dma_addr and size
+ * must match what was provided for in a previous pci_map_single call.  All
+ * other usages are undefined.
+ *
+ * After this call, reads by the cpu to the buffer are guaranteed to see
+ * whatever the device wrote there.
+ */
+extern void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t dma_addr, size_t size, int direction);
+
+/* pci_unmap_{single,page} is not a nop, thus... */
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
+	dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
+	__u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)			\
+	((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
+	(((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)			\
+	((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
+	(((PTR)->LEN_NAME) = (VAL))
+
+/*
+ * Same as above, only with pages instead of mapped addresses.
+ */
+extern dma_addr_t pci_map_page(struct pci_dev *hwdev, struct page *page,
+			unsigned long offset, size_t size, int direction);
+extern void pci_unmap_page(struct pci_dev *hwdev,
+			dma_addr_t dma_address, size_t size, int direction);
+
+/* Map a set of buffers described by scatterlist in streaming
+ * mode for DMA.  This is the scather-gather version of the
+ * above pci_map_single interface.  Here the scatter gather list
+ * elements are each tagged with the appropriate dma address
+ * and length.  They are obtained via sg_dma_{address,length}(SG).
+ *
+ * NOTE: An implementation may be able to use a smaller number of
+ *       DMA address/length pairs than there are SG table elements.
+ *       (for example via virtual mapping capabilities)
+ *       The routine returns the number of addr/length pairs actually
+ *       used, at most nents.
+ *
+ * Device ownership issues as mentioned above for pci_map_single are
+ * the same here.
+ */
+extern int pci_map_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nents, int direction);
+
+/* Unmap a set of streaming mode DMA translations.
+ * Again, cpu read rules concerning calls here are the same as for
+ * pci_unmap_single() above.
+ */
+extern void pci_unmap_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nhwents, int direction);
+
+/* Make physical memory consistent for a single
+ * streaming mode DMA translation after a transfer.
+ *
+ * If you perform a pci_map_single() but wish to interrogate the
+ * buffer using the cpu, yet do not wish to teardown the PCI dma
+ * mapping, you must call this function before doing so.  At the
+ * next point you give the PCI dma address back to the card, you
+ * must first perform a pci_dma_sync_for_device, and then the device
+ * again owns the buffer.
+ */
+extern void pci_dma_sync_single_for_cpu(struct pci_dev *hwdev, dma_addr_t dma_handle, size_t size, int direction);
+extern void pci_dma_sync_single_for_device(struct pci_dev *hwdev, dma_addr_t dma_handle, size_t size, int direction);
+
+/* Make physical memory consistent for a set of streaming
+ * mode DMA translations after a transfer.
+ *
+ * The same as pci_dma_sync_single_* but for a scatter-gather list,
+ * same rules and usage.
+ */
+extern void pci_dma_sync_sg_for_cpu(struct pci_dev *hwdev, struct scatterlist *sg, int nelems, int direction);
+extern void pci_dma_sync_sg_for_device(struct pci_dev *hwdev, struct scatterlist *sg, int nelems, int direction);
+
+/* Return whether the given PCI device DMA address mask can
+ * be supported properly.  For example, if your device can
+ * only drive the low 24-bits during PCI bus mastering, then
+ * you would pass 0x00ffffff as the mask to this function.
+ */
+static inline int pci_dma_supported(struct pci_dev *hwdev, u64 mask)
+{
+	return 1;
+}
+
+#ifdef CONFIG_PCI
+static inline void pci_dma_burst_advice(struct pci_dev *pdev,
+					enum pci_dma_burst_strategy *strat,
+					unsigned long *strategy_parameter)
+{
+	*strat = PCI_DMA_BURST_INFINITY;
+	*strategy_parameter = ~0UL;
+}
+#endif
+
+#define PCI_DMA_ERROR_CODE      (~(dma_addr_t)0x0)
+
+static inline int pci_dma_mapping_error(dma_addr_t dma_addr)
+{
+        return (dma_addr == PCI_DMA_ERROR_CODE);
+}
+
+struct device_node;
+extern struct device_node *pci_device_to_OF_node(struct pci_dev *pdev);
+
+#endif /* __KERNEL__ */
+
+/* generic pci stuff */
+#include <asm-generic/pci.h>
+
+#endif /* __SPARC_PCI_H */
diff --git a/include/asm-sparc/pci_64.h b/include/asm-sparc/pci_64.h
new file mode 100644
index 00000000000..f59f2571295
--- /dev/null
+++ b/include/asm-sparc/pci_64.h
@@ -0,0 +1,209 @@
+#ifndef __SPARC64_PCI_H
+#define __SPARC64_PCI_H
+
+#ifdef __KERNEL__
+
+#include <linux/dma-mapping.h>
+
+/* Can be used to override the logic in pci_scan_bus for skipping
+ * already-configured bus numbers - to be used for buggy BIOSes
+ * or architectures with incomplete PCI setup by the loader.
+ */
+#define pcibios_assign_all_busses()	0
+#define pcibios_scan_all_fns(a, b)	0
+
+#define PCIBIOS_MIN_IO		0UL
+#define PCIBIOS_MIN_MEM		0UL
+
+#define PCI_IRQ_NONE		0xffffffff
+
+#define PCI_CACHE_LINE_BYTES	64
+
+static inline void pcibios_set_master(struct pci_dev *dev)
+{
+	/* No special bus mastering setup handling */
+}
+
+static inline void pcibios_penalize_isa_irq(int irq, int active)
+{
+	/* We don't do dynamic PCI IRQ allocation */
+}
+
+/* The PCI address space does not equal the physical memory
+ * address space.  The networking and block device layers use
+ * this boolean for bounce buffer decisions.
+ */
+#define PCI_DMA_BUS_IS_PHYS	(0)
+
+static inline void *pci_alloc_consistent(struct pci_dev *pdev, size_t size,
+					 dma_addr_t *dma_handle)
+{
+	return dma_alloc_coherent(&pdev->dev, size, dma_handle, GFP_ATOMIC);
+}
+
+static inline void pci_free_consistent(struct pci_dev *pdev, size_t size,
+				       void *vaddr, dma_addr_t dma_handle)
+{
+	return dma_free_coherent(&pdev->dev, size, vaddr, dma_handle);
+}
+
+static inline dma_addr_t pci_map_single(struct pci_dev *pdev, void *ptr,
+					size_t size, int direction)
+{
+	return dma_map_single(&pdev->dev, ptr, size,
+			      (enum dma_data_direction) direction);
+}
+
+static inline void pci_unmap_single(struct pci_dev *pdev, dma_addr_t dma_addr,
+				    size_t size, int direction)
+{
+	dma_unmap_single(&pdev->dev, dma_addr, size,
+			 (enum dma_data_direction) direction);
+}
+
+#define pci_map_page(dev, page, off, size, dir) \
+	pci_map_single(dev, (page_address(page) + (off)), size, dir)
+#define pci_unmap_page(dev,addr,sz,dir) \
+	pci_unmap_single(dev,addr,sz,dir)
+
+/* pci_unmap_{single,page} is not a nop, thus... */
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
+	dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
+	__u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)			\
+	((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
+	(((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)			\
+	((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
+	(((PTR)->LEN_NAME) = (VAL))
+
+static inline int pci_map_sg(struct pci_dev *pdev, struct scatterlist *sg,
+			     int nents, int direction)
+{
+	return dma_map_sg(&pdev->dev, sg, nents,
+			  (enum dma_data_direction) direction);
+}
+
+static inline void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sg,
+				int nents, int direction)
+{
+	dma_unmap_sg(&pdev->dev, sg, nents,
+		     (enum dma_data_direction) direction);
+}
+
+static inline void pci_dma_sync_single_for_cpu(struct pci_dev *pdev,
+					       dma_addr_t dma_handle,
+					       size_t size, int direction)
+{
+	dma_sync_single_for_cpu(&pdev->dev, dma_handle, size,
+				(enum dma_data_direction) direction);
+}
+
+static inline void pci_dma_sync_single_for_device(struct pci_dev *pdev,
+						  dma_addr_t dma_handle,
+						  size_t size, int direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+static inline void pci_dma_sync_sg_for_cpu(struct pci_dev *pdev,
+					   struct scatterlist *sg,
+					   int nents, int direction)
+{
+	dma_sync_sg_for_cpu(&pdev->dev, sg, nents,
+			    (enum dma_data_direction) direction);
+}
+
+static inline void pci_dma_sync_sg_for_device(struct pci_dev *pdev,
+					      struct scatterlist *sg,
+					      int nelems, int direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+/* Return whether the given PCI device DMA address mask can
+ * be supported properly.  For example, if your device can
+ * only drive the low 24-bits during PCI bus mastering, then
+ * you would pass 0x00ffffff as the mask to this function.
+ */
+extern int pci_dma_supported(struct pci_dev *hwdev, u64 mask);
+
+/* PCI IOMMU mapping bypass support. */
+
+/* PCI 64-bit addressing works for all slots on all controller
+ * types on sparc64.  However, it requires that the device
+ * can drive enough of the 64 bits.
+ */
+#define PCI64_REQUIRED_MASK	(~(dma64_addr_t)0)
+#define PCI64_ADDR_BASE		0xfffc000000000000UL
+
+static inline int pci_dma_mapping_error(dma_addr_t dma_addr)
+{
+	return dma_mapping_error(dma_addr);
+}
+
+#ifdef CONFIG_PCI
+static inline void pci_dma_burst_advice(struct pci_dev *pdev,
+					enum pci_dma_burst_strategy *strat,
+					unsigned long *strategy_parameter)
+{
+	unsigned long cacheline_size;
+	u8 byte;
+
+	pci_read_config_byte(pdev, PCI_CACHE_LINE_SIZE, &byte);
+	if (byte == 0)
+		cacheline_size = 1024;
+	else
+		cacheline_size = (int) byte * 4;
+
+	*strat = PCI_DMA_BURST_BOUNDARY;
+	*strategy_parameter = cacheline_size;
+}
+#endif
+
+/* Return the index of the PCI controller for device PDEV. */
+
+extern int pci_domain_nr(struct pci_bus *bus);
+static inline int pci_proc_domain(struct pci_bus *bus)
+{
+	return 1;
+}
+
+/* Platform support for /proc/bus/pci/X/Y mmap()s. */
+
+#define HAVE_PCI_MMAP
+#define HAVE_ARCH_PCI_GET_UNMAPPED_AREA
+#define get_pci_unmapped_area get_fb_unmapped_area
+
+extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+			       enum pci_mmap_state mmap_state,
+			       int write_combine);
+
+extern void
+pcibios_resource_to_bus(struct pci_dev *dev, struct pci_bus_region *region,
+			struct resource *res);
+
+extern void
+pcibios_bus_to_resource(struct pci_dev *dev, struct resource *res,
+			struct pci_bus_region *region);
+
+extern struct resource *pcibios_select_root(struct pci_dev *, struct resource *);
+
+static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
+{
+	return PCI_IRQ_NONE;
+}
+
+struct device_node;
+extern struct device_node *pci_device_to_OF_node(struct pci_dev *pdev);
+
+#define HAVE_ARCH_PCI_RESOURCE_TO_USER
+extern void pci_resource_to_user(const struct pci_dev *dev, int bar,
+				 const struct resource *rsrc,
+				 resource_size_t *start, resource_size_t *end);
+#endif /* __KERNEL__ */
+
+#endif /* __SPARC64_PCI_H */
diff --git a/include/asm-sparc/percpu.h b/include/asm-sparc/percpu.h
index 06066a7aaec..d98ed6cf2e3 100644
--- a/include/asm-sparc/percpu.h
+++ b/include/asm-sparc/percpu.h
@@ -1,6 +1,8 @@
-#ifndef __ARCH_SPARC_PERCPU__
-#define __ARCH_SPARC_PERCPU__
-
-#include <asm-generic/percpu.h>
-
-#endif /* __ARCH_SPARC_PERCPU__ */
+#ifndef ___ASM_SPARC_PERCPU_H
+#define ___ASM_SPARC_PERCPU_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/percpu_64.h>
+#else
+#include <asm-sparc/percpu_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/percpu_32.h b/include/asm-sparc/percpu_32.h
new file mode 100644
index 00000000000..06066a7aaec
--- /dev/null
+++ b/include/asm-sparc/percpu_32.h
@@ -0,0 +1,6 @@
+#ifndef __ARCH_SPARC_PERCPU__
+#define __ARCH_SPARC_PERCPU__
+
+#include <asm-generic/percpu.h>
+
+#endif /* __ARCH_SPARC_PERCPU__ */
diff --git a/include/asm-sparc/percpu_64.h b/include/asm-sparc/percpu_64.h
new file mode 100644
index 00000000000..bee64593023
--- /dev/null
+++ b/include/asm-sparc/percpu_64.h
@@ -0,0 +1,28 @@
+#ifndef __ARCH_SPARC64_PERCPU__
+#define __ARCH_SPARC64_PERCPU__
+
+#include <linux/compiler.h>
+
+register unsigned long __local_per_cpu_offset asm("g5");
+
+#ifdef CONFIG_SMP
+
+extern void real_setup_per_cpu_areas(void);
+
+extern unsigned long __per_cpu_base;
+extern unsigned long __per_cpu_shift;
+#define __per_cpu_offset(__cpu) \
+	(__per_cpu_base + ((unsigned long)(__cpu) << __per_cpu_shift))
+#define per_cpu_offset(x) (__per_cpu_offset(x))
+
+#define __my_cpu_offset __local_per_cpu_offset
+
+#else /* ! SMP */
+
+#define real_setup_per_cpu_areas()		do { } while (0)
+
+#endif	/* SMP */
+
+#include <asm-generic/percpu.h>
+
+#endif /* __ARCH_SPARC64_PERCPU__ */
diff --git a/include/asm-sparc/pgalloc.h b/include/asm-sparc/pgalloc.h
index 681582d2696..7fa02b53d39 100644
--- a/include/asm-sparc/pgalloc.h
+++ b/include/asm-sparc/pgalloc.h
@@ -1,68 +1,8 @@
-#ifndef _SPARC_PGALLOC_H
-#define _SPARC_PGALLOC_H
-
-#include <linux/kernel.h>
-#include <linux/sched.h>
-
-#include <asm/page.h>
-#include <asm/btfixup.h>
-
-struct page;
-
-extern struct pgtable_cache_struct {
-	unsigned long *pgd_cache;
-	unsigned long *pte_cache;
-	unsigned long pgtable_cache_sz;
-	unsigned long pgd_cache_sz;
-} pgt_quicklists;
-#define pgd_quicklist           (pgt_quicklists.pgd_cache)
-#define pmd_quicklist           ((unsigned long *)0)
-#define pte_quicklist           (pgt_quicklists.pte_cache)
-#define pgtable_cache_size      (pgt_quicklists.pgtable_cache_sz)
-#define pgd_cache_size		(pgt_quicklists.pgd_cache_sz)
-
-extern void check_pgt_cache(void);
-BTFIXUPDEF_CALL(void,	 do_check_pgt_cache, int, int)
-#define do_check_pgt_cache(low,high) BTFIXUP_CALL(do_check_pgt_cache)(low,high)
-
-BTFIXUPDEF_CALL(pgd_t *, get_pgd_fast, void)
-#define get_pgd_fast()		BTFIXUP_CALL(get_pgd_fast)()
-
-BTFIXUPDEF_CALL(void, free_pgd_fast, pgd_t *)
-#define free_pgd_fast(pgd)	BTFIXUP_CALL(free_pgd_fast)(pgd)
-
-#define pgd_free(mm, pgd)	free_pgd_fast(pgd)
-#define pgd_alloc(mm)	get_pgd_fast()
-
-BTFIXUPDEF_CALL(void, pgd_set, pgd_t *, pmd_t *)
-#define pgd_set(pgdp,pmdp) BTFIXUP_CALL(pgd_set)(pgdp,pmdp)
-#define pgd_populate(MM, PGD, PMD)      pgd_set(PGD, PMD)
-
-BTFIXUPDEF_CALL(pmd_t *, pmd_alloc_one, struct mm_struct *, unsigned long)
-#define pmd_alloc_one(mm, address)	BTFIXUP_CALL(pmd_alloc_one)(mm, address)
-
-BTFIXUPDEF_CALL(void, free_pmd_fast, pmd_t *)
-#define free_pmd_fast(pmd)	BTFIXUP_CALL(free_pmd_fast)(pmd)
-
-#define pmd_free(mm, pmd)	free_pmd_fast(pmd)
-#define __pmd_free_tlb(tlb, pmd) pmd_free((tlb)->mm, pmd)
-
-BTFIXUPDEF_CALL(void, pmd_populate, pmd_t *, struct page *)
-#define pmd_populate(MM, PMD, PTE)        BTFIXUP_CALL(pmd_populate)(PMD, PTE)
-#define pmd_pgtable(pmd) pmd_page(pmd)
-BTFIXUPDEF_CALL(void, pmd_set, pmd_t *, pte_t *)
-#define pmd_populate_kernel(MM, PMD, PTE) BTFIXUP_CALL(pmd_set)(PMD, PTE)
-
-BTFIXUPDEF_CALL(pgtable_t , pte_alloc_one, struct mm_struct *, unsigned long)
-#define pte_alloc_one(mm, address)	BTFIXUP_CALL(pte_alloc_one)(mm, address)
-BTFIXUPDEF_CALL(pte_t *, pte_alloc_one_kernel, struct mm_struct *, unsigned long)
-#define pte_alloc_one_kernel(mm, addr)	BTFIXUP_CALL(pte_alloc_one_kernel)(mm, addr)
-
-BTFIXUPDEF_CALL(void, free_pte_fast, pte_t *)
-#define pte_free_kernel(mm, pte)	BTFIXUP_CALL(free_pte_fast)(pte)
-
-BTFIXUPDEF_CALL(void, pte_free, pgtable_t )
-#define pte_free(mm, pte)	BTFIXUP_CALL(pte_free)(pte)
-#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, pte)
-
-#endif /* _SPARC_PGALLOC_H */
+#ifndef ___ASM_SPARC_PGALLOC_H
+#define ___ASM_SPARC_PGALLOC_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/pgalloc_64.h>
+#else
+#include <asm-sparc/pgalloc_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/pgalloc_32.h b/include/asm-sparc/pgalloc_32.h
new file mode 100644
index 00000000000..681582d2696
--- /dev/null
+++ b/include/asm-sparc/pgalloc_32.h
@@ -0,0 +1,68 @@
+#ifndef _SPARC_PGALLOC_H
+#define _SPARC_PGALLOC_H
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+
+#include <asm/page.h>
+#include <asm/btfixup.h>
+
+struct page;
+
+extern struct pgtable_cache_struct {
+	unsigned long *pgd_cache;
+	unsigned long *pte_cache;
+	unsigned long pgtable_cache_sz;
+	unsigned long pgd_cache_sz;
+} pgt_quicklists;
+#define pgd_quicklist           (pgt_quicklists.pgd_cache)
+#define pmd_quicklist           ((unsigned long *)0)
+#define pte_quicklist           (pgt_quicklists.pte_cache)
+#define pgtable_cache_size      (pgt_quicklists.pgtable_cache_sz)
+#define pgd_cache_size		(pgt_quicklists.pgd_cache_sz)
+
+extern void check_pgt_cache(void);
+BTFIXUPDEF_CALL(void,	 do_check_pgt_cache, int, int)
+#define do_check_pgt_cache(low,high) BTFIXUP_CALL(do_check_pgt_cache)(low,high)
+
+BTFIXUPDEF_CALL(pgd_t *, get_pgd_fast, void)
+#define get_pgd_fast()		BTFIXUP_CALL(get_pgd_fast)()
+
+BTFIXUPDEF_CALL(void, free_pgd_fast, pgd_t *)
+#define free_pgd_fast(pgd)	BTFIXUP_CALL(free_pgd_fast)(pgd)
+
+#define pgd_free(mm, pgd)	free_pgd_fast(pgd)
+#define pgd_alloc(mm)	get_pgd_fast()
+
+BTFIXUPDEF_CALL(void, pgd_set, pgd_t *, pmd_t *)
+#define pgd_set(pgdp,pmdp) BTFIXUP_CALL(pgd_set)(pgdp,pmdp)
+#define pgd_populate(MM, PGD, PMD)      pgd_set(PGD, PMD)
+
+BTFIXUPDEF_CALL(pmd_t *, pmd_alloc_one, struct mm_struct *, unsigned long)
+#define pmd_alloc_one(mm, address)	BTFIXUP_CALL(pmd_alloc_one)(mm, address)
+
+BTFIXUPDEF_CALL(void, free_pmd_fast, pmd_t *)
+#define free_pmd_fast(pmd)	BTFIXUP_CALL(free_pmd_fast)(pmd)
+
+#define pmd_free(mm, pmd)	free_pmd_fast(pmd)
+#define __pmd_free_tlb(tlb, pmd) pmd_free((tlb)->mm, pmd)
+
+BTFIXUPDEF_CALL(void, pmd_populate, pmd_t *, struct page *)
+#define pmd_populate(MM, PMD, PTE)        BTFIXUP_CALL(pmd_populate)(PMD, PTE)
+#define pmd_pgtable(pmd) pmd_page(pmd)
+BTFIXUPDEF_CALL(void, pmd_set, pmd_t *, pte_t *)
+#define pmd_populate_kernel(MM, PMD, PTE) BTFIXUP_CALL(pmd_set)(PMD, PTE)
+
+BTFIXUPDEF_CALL(pgtable_t , pte_alloc_one, struct mm_struct *, unsigned long)
+#define pte_alloc_one(mm, address)	BTFIXUP_CALL(pte_alloc_one)(mm, address)
+BTFIXUPDEF_CALL(pte_t *, pte_alloc_one_kernel, struct mm_struct *, unsigned long)
+#define pte_alloc_one_kernel(mm, addr)	BTFIXUP_CALL(pte_alloc_one_kernel)(mm, addr)
+
+BTFIXUPDEF_CALL(void, free_pte_fast, pte_t *)
+#define pte_free_kernel(mm, pte)	BTFIXUP_CALL(free_pte_fast)(pte)
+
+BTFIXUPDEF_CALL(void, pte_free, pgtable_t )
+#define pte_free(mm, pte)	BTFIXUP_CALL(pte_free)(pte)
+#define __pte_free_tlb(tlb, pte)	pte_free((tlb)->mm, pte)
+
+#endif /* _SPARC_PGALLOC_H */
diff --git a/include/asm-sparc/pgalloc_64.h b/include/asm-sparc/pgalloc_64.h
new file mode 100644
index 00000000000..5bdfa2c6e40
--- /dev/null
+++ b/include/asm-sparc/pgalloc_64.h
@@ -0,0 +1,81 @@
+#ifndef _SPARC64_PGALLOC_H
+#define _SPARC64_PGALLOC_H
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/quicklist.h>
+
+#include <asm/spitfire.h>
+#include <asm/cpudata.h>
+#include <asm/cacheflush.h>
+#include <asm/page.h>
+
+/* Page table allocation/freeing. */
+
+static inline pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+	return quicklist_alloc(0, GFP_KERNEL, NULL);
+}
+
+static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
+{
+	quicklist_free(0, NULL, pgd);
+}
+
+#define pud_populate(MM, PUD, PMD)	pud_set(PUD, PMD)
+
+static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long addr)
+{
+	return quicklist_alloc(0, GFP_KERNEL, NULL);
+}
+
+static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
+{
+	quicklist_free(0, NULL, pmd);
+}
+
+static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
+					  unsigned long address)
+{
+	return quicklist_alloc(0, GFP_KERNEL, NULL);
+}
+
+static inline pgtable_t pte_alloc_one(struct mm_struct *mm,
+					unsigned long address)
+{
+	struct page *page;
+	void *pg;
+
+	pg = quicklist_alloc(0, GFP_KERNEL, NULL);
+	if (!pg)
+		return NULL;
+	page = virt_to_page(pg);
+	pgtable_page_ctor(page);
+	return page;
+}
+
+static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
+{
+	quicklist_free(0, NULL, pte);
+}
+
+static inline void pte_free(struct mm_struct *mm, pgtable_t ptepage)
+{
+	pgtable_page_dtor(ptepage);
+	quicklist_free_page(0, NULL, ptepage);
+}
+
+
+#define pmd_populate_kernel(MM, PMD, PTE)	pmd_set(PMD, PTE)
+#define pmd_populate(MM,PMD,PTE_PAGE)		\
+	pmd_populate_kernel(MM,PMD,page_address(PTE_PAGE))
+#define pmd_pgtable(pmd) pmd_page(pmd)
+
+static inline void check_pgt_cache(void)
+{
+	quicklist_trim(0, NULL, 25, 16);
+}
+
+#endif /* _SPARC64_PGALLOC_H */
diff --git a/include/asm-sparc/pgtable.h b/include/asm-sparc/pgtable.h
index 60512296b2c..63cdef53bc5 100644
--- a/include/asm-sparc/pgtable.h
+++ b/include/asm-sparc/pgtable.h
@@ -1,475 +1,8 @@
-#ifndef _SPARC_PGTABLE_H
-#define _SPARC_PGTABLE_H
-
-/*  asm-sparc/pgtable.h:  Defines and functions used to work
- *                        with Sparc page tables.
- *
- *  Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- *  Copyright (C) 1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-
-#include <asm-generic/4level-fixup.h>
-
-#include <linux/spinlock.h>
-#include <linux/swap.h>
-#include <asm/types.h>
-#ifdef CONFIG_SUN4
-#include <asm/pgtsun4.h>
+#ifndef ___ASM_SPARC_PGTABLE_H
+#define ___ASM_SPARC_PGTABLE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/pgtable_64.h>
 #else
-#include <asm/pgtsun4c.h>
+#include <asm-sparc/pgtable_32.h>
 #endif
-#include <asm/pgtsrmmu.h>
-#include <asm/vac-ops.h>
-#include <asm/oplib.h>
-#include <asm/btfixup.h>
-#include <asm/system.h>
-
-#ifndef __ASSEMBLY__
-
-struct vm_area_struct;
-struct page;
-
-extern void load_mmu(void);
-extern unsigned long calc_highpages(void);
-
-BTFIXUPDEF_SIMM13(pgdir_shift)
-BTFIXUPDEF_SETHI(pgdir_size)
-BTFIXUPDEF_SETHI(pgdir_mask)
-
-BTFIXUPDEF_SIMM13(ptrs_per_pmd)
-BTFIXUPDEF_SIMM13(ptrs_per_pgd)
-BTFIXUPDEF_SIMM13(user_ptrs_per_pgd)
-
-#define pte_ERROR(e)   __builtin_trap()
-#define pmd_ERROR(e)   __builtin_trap()
-#define pgd_ERROR(e)   __builtin_trap()
-
-BTFIXUPDEF_INT(page_none)
-BTFIXUPDEF_INT(page_copy)
-BTFIXUPDEF_INT(page_readonly)
-BTFIXUPDEF_INT(page_kernel)
-
-#define PMD_SHIFT		SUN4C_PMD_SHIFT
-#define PMD_SIZE        	(1UL << PMD_SHIFT)
-#define PMD_MASK        	(~(PMD_SIZE-1))
-#define PMD_ALIGN(__addr) 	(((__addr) + ~PMD_MASK) & PMD_MASK)
-#define PGDIR_SHIFT     	BTFIXUP_SIMM13(pgdir_shift)
-#define PGDIR_SIZE      	BTFIXUP_SETHI(pgdir_size)
-#define PGDIR_MASK      	BTFIXUP_SETHI(pgdir_mask)
-#define PTRS_PER_PTE    	1024
-#define PTRS_PER_PMD    	BTFIXUP_SIMM13(ptrs_per_pmd)
-#define PTRS_PER_PGD    	BTFIXUP_SIMM13(ptrs_per_pgd)
-#define USER_PTRS_PER_PGD	BTFIXUP_SIMM13(user_ptrs_per_pgd)
-#define FIRST_USER_ADDRESS	0
-#define PTE_SIZE		(PTRS_PER_PTE*4)
-
-#define PAGE_NONE      __pgprot(BTFIXUP_INT(page_none))
-extern pgprot_t PAGE_SHARED;
-#define PAGE_COPY      __pgprot(BTFIXUP_INT(page_copy))
-#define PAGE_READONLY  __pgprot(BTFIXUP_INT(page_readonly))
-
-extern unsigned long page_kernel;
-
-#ifdef MODULE
-#define PAGE_KERNEL	page_kernel
-#else
-#define PAGE_KERNEL    __pgprot(BTFIXUP_INT(page_kernel))
 #endif
-
-/* Top-level page directory */
-extern pgd_t swapper_pg_dir[1024];
-
-extern void paging_init(void);
-
-/* Page table for 0-4MB for everybody, on the Sparc this
- * holds the same as on the i386.
- */
-extern pte_t pg0[1024];
-extern pte_t pg1[1024];
-extern pte_t pg2[1024];
-extern pte_t pg3[1024];
-
-extern unsigned long ptr_in_current_pgd;
-
-/* Here is a trick, since mmap.c need the initializer elements for
- * protection_map[] to be constant at compile time, I set the following
- * to all zeros.  I set it to the real values after I link in the
- * appropriate MMU page table routines at boot time.
- */
-#define __P000  __pgprot(0)
-#define __P001  __pgprot(0)
-#define __P010  __pgprot(0)
-#define __P011  __pgprot(0)
-#define __P100  __pgprot(0)
-#define __P101  __pgprot(0)
-#define __P110  __pgprot(0)
-#define __P111  __pgprot(0)
-
-#define __S000	__pgprot(0)
-#define __S001	__pgprot(0)
-#define __S010	__pgprot(0)
-#define __S011	__pgprot(0)
-#define __S100	__pgprot(0)
-#define __S101	__pgprot(0)
-#define __S110	__pgprot(0)
-#define __S111	__pgprot(0)
-
-extern int num_contexts;
-
-/* First physical page can be anywhere, the following is needed so that
- * va-->pa and vice versa conversions work properly without performance
- * hit for all __pa()/__va() operations.
- */
-extern unsigned long phys_base;
-extern unsigned long pfn_base;
-
-/*
- * BAD_PAGETABLE is used when we need a bogus page-table, while
- * BAD_PAGE is used for a bogus page.
- *
- * ZERO_PAGE is a global shared page that is always zero: used
- * for zero-mapped memory areas etc..
- */
-extern pte_t * __bad_pagetable(void);
-extern pte_t __bad_page(void);
-extern unsigned long empty_zero_page;
-
-#define BAD_PAGETABLE __bad_pagetable()
-#define BAD_PAGE __bad_page()
-#define ZERO_PAGE(vaddr) (virt_to_page(&empty_zero_page))
-
-/*
- */
-BTFIXUPDEF_CALL_CONST(struct page *, pmd_page, pmd_t)
-BTFIXUPDEF_CALL_CONST(unsigned long, pgd_page_vaddr, pgd_t)
-
-#define pmd_page(pmd) BTFIXUP_CALL(pmd_page)(pmd)
-#define pgd_page_vaddr(pgd) BTFIXUP_CALL(pgd_page_vaddr)(pgd)
-
-BTFIXUPDEF_SETHI(none_mask)
-BTFIXUPDEF_CALL_CONST(int, pte_present, pte_t)
-BTFIXUPDEF_CALL(void, pte_clear, pte_t *)
-
-static inline int pte_none(pte_t pte)
-{
-	return !(pte_val(pte) & ~BTFIXUP_SETHI(none_mask));
-}
-
-#define pte_present(pte) BTFIXUP_CALL(pte_present)(pte)
-#define pte_clear(mm,addr,pte) BTFIXUP_CALL(pte_clear)(pte)
-
-BTFIXUPDEF_CALL_CONST(int, pmd_bad, pmd_t)
-BTFIXUPDEF_CALL_CONST(int, pmd_present, pmd_t)
-BTFIXUPDEF_CALL(void, pmd_clear, pmd_t *)
-
-static inline int pmd_none(pmd_t pmd)
-{
-	return !(pmd_val(pmd) & ~BTFIXUP_SETHI(none_mask));
-}
-
-#define pmd_bad(pmd) BTFIXUP_CALL(pmd_bad)(pmd)
-#define pmd_present(pmd) BTFIXUP_CALL(pmd_present)(pmd)
-#define pmd_clear(pmd) BTFIXUP_CALL(pmd_clear)(pmd)
-
-BTFIXUPDEF_CALL_CONST(int, pgd_none, pgd_t)
-BTFIXUPDEF_CALL_CONST(int, pgd_bad, pgd_t)
-BTFIXUPDEF_CALL_CONST(int, pgd_present, pgd_t)
-BTFIXUPDEF_CALL(void, pgd_clear, pgd_t *)
-
-#define pgd_none(pgd) BTFIXUP_CALL(pgd_none)(pgd)
-#define pgd_bad(pgd) BTFIXUP_CALL(pgd_bad)(pgd)
-#define pgd_present(pgd) BTFIXUP_CALL(pgd_present)(pgd)
-#define pgd_clear(pgd) BTFIXUP_CALL(pgd_clear)(pgd)
-
-/*
- * The following only work if pte_present() is true.
- * Undefined behaviour if not..
- */
-BTFIXUPDEF_HALF(pte_writei)
-BTFIXUPDEF_HALF(pte_dirtyi)
-BTFIXUPDEF_HALF(pte_youngi)
-
-static int pte_write(pte_t pte) __attribute_const__;
-static inline int pte_write(pte_t pte)
-{
-	return pte_val(pte) & BTFIXUP_HALF(pte_writei);
-}
-
-static int pte_dirty(pte_t pte) __attribute_const__;
-static inline int pte_dirty(pte_t pte)
-{
-	return pte_val(pte) & BTFIXUP_HALF(pte_dirtyi);
-}
-
-static int pte_young(pte_t pte) __attribute_const__;
-static inline int pte_young(pte_t pte)
-{
-	return pte_val(pte) & BTFIXUP_HALF(pte_youngi);
-}
-
-/*
- * The following only work if pte_present() is not true.
- */
-BTFIXUPDEF_HALF(pte_filei)
-
-static int pte_file(pte_t pte) __attribute_const__;
-static inline int pte_file(pte_t pte)
-{
-	return pte_val(pte) & BTFIXUP_HALF(pte_filei);
-}
-
-static inline int pte_special(pte_t pte)
-{
-	return 0;
-}
-
-/*
- */
-BTFIXUPDEF_HALF(pte_wrprotecti)
-BTFIXUPDEF_HALF(pte_mkcleani)
-BTFIXUPDEF_HALF(pte_mkoldi)
-
-static pte_t pte_wrprotect(pte_t pte) __attribute_const__;
-static inline pte_t pte_wrprotect(pte_t pte)
-{
-	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_wrprotecti));
-}
-
-static pte_t pte_mkclean(pte_t pte) __attribute_const__;
-static inline pte_t pte_mkclean(pte_t pte)
-{
-	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_mkcleani));
-}
-
-static pte_t pte_mkold(pte_t pte) __attribute_const__;
-static inline pte_t pte_mkold(pte_t pte)
-{
-	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_mkoldi));
-}
-
-BTFIXUPDEF_CALL_CONST(pte_t, pte_mkwrite, pte_t)
-BTFIXUPDEF_CALL_CONST(pte_t, pte_mkdirty, pte_t)
-BTFIXUPDEF_CALL_CONST(pte_t, pte_mkyoung, pte_t)
-
-#define pte_mkwrite(pte) BTFIXUP_CALL(pte_mkwrite)(pte)
-#define pte_mkdirty(pte) BTFIXUP_CALL(pte_mkdirty)(pte)
-#define pte_mkyoung(pte) BTFIXUP_CALL(pte_mkyoung)(pte)
-
-#define pte_mkspecial(pte)    (pte)
-
-#define pfn_pte(pfn, prot)		mk_pte(pfn_to_page(pfn), prot)
-
-BTFIXUPDEF_CALL(unsigned long,	 pte_pfn, pte_t)
-#define pte_pfn(pte) BTFIXUP_CALL(pte_pfn)(pte)
-#define pte_page(pte)	pfn_to_page(pte_pfn(pte))
-
-/*
- * Conversion functions: convert a page and protection to a page entry,
- * and a page entry and page directory to the page they refer to.
- */
-BTFIXUPDEF_CALL_CONST(pte_t, mk_pte, struct page *, pgprot_t)
-
-BTFIXUPDEF_CALL_CONST(pte_t, mk_pte_phys, unsigned long, pgprot_t)
-BTFIXUPDEF_CALL_CONST(pte_t, mk_pte_io, unsigned long, pgprot_t, int)
-BTFIXUPDEF_CALL_CONST(pgprot_t, pgprot_noncached, pgprot_t)
-
-#define mk_pte(page,pgprot) BTFIXUP_CALL(mk_pte)(page,pgprot)
-#define mk_pte_phys(page,pgprot) BTFIXUP_CALL(mk_pte_phys)(page,pgprot)
-#define mk_pte_io(page,pgprot,space) BTFIXUP_CALL(mk_pte_io)(page,pgprot,space)
-
-#define pgprot_noncached(pgprot) BTFIXUP_CALL(pgprot_noncached)(pgprot)
-
-BTFIXUPDEF_INT(pte_modify_mask)
-
-static pte_t pte_modify(pte_t pte, pgprot_t newprot) __attribute_const__;
-static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
-{
-	return __pte((pte_val(pte) & BTFIXUP_INT(pte_modify_mask)) |
-		pgprot_val(newprot));
-}
-
-#define pgd_index(address) ((address) >> PGDIR_SHIFT)
-
-/* to find an entry in a page-table-directory */
-#define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address))
-
-/* to find an entry in a kernel page-table-directory */
-#define pgd_offset_k(address) pgd_offset(&init_mm, address)
-
-/* Find an entry in the second-level page table.. */
-BTFIXUPDEF_CALL(pmd_t *, pmd_offset, pgd_t *, unsigned long)
-#define pmd_offset(dir,addr) BTFIXUP_CALL(pmd_offset)(dir,addr)
-
-/* Find an entry in the third-level page table.. */ 
-BTFIXUPDEF_CALL(pte_t *, pte_offset_kernel, pmd_t *, unsigned long)
-#define pte_offset_kernel(dir,addr) BTFIXUP_CALL(pte_offset_kernel)(dir,addr)
-
-/*
- * This shortcut works on sun4m (and sun4d) because the nocache area is static,
- * and sun4c is guaranteed to have no highmem anyway.
- */
-#define pte_offset_map(d, a)		pte_offset_kernel(d,a)
-#define pte_offset_map_nested(d, a)	pte_offset_kernel(d,a)
-
-#define pte_unmap(pte)		do{}while(0)
-#define pte_unmap_nested(pte)	do{}while(0)
-
-/* Certain architectures need to do special things when pte's
- * within a page table are directly modified.  Thus, the following
- * hook is made available.
- */
-
-BTFIXUPDEF_CALL(void, set_pte, pte_t *, pte_t)
-
-#define set_pte(ptep,pteval) BTFIXUP_CALL(set_pte)(ptep,pteval)
-#define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval)
-
-struct seq_file;
-BTFIXUPDEF_CALL(void, mmu_info, struct seq_file *)
-
-#define mmu_info(p) BTFIXUP_CALL(mmu_info)(p)
-
-/* Fault handler stuff... */
-#define FAULT_CODE_PROT     0x1
-#define FAULT_CODE_WRITE    0x2
-#define FAULT_CODE_USER     0x4
-
-BTFIXUPDEF_CALL(void, update_mmu_cache, struct vm_area_struct *, unsigned long, pte_t)
-
-#define update_mmu_cache(vma,addr,pte) BTFIXUP_CALL(update_mmu_cache)(vma,addr,pte)
-
-BTFIXUPDEF_CALL(void, sparc_mapiorange, unsigned int, unsigned long,
-    unsigned long, unsigned int)
-BTFIXUPDEF_CALL(void, sparc_unmapiorange, unsigned long, unsigned int)
-#define sparc_mapiorange(bus,pa,va,len) BTFIXUP_CALL(sparc_mapiorange)(bus,pa,va,len)
-#define sparc_unmapiorange(va,len) BTFIXUP_CALL(sparc_unmapiorange)(va,len)
-
-extern int invalid_segment;
-
-/* Encode and de-code a swap entry */
-BTFIXUPDEF_CALL(unsigned long, __swp_type, swp_entry_t)
-BTFIXUPDEF_CALL(unsigned long, __swp_offset, swp_entry_t)
-BTFIXUPDEF_CALL(swp_entry_t, __swp_entry, unsigned long, unsigned long)
-
-#define __swp_type(__x)			BTFIXUP_CALL(__swp_type)(__x)
-#define __swp_offset(__x)		BTFIXUP_CALL(__swp_offset)(__x)
-#define __swp_entry(__type,__off)	BTFIXUP_CALL(__swp_entry)(__type,__off)
-
-#define __pte_to_swp_entry(pte)		((swp_entry_t) { pte_val(pte) })
-#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
-
-/* file-offset-in-pte helpers */
-BTFIXUPDEF_CALL(unsigned long, pte_to_pgoff, pte_t pte);
-BTFIXUPDEF_CALL(pte_t, pgoff_to_pte, unsigned long pgoff);
-
-#define pte_to_pgoff(pte) BTFIXUP_CALL(pte_to_pgoff)(pte)
-#define pgoff_to_pte(off) BTFIXUP_CALL(pgoff_to_pte)(off)
-
-/*
- * This is made a constant because mm/fremap.c required a constant.
- * Note that layout of these bits is different between sun4c.c and srmmu.c.
- */
-#define PTE_FILE_MAX_BITS 24
-
-/*
- */
-struct ctx_list {
-	struct ctx_list *next;
-	struct ctx_list *prev;
-	unsigned int ctx_number;
-	struct mm_struct *ctx_mm;
-};
-
-extern struct ctx_list *ctx_list_pool;  /* Dynamically allocated */
-extern struct ctx_list ctx_free;        /* Head of free list */
-extern struct ctx_list ctx_used;        /* Head of used contexts list */
-
-#define NO_CONTEXT     -1
-
-static inline void remove_from_ctx_list(struct ctx_list *entry)
-{
-	entry->next->prev = entry->prev;
-	entry->prev->next = entry->next;
-}
-
-static inline void add_to_ctx_list(struct ctx_list *head, struct ctx_list *entry)
-{
-	entry->next = head;
-	(entry->prev = head->prev)->next = entry;
-	head->prev = entry;
-}
-#define add_to_free_ctxlist(entry) add_to_ctx_list(&ctx_free, entry)
-#define add_to_used_ctxlist(entry) add_to_ctx_list(&ctx_used, entry)
-
-static inline unsigned long
-__get_phys (unsigned long addr)
-{
-	switch (sparc_cpu_model){
-	case sun4:
-	case sun4c:
-		return sun4c_get_pte (addr) << PAGE_SHIFT;
-	case sun4m:
-	case sun4d:
-		return ((srmmu_get_pte (addr) & 0xffffff00) << 4);
-	default:
-		return 0;
-	}
-}
-
-static inline int
-__get_iospace (unsigned long addr)
-{
-	switch (sparc_cpu_model){
-	case sun4:
-	case sun4c:
-		return -1; /* Don't check iospace on sun4c */
-	case sun4m:
-	case sun4d:
-		return (srmmu_get_pte (addr) >> 28);
-	default:
-		return -1;
-	}
-}
-
-extern unsigned long *sparc_valid_addr_bitmap;
-
-/* Needs to be defined here and not in linux/mm.h, as it is arch dependent */
-#define kern_addr_valid(addr) \
-	(test_bit(__pa((unsigned long)(addr))>>20, sparc_valid_addr_bitmap))
-
-extern int io_remap_pfn_range(struct vm_area_struct *vma,
-			      unsigned long from, unsigned long pfn,
-			      unsigned long size, pgprot_t prot);
-
-/*
- * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
- * its high 4 bits.  These macros/functions put it there or get it from there.
- */
-#define MK_IOSPACE_PFN(space, pfn)	(pfn | (space << (BITS_PER_LONG - 4)))
-#define GET_IOSPACE(pfn)		(pfn >> (BITS_PER_LONG - 4))
-#define GET_PFN(pfn)			(pfn & 0x0fffffffUL)
-
-#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
-#define ptep_set_access_flags(__vma, __address, __ptep, __entry, __dirty) \
-({									  \
-	int __changed = !pte_same(*(__ptep), __entry);			  \
-	if (__changed) {						  \
-		set_pte_at((__vma)->vm_mm, (__address), __ptep, __entry); \
-		flush_tlb_page(__vma, __address);			  \
-	}								  \
-	(sparc_cpu_model == sun4c) || __changed;			  \
-})
-
-#include <asm-generic/pgtable.h>
-
-#endif /* !(__ASSEMBLY__) */
-
-/* We provide our own get_unmapped_area to cope with VA holes for userland */
-#define HAVE_ARCH_UNMAPPED_AREA
-
-/*
- * No page table caches to initialise
- */
-#define pgtable_cache_init()	do { } while (0)
-
-#endif /* !(_SPARC_PGTABLE_H) */
diff --git a/include/asm-sparc/pgtable_32.h b/include/asm-sparc/pgtable_32.h
new file mode 100644
index 00000000000..781bd4694a1
--- /dev/null
+++ b/include/asm-sparc/pgtable_32.h
@@ -0,0 +1,480 @@
+#ifndef _SPARC_PGTABLE_H
+#define _SPARC_PGTABLE_H
+
+/*  asm-sparc/pgtable.h:  Defines and functions used to work
+ *                        with Sparc page tables.
+ *
+ *  Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ *  Copyright (C) 1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+
+#ifndef __ASSEMBLY__
+#include <asm-generic/4level-fixup.h>
+
+#include <linux/spinlock.h>
+#include <linux/swap.h>
+#include <asm/types.h>
+#ifdef CONFIG_SUN4
+#include <asm/pgtsun4.h>
+#else
+#include <asm/pgtsun4c.h>
+#endif
+#include <asm/pgtsrmmu.h>
+#include <asm/vac-ops.h>
+#include <asm/oplib.h>
+#include <asm/btfixup.h>
+#include <asm/system.h>
+
+
+struct vm_area_struct;
+struct page;
+
+extern void load_mmu(void);
+extern unsigned long calc_highpages(void);
+
+BTFIXUPDEF_SIMM13(pgdir_shift)
+BTFIXUPDEF_SETHI(pgdir_size)
+BTFIXUPDEF_SETHI(pgdir_mask)
+
+BTFIXUPDEF_SIMM13(ptrs_per_pmd)
+BTFIXUPDEF_SIMM13(ptrs_per_pgd)
+BTFIXUPDEF_SIMM13(user_ptrs_per_pgd)
+
+#define pte_ERROR(e)   __builtin_trap()
+#define pmd_ERROR(e)   __builtin_trap()
+#define pgd_ERROR(e)   __builtin_trap()
+
+BTFIXUPDEF_INT(page_none)
+BTFIXUPDEF_INT(page_copy)
+BTFIXUPDEF_INT(page_readonly)
+BTFIXUPDEF_INT(page_kernel)
+
+#define PMD_SHIFT		SUN4C_PMD_SHIFT
+#define PMD_SIZE        	(1UL << PMD_SHIFT)
+#define PMD_MASK        	(~(PMD_SIZE-1))
+#define PMD_ALIGN(__addr) 	(((__addr) + ~PMD_MASK) & PMD_MASK)
+#define PGDIR_SHIFT     	BTFIXUP_SIMM13(pgdir_shift)
+#define PGDIR_SIZE      	BTFIXUP_SETHI(pgdir_size)
+#define PGDIR_MASK      	BTFIXUP_SETHI(pgdir_mask)
+#define PTRS_PER_PTE    	1024
+#define PTRS_PER_PMD    	BTFIXUP_SIMM13(ptrs_per_pmd)
+#define PTRS_PER_PGD    	BTFIXUP_SIMM13(ptrs_per_pgd)
+#define USER_PTRS_PER_PGD	BTFIXUP_SIMM13(user_ptrs_per_pgd)
+#define FIRST_USER_ADDRESS	0
+#define PTE_SIZE		(PTRS_PER_PTE*4)
+
+#define PAGE_NONE      __pgprot(BTFIXUP_INT(page_none))
+extern pgprot_t PAGE_SHARED;
+#define PAGE_COPY      __pgprot(BTFIXUP_INT(page_copy))
+#define PAGE_READONLY  __pgprot(BTFIXUP_INT(page_readonly))
+
+extern unsigned long page_kernel;
+
+#ifdef MODULE
+#define PAGE_KERNEL	page_kernel
+#else
+#define PAGE_KERNEL    __pgprot(BTFIXUP_INT(page_kernel))
+#endif
+
+/* Top-level page directory */
+extern pgd_t swapper_pg_dir[1024];
+
+extern void paging_init(void);
+
+/* Page table for 0-4MB for everybody, on the Sparc this
+ * holds the same as on the i386.
+ */
+extern pte_t pg0[1024];
+extern pte_t pg1[1024];
+extern pte_t pg2[1024];
+extern pte_t pg3[1024];
+
+extern unsigned long ptr_in_current_pgd;
+
+/* Here is a trick, since mmap.c need the initializer elements for
+ * protection_map[] to be constant at compile time, I set the following
+ * to all zeros.  I set it to the real values after I link in the
+ * appropriate MMU page table routines at boot time.
+ */
+#define __P000  __pgprot(0)
+#define __P001  __pgprot(0)
+#define __P010  __pgprot(0)
+#define __P011  __pgprot(0)
+#define __P100  __pgprot(0)
+#define __P101  __pgprot(0)
+#define __P110  __pgprot(0)
+#define __P111  __pgprot(0)
+
+#define __S000	__pgprot(0)
+#define __S001	__pgprot(0)
+#define __S010	__pgprot(0)
+#define __S011	__pgprot(0)
+#define __S100	__pgprot(0)
+#define __S101	__pgprot(0)
+#define __S110	__pgprot(0)
+#define __S111	__pgprot(0)
+
+extern int num_contexts;
+
+/* First physical page can be anywhere, the following is needed so that
+ * va-->pa and vice versa conversions work properly without performance
+ * hit for all __pa()/__va() operations.
+ */
+extern unsigned long phys_base;
+extern unsigned long pfn_base;
+
+/*
+ * BAD_PAGETABLE is used when we need a bogus page-table, while
+ * BAD_PAGE is used for a bogus page.
+ *
+ * ZERO_PAGE is a global shared page that is always zero: used
+ * for zero-mapped memory areas etc..
+ */
+extern pte_t * __bad_pagetable(void);
+extern pte_t __bad_page(void);
+extern unsigned long empty_zero_page;
+
+#define BAD_PAGETABLE __bad_pagetable()
+#define BAD_PAGE __bad_page()
+#define ZERO_PAGE(vaddr) (virt_to_page(&empty_zero_page))
+
+/*
+ */
+BTFIXUPDEF_CALL_CONST(struct page *, pmd_page, pmd_t)
+BTFIXUPDEF_CALL_CONST(unsigned long, pgd_page_vaddr, pgd_t)
+
+#define pmd_page(pmd) BTFIXUP_CALL(pmd_page)(pmd)
+#define pgd_page_vaddr(pgd) BTFIXUP_CALL(pgd_page_vaddr)(pgd)
+
+BTFIXUPDEF_SETHI(none_mask)
+BTFIXUPDEF_CALL_CONST(int, pte_present, pte_t)
+BTFIXUPDEF_CALL(void, pte_clear, pte_t *)
+
+static inline int pte_none(pte_t pte)
+{
+	return !(pte_val(pte) & ~BTFIXUP_SETHI(none_mask));
+}
+
+#define pte_present(pte) BTFIXUP_CALL(pte_present)(pte)
+#define pte_clear(mm,addr,pte) BTFIXUP_CALL(pte_clear)(pte)
+
+BTFIXUPDEF_CALL_CONST(int, pmd_bad, pmd_t)
+BTFIXUPDEF_CALL_CONST(int, pmd_present, pmd_t)
+BTFIXUPDEF_CALL(void, pmd_clear, pmd_t *)
+
+static inline int pmd_none(pmd_t pmd)
+{
+	return !(pmd_val(pmd) & ~BTFIXUP_SETHI(none_mask));
+}
+
+#define pmd_bad(pmd) BTFIXUP_CALL(pmd_bad)(pmd)
+#define pmd_present(pmd) BTFIXUP_CALL(pmd_present)(pmd)
+#define pmd_clear(pmd) BTFIXUP_CALL(pmd_clear)(pmd)
+
+BTFIXUPDEF_CALL_CONST(int, pgd_none, pgd_t)
+BTFIXUPDEF_CALL_CONST(int, pgd_bad, pgd_t)
+BTFIXUPDEF_CALL_CONST(int, pgd_present, pgd_t)
+BTFIXUPDEF_CALL(void, pgd_clear, pgd_t *)
+
+#define pgd_none(pgd) BTFIXUP_CALL(pgd_none)(pgd)
+#define pgd_bad(pgd) BTFIXUP_CALL(pgd_bad)(pgd)
+#define pgd_present(pgd) BTFIXUP_CALL(pgd_present)(pgd)
+#define pgd_clear(pgd) BTFIXUP_CALL(pgd_clear)(pgd)
+
+/*
+ * The following only work if pte_present() is true.
+ * Undefined behaviour if not..
+ */
+BTFIXUPDEF_HALF(pte_writei)
+BTFIXUPDEF_HALF(pte_dirtyi)
+BTFIXUPDEF_HALF(pte_youngi)
+
+static int pte_write(pte_t pte) __attribute_const__;
+static inline int pte_write(pte_t pte)
+{
+	return pte_val(pte) & BTFIXUP_HALF(pte_writei);
+}
+
+static int pte_dirty(pte_t pte) __attribute_const__;
+static inline int pte_dirty(pte_t pte)
+{
+	return pte_val(pte) & BTFIXUP_HALF(pte_dirtyi);
+}
+
+static int pte_young(pte_t pte) __attribute_const__;
+static inline int pte_young(pte_t pte)
+{
+	return pte_val(pte) & BTFIXUP_HALF(pte_youngi);
+}
+
+/*
+ * The following only work if pte_present() is not true.
+ */
+BTFIXUPDEF_HALF(pte_filei)
+
+static int pte_file(pte_t pte) __attribute_const__;
+static inline int pte_file(pte_t pte)
+{
+	return pte_val(pte) & BTFIXUP_HALF(pte_filei);
+}
+
+static inline int pte_special(pte_t pte)
+{
+	return 0;
+}
+
+/*
+ */
+BTFIXUPDEF_HALF(pte_wrprotecti)
+BTFIXUPDEF_HALF(pte_mkcleani)
+BTFIXUPDEF_HALF(pte_mkoldi)
+
+static pte_t pte_wrprotect(pte_t pte) __attribute_const__;
+static inline pte_t pte_wrprotect(pte_t pte)
+{
+	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_wrprotecti));
+}
+
+static pte_t pte_mkclean(pte_t pte) __attribute_const__;
+static inline pte_t pte_mkclean(pte_t pte)
+{
+	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_mkcleani));
+}
+
+static pte_t pte_mkold(pte_t pte) __attribute_const__;
+static inline pte_t pte_mkold(pte_t pte)
+{
+	return __pte(pte_val(pte) & ~BTFIXUP_HALF(pte_mkoldi));
+}
+
+BTFIXUPDEF_CALL_CONST(pte_t, pte_mkwrite, pte_t)
+BTFIXUPDEF_CALL_CONST(pte_t, pte_mkdirty, pte_t)
+BTFIXUPDEF_CALL_CONST(pte_t, pte_mkyoung, pte_t)
+
+#define pte_mkwrite(pte) BTFIXUP_CALL(pte_mkwrite)(pte)
+#define pte_mkdirty(pte) BTFIXUP_CALL(pte_mkdirty)(pte)
+#define pte_mkyoung(pte) BTFIXUP_CALL(pte_mkyoung)(pte)
+
+#define pte_mkspecial(pte)    (pte)
+
+#define pfn_pte(pfn, prot)		mk_pte(pfn_to_page(pfn), prot)
+
+BTFIXUPDEF_CALL(unsigned long,	 pte_pfn, pte_t)
+#define pte_pfn(pte) BTFIXUP_CALL(pte_pfn)(pte)
+#define pte_page(pte)	pfn_to_page(pte_pfn(pte))
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+BTFIXUPDEF_CALL_CONST(pte_t, mk_pte, struct page *, pgprot_t)
+
+BTFIXUPDEF_CALL_CONST(pte_t, mk_pte_phys, unsigned long, pgprot_t)
+BTFIXUPDEF_CALL_CONST(pte_t, mk_pte_io, unsigned long, pgprot_t, int)
+BTFIXUPDEF_CALL_CONST(pgprot_t, pgprot_noncached, pgprot_t)
+
+#define mk_pte(page,pgprot) BTFIXUP_CALL(mk_pte)(page,pgprot)
+#define mk_pte_phys(page,pgprot) BTFIXUP_CALL(mk_pte_phys)(page,pgprot)
+#define mk_pte_io(page,pgprot,space) BTFIXUP_CALL(mk_pte_io)(page,pgprot,space)
+
+#define pgprot_noncached(pgprot) BTFIXUP_CALL(pgprot_noncached)(pgprot)
+
+BTFIXUPDEF_INT(pte_modify_mask)
+
+static pte_t pte_modify(pte_t pte, pgprot_t newprot) __attribute_const__;
+static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
+{
+	return __pte((pte_val(pte) & BTFIXUP_INT(pte_modify_mask)) |
+		pgprot_val(newprot));
+}
+
+#define pgd_index(address) ((address) >> PGDIR_SHIFT)
+
+/* to find an entry in a page-table-directory */
+#define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address))
+
+/* to find an entry in a kernel page-table-directory */
+#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+
+/* Find an entry in the second-level page table.. */
+BTFIXUPDEF_CALL(pmd_t *, pmd_offset, pgd_t *, unsigned long)
+#define pmd_offset(dir,addr) BTFIXUP_CALL(pmd_offset)(dir,addr)
+
+/* Find an entry in the third-level page table.. */
+BTFIXUPDEF_CALL(pte_t *, pte_offset_kernel, pmd_t *, unsigned long)
+#define pte_offset_kernel(dir,addr) BTFIXUP_CALL(pte_offset_kernel)(dir,addr)
+
+/*
+ * This shortcut works on sun4m (and sun4d) because the nocache area is static,
+ * and sun4c is guaranteed to have no highmem anyway.
+ */
+#define pte_offset_map(d, a)		pte_offset_kernel(d,a)
+#define pte_offset_map_nested(d, a)	pte_offset_kernel(d,a)
+
+#define pte_unmap(pte)		do{}while(0)
+#define pte_unmap_nested(pte)	do{}while(0)
+
+/* Certain architectures need to do special things when pte's
+ * within a page table are directly modified.  Thus, the following
+ * hook is made available.
+ */
+
+BTFIXUPDEF_CALL(void, set_pte, pte_t *, pte_t)
+
+#define set_pte(ptep,pteval) BTFIXUP_CALL(set_pte)(ptep,pteval)
+#define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval)
+
+struct seq_file;
+BTFIXUPDEF_CALL(void, mmu_info, struct seq_file *)
+
+#define mmu_info(p) BTFIXUP_CALL(mmu_info)(p)
+
+/* Fault handler stuff... */
+#define FAULT_CODE_PROT     0x1
+#define FAULT_CODE_WRITE    0x2
+#define FAULT_CODE_USER     0x4
+
+BTFIXUPDEF_CALL(void, update_mmu_cache, struct vm_area_struct *, unsigned long, pte_t)
+
+#define update_mmu_cache(vma,addr,pte) BTFIXUP_CALL(update_mmu_cache)(vma,addr,pte)
+
+BTFIXUPDEF_CALL(void, sparc_mapiorange, unsigned int, unsigned long,
+    unsigned long, unsigned int)
+BTFIXUPDEF_CALL(void, sparc_unmapiorange, unsigned long, unsigned int)
+#define sparc_mapiorange(bus,pa,va,len) BTFIXUP_CALL(sparc_mapiorange)(bus,pa,va,len)
+#define sparc_unmapiorange(va,len) BTFIXUP_CALL(sparc_unmapiorange)(va,len)
+
+extern int invalid_segment;
+
+/* Encode and de-code a swap entry */
+BTFIXUPDEF_CALL(unsigned long, __swp_type, swp_entry_t)
+BTFIXUPDEF_CALL(unsigned long, __swp_offset, swp_entry_t)
+BTFIXUPDEF_CALL(swp_entry_t, __swp_entry, unsigned long, unsigned long)
+
+#define __swp_type(__x)			BTFIXUP_CALL(__swp_type)(__x)
+#define __swp_offset(__x)		BTFIXUP_CALL(__swp_offset)(__x)
+#define __swp_entry(__type,__off)	BTFIXUP_CALL(__swp_entry)(__type,__off)
+
+#define __pte_to_swp_entry(pte)		((swp_entry_t) { pte_val(pte) })
+#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
+
+/* file-offset-in-pte helpers */
+BTFIXUPDEF_CALL(unsigned long, pte_to_pgoff, pte_t pte);
+BTFIXUPDEF_CALL(pte_t, pgoff_to_pte, unsigned long pgoff);
+
+#define pte_to_pgoff(pte) BTFIXUP_CALL(pte_to_pgoff)(pte)
+#define pgoff_to_pte(off) BTFIXUP_CALL(pgoff_to_pte)(off)
+
+/*
+ * This is made a constant because mm/fremap.c required a constant.
+ * Note that layout of these bits is different between sun4c.c and srmmu.c.
+ */
+#define PTE_FILE_MAX_BITS 24
+
+/*
+ */
+struct ctx_list {
+	struct ctx_list *next;
+	struct ctx_list *prev;
+	unsigned int ctx_number;
+	struct mm_struct *ctx_mm;
+};
+
+extern struct ctx_list *ctx_list_pool;  /* Dynamically allocated */
+extern struct ctx_list ctx_free;        /* Head of free list */
+extern struct ctx_list ctx_used;        /* Head of used contexts list */
+
+#define NO_CONTEXT     -1
+
+static inline void remove_from_ctx_list(struct ctx_list *entry)
+{
+	entry->next->prev = entry->prev;
+	entry->prev->next = entry->next;
+}
+
+static inline void add_to_ctx_list(struct ctx_list *head, struct ctx_list *entry)
+{
+	entry->next = head;
+	(entry->prev = head->prev)->next = entry;
+	head->prev = entry;
+}
+#define add_to_free_ctxlist(entry) add_to_ctx_list(&ctx_free, entry)
+#define add_to_used_ctxlist(entry) add_to_ctx_list(&ctx_used, entry)
+
+static inline unsigned long
+__get_phys (unsigned long addr)
+{
+	switch (sparc_cpu_model){
+	case sun4:
+	case sun4c:
+		return sun4c_get_pte (addr) << PAGE_SHIFT;
+	case sun4m:
+	case sun4d:
+		return ((srmmu_get_pte (addr) & 0xffffff00) << 4);
+	default:
+		return 0;
+	}
+}
+
+static inline int
+__get_iospace (unsigned long addr)
+{
+	switch (sparc_cpu_model){
+	case sun4:
+	case sun4c:
+		return -1; /* Don't check iospace on sun4c */
+	case sun4m:
+	case sun4d:
+		return (srmmu_get_pte (addr) >> 28);
+	default:
+		return -1;
+	}
+}
+
+extern unsigned long *sparc_valid_addr_bitmap;
+
+/* Needs to be defined here and not in linux/mm.h, as it is arch dependent */
+#define kern_addr_valid(addr) \
+	(test_bit(__pa((unsigned long)(addr))>>20, sparc_valid_addr_bitmap))
+
+extern int io_remap_pfn_range(struct vm_area_struct *vma,
+			      unsigned long from, unsigned long pfn,
+			      unsigned long size, pgprot_t prot);
+
+/*
+ * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
+ * its high 4 bits.  These macros/functions put it there or get it from there.
+ */
+#define MK_IOSPACE_PFN(space, pfn)	(pfn | (space << (BITS_PER_LONG - 4)))
+#define GET_IOSPACE(pfn)		(pfn >> (BITS_PER_LONG - 4))
+#define GET_PFN(pfn)			(pfn & 0x0fffffffUL)
+
+#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
+#define ptep_set_access_flags(__vma, __address, __ptep, __entry, __dirty) \
+({									  \
+	int __changed = !pte_same(*(__ptep), __entry);			  \
+	if (__changed) {						  \
+		set_pte_at((__vma)->vm_mm, (__address), __ptep, __entry); \
+		flush_tlb_page(__vma, __address);			  \
+	}								  \
+	(sparc_cpu_model == sun4c) || __changed;			  \
+})
+
+#include <asm-generic/pgtable.h>
+
+#endif /* !(__ASSEMBLY__) */
+
+#define VMALLOC_START           0xfe600000
+/* XXX Alter this when I get around to fixing sun4c - Anton */
+#define VMALLOC_END             0xffc00000
+
+
+/* We provide our own get_unmapped_area to cope with VA holes for userland */
+#define HAVE_ARCH_UNMAPPED_AREA
+
+/*
+ * No page table caches to initialise
+ */
+#define pgtable_cache_init()	do { } while (0)
+
+#endif /* !(_SPARC_PGTABLE_H) */
diff --git a/include/asm-sparc/pgtable_64.h b/include/asm-sparc/pgtable_64.h
new file mode 100644
index 00000000000..bb9ec2cce35
--- /dev/null
+++ b/include/asm-sparc/pgtable_64.h
@@ -0,0 +1,775 @@
+/*
+ * pgtable.h: SpitFire page table operations.
+ *
+ * Copyright 1996,1997 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+
+#ifndef _SPARC64_PGTABLE_H
+#define _SPARC64_PGTABLE_H
+
+/* This file contains the functions and defines necessary to modify and use
+ * the SpitFire page tables.
+ */
+
+#include <asm-generic/pgtable-nopud.h>
+
+#include <linux/compiler.h>
+#include <linux/const.h>
+#include <asm/types.h>
+#include <asm/spitfire.h>
+#include <asm/asi.h>
+#include <asm/system.h>
+#include <asm/page.h>
+#include <asm/processor.h>
+
+/* The kernel image occupies 0x4000000 to 0x6000000 (4MB --> 96MB).
+ * The page copy blockops can use 0x6000000 to 0x8000000.
+ * The TSB is mapped in the 0x8000000 to 0xa000000 range.
+ * The PROM resides in an area spanning 0xf0000000 to 0x100000000.
+ * The vmalloc area spans 0x100000000 to 0x200000000.
+ * Since modules need to be in the lowest 32-bits of the address space,
+ * we place them right before the OBP area from 0x10000000 to 0xf0000000.
+ * There is a single static kernel PMD which maps from 0x0 to address
+ * 0x400000000.
+ */
+#define	TLBTEMP_BASE		_AC(0x0000000006000000,UL)
+#define	TSBMAP_BASE		_AC(0x0000000008000000,UL)
+#define MODULES_VADDR		_AC(0x0000000010000000,UL)
+#define MODULES_LEN		_AC(0x00000000e0000000,UL)
+#define MODULES_END		_AC(0x00000000f0000000,UL)
+#define LOW_OBP_ADDRESS		_AC(0x00000000f0000000,UL)
+#define HI_OBP_ADDRESS		_AC(0x0000000100000000,UL)
+#define VMALLOC_START		_AC(0x0000000100000000,UL)
+#define VMALLOC_END		_AC(0x0000000200000000,UL)
+#define VMEMMAP_BASE		_AC(0x0000000200000000,UL)
+
+#define vmemmap			((struct page *)VMEMMAP_BASE)
+
+/* XXX All of this needs to be rethought so we can take advantage
+ * XXX cheetah's full 64-bit virtual address space, ie. no more hole
+ * XXX in the middle like on spitfire. -DaveM
+ */
+/*
+ * Given a virtual address, the lowest PAGE_SHIFT bits determine offset
+ * into the page; the next higher PAGE_SHIFT-3 bits determine the pte#
+ * in the proper pagetable (the -3 is from the 8 byte ptes, and each page
+ * table is a single page long). The next higher PMD_BITS determine pmd#
+ * in the proper pmdtable (where we must have PMD_BITS <= (PAGE_SHIFT-2)
+ * since the pmd entries are 4 bytes, and each pmd page is a single page
+ * long). Finally, the higher few bits determine pgde#.
+ */
+
+/* PMD_SHIFT determines the size of the area a second-level page
+ * table can map
+ */
+#define PMD_SHIFT	(PAGE_SHIFT + (PAGE_SHIFT-3))
+#define PMD_SIZE	(_AC(1,UL) << PMD_SHIFT)
+#define PMD_MASK	(~(PMD_SIZE-1))
+#define PMD_BITS	(PAGE_SHIFT - 2)
+
+/* PGDIR_SHIFT determines what a third-level page table entry can map */
+#define PGDIR_SHIFT	(PAGE_SHIFT + (PAGE_SHIFT-3) + PMD_BITS)
+#define PGDIR_SIZE	(_AC(1,UL) << PGDIR_SHIFT)
+#define PGDIR_MASK	(~(PGDIR_SIZE-1))
+#define PGDIR_BITS	(PAGE_SHIFT - 2)
+
+#ifndef __ASSEMBLY__
+
+#include <linux/sched.h>
+
+/* Entries per page directory level. */
+#define PTRS_PER_PTE	(1UL << (PAGE_SHIFT-3))
+#define PTRS_PER_PMD	(1UL << PMD_BITS)
+#define PTRS_PER_PGD	(1UL << PGDIR_BITS)
+
+/* Kernel has a separate 44bit address space. */
+#define FIRST_USER_ADDRESS	0
+
+#define pte_ERROR(e)	__builtin_trap()
+#define pmd_ERROR(e)	__builtin_trap()
+#define pgd_ERROR(e)	__builtin_trap()
+
+#endif /* !(__ASSEMBLY__) */
+
+/* PTE bits which are the same in SUN4U and SUN4V format.  */
+#define _PAGE_VALID	  _AC(0x8000000000000000,UL) /* Valid TTE            */
+#define _PAGE_R	  	  _AC(0x8000000000000000,UL) /* Keep ref bit uptodate*/
+
+/* SUN4U pte bits... */
+#define _PAGE_SZ4MB_4U	  _AC(0x6000000000000000,UL) /* 4MB Page             */
+#define _PAGE_SZ512K_4U	  _AC(0x4000000000000000,UL) /* 512K Page            */
+#define _PAGE_SZ64K_4U	  _AC(0x2000000000000000,UL) /* 64K Page             */
+#define _PAGE_SZ8K_4U	  _AC(0x0000000000000000,UL) /* 8K Page              */
+#define _PAGE_NFO_4U	  _AC(0x1000000000000000,UL) /* No Fault Only        */
+#define _PAGE_IE_4U	  _AC(0x0800000000000000,UL) /* Invert Endianness    */
+#define _PAGE_SOFT2_4U	  _AC(0x07FC000000000000,UL) /* Software bits, set 2 */
+#define _PAGE_RES1_4U	  _AC(0x0002000000000000,UL) /* Reserved             */
+#define _PAGE_SZ32MB_4U	  _AC(0x0001000000000000,UL) /* (Panther) 32MB page  */
+#define _PAGE_SZ256MB_4U  _AC(0x2001000000000000,UL) /* (Panther) 256MB page */
+#define _PAGE_SZALL_4U	  _AC(0x6001000000000000,UL) /* All pgsz bits        */
+#define _PAGE_SN_4U	  _AC(0x0000800000000000,UL) /* (Cheetah) Snoop      */
+#define _PAGE_RES2_4U	  _AC(0x0000780000000000,UL) /* Reserved             */
+#define _PAGE_PADDR_4U	  _AC(0x000007FFFFFFE000,UL) /* (Cheetah) pa[42:13]  */
+#define _PAGE_SOFT_4U	  _AC(0x0000000000001F80,UL) /* Software bits:       */
+#define _PAGE_EXEC_4U	  _AC(0x0000000000001000,UL) /* Executable SW bit    */
+#define _PAGE_MODIFIED_4U _AC(0x0000000000000800,UL) /* Modified (dirty)     */
+#define _PAGE_FILE_4U	  _AC(0x0000000000000800,UL) /* Pagecache page       */
+#define _PAGE_ACCESSED_4U _AC(0x0000000000000400,UL) /* Accessed (ref'd)     */
+#define _PAGE_READ_4U	  _AC(0x0000000000000200,UL) /* Readable SW Bit      */
+#define _PAGE_WRITE_4U	  _AC(0x0000000000000100,UL) /* Writable SW Bit      */
+#define _PAGE_PRESENT_4U  _AC(0x0000000000000080,UL) /* Present              */
+#define _PAGE_L_4U	  _AC(0x0000000000000040,UL) /* Locked TTE           */
+#define _PAGE_CP_4U	  _AC(0x0000000000000020,UL) /* Cacheable in P-Cache */
+#define _PAGE_CV_4U	  _AC(0x0000000000000010,UL) /* Cacheable in V-Cache */
+#define _PAGE_E_4U	  _AC(0x0000000000000008,UL) /* side-Effect          */
+#define _PAGE_P_4U	  _AC(0x0000000000000004,UL) /* Privileged Page      */
+#define _PAGE_W_4U	  _AC(0x0000000000000002,UL) /* Writable             */
+
+/* SUN4V pte bits... */
+#define _PAGE_NFO_4V	  _AC(0x4000000000000000,UL) /* No Fault Only        */
+#define _PAGE_SOFT2_4V	  _AC(0x3F00000000000000,UL) /* Software bits, set 2 */
+#define _PAGE_MODIFIED_4V _AC(0x2000000000000000,UL) /* Modified (dirty)     */
+#define _PAGE_ACCESSED_4V _AC(0x1000000000000000,UL) /* Accessed (ref'd)     */
+#define _PAGE_READ_4V	  _AC(0x0800000000000000,UL) /* Readable SW Bit      */
+#define _PAGE_WRITE_4V	  _AC(0x0400000000000000,UL) /* Writable SW Bit      */
+#define _PAGE_PADDR_4V	  _AC(0x00FFFFFFFFFFE000,UL) /* paddr[55:13]         */
+#define _PAGE_IE_4V	  _AC(0x0000000000001000,UL) /* Invert Endianness    */
+#define _PAGE_E_4V	  _AC(0x0000000000000800,UL) /* side-Effect          */
+#define _PAGE_CP_4V	  _AC(0x0000000000000400,UL) /* Cacheable in P-Cache */
+#define _PAGE_CV_4V	  _AC(0x0000000000000200,UL) /* Cacheable in V-Cache */
+#define _PAGE_P_4V	  _AC(0x0000000000000100,UL) /* Privileged Page      */
+#define _PAGE_EXEC_4V	  _AC(0x0000000000000080,UL) /* Executable Page      */
+#define _PAGE_W_4V	  _AC(0x0000000000000040,UL) /* Writable             */
+#define _PAGE_SOFT_4V	  _AC(0x0000000000000030,UL) /* Software bits        */
+#define _PAGE_FILE_4V	  _AC(0x0000000000000020,UL) /* Pagecache page       */
+#define _PAGE_PRESENT_4V  _AC(0x0000000000000010,UL) /* Present              */
+#define _PAGE_RESV_4V	  _AC(0x0000000000000008,UL) /* Reserved             */
+#define _PAGE_SZ16GB_4V	  _AC(0x0000000000000007,UL) /* 16GB Page            */
+#define _PAGE_SZ2GB_4V	  _AC(0x0000000000000006,UL) /* 2GB Page             */
+#define _PAGE_SZ256MB_4V  _AC(0x0000000000000005,UL) /* 256MB Page           */
+#define _PAGE_SZ32MB_4V	  _AC(0x0000000000000004,UL) /* 32MB Page            */
+#define _PAGE_SZ4MB_4V	  _AC(0x0000000000000003,UL) /* 4MB Page             */
+#define _PAGE_SZ512K_4V	  _AC(0x0000000000000002,UL) /* 512K Page            */
+#define _PAGE_SZ64K_4V	  _AC(0x0000000000000001,UL) /* 64K Page             */
+#define _PAGE_SZ8K_4V	  _AC(0x0000000000000000,UL) /* 8K Page              */
+#define _PAGE_SZALL_4V	  _AC(0x0000000000000007,UL) /* All pgsz bits        */
+
+#if PAGE_SHIFT == 13
+#define _PAGE_SZBITS_4U	_PAGE_SZ8K_4U
+#define _PAGE_SZBITS_4V	_PAGE_SZ8K_4V
+#elif PAGE_SHIFT == 16
+#define _PAGE_SZBITS_4U	_PAGE_SZ64K_4U
+#define _PAGE_SZBITS_4V	_PAGE_SZ64K_4V
+#else
+#error Wrong PAGE_SHIFT specified
+#endif
+
+#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
+#define _PAGE_SZHUGE_4U	_PAGE_SZ4MB_4U
+#define _PAGE_SZHUGE_4V	_PAGE_SZ4MB_4V
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
+#define _PAGE_SZHUGE_4U	_PAGE_SZ512K_4U
+#define _PAGE_SZHUGE_4V	_PAGE_SZ512K_4V
+#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
+#define _PAGE_SZHUGE_4U	_PAGE_SZ64K_4U
+#define _PAGE_SZHUGE_4V	_PAGE_SZ64K_4V
+#endif
+
+/* These are actually filled in at boot time by sun4{u,v}_pgprot_init() */
+#define __P000	__pgprot(0)
+#define __P001	__pgprot(0)
+#define __P010	__pgprot(0)
+#define __P011	__pgprot(0)
+#define __P100	__pgprot(0)
+#define __P101	__pgprot(0)
+#define __P110	__pgprot(0)
+#define __P111	__pgprot(0)
+
+#define __S000	__pgprot(0)
+#define __S001	__pgprot(0)
+#define __S010	__pgprot(0)
+#define __S011	__pgprot(0)
+#define __S100	__pgprot(0)
+#define __S101	__pgprot(0)
+#define __S110	__pgprot(0)
+#define __S111	__pgprot(0)
+
+#ifndef __ASSEMBLY__
+
+extern pte_t mk_pte_io(unsigned long, pgprot_t, int, unsigned long);
+
+extern unsigned long pte_sz_bits(unsigned long size);
+
+extern pgprot_t PAGE_KERNEL;
+extern pgprot_t PAGE_KERNEL_LOCKED;
+extern pgprot_t PAGE_COPY;
+extern pgprot_t PAGE_SHARED;
+
+/* XXX This uglyness is for the atyfb driver's sparc mmap() support. XXX */
+extern unsigned long _PAGE_IE;
+extern unsigned long _PAGE_E;
+extern unsigned long _PAGE_CACHE;
+
+extern unsigned long pg_iobits;
+extern unsigned long _PAGE_ALL_SZ_BITS;
+extern unsigned long _PAGE_SZBITS;
+
+extern struct page *mem_map_zero;
+#define ZERO_PAGE(vaddr)	(mem_map_zero)
+
+/* PFNs are real physical page numbers.  However, mem_map only begins to record
+ * per-page information starting at pfn_base.  This is to handle systems where
+ * the first physical page in the machine is at some huge physical address,
+ * such as 4GB.   This is common on a partitioned E10000, for example.
+ */
+static inline pte_t pfn_pte(unsigned long pfn, pgprot_t prot)
+{
+	unsigned long paddr = pfn << PAGE_SHIFT;
+	unsigned long sz_bits;
+
+	sz_bits = 0UL;
+	if (_PAGE_SZBITS_4U != 0UL || _PAGE_SZBITS_4V != 0UL) {
+		__asm__ __volatile__(
+		"\n661:	sethi		%%uhi(%1), %0\n"
+		"	sllx		%0, 32, %0\n"
+		"	.section	.sun4v_2insn_patch, \"ax\"\n"
+		"	.word		661b\n"
+		"	mov		%2, %0\n"
+		"	nop\n"
+		"	.previous\n"
+		: "=r" (sz_bits)
+		: "i" (_PAGE_SZBITS_4U), "i" (_PAGE_SZBITS_4V));
+	}
+	return __pte(paddr | sz_bits | pgprot_val(prot));
+}
+#define mk_pte(page, pgprot)	pfn_pte(page_to_pfn(page), (pgprot))
+
+/* This one can be done with two shifts.  */
+static inline unsigned long pte_pfn(pte_t pte)
+{
+	unsigned long ret;
+
+	__asm__ __volatile__(
+	"\n661:	sllx		%1, %2, %0\n"
+	"	srlx		%0, %3, %0\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sllx		%1, %4, %0\n"
+	"	srlx		%0, %5, %0\n"
+	"	.previous\n"
+	: "=r" (ret)
+	: "r" (pte_val(pte)),
+	  "i" (21), "i" (21 + PAGE_SHIFT),
+	  "i" (8), "i" (8 + PAGE_SHIFT));
+
+	return ret;
+}
+#define pte_page(x) pfn_to_page(pte_pfn(x))
+
+static inline pte_t pte_modify(pte_t pte, pgprot_t prot)
+{
+	unsigned long mask, tmp;
+
+	/* SUN4U: 0x600307ffffffecb8 (negated == 0x9ffcf80000001347)
+	 * SUN4V: 0x30ffffffffffee17 (negated == 0xcf000000000011e8)
+	 *
+	 * Even if we use negation tricks the result is still a 6
+	 * instruction sequence, so don't try to play fancy and just
+	 * do the most straightforward implementation.
+	 *
+	 * Note: We encode this into 3 sun4v 2-insn patch sequences.
+	 */
+
+	__asm__ __volatile__(
+	"\n661:	sethi		%%uhi(%2), %1\n"
+	"	sethi		%%hi(%2), %0\n"
+	"\n662:	or		%1, %%ulo(%2), %1\n"
+	"	or		%0, %%lo(%2), %0\n"
+	"\n663:	sllx		%1, 32, %1\n"
+	"	or		%0, %1, %0\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%3), %1\n"
+	"	sethi		%%hi(%3), %0\n"
+	"	.word		662b\n"
+	"	or		%1, %%ulo(%3), %1\n"
+	"	or		%0, %%lo(%3), %0\n"
+	"	.word		663b\n"
+	"	sllx		%1, 32, %1\n"
+	"	or		%0, %1, %0\n"
+	"	.previous\n"
+	: "=r" (mask), "=r" (tmp)
+	: "i" (_PAGE_PADDR_4U | _PAGE_MODIFIED_4U | _PAGE_ACCESSED_4U |
+	       _PAGE_CP_4U | _PAGE_CV_4U | _PAGE_E_4U | _PAGE_PRESENT_4U |
+	       _PAGE_SZBITS_4U),
+	  "i" (_PAGE_PADDR_4V | _PAGE_MODIFIED_4V | _PAGE_ACCESSED_4V |
+	       _PAGE_CP_4V | _PAGE_CV_4V | _PAGE_E_4V | _PAGE_PRESENT_4V |
+	       _PAGE_SZBITS_4V));
+
+	return __pte((pte_val(pte) & mask) | (pgprot_val(prot) & ~mask));
+}
+
+static inline pte_t pgoff_to_pte(unsigned long off)
+{
+	off <<= PAGE_SHIFT;
+
+	__asm__ __volatile__(
+	"\n661:	or		%0, %2, %0\n"
+	"	.section	.sun4v_1insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	or		%0, %3, %0\n"
+	"	.previous\n"
+	: "=r" (off)
+	: "0" (off), "i" (_PAGE_FILE_4U), "i" (_PAGE_FILE_4V));
+
+	return __pte(off);
+}
+
+static inline pgprot_t pgprot_noncached(pgprot_t prot)
+{
+	unsigned long val = pgprot_val(prot);
+
+	__asm__ __volatile__(
+	"\n661:	andn		%0, %2, %0\n"
+	"	or		%0, %3, %0\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	andn		%0, %4, %0\n"
+	"	or		%0, %5, %0\n"
+	"	.previous\n"
+	: "=r" (val)
+	: "0" (val), "i" (_PAGE_CP_4U | _PAGE_CV_4U), "i" (_PAGE_E_4U),
+	             "i" (_PAGE_CP_4V | _PAGE_CV_4V), "i" (_PAGE_E_4V));
+
+	return __pgprot(val);
+}
+/* Various pieces of code check for platform support by ifdef testing
+ * on "pgprot_noncached".  That's broken and should be fixed, but for
+ * now...
+ */
+#define pgprot_noncached pgprot_noncached
+
+#ifdef CONFIG_HUGETLB_PAGE
+static inline pte_t pte_mkhuge(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	sethi		%%uhi(%1), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	mov		%2, %0\n"
+	"	nop\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_SZHUGE_4U), "i" (_PAGE_SZHUGE_4V));
+
+	return __pte(pte_val(pte) | mask);
+}
+#endif
+
+static inline pte_t pte_mkdirty(pte_t pte)
+{
+	unsigned long val = pte_val(pte), tmp;
+
+	__asm__ __volatile__(
+	"\n661:	or		%0, %3, %0\n"
+	"	nop\n"
+	"\n662:	nop\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%4), %1\n"
+	"	sllx		%1, 32, %1\n"
+	"	.word		662b\n"
+	"	or		%1, %%lo(%4), %1\n"
+	"	or		%0, %1, %0\n"
+	"	.previous\n"
+	: "=r" (val), "=r" (tmp)
+	: "0" (val), "i" (_PAGE_MODIFIED_4U | _PAGE_W_4U),
+	  "i" (_PAGE_MODIFIED_4V | _PAGE_W_4V));
+
+	return __pte(val);
+}
+
+static inline pte_t pte_mkclean(pte_t pte)
+{
+	unsigned long val = pte_val(pte), tmp;
+
+	__asm__ __volatile__(
+	"\n661:	andn		%0, %3, %0\n"
+	"	nop\n"
+	"\n662:	nop\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%4), %1\n"
+	"	sllx		%1, 32, %1\n"
+	"	.word		662b\n"
+	"	or		%1, %%lo(%4), %1\n"
+	"	andn		%0, %1, %0\n"
+	"	.previous\n"
+	: "=r" (val), "=r" (tmp)
+	: "0" (val), "i" (_PAGE_MODIFIED_4U | _PAGE_W_4U),
+	  "i" (_PAGE_MODIFIED_4V | _PAGE_W_4V));
+
+	return __pte(val);
+}
+
+static inline pte_t pte_mkwrite(pte_t pte)
+{
+	unsigned long val = pte_val(pte), mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_WRITE_4U), "i" (_PAGE_WRITE_4V));
+
+	return __pte(val | mask);
+}
+
+static inline pte_t pte_wrprotect(pte_t pte)
+{
+	unsigned long val = pte_val(pte), tmp;
+
+	__asm__ __volatile__(
+	"\n661:	andn		%0, %3, %0\n"
+	"	nop\n"
+	"\n662:	nop\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%4), %1\n"
+	"	sllx		%1, 32, %1\n"
+	"	.word		662b\n"
+	"	or		%1, %%lo(%4), %1\n"
+	"	andn		%0, %1, %0\n"
+	"	.previous\n"
+	: "=r" (val), "=r" (tmp)
+	: "0" (val), "i" (_PAGE_WRITE_4U | _PAGE_W_4U),
+	  "i" (_PAGE_WRITE_4V | _PAGE_W_4V));
+
+	return __pte(val);
+}
+
+static inline pte_t pte_mkold(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
+
+	mask |= _PAGE_R;
+
+	return __pte(pte_val(pte) & ~mask);
+}
+
+static inline pte_t pte_mkyoung(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
+
+	mask |= _PAGE_R;
+
+	return __pte(pte_val(pte) | mask);
+}
+
+static inline pte_t pte_mkspecial(pte_t pte)
+{
+	return pte;
+}
+
+static inline unsigned long pte_young(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
+
+	return (pte_val(pte) & mask);
+}
+
+static inline unsigned long pte_dirty(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_MODIFIED_4U), "i" (_PAGE_MODIFIED_4V));
+
+	return (pte_val(pte) & mask);
+}
+
+static inline unsigned long pte_write(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	mov		%1, %0\n"
+	"	nop\n"
+	"	.section	.sun4v_2insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	sethi		%%uhi(%2), %0\n"
+	"	sllx		%0, 32, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_WRITE_4U), "i" (_PAGE_WRITE_4V));
+
+	return (pte_val(pte) & mask);
+}
+
+static inline unsigned long pte_exec(pte_t pte)
+{
+	unsigned long mask;
+
+	__asm__ __volatile__(
+	"\n661:	sethi		%%hi(%1), %0\n"
+	"	.section	.sun4v_1insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	mov		%2, %0\n"
+	"	.previous\n"
+	: "=r" (mask)
+	: "i" (_PAGE_EXEC_4U), "i" (_PAGE_EXEC_4V));
+
+	return (pte_val(pte) & mask);
+}
+
+static inline unsigned long pte_file(pte_t pte)
+{
+	unsigned long val = pte_val(pte);
+
+	__asm__ __volatile__(
+	"\n661:	and		%0, %2, %0\n"
+	"	.section	.sun4v_1insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	and		%0, %3, %0\n"
+	"	.previous\n"
+	: "=r" (val)
+	: "0" (val), "i" (_PAGE_FILE_4U), "i" (_PAGE_FILE_4V));
+
+	return val;
+}
+
+static inline unsigned long pte_present(pte_t pte)
+{
+	unsigned long val = pte_val(pte);
+
+	__asm__ __volatile__(
+	"\n661:	and		%0, %2, %0\n"
+	"	.section	.sun4v_1insn_patch, \"ax\"\n"
+	"	.word		661b\n"
+	"	and		%0, %3, %0\n"
+	"	.previous\n"
+	: "=r" (val)
+	: "0" (val), "i" (_PAGE_PRESENT_4U), "i" (_PAGE_PRESENT_4V));
+
+	return val;
+}
+
+static inline int pte_special(pte_t pte)
+{
+	return 0;
+}
+
+#define pmd_set(pmdp, ptep)	\
+	(pmd_val(*(pmdp)) = (__pa((unsigned long) (ptep)) >> 11UL))
+#define pud_set(pudp, pmdp)	\
+	(pud_val(*(pudp)) = (__pa((unsigned long) (pmdp)) >> 11UL))
+#define __pmd_page(pmd)		\
+	((unsigned long) __va((((unsigned long)pmd_val(pmd))<<11UL)))
+#define pmd_page(pmd) 			virt_to_page((void *)__pmd_page(pmd))
+#define pud_page_vaddr(pud)		\
+	((unsigned long) __va((((unsigned long)pud_val(pud))<<11UL)))
+#define pud_page(pud) 			virt_to_page((void *)pud_page_vaddr(pud))
+#define pmd_none(pmd)			(!pmd_val(pmd))
+#define pmd_bad(pmd)			(0)
+#define pmd_present(pmd)		(pmd_val(pmd) != 0U)
+#define pmd_clear(pmdp)			(pmd_val(*(pmdp)) = 0U)
+#define pud_none(pud)			(!pud_val(pud))
+#define pud_bad(pud)			(0)
+#define pud_present(pud)		(pud_val(pud) != 0U)
+#define pud_clear(pudp)			(pud_val(*(pudp)) = 0U)
+
+/* Same in both SUN4V and SUN4U.  */
+#define pte_none(pte) 			(!pte_val(pte))
+
+/* to find an entry in a page-table-directory. */
+#define pgd_index(address)	(((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD - 1))
+#define pgd_offset(mm, address)	((mm)->pgd + pgd_index(address))
+
+/* to find an entry in a kernel page-table-directory */
+#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+
+/* Find an entry in the second-level page table.. */
+#define pmd_offset(pudp, address)	\
+	((pmd_t *) pud_page_vaddr(*(pudp)) + \
+	 (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1)))
+
+/* Find an entry in the third-level page table.. */
+#define pte_index(dir, address)	\
+	((pte_t *) __pmd_page(*(dir)) + \
+	 ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
+#define pte_offset_kernel		pte_index
+#define pte_offset_map			pte_index
+#define pte_offset_map_nested		pte_index
+#define pte_unmap(pte)			do { } while (0)
+#define pte_unmap_nested(pte)		do { } while (0)
+
+/* Actual page table PTE updates.  */
+extern void tlb_batch_add(struct mm_struct *mm, unsigned long vaddr, pte_t *ptep, pte_t orig);
+
+static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte)
+{
+	pte_t orig = *ptep;
+
+	*ptep = pte;
+
+	/* It is more efficient to let flush_tlb_kernel_range()
+	 * handle init_mm tlb flushes.
+	 *
+	 * SUN4V NOTE: _PAGE_VALID is the same value in both the SUN4U
+	 *             and SUN4V pte layout, so this inline test is fine.
+	 */
+	if (likely(mm != &init_mm) && (pte_val(orig) & _PAGE_VALID))
+		tlb_batch_add(mm, addr, ptep, orig);
+}
+
+#define pte_clear(mm,addr,ptep)		\
+	set_pte_at((mm), (addr), (ptep), __pte(0UL))
+
+#ifdef DCACHE_ALIASING_POSSIBLE
+#define __HAVE_ARCH_MOVE_PTE
+#define move_pte(pte, prot, old_addr, new_addr)				\
+({									\
+	pte_t newpte = (pte);						\
+	if (tlb_type != hypervisor && pte_present(pte)) {		\
+		unsigned long this_pfn = pte_pfn(pte);			\
+									\
+		if (pfn_valid(this_pfn) &&				\
+		    (((old_addr) ^ (new_addr)) & (1 << 13)))		\
+			flush_dcache_page_all(current->mm,		\
+					      pfn_to_page(this_pfn));	\
+	}								\
+	newpte;								\
+})
+#endif
+
+extern pgd_t swapper_pg_dir[2048];
+extern pmd_t swapper_low_pmd_dir[2048];
+
+extern void paging_init(void);
+extern unsigned long find_ecache_flush_span(unsigned long size);
+
+/* These do nothing with the way I have things setup. */
+#define mmu_lockarea(vaddr, len)		(vaddr)
+#define mmu_unlockarea(vaddr, len)		do { } while(0)
+
+struct vm_area_struct;
+extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t);
+
+/* Encode and de-code a swap entry */
+#define __swp_type(entry)	(((entry).val >> PAGE_SHIFT) & 0xffUL)
+#define __swp_offset(entry)	((entry).val >> (PAGE_SHIFT + 8UL))
+#define __swp_entry(type, offset)	\
+	( (swp_entry_t) \
+	  { \
+		(((long)(type) << PAGE_SHIFT) | \
+                 ((long)(offset) << (PAGE_SHIFT + 8UL))) \
+	  } )
+#define __pte_to_swp_entry(pte)		((swp_entry_t) { pte_val(pte) })
+#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
+
+/* File offset in PTE support. */
+extern unsigned long pte_file(pte_t);
+#define pte_to_pgoff(pte)	(pte_val(pte) >> PAGE_SHIFT)
+extern pte_t pgoff_to_pte(unsigned long);
+#define PTE_FILE_MAX_BITS	(64UL - PAGE_SHIFT - 1UL)
+
+extern unsigned long *sparc64_valid_addr_bitmap;
+
+/* Needs to be defined here and not in linux/mm.h, as it is arch dependent */
+#define kern_addr_valid(addr)	\
+	(test_bit(__pa((unsigned long)(addr))>>22, sparc64_valid_addr_bitmap))
+
+extern int page_in_phys_avail(unsigned long paddr);
+
+extern int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
+			       unsigned long pfn,
+			       unsigned long size, pgprot_t prot);
+
+/*
+ * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
+ * its high 4 bits.  These macros/functions put it there or get it from there.
+ */
+#define MK_IOSPACE_PFN(space, pfn)	(pfn | (space << (BITS_PER_LONG - 4)))
+#define GET_IOSPACE(pfn)		(pfn >> (BITS_PER_LONG - 4))
+#define GET_PFN(pfn)			(pfn & 0x0fffffffffffffffUL)
+
+#include <asm-generic/pgtable.h>
+
+/* We provide our own get_unmapped_area to cope with VA holes and
+ * SHM area cache aliasing for userland.
+ */
+#define HAVE_ARCH_UNMAPPED_AREA
+#define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
+
+/* We provide a special get_unmapped_area for framebuffer mmaps to try and use
+ * the largest alignment possible such that larget PTEs can be used.
+ */
+extern unsigned long get_fb_unmapped_area(struct file *filp, unsigned long,
+					  unsigned long, unsigned long,
+					  unsigned long);
+#define HAVE_ARCH_FB_UNMAPPED_AREA
+
+extern void pgtable_cache_init(void);
+extern void sun4v_register_fault_status(void);
+extern void sun4v_ktsb_register(void);
+extern void __init cheetah_ecache_flush_init(void);
+extern void sun4v_patch_tlb_handlers(void);
+
+extern unsigned long cmdline_memory_size;
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(_SPARC64_PGTABLE_H) */
diff --git a/include/asm-sparc/pil.h b/include/asm-sparc/pil.h
new file mode 100644
index 00000000000..71819bb943f
--- /dev/null
+++ b/include/asm-sparc/pil.h
@@ -0,0 +1,22 @@
+#ifndef _SPARC64_PIL_H
+#define _SPARC64_PIL_H
+
+/* To avoid some locking problems, we hard allocate certain PILs
+ * for SMP cross call messages that must do a etrap/rtrap.
+ *
+ * A local_irq_disable() does not block the cross call delivery, so
+ * when SMP locking is an issue we reschedule the event into a PIL
+ * interrupt which is blocked by local_irq_disable().
+ *
+ * In fact any XCALL which has to etrap/rtrap has a problem because
+ * it is difficult to prevent rtrap from running BH's, and that would
+ * need to be done if the XCALL arrived while %pil==15.
+ */
+#define PIL_SMP_CALL_FUNC	1
+#define PIL_SMP_RECEIVE_SIGNAL	2
+#define PIL_SMP_CAPTURE		3
+#define PIL_SMP_CTX_NEW_VERSION	4
+#define PIL_DEVICE_IRQ		5
+#define PIL_SMP_CALL_FUNC_SNGL	6
+
+#endif /* !(_SPARC64_PIL_H) */
diff --git a/include/asm-sparc/posix_types.h b/include/asm-sparc/posix_types.h
index dcc07eb5e18..58c820d75e8 100644
--- a/include/asm-sparc/posix_types.h
+++ b/include/asm-sparc/posix_types.h
@@ -1,118 +1,8 @@
-#ifndef __ARCH_SPARC_POSIX_TYPES_H
-#define __ARCH_SPARC_POSIX_TYPES_H
-
-/*
- * This file is generally used by user-level software, so you need to
- * be a little careful about namespace pollution etc.  Also, we cannot
- * assume GCC is being used.
- */
-
-typedef unsigned int           __kernel_size_t;
-typedef int                    __kernel_ssize_t;
-typedef long int               __kernel_ptrdiff_t;
-typedef long                   __kernel_time_t;
-typedef long		       __kernel_suseconds_t;
-typedef long                   __kernel_clock_t;
-typedef int                    __kernel_pid_t;
-typedef unsigned short         __kernel_ipc_pid_t;
-typedef unsigned short         __kernel_uid_t;
-typedef unsigned short         __kernel_gid_t;
-typedef unsigned long          __kernel_ino_t;
-typedef unsigned short         __kernel_mode_t;
-typedef unsigned short         __kernel_umode_t;
-typedef short                  __kernel_nlink_t;
-typedef long                   __kernel_daddr_t;
-typedef long                   __kernel_off_t;
-typedef char *                 __kernel_caddr_t;
-typedef unsigned short	       __kernel_uid16_t;
-typedef unsigned short	       __kernel_gid16_t;
-typedef unsigned int	       __kernel_uid32_t;
-typedef unsigned int	       __kernel_gid32_t;
-typedef unsigned short	       __kernel_old_uid_t;
-typedef unsigned short	       __kernel_old_gid_t;
-typedef unsigned short	       __kernel_old_dev_t;
-typedef int                    __kernel_clockid_t;
-typedef int                    __kernel_timer_t;
-
-#ifdef __GNUC__
-typedef long long	__kernel_loff_t;
+#ifndef ___ASM_SPARC_POSIX_TYPES_H
+#define ___ASM_SPARC_POSIX_TYPES_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/posix_types_64.h>
+#else
+#include <asm-sparc/posix_types_32.h>
+#endif
 #endif
-
-typedef struct {
-	int	val[2];
-} __kernel_fsid_t;
-
-#if defined(__KERNEL__)
-
-#undef __FD_SET
-static inline void __FD_SET(unsigned long fd, __kernel_fd_set *fdsetp)
-{
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	fdsetp->fds_bits[_tmp] |= (1UL<<_rem);
-}
-
-#undef __FD_CLR
-static inline void __FD_CLR(unsigned long fd, __kernel_fd_set *fdsetp)
-{
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	fdsetp->fds_bits[_tmp] &= ~(1UL<<_rem);
-}
-
-#undef __FD_ISSET
-static inline int __FD_ISSET(unsigned long fd, __const__ __kernel_fd_set *p)
-{ 
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	return (p->fds_bits[_tmp] & (1UL<<_rem)) != 0;
-}
-
-/*
- * This will unroll the loop for the normal constant cases (8 or 32 longs,
- * for 256 and 1024-bit fd_sets respectively)
- */
-#undef __FD_ZERO
-static inline void __FD_ZERO(__kernel_fd_set *p)
-{
-	unsigned long *tmp = p->fds_bits;
-	int i;
-
-	if (__builtin_constant_p(__FDSET_LONGS)) {
-		switch (__FDSET_LONGS) {
-			case 32:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
-			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
-			  tmp[16] = 0; tmp[17] = 0; tmp[18] = 0; tmp[19] = 0;
-			  tmp[20] = 0; tmp[21] = 0; tmp[22] = 0; tmp[23] = 0;
-			  tmp[24] = 0; tmp[25] = 0; tmp[26] = 0; tmp[27] = 0;
-			  tmp[28] = 0; tmp[29] = 0; tmp[30] = 0; tmp[31] = 0;
-			  return;
-			case 16:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
-			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
-			  return;
-			case 8:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  return;
-			case 4:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  return;
-		}
-	}
-	i = __FDSET_LONGS;
-	while (i) {
-		i--;
-		*tmp = 0;
-		tmp++;
-	}
-}
-
-#endif /* defined(__KERNEL__) */
-
-#endif /* !(__ARCH_SPARC_POSIX_TYPES_H) */
diff --git a/include/asm-sparc/posix_types_32.h b/include/asm-sparc/posix_types_32.h
new file mode 100644
index 00000000000..6bb6eb1ca0f
--- /dev/null
+++ b/include/asm-sparc/posix_types_32.h
@@ -0,0 +1,118 @@
+#ifndef __ARCH_SPARC_POSIX_TYPES_H
+#define __ARCH_SPARC_POSIX_TYPES_H
+
+/*
+ * This file is generally used by user-level software, so you need to
+ * be a little careful about namespace pollution etc.  Also, we cannot
+ * assume GCC is being used.
+ */
+
+typedef unsigned int           __kernel_size_t;
+typedef int                    __kernel_ssize_t;
+typedef long int               __kernel_ptrdiff_t;
+typedef long                   __kernel_time_t;
+typedef long		       __kernel_suseconds_t;
+typedef long                   __kernel_clock_t;
+typedef int                    __kernel_pid_t;
+typedef unsigned short         __kernel_ipc_pid_t;
+typedef unsigned short         __kernel_uid_t;
+typedef unsigned short         __kernel_gid_t;
+typedef unsigned long          __kernel_ino_t;
+typedef unsigned short         __kernel_mode_t;
+typedef unsigned short         __kernel_umode_t;
+typedef short                  __kernel_nlink_t;
+typedef long                   __kernel_daddr_t;
+typedef long                   __kernel_off_t;
+typedef char *                 __kernel_caddr_t;
+typedef unsigned short	       __kernel_uid16_t;
+typedef unsigned short	       __kernel_gid16_t;
+typedef unsigned int	       __kernel_uid32_t;
+typedef unsigned int	       __kernel_gid32_t;
+typedef unsigned short	       __kernel_old_uid_t;
+typedef unsigned short	       __kernel_old_gid_t;
+typedef unsigned short	       __kernel_old_dev_t;
+typedef int                    __kernel_clockid_t;
+typedef int                    __kernel_timer_t;
+
+#ifdef __GNUC__
+typedef long long	__kernel_loff_t;
+#endif
+
+typedef struct {
+	int	val[2];
+} __kernel_fsid_t;
+
+#if defined(__KERNEL__)
+
+#undef __FD_SET
+static inline void __FD_SET(unsigned long fd, __kernel_fd_set *fdsetp)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	fdsetp->fds_bits[_tmp] |= (1UL<<_rem);
+}
+
+#undef __FD_CLR
+static inline void __FD_CLR(unsigned long fd, __kernel_fd_set *fdsetp)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	fdsetp->fds_bits[_tmp] &= ~(1UL<<_rem);
+}
+
+#undef __FD_ISSET
+static inline int __FD_ISSET(unsigned long fd, __const__ __kernel_fd_set *p)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	return (p->fds_bits[_tmp] & (1UL<<_rem)) != 0;
+}
+
+/*
+ * This will unroll the loop for the normal constant cases (8 or 32 longs,
+ * for 256 and 1024-bit fd_sets respectively)
+ */
+#undef __FD_ZERO
+static inline void __FD_ZERO(__kernel_fd_set *p)
+{
+	unsigned long *tmp = p->fds_bits;
+	int i;
+
+	if (__builtin_constant_p(__FDSET_LONGS)) {
+		switch (__FDSET_LONGS) {
+			case 32:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
+			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
+			  tmp[16] = 0; tmp[17] = 0; tmp[18] = 0; tmp[19] = 0;
+			  tmp[20] = 0; tmp[21] = 0; tmp[22] = 0; tmp[23] = 0;
+			  tmp[24] = 0; tmp[25] = 0; tmp[26] = 0; tmp[27] = 0;
+			  tmp[28] = 0; tmp[29] = 0; tmp[30] = 0; tmp[31] = 0;
+			  return;
+			case 16:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
+			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
+			  return;
+			case 8:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  return;
+			case 4:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  return;
+		}
+	}
+	i = __FDSET_LONGS;
+	while (i) {
+		i--;
+		*tmp = 0;
+		tmp++;
+	}
+}
+
+#endif /* defined(__KERNEL__) */
+
+#endif /* !(__ARCH_SPARC_POSIX_TYPES_H) */
diff --git a/include/asm-sparc/posix_types_64.h b/include/asm-sparc/posix_types_64.h
new file mode 100644
index 00000000000..ba8f9329576
--- /dev/null
+++ b/include/asm-sparc/posix_types_64.h
@@ -0,0 +1,122 @@
+#ifndef __ARCH_SPARC64_POSIX_TYPES_H
+#define __ARCH_SPARC64_POSIX_TYPES_H
+
+/*
+ * This file is generally used by user-level software, so you need to
+ * be a little careful about namespace pollution etc.  Also, we cannot
+ * assume GCC is being used.
+ */
+
+typedef unsigned long          __kernel_size_t;
+typedef long                   __kernel_ssize_t;
+typedef long                   __kernel_ptrdiff_t;
+typedef long                   __kernel_time_t;
+typedef long                   __kernel_clock_t;
+typedef int                    __kernel_pid_t;
+typedef int                    __kernel_ipc_pid_t;
+typedef unsigned int           __kernel_uid_t;
+typedef unsigned int           __kernel_gid_t;
+typedef unsigned long          __kernel_ino_t;
+typedef unsigned int           __kernel_mode_t;
+typedef unsigned short         __kernel_umode_t;
+typedef unsigned int           __kernel_nlink_t;
+typedef int                    __kernel_daddr_t;
+typedef long                   __kernel_off_t;
+typedef char *                 __kernel_caddr_t;
+typedef unsigned short	       __kernel_uid16_t;
+typedef unsigned short	       __kernel_gid16_t;
+typedef int                    __kernel_clockid_t;
+typedef int                    __kernel_timer_t;
+
+typedef unsigned short 	       __kernel_old_uid_t;
+typedef unsigned short         __kernel_old_gid_t;
+typedef __kernel_uid_t	       __kernel_uid32_t;
+typedef __kernel_gid_t	       __kernel_gid32_t;
+
+typedef unsigned int	       __kernel_old_dev_t;
+
+/* Note this piece of asymmetry from the v9 ABI.  */
+typedef int		       __kernel_suseconds_t;
+
+#ifdef __GNUC__
+typedef long long              __kernel_loff_t;
+#endif
+
+typedef struct {
+	int	val[2];
+} __kernel_fsid_t;
+
+#if defined(__KERNEL__)
+
+#undef __FD_SET
+static inline void __FD_SET(unsigned long fd, __kernel_fd_set *fdsetp)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	fdsetp->fds_bits[_tmp] |= (1UL<<_rem);
+}
+
+#undef __FD_CLR
+static inline void __FD_CLR(unsigned long fd, __kernel_fd_set *fdsetp)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	fdsetp->fds_bits[_tmp] &= ~(1UL<<_rem);
+}
+
+#undef __FD_ISSET
+static inline int __FD_ISSET(unsigned long fd, __const__ __kernel_fd_set *p)
+{
+	unsigned long _tmp = fd / __NFDBITS;
+	unsigned long _rem = fd % __NFDBITS;
+	return (p->fds_bits[_tmp] & (1UL<<_rem)) != 0;
+}
+
+/*
+ * This will unroll the loop for the normal constant cases (8 or 32 longs,
+ * for 256 and 1024-bit fd_sets respectively)
+ */
+#undef __FD_ZERO
+static inline void __FD_ZERO(__kernel_fd_set *p)
+{
+	unsigned long *tmp = p->fds_bits;
+	int i;
+
+	if (__builtin_constant_p(__FDSET_LONGS)) {
+		switch (__FDSET_LONGS) {
+			case 32:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
+			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
+			  tmp[16] = 0; tmp[17] = 0; tmp[18] = 0; tmp[19] = 0;
+			  tmp[20] = 0; tmp[21] = 0; tmp[22] = 0; tmp[23] = 0;
+			  tmp[24] = 0; tmp[25] = 0; tmp[26] = 0; tmp[27] = 0;
+			  tmp[28] = 0; tmp[29] = 0; tmp[30] = 0; tmp[31] = 0;
+			  return;
+			case 16:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
+			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
+			  return;
+			case 8:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
+			  return;
+			case 4:
+			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
+			  return;
+		}
+	}
+	i = __FDSET_LONGS;
+	while (i) {
+		i--;
+		*tmp = 0;
+		tmp++;
+	}
+}
+
+#endif /* defined(__KERNEL__) */
+
+#endif /* !(__ARCH_SPARC64_POSIX_TYPES_H) */
diff --git a/include/asm-sparc/processor.h b/include/asm-sparc/processor.h
index 8898efbbbe0..11a66bb02ea 100644
--- a/include/asm-sparc/processor.h
+++ b/include/asm-sparc/processor.h
@@ -1,128 +1,8 @@
-/* include/asm-sparc/processor.h
- *
- * Copyright (C) 1994 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __ASM_SPARC_PROCESSOR_H
-#define __ASM_SPARC_PROCESSOR_H
-
-/*
- * Sparc32 implementation of macro that returns current
- * instruction pointer ("program counter").
- */
-#define current_text_addr() ({ void *pc; __asm__("sethi %%hi(1f), %0; or %0, %%lo(1f), %0;\n1:" : "=r" (pc)); pc; })
-
-#include <asm/psr.h>
-#include <asm/ptrace.h>
-#include <asm/head.h>
-#include <asm/signal.h>
-#include <asm/btfixup.h>
-#include <asm/page.h>
-
-/*
- * The sparc has no problems with write protection
- */
-#define wp_works_ok 1
-#define wp_works_ok__is_a_macro /* for versions in ksyms.c */
-
-/* Whee, this is STACK_TOP + PAGE_SIZE and the lowest kernel address too... 
- * That one page is used to protect kernel from intruders, so that
- * we can make our access_ok test faster
- */
-#define TASK_SIZE	PAGE_OFFSET
-#ifdef __KERNEL__
-#define STACK_TOP	(PAGE_OFFSET - PAGE_SIZE)
-#define STACK_TOP_MAX	STACK_TOP
-#endif /* __KERNEL__ */
-
-struct task_struct;
-
-#ifdef __KERNEL__
-struct fpq {
-	unsigned long *insn_addr;
-	unsigned long insn;
-};
+#ifndef ___ASM_SPARC_PROCESSOR_H
+#define ___ASM_SPARC_PROCESSOR_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/processor_64.h>
+#else
+#include <asm-sparc/processor_32.h>
 #endif
-
-typedef struct {
-	int seg;
-} mm_segment_t;
-
-/* The Sparc processor specific thread struct. */
-struct thread_struct {
-	struct pt_regs *kregs;
-	unsigned int _pad1;
-
-	/* Special child fork kpsr/kwim values. */
-	unsigned long fork_kpsr __attribute__ ((aligned (8)));
-	unsigned long fork_kwim;
-
-	/* Floating point regs */
-	unsigned long   float_regs[32] __attribute__ ((aligned (8)));
-	unsigned long   fsr;
-	unsigned long   fpqdepth;
-	struct fpq	fpqueue[16];
-	unsigned long flags;
-	mm_segment_t current_ds;
-};
-
-#define SPARC_FLAG_KTHREAD      0x1    /* task is a kernel thread */
-#define SPARC_FLAG_UNALIGNED    0x2    /* is allowed to do unaligned accesses */
-
-#define INIT_THREAD  { \
-	.flags = SPARC_FLAG_KTHREAD, \
-	.current_ds = KERNEL_DS, \
-}
-
-/* Return saved PC of a blocked thread. */
-extern unsigned long thread_saved_pc(struct task_struct *t);
-
-/* Do necessary setup to start up a newly executed thread. */
-static inline void start_thread(struct pt_regs * regs, unsigned long pc,
-				    unsigned long sp)
-{
-	register unsigned long zero asm("g1");
-
-	regs->psr = (regs->psr & (PSR_CWP)) | PSR_S;
-	regs->pc = ((pc & (~3)) - 4);
-	regs->npc = regs->pc + 4;
-	regs->y = 0;
-	zero = 0;
-	__asm__ __volatile__("std\t%%g0, [%0 + %3 + 0x00]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x08]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x10]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x18]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x20]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x28]\n\t"
-			     "std\t%%g0, [%0 + %3 + 0x30]\n\t"
-			     "st\t%1, [%0 + %3 + 0x38]\n\t"
-			     "st\t%%g0, [%0 + %3 + 0x3c]"
-			     : /* no outputs */
-			     : "r" (regs),
-			       "r" (sp - sizeof(struct reg_window)),
-			       "r" (zero),
-			       "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))
-			     : "memory");
-}
-
-/* Free all resources held by a thread. */
-#define release_thread(tsk)		do { } while(0)
-extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
-/* Prepare to copy thread state - unlazy all lazy status */
-#define prepare_to_copy(tsk)	do { } while (0)
-
-extern unsigned long get_wchan(struct task_struct *);
-
-#define KSTK_EIP(tsk)  ((tsk)->thread.kregs->pc)
-#define KSTK_ESP(tsk)  ((tsk)->thread.kregs->u_regs[UREG_FP])
-
-#ifdef __KERNEL__
-
-extern struct task_struct *last_task_used_math;
-
-#define cpu_relax()	barrier()
-
 #endif
-
-#endif /* __ASM_SPARC_PROCESSOR_H */
diff --git a/include/asm-sparc/processor_32.h b/include/asm-sparc/processor_32.h
new file mode 100644
index 00000000000..562c0d69c53
--- /dev/null
+++ b/include/asm-sparc/processor_32.h
@@ -0,0 +1,128 @@
+/* include/asm-sparc/processor.h
+ *
+ * Copyright (C) 1994 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __ASM_SPARC_PROCESSOR_H
+#define __ASM_SPARC_PROCESSOR_H
+
+/*
+ * Sparc32 implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ void *pc; __asm__("sethi %%hi(1f), %0; or %0, %%lo(1f), %0;\n1:" : "=r" (pc)); pc; })
+
+#include <asm/psr.h>
+#include <asm/ptrace.h>
+#include <asm/head.h>
+#include <asm/signal.h>
+#include <asm/btfixup.h>
+#include <asm/page.h>
+
+/*
+ * The sparc has no problems with write protection
+ */
+#define wp_works_ok 1
+#define wp_works_ok__is_a_macro /* for versions in ksyms.c */
+
+/* Whee, this is STACK_TOP + PAGE_SIZE and the lowest kernel address too...
+ * That one page is used to protect kernel from intruders, so that
+ * we can make our access_ok test faster
+ */
+#define TASK_SIZE	PAGE_OFFSET
+#ifdef __KERNEL__
+#define STACK_TOP	(PAGE_OFFSET - PAGE_SIZE)
+#define STACK_TOP_MAX	STACK_TOP
+#endif /* __KERNEL__ */
+
+struct task_struct;
+
+#ifdef __KERNEL__
+struct fpq {
+	unsigned long *insn_addr;
+	unsigned long insn;
+};
+#endif
+
+typedef struct {
+	int seg;
+} mm_segment_t;
+
+/* The Sparc processor specific thread struct. */
+struct thread_struct {
+	struct pt_regs *kregs;
+	unsigned int _pad1;
+
+	/* Special child fork kpsr/kwim values. */
+	unsigned long fork_kpsr __attribute__ ((aligned (8)));
+	unsigned long fork_kwim;
+
+	/* Floating point regs */
+	unsigned long   float_regs[32] __attribute__ ((aligned (8)));
+	unsigned long   fsr;
+	unsigned long   fpqdepth;
+	struct fpq	fpqueue[16];
+	unsigned long flags;
+	mm_segment_t current_ds;
+};
+
+#define SPARC_FLAG_KTHREAD      0x1    /* task is a kernel thread */
+#define SPARC_FLAG_UNALIGNED    0x2    /* is allowed to do unaligned accesses */
+
+#define INIT_THREAD  { \
+	.flags = SPARC_FLAG_KTHREAD, \
+	.current_ds = KERNEL_DS, \
+}
+
+/* Return saved PC of a blocked thread. */
+extern unsigned long thread_saved_pc(struct task_struct *t);
+
+/* Do necessary setup to start up a newly executed thread. */
+static inline void start_thread(struct pt_regs * regs, unsigned long pc,
+				    unsigned long sp)
+{
+	register unsigned long zero asm("g1");
+
+	regs->psr = (regs->psr & (PSR_CWP)) | PSR_S;
+	regs->pc = ((pc & (~3)) - 4);
+	regs->npc = regs->pc + 4;
+	regs->y = 0;
+	zero = 0;
+	__asm__ __volatile__("std\t%%g0, [%0 + %3 + 0x00]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x08]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x10]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x18]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x20]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x28]\n\t"
+			     "std\t%%g0, [%0 + %3 + 0x30]\n\t"
+			     "st\t%1, [%0 + %3 + 0x38]\n\t"
+			     "st\t%%g0, [%0 + %3 + 0x3c]"
+			     : /* no outputs */
+			     : "r" (regs),
+			       "r" (sp - sizeof(struct reg_window)),
+			       "r" (zero),
+			       "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))
+			     : "memory");
+}
+
+/* Free all resources held by a thread. */
+#define release_thread(tsk)		do { } while(0)
+extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
+
+/* Prepare to copy thread state - unlazy all lazy status */
+#define prepare_to_copy(tsk)	do { } while (0)
+
+extern unsigned long get_wchan(struct task_struct *);
+
+#define KSTK_EIP(tsk)  ((tsk)->thread.kregs->pc)
+#define KSTK_ESP(tsk)  ((tsk)->thread.kregs->u_regs[UREG_FP])
+
+#ifdef __KERNEL__
+
+extern struct task_struct *last_task_used_math;
+
+#define cpu_relax()	barrier()
+
+#endif
+
+#endif /* __ASM_SPARC_PROCESSOR_H */
diff --git a/include/asm-sparc/processor_64.h b/include/asm-sparc/processor_64.h
new file mode 100644
index 00000000000..70d42801a0d
--- /dev/null
+++ b/include/asm-sparc/processor_64.h
@@ -0,0 +1,237 @@
+/*
+ * include/asm-sparc64/processor.h
+ *
+ * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __ASM_SPARC64_PROCESSOR_H
+#define __ASM_SPARC64_PROCESSOR_H
+
+/*
+ * Sparc64 implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ void *pc; __asm__("rd %%pc, %0" : "=r" (pc)); pc; })
+
+#include <asm/asi.h>
+#include <asm/pstate.h>
+#include <asm/ptrace.h>
+#include <asm/page.h>
+
+/* The sparc has no problems with write protection */
+#define wp_works_ok 1
+#define wp_works_ok__is_a_macro /* for versions in ksyms.c */
+
+/*
+ * User lives in his very own context, and cannot reference us. Note
+ * that TASK_SIZE is a misnomer, it really gives maximum user virtual
+ * address that the kernel will allocate out.
+ *
+ * XXX No longer using virtual page tables, kill this upper limit...
+ */
+#define VA_BITS		44
+#ifndef __ASSEMBLY__
+#define VPTE_SIZE	(1UL << (VA_BITS - PAGE_SHIFT + 3))
+#else
+#define VPTE_SIZE	(1 << (VA_BITS - PAGE_SHIFT + 3))
+#endif
+
+#define TASK_SIZE	((unsigned long)-VPTE_SIZE)
+#define TASK_SIZE_OF(tsk) \
+	(test_tsk_thread_flag(tsk,TIF_32BIT) ? \
+	 (1UL << 32UL) : TASK_SIZE)
+#ifdef __KERNEL__
+
+#define STACK_TOP32	((1UL << 32UL) - PAGE_SIZE)
+#define STACK_TOP64	(0x0000080000000000UL - (1UL << 32UL))
+
+#define STACK_TOP	(test_thread_flag(TIF_32BIT) ? \
+			 STACK_TOP32 : STACK_TOP64)
+
+#define STACK_TOP_MAX	STACK_TOP64
+
+#endif
+
+#ifndef __ASSEMBLY__
+
+typedef struct {
+	unsigned char seg;
+} mm_segment_t;
+
+/* The Sparc processor specific thread struct. */
+/* XXX This should die, everything can go into thread_info now. */
+struct thread_struct {
+#ifdef CONFIG_DEBUG_SPINLOCK
+	/* How many spinlocks held by this thread.
+	 * Used with spin lock debugging to catch tasks
+	 * sleeping illegally with locks held.
+	 */
+	int smp_lock_count;
+	unsigned int smp_lock_pc;
+#else
+	int dummy; /* f'in gcc bug... */
+#endif
+};
+
+#endif /* !(__ASSEMBLY__) */
+
+#ifndef CONFIG_DEBUG_SPINLOCK
+#define INIT_THREAD  {			\
+	0,				\
+}
+#else /* CONFIG_DEBUG_SPINLOCK */
+#define INIT_THREAD  {					\
+/* smp_lock_count, smp_lock_pc, */			\
+   0,		   0,					\
+}
+#endif /* !(CONFIG_DEBUG_SPINLOCK) */
+
+#ifndef __ASSEMBLY__
+
+#include <linux/types.h>
+
+/* Return saved PC of a blocked thread. */
+struct task_struct;
+extern unsigned long thread_saved_pc(struct task_struct *);
+
+/* On Uniprocessor, even in RMO processes see TSO semantics */
+#ifdef CONFIG_SMP
+#define TSTATE_INITIAL_MM	TSTATE_TSO
+#else
+#define TSTATE_INITIAL_MM	TSTATE_RMO
+#endif
+
+/* Do necessary setup to start up a newly executed thread. */
+#define start_thread(regs, pc, sp) \
+do { \
+	unsigned long __asi = ASI_PNF; \
+	regs->tstate = (regs->tstate & (TSTATE_CWP)) | (TSTATE_INITIAL_MM|TSTATE_IE) | (__asi << 24UL); \
+	regs->tpc = ((pc & (~3)) - 4); \
+	regs->tnpc = regs->tpc + 4; \
+	regs->y = 0; \
+	set_thread_wstate(1 << 3); \
+	if (current_thread_info()->utraps) { \
+		if (*(current_thread_info()->utraps) < 2) \
+			kfree(current_thread_info()->utraps); \
+		else \
+			(*(current_thread_info()->utraps))--; \
+		current_thread_info()->utraps = NULL; \
+	} \
+	__asm__ __volatile__( \
+	"stx		%%g0, [%0 + %2 + 0x00]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x08]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x10]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x18]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x20]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x28]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x30]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x38]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x40]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x48]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x50]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x58]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x60]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x68]\n\t" \
+	"stx		%1,   [%0 + %2 + 0x70]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x78]\n\t" \
+	"wrpr		%%g0, (1 << 3), %%wstate\n\t" \
+	: \
+	: "r" (regs), "r" (sp - sizeof(struct reg_window) - STACK_BIAS), \
+	  "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))); \
+} while (0)
+
+#define start_thread32(regs, pc, sp) \
+do { \
+	unsigned long __asi = ASI_PNF; \
+	pc &= 0x00000000ffffffffUL; \
+	sp &= 0x00000000ffffffffUL; \
+	regs->tstate = (regs->tstate & (TSTATE_CWP))|(TSTATE_INITIAL_MM|TSTATE_IE|TSTATE_AM) | (__asi << 24UL); \
+	regs->tpc = ((pc & (~3)) - 4); \
+	regs->tnpc = regs->tpc + 4; \
+	regs->y = 0; \
+	set_thread_wstate(2 << 3); \
+	if (current_thread_info()->utraps) { \
+		if (*(current_thread_info()->utraps) < 2) \
+			kfree(current_thread_info()->utraps); \
+		else \
+			(*(current_thread_info()->utraps))--; \
+		current_thread_info()->utraps = NULL; \
+	} \
+	__asm__ __volatile__( \
+	"stx		%%g0, [%0 + %2 + 0x00]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x08]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x10]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x18]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x20]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x28]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x30]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x38]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x40]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x48]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x50]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x58]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x60]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x68]\n\t" \
+	"stx		%1,   [%0 + %2 + 0x70]\n\t" \
+	"stx		%%g0, [%0 + %2 + 0x78]\n\t" \
+	"wrpr		%%g0, (2 << 3), %%wstate\n\t" \
+	: \
+	: "r" (regs), "r" (sp - sizeof(struct reg_window32)), \
+	  "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))); \
+} while (0)
+
+/* Free all resources held by a thread. */
+#define release_thread(tsk)		do { } while (0)
+
+/* Prepare to copy thread state - unlazy all lazy status */
+#define prepare_to_copy(tsk)	do { } while (0)
+
+extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
+
+extern unsigned long get_wchan(struct task_struct *task);
+
+#define task_pt_regs(tsk) (task_thread_info(tsk)->kregs)
+#define KSTK_EIP(tsk)  (task_pt_regs(tsk)->tpc)
+#define KSTK_ESP(tsk)  (task_pt_regs(tsk)->u_regs[UREG_FP])
+
+#define cpu_relax()	barrier()
+
+/* Prefetch support.  This is tuned for UltraSPARC-III and later.
+ * UltraSPARC-I will treat these as nops, and UltraSPARC-II has
+ * a shallower prefetch queue than later chips.
+ */
+#define ARCH_HAS_PREFETCH
+#define ARCH_HAS_PREFETCHW
+#define ARCH_HAS_SPINLOCK_PREFETCH
+
+static inline void prefetch(const void *x)
+{
+	/* We do not use the read prefetch mnemonic because that
+	 * prefetches into the prefetch-cache which only is accessible
+	 * by floating point operations in UltraSPARC-III and later.
+	 * By contrast, "#one_write" prefetches into the L2 cache
+	 * in shared state.
+	 */
+	__asm__ __volatile__("prefetch [%0], #one_write"
+			     : /* no outputs */
+			     : "r" (x));
+}
+
+static inline void prefetchw(const void *x)
+{
+	/* The most optimal prefetch to use for writes is
+	 * "#n_writes".  This brings the cacheline into the
+	 * L2 cache in "owned" state.
+	 */
+	__asm__ __volatile__("prefetch [%0], #n_writes"
+			     : /* no outputs */
+			     : "r" (x));
+}
+
+#define spin_lock_prefetch(x)	prefetchw(x)
+
+#define HAVE_ARCH_PICK_MMAP_LAYOUT
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__ASM_SPARC64_PROCESSOR_H) */
diff --git a/include/asm-sparc/psrcompat.h b/include/asm-sparc/psrcompat.h
new file mode 100644
index 00000000000..44b6327dbbf
--- /dev/null
+++ b/include/asm-sparc/psrcompat.h
@@ -0,0 +1,45 @@
+#ifndef _SPARC64_PSRCOMPAT_H
+#define _SPARC64_PSRCOMPAT_H
+
+#include <asm/pstate.h>
+
+/* Old 32-bit PSR fields for the compatibility conversion code. */
+#define PSR_CWP     0x0000001f         /* current window pointer     */
+#define PSR_ET      0x00000020         /* enable traps field         */
+#define PSR_PS      0x00000040         /* previous privilege level   */
+#define PSR_S       0x00000080         /* current privilege level    */
+#define PSR_PIL     0x00000f00         /* processor interrupt level  */
+#define PSR_EF      0x00001000         /* enable floating point      */
+#define PSR_EC      0x00002000         /* enable co-processor        */
+#define PSR_SYSCALL 0x00004000         /* inside of a syscall        */
+#define PSR_LE      0x00008000         /* SuperSparcII little-endian */
+#define PSR_ICC     0x00f00000         /* integer condition codes    */
+#define PSR_C       0x00100000         /* carry bit                  */
+#define PSR_V       0x00200000         /* overflow bit               */
+#define PSR_Z       0x00400000         /* zero bit                   */
+#define PSR_N       0x00800000         /* negative bit               */
+#define PSR_VERS    0x0f000000         /* cpu-version field          */
+#define PSR_IMPL    0xf0000000         /* cpu-implementation field   */
+
+#define PSR_V8PLUS  0xff000000         /* fake impl/ver, meaning a 64bit CPU is present */
+#define PSR_XCC	    0x000f0000         /* if PSR_V8PLUS, this is %xcc */
+
+static inline unsigned int tstate_to_psr(unsigned long tstate)
+{
+	return ((tstate & TSTATE_CWP)			|
+		PSR_S					|
+		((tstate & TSTATE_ICC) >> 12)		|
+		((tstate & TSTATE_XCC) >> 20)		|
+		((tstate & TSTATE_SYSCALL) ? PSR_SYSCALL : 0) |
+		PSR_V8PLUS);
+}
+
+static inline unsigned long psr_to_tstate_icc(unsigned int psr)
+{
+	unsigned long tstate = ((unsigned long)(psr & PSR_ICC)) << 12;
+	if ((psr & (PSR_VERS|PSR_IMPL)) == PSR_V8PLUS)
+		tstate |= ((unsigned long)(psr & PSR_XCC)) << 20;
+	return tstate;
+}
+
+#endif /* !(_SPARC64_PSRCOMPAT_H) */
diff --git a/include/asm-sparc/pstate.h b/include/asm-sparc/pstate.h
new file mode 100644
index 00000000000..a26a53777bb
--- /dev/null
+++ b/include/asm-sparc/pstate.h
@@ -0,0 +1,91 @@
+#ifndef _SPARC64_PSTATE_H
+#define _SPARC64_PSTATE_H
+
+#include <linux/const.h>
+
+/* The V9 PSTATE Register (with SpitFire extensions).
+ *
+ * -----------------------------------------------------------------------
+ * | Resv | IG | MG | CLE | TLE |  MM  | RED | PEF | AM | PRIV | IE | AG |
+ * -----------------------------------------------------------------------
+ *  63  12  11   10    9     8    7   6   5     4     3     2     1    0
+ */
+#define PSTATE_IG   _AC(0x0000000000000800,UL) /* Interrupt Globals.	*/
+#define PSTATE_MG   _AC(0x0000000000000400,UL) /* MMU Globals.		*/
+#define PSTATE_CLE  _AC(0x0000000000000200,UL) /* Current Little Endian.*/
+#define PSTATE_TLE  _AC(0x0000000000000100,UL) /* Trap Little Endian.	*/
+#define PSTATE_MM   _AC(0x00000000000000c0,UL) /* Memory Model.		*/
+#define PSTATE_TSO  _AC(0x0000000000000000,UL) /* MM: TotalStoreOrder	*/
+#define PSTATE_PSO  _AC(0x0000000000000040,UL) /* MM: PartialStoreOrder	*/
+#define PSTATE_RMO  _AC(0x0000000000000080,UL) /* MM: RelaxedMemoryOrder*/
+#define PSTATE_RED  _AC(0x0000000000000020,UL) /* Reset Error Debug.	*/
+#define PSTATE_PEF  _AC(0x0000000000000010,UL) /* Floating Point Enable.*/
+#define PSTATE_AM   _AC(0x0000000000000008,UL) /* Address Mask.		*/
+#define PSTATE_PRIV _AC(0x0000000000000004,UL) /* Privilege.		*/
+#define PSTATE_IE   _AC(0x0000000000000002,UL) /* Interrupt Enable.	*/
+#define PSTATE_AG   _AC(0x0000000000000001,UL) /* Alternate Globals.	*/
+
+/* The V9 TSTATE Register (with SpitFire and Linux extensions).
+ *
+ * ---------------------------------------------------------------------
+ * |  Resv |  GL  |  CCR  |  ASI  |  %pil  |  PSTATE  |  Resv  |  CWP  |
+ * ---------------------------------------------------------------------
+ *  63   43 42  40 39   32 31   24 23    20 19       8 7      5 4     0
+ */
+#define TSTATE_GL	_AC(0x0000070000000000,UL) /* Global reg level  */
+#define TSTATE_CCR	_AC(0x000000ff00000000,UL) /* Condition Codes.	*/
+#define TSTATE_XCC	_AC(0x000000f000000000,UL) /* Condition Codes.	*/
+#define TSTATE_XNEG	_AC(0x0000008000000000,UL) /* %xcc Negative.	*/
+#define TSTATE_XZERO	_AC(0x0000004000000000,UL) /* %xcc Zero.	*/
+#define TSTATE_XOVFL	_AC(0x0000002000000000,UL) /* %xcc Overflow.	*/
+#define TSTATE_XCARRY	_AC(0x0000001000000000,UL) /* %xcc Carry.	*/
+#define TSTATE_ICC	_AC(0x0000000f00000000,UL) /* Condition Codes.	*/
+#define TSTATE_INEG	_AC(0x0000000800000000,UL) /* %icc Negative.	*/
+#define TSTATE_IZERO	_AC(0x0000000400000000,UL) /* %icc Zero.	*/
+#define TSTATE_IOVFL	_AC(0x0000000200000000,UL) /* %icc Overflow.	*/
+#define TSTATE_ICARRY	_AC(0x0000000100000000,UL) /* %icc Carry.	*/
+#define TSTATE_ASI	_AC(0x00000000ff000000,UL) /* AddrSpace ID.	*/
+#define TSTATE_PIL	_AC(0x0000000000f00000,UL) /* %pil (Linux traps)*/
+#define TSTATE_PSTATE	_AC(0x00000000000fff00,UL) /* PSTATE.		*/
+#define TSTATE_IG	_AC(0x0000000000080000,UL) /* Interrupt Globals.*/
+#define TSTATE_MG	_AC(0x0000000000040000,UL) /* MMU Globals.	*/
+#define TSTATE_CLE	_AC(0x0000000000020000,UL) /* CurrLittleEndian.	*/
+#define TSTATE_TLE	_AC(0x0000000000010000,UL) /* TrapLittleEndian.	*/
+#define TSTATE_MM	_AC(0x000000000000c000,UL) /* Memory Model.	*/
+#define TSTATE_TSO	_AC(0x0000000000000000,UL) /* MM: TSO		*/
+#define TSTATE_PSO	_AC(0x0000000000004000,UL) /* MM: PSO		*/
+#define TSTATE_RMO	_AC(0x0000000000008000,UL) /* MM: RMO		*/
+#define TSTATE_RED	_AC(0x0000000000002000,UL) /* Reset Error Debug.*/
+#define TSTATE_PEF	_AC(0x0000000000001000,UL) /* FPU Enable.	*/
+#define TSTATE_AM	_AC(0x0000000000000800,UL) /* Address Mask.	*/
+#define TSTATE_PRIV	_AC(0x0000000000000400,UL) /* Privilege.	*/
+#define TSTATE_IE	_AC(0x0000000000000200,UL) /* Interrupt Enable.	*/
+#define TSTATE_AG	_AC(0x0000000000000100,UL) /* Alternate Globals.*/
+#define TSTATE_SYSCALL	_AC(0x0000000000000020,UL) /* in syscall trap   */
+#define TSTATE_CWP	_AC(0x000000000000001f,UL) /* Curr Win-Pointer.	*/
+
+/* Floating-Point Registers State Register.
+ *
+ * --------------------------------
+ * |  Resv  |  FEF  |  DU  |  DL  |
+ * --------------------------------
+ *  63     3    2       1      0
+ */
+#define FPRS_FEF	_AC(0x0000000000000004,UL) /* FPU Enable.	*/
+#define FPRS_DU		_AC(0x0000000000000002,UL) /* Dirty Upper.	*/
+#define FPRS_DL		_AC(0x0000000000000001,UL) /* Dirty Lower.	*/
+
+/* Version Register.
+ *
+ * ------------------------------------------------------
+ * | MANUF | IMPL | MASK | Resv | MAXTL | Resv | MAXWIN |
+ * ------------------------------------------------------
+ *  63   48 47  32 31  24 23  16 15    8 7    5 4      0
+ */
+#define VERS_MANUF	_AC(0xffff000000000000,UL) /* Manufacturer.	*/
+#define VERS_IMPL	_AC(0x0000ffff00000000,UL) /* Implementation.	*/
+#define VERS_MASK	_AC(0x00000000ff000000,UL) /* Mask Set Revision.*/
+#define VERS_MAXTL	_AC(0x000000000000ff00,UL) /* Max Trap Level.	*/
+#define VERS_MAXWIN	_AC(0x000000000000001f,UL) /* Max RegWindow Idx.*/
+
+#endif /* !(_SPARC64_PSTATE_H) */
diff --git a/include/asm-sparc/ptrace.h b/include/asm-sparc/ptrace.h
index 11f3bc2bb3f..f36ab6c30ff 100644
--- a/include/asm-sparc/ptrace.h
+++ b/include/asm-sparc/ptrace.h
@@ -1,175 +1,8 @@
-#ifndef _SPARC_PTRACE_H
-#define _SPARC_PTRACE_H
-
-#include <asm/psr.h>
-
-/* This struct defines the way the registers are stored on the 
- * stack during a system call and basically all traps.
- */
-
-#ifndef __ASSEMBLY__
-
-#include <linux/types.h>
-
-struct pt_regs {
-	unsigned long psr;
-	unsigned long pc;
-	unsigned long npc;
-	unsigned long y;
-	unsigned long u_regs[16]; /* globals and ins */
-};
-
-#define UREG_G0        0
-#define UREG_G1        1
-#define UREG_G2        2
-#define UREG_G3        3
-#define UREG_G4        4
-#define UREG_G5        5
-#define UREG_G6        6
-#define UREG_G7        7
-#define UREG_I0        8
-#define UREG_I1        9
-#define UREG_I2        10
-#define UREG_I3        11
-#define UREG_I4        12
-#define UREG_I5        13
-#define UREG_I6        14
-#define UREG_I7        15
-#define UREG_WIM       UREG_G0
-#define UREG_FADDR     UREG_G0
-#define UREG_FP        UREG_I6
-#define UREG_RETPC     UREG_I7
-
-static inline bool pt_regs_is_syscall(struct pt_regs *regs)
-{
-	return (regs->psr & PSR_SYSCALL);
-}
-
-static inline bool pt_regs_clear_syscall(struct pt_regs *regs)
-{
-	return (regs->psr &= ~PSR_SYSCALL);
-}
-
-/* A register window */
-struct reg_window {
-	unsigned long locals[8];
-	unsigned long ins[8];
-};
-
-/* A Sparc stack frame */
-struct sparc_stackf {
-	unsigned long locals[8];
-        unsigned long ins[6];
-	struct sparc_stackf *fp;
-	unsigned long callers_pc;
-	char *structptr;
-	unsigned long xargs[6];
-	unsigned long xxargs[1];
-};	
-
-#define TRACEREG_SZ   sizeof(struct pt_regs)
-#define STACKFRAME_SZ sizeof(struct sparc_stackf)
-
-#ifdef __KERNEL__
-
-#define user_mode(regs) (!((regs)->psr & PSR_PS))
-#define instruction_pointer(regs) ((regs)->pc)
-unsigned long profile_pc(struct pt_regs *);
-extern void show_regs(struct pt_regs *);
+#ifndef ___ASM_SPARC_PTRACE_H
+#define ___ASM_SPARC_PTRACE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/ptrace_64.h>
+#else
+#include <asm-sparc/ptrace_32.h>
 #endif
-
-#else /* __ASSEMBLY__ */
-/* For assembly code. */
-#define TRACEREG_SZ       0x50
-#define STACKFRAME_SZ     0x60
 #endif
-
-/*
- * The asm-offsets.h is a generated file, so we cannot include it.
- * It may be OK for glibc headers, but it's utterly pointless for C code.
- * The assembly code using those offsets has to include it explicitly.
- */
-/* #include <asm/asm-offsets.h> */
-
-/* These are for pt_regs. */
-#define PT_PSR    0x0
-#define PT_PC     0x4
-#define PT_NPC    0x8
-#define PT_Y      0xc
-#define PT_G0     0x10
-#define PT_WIM    PT_G0
-#define PT_G1     0x14
-#define PT_G2     0x18
-#define PT_G3     0x1c
-#define PT_G4     0x20
-#define PT_G5     0x24
-#define PT_G6     0x28
-#define PT_G7     0x2c
-#define PT_I0     0x30
-#define PT_I1     0x34
-#define PT_I2     0x38
-#define PT_I3     0x3c
-#define PT_I4     0x40
-#define PT_I5     0x44
-#define PT_I6     0x48
-#define PT_FP     PT_I6
-#define PT_I7     0x4c
-
-/* Reg_window offsets */
-#define RW_L0     0x00
-#define RW_L1     0x04
-#define RW_L2     0x08
-#define RW_L3     0x0c
-#define RW_L4     0x10
-#define RW_L5     0x14
-#define RW_L6     0x18
-#define RW_L7     0x1c
-#define RW_I0     0x20
-#define RW_I1     0x24
-#define RW_I2     0x28
-#define RW_I3     0x2c
-#define RW_I4     0x30
-#define RW_I5     0x34
-#define RW_I6     0x38
-#define RW_I7     0x3c
-
-/* Stack_frame offsets */
-#define SF_L0     0x00
-#define SF_L1     0x04
-#define SF_L2     0x08
-#define SF_L3     0x0c
-#define SF_L4     0x10
-#define SF_L5     0x14
-#define SF_L6     0x18
-#define SF_L7     0x1c
-#define SF_I0     0x20
-#define SF_I1     0x24
-#define SF_I2     0x28
-#define SF_I3     0x2c
-#define SF_I4     0x30
-#define SF_I5     0x34
-#define SF_FP     0x38
-#define SF_PC     0x3c
-#define SF_RETP   0x40
-#define SF_XARG0  0x44
-#define SF_XARG1  0x48
-#define SF_XARG2  0x4c
-#define SF_XARG3  0x50
-#define SF_XARG4  0x54
-#define SF_XARG5  0x58
-#define SF_XXARG  0x5c
-
-/* Stuff for the ptrace system call */
-#define PTRACE_SPARC_DETACH       11
-#define PTRACE_GETREGS            12
-#define PTRACE_SETREGS            13
-#define PTRACE_GETFPREGS          14
-#define PTRACE_SETFPREGS          15
-#define PTRACE_READDATA           16
-#define PTRACE_WRITEDATA          17
-#define PTRACE_READTEXT           18
-#define PTRACE_WRITETEXT          19
-#define PTRACE_GETFPAREGS         20
-#define PTRACE_SETFPAREGS         21
-
-#endif /* !(_SPARC_PTRACE_H) */
diff --git a/include/asm-sparc/ptrace_32.h b/include/asm-sparc/ptrace_32.h
new file mode 100644
index 00000000000..0401cc7ec38
--- /dev/null
+++ b/include/asm-sparc/ptrace_32.h
@@ -0,0 +1,175 @@
+#ifndef _SPARC_PTRACE_H
+#define _SPARC_PTRACE_H
+
+#include <asm/psr.h>
+
+/* This struct defines the way the registers are stored on the
+ * stack during a system call and basically all traps.
+ */
+
+#ifndef __ASSEMBLY__
+
+#include <linux/types.h>
+
+struct pt_regs {
+	unsigned long psr;
+	unsigned long pc;
+	unsigned long npc;
+	unsigned long y;
+	unsigned long u_regs[16]; /* globals and ins */
+};
+
+#define UREG_G0        0
+#define UREG_G1        1
+#define UREG_G2        2
+#define UREG_G3        3
+#define UREG_G4        4
+#define UREG_G5        5
+#define UREG_G6        6
+#define UREG_G7        7
+#define UREG_I0        8
+#define UREG_I1        9
+#define UREG_I2        10
+#define UREG_I3        11
+#define UREG_I4        12
+#define UREG_I5        13
+#define UREG_I6        14
+#define UREG_I7        15
+#define UREG_WIM       UREG_G0
+#define UREG_FADDR     UREG_G0
+#define UREG_FP        UREG_I6
+#define UREG_RETPC     UREG_I7
+
+static inline bool pt_regs_is_syscall(struct pt_regs *regs)
+{
+	return (regs->psr & PSR_SYSCALL);
+}
+
+static inline bool pt_regs_clear_syscall(struct pt_regs *regs)
+{
+	return (regs->psr &= ~PSR_SYSCALL);
+}
+
+/* A register window */
+struct reg_window {
+	unsigned long locals[8];
+	unsigned long ins[8];
+};
+
+/* A Sparc stack frame */
+struct sparc_stackf {
+	unsigned long locals[8];
+        unsigned long ins[6];
+	struct sparc_stackf *fp;
+	unsigned long callers_pc;
+	char *structptr;
+	unsigned long xargs[6];
+	unsigned long xxargs[1];
+};
+
+#define TRACEREG_SZ   sizeof(struct pt_regs)
+#define STACKFRAME_SZ sizeof(struct sparc_stackf)
+
+#ifdef __KERNEL__
+
+#define user_mode(regs) (!((regs)->psr & PSR_PS))
+#define instruction_pointer(regs) ((regs)->pc)
+unsigned long profile_pc(struct pt_regs *);
+extern void show_regs(struct pt_regs *);
+#endif
+
+#else /* __ASSEMBLY__ */
+/* For assembly code. */
+#define TRACEREG_SZ       0x50
+#define STACKFRAME_SZ     0x60
+#endif
+
+/*
+ * The asm-offsets.h is a generated file, so we cannot include it.
+ * It may be OK for glibc headers, but it's utterly pointless for C code.
+ * The assembly code using those offsets has to include it explicitly.
+ */
+/* #include <asm/asm-offsets.h> */
+
+/* These are for pt_regs. */
+#define PT_PSR    0x0
+#define PT_PC     0x4
+#define PT_NPC    0x8
+#define PT_Y      0xc
+#define PT_G0     0x10
+#define PT_WIM    PT_G0
+#define PT_G1     0x14
+#define PT_G2     0x18
+#define PT_G3     0x1c
+#define PT_G4     0x20
+#define PT_G5     0x24
+#define PT_G6     0x28
+#define PT_G7     0x2c
+#define PT_I0     0x30
+#define PT_I1     0x34
+#define PT_I2     0x38
+#define PT_I3     0x3c
+#define PT_I4     0x40
+#define PT_I5     0x44
+#define PT_I6     0x48
+#define PT_FP     PT_I6
+#define PT_I7     0x4c
+
+/* Reg_window offsets */
+#define RW_L0     0x00
+#define RW_L1     0x04
+#define RW_L2     0x08
+#define RW_L3     0x0c
+#define RW_L4     0x10
+#define RW_L5     0x14
+#define RW_L6     0x18
+#define RW_L7     0x1c
+#define RW_I0     0x20
+#define RW_I1     0x24
+#define RW_I2     0x28
+#define RW_I3     0x2c
+#define RW_I4     0x30
+#define RW_I5     0x34
+#define RW_I6     0x38
+#define RW_I7     0x3c
+
+/* Stack_frame offsets */
+#define SF_L0     0x00
+#define SF_L1     0x04
+#define SF_L2     0x08
+#define SF_L3     0x0c
+#define SF_L4     0x10
+#define SF_L5     0x14
+#define SF_L6     0x18
+#define SF_L7     0x1c
+#define SF_I0     0x20
+#define SF_I1     0x24
+#define SF_I2     0x28
+#define SF_I3     0x2c
+#define SF_I4     0x30
+#define SF_I5     0x34
+#define SF_FP     0x38
+#define SF_PC     0x3c
+#define SF_RETP   0x40
+#define SF_XARG0  0x44
+#define SF_XARG1  0x48
+#define SF_XARG2  0x4c
+#define SF_XARG3  0x50
+#define SF_XARG4  0x54
+#define SF_XARG5  0x58
+#define SF_XXARG  0x5c
+
+/* Stuff for the ptrace system call */
+#define PTRACE_SPARC_DETACH       11
+#define PTRACE_GETREGS            12
+#define PTRACE_SETREGS            13
+#define PTRACE_GETFPREGS          14
+#define PTRACE_SETFPREGS          15
+#define PTRACE_READDATA           16
+#define PTRACE_WRITEDATA          17
+#define PTRACE_READTEXT           18
+#define PTRACE_WRITETEXT          19
+#define PTRACE_GETFPAREGS         20
+#define PTRACE_SETFPAREGS         21
+
+#endif /* !(_SPARC_PTRACE_H) */
diff --git a/include/asm-sparc/ptrace_64.h b/include/asm-sparc/ptrace_64.h
new file mode 100644
index 00000000000..a682e66d5c4
--- /dev/null
+++ b/include/asm-sparc/ptrace_64.h
@@ -0,0 +1,346 @@
+#ifndef _SPARC64_PTRACE_H
+#define _SPARC64_PTRACE_H
+
+#include <asm/pstate.h>
+
+/* This struct defines the way the registers are stored on the
+ * stack during a system call and basically all traps.
+ */
+
+/* This magic value must have the low 9 bits clear,
+ * as that is where we encode the %tt value, see below.
+ */
+#define PT_REGS_MAGIC 0x57ac6c00
+
+#ifndef __ASSEMBLY__
+
+#include <linux/types.h>
+
+struct pt_regs {
+	unsigned long u_regs[16]; /* globals and ins */
+	unsigned long tstate;
+	unsigned long tpc;
+	unsigned long tnpc;
+	unsigned int y;
+
+	/* We encode a magic number, PT_REGS_MAGIC, along
+	 * with the %tt (trap type) register value at trap
+	 * entry time.  The magic number allows us to identify
+	 * accurately a trap stack frame in the stack
+	 * unwinder, and the %tt value allows us to test
+	 * things like "in a system call" etc. for an arbitray
+	 * process.
+	 *
+	 * The PT_REGS_MAGIC is choosen such that it can be
+	 * loaded completely using just a sethi instruction.
+	 */
+	unsigned int magic;
+};
+
+static inline int pt_regs_trap_type(struct pt_regs *regs)
+{
+	return regs->magic & 0x1ff;
+}
+
+static inline bool pt_regs_is_syscall(struct pt_regs *regs)
+{
+	return (regs->tstate & TSTATE_SYSCALL);
+}
+
+static inline bool pt_regs_clear_syscall(struct pt_regs *regs)
+{
+	return (regs->tstate &= ~TSTATE_SYSCALL);
+}
+
+struct pt_regs32 {
+	unsigned int psr;
+	unsigned int pc;
+	unsigned int npc;
+	unsigned int y;
+	unsigned int u_regs[16]; /* globals and ins */
+};
+
+#define UREG_G0        0
+#define UREG_G1        1
+#define UREG_G2        2
+#define UREG_G3        3
+#define UREG_G4        4
+#define UREG_G5        5
+#define UREG_G6        6
+#define UREG_G7        7
+#define UREG_I0        8
+#define UREG_I1        9
+#define UREG_I2        10
+#define UREG_I3        11
+#define UREG_I4        12
+#define UREG_I5        13
+#define UREG_I6        14
+#define UREG_I7        15
+#define UREG_FP        UREG_I6
+#define UREG_RETPC     UREG_I7
+
+/* A V9 register window */
+struct reg_window {
+	unsigned long locals[8];
+	unsigned long ins[8];
+};
+
+/* A 32-bit register window. */
+struct reg_window32 {
+	unsigned int locals[8];
+	unsigned int ins[8];
+};
+
+/* A V9 Sparc stack frame */
+struct sparc_stackf {
+	unsigned long locals[8];
+        unsigned long ins[6];
+	struct sparc_stackf *fp;
+	unsigned long callers_pc;
+	char *structptr;
+	unsigned long xargs[6];
+	unsigned long xxargs[1];
+};
+
+/* A 32-bit Sparc stack frame */
+struct sparc_stackf32 {
+	unsigned int locals[8];
+        unsigned int ins[6];
+	unsigned int fp;
+	unsigned int callers_pc;
+	unsigned int structptr;
+	unsigned int xargs[6];
+	unsigned int xxargs[1];
+};
+
+struct sparc_trapf {
+	unsigned long locals[8];
+	unsigned long ins[8];
+	unsigned long _unused;
+	struct pt_regs *regs;
+};
+
+#define TRACEREG_SZ	sizeof(struct pt_regs)
+#define STACKFRAME_SZ	sizeof(struct sparc_stackf)
+
+#define TRACEREG32_SZ	sizeof(struct pt_regs32)
+#define STACKFRAME32_SZ	sizeof(struct sparc_stackf32)
+
+#ifdef __KERNEL__
+
+struct global_reg_snapshot {
+	unsigned long		tstate;
+	unsigned long		tpc;
+	unsigned long		tnpc;
+	unsigned long		o7;
+	unsigned long		i7;
+	struct thread_info	*thread;
+	unsigned long		pad1;
+	unsigned long		pad2;
+};
+
+#define __ARCH_WANT_COMPAT_SYS_PTRACE
+
+#define force_successful_syscall_return()	    \
+do {	current_thread_info()->syscall_noerror = 1; \
+} while (0)
+#define user_mode(regs) (!((regs)->tstate & TSTATE_PRIV))
+#define instruction_pointer(regs) ((regs)->tpc)
+#define regs_return_value(regs) ((regs)->u_regs[UREG_I0])
+#ifdef CONFIG_SMP
+extern unsigned long profile_pc(struct pt_regs *);
+#else
+#define profile_pc(regs) instruction_pointer(regs)
+#endif
+extern void show_regs(struct pt_regs *);
+extern void __show_regs(struct pt_regs *);
+#endif
+
+#else /* __ASSEMBLY__ */
+/* For assembly code. */
+#define TRACEREG_SZ		0xa0
+#define STACKFRAME_SZ		0xc0
+
+#define TRACEREG32_SZ		0x50
+#define STACKFRAME32_SZ		0x60
+#endif
+
+#ifdef __KERNEL__
+#define STACK_BIAS		2047
+#endif
+
+/* These are for pt_regs. */
+#define PT_V9_G0     0x00
+#define PT_V9_G1     0x08
+#define PT_V9_G2     0x10
+#define PT_V9_G3     0x18
+#define PT_V9_G4     0x20
+#define PT_V9_G5     0x28
+#define PT_V9_G6     0x30
+#define PT_V9_G7     0x38
+#define PT_V9_I0     0x40
+#define PT_V9_I1     0x48
+#define PT_V9_I2     0x50
+#define PT_V9_I3     0x58
+#define PT_V9_I4     0x60
+#define PT_V9_I5     0x68
+#define PT_V9_I6     0x70
+#define PT_V9_FP     PT_V9_I6
+#define PT_V9_I7     0x78
+#define PT_V9_TSTATE 0x80
+#define PT_V9_TPC    0x88
+#define PT_V9_TNPC   0x90
+#define PT_V9_Y      0x98
+#define PT_V9_MAGIC  0x9c
+#define PT_TSTATE	PT_V9_TSTATE
+#define PT_TPC		PT_V9_TPC
+#define PT_TNPC		PT_V9_TNPC
+
+/* These for pt_regs32. */
+#define PT_PSR    0x0
+#define PT_PC     0x4
+#define PT_NPC    0x8
+#define PT_Y      0xc
+#define PT_G0     0x10
+#define PT_WIM    PT_G0
+#define PT_G1     0x14
+#define PT_G2     0x18
+#define PT_G3     0x1c
+#define PT_G4     0x20
+#define PT_G5     0x24
+#define PT_G6     0x28
+#define PT_G7     0x2c
+#define PT_I0     0x30
+#define PT_I1     0x34
+#define PT_I2     0x38
+#define PT_I3     0x3c
+#define PT_I4     0x40
+#define PT_I5     0x44
+#define PT_I6     0x48
+#define PT_FP     PT_I6
+#define PT_I7     0x4c
+
+/* Reg_window offsets */
+#define RW_V9_L0     0x00
+#define RW_V9_L1     0x08
+#define RW_V9_L2     0x10
+#define RW_V9_L3     0x18
+#define RW_V9_L4     0x20
+#define RW_V9_L5     0x28
+#define RW_V9_L6     0x30
+#define RW_V9_L7     0x38
+#define RW_V9_I0     0x40
+#define RW_V9_I1     0x48
+#define RW_V9_I2     0x50
+#define RW_V9_I3     0x58
+#define RW_V9_I4     0x60
+#define RW_V9_I5     0x68
+#define RW_V9_I6     0x70
+#define RW_V9_I7     0x78
+
+#define RW_L0     0x00
+#define RW_L1     0x04
+#define RW_L2     0x08
+#define RW_L3     0x0c
+#define RW_L4     0x10
+#define RW_L5     0x14
+#define RW_L6     0x18
+#define RW_L7     0x1c
+#define RW_I0     0x20
+#define RW_I1     0x24
+#define RW_I2     0x28
+#define RW_I3     0x2c
+#define RW_I4     0x30
+#define RW_I5     0x34
+#define RW_I6     0x38
+#define RW_I7     0x3c
+
+/* Stack_frame offsets */
+#define SF_V9_L0     0x00
+#define SF_V9_L1     0x08
+#define SF_V9_L2     0x10
+#define SF_V9_L3     0x18
+#define SF_V9_L4     0x20
+#define SF_V9_L5     0x28
+#define SF_V9_L6     0x30
+#define SF_V9_L7     0x38
+#define SF_V9_I0     0x40
+#define SF_V9_I1     0x48
+#define SF_V9_I2     0x50
+#define SF_V9_I3     0x58
+#define SF_V9_I4     0x60
+#define SF_V9_I5     0x68
+#define SF_V9_FP     0x70
+#define SF_V9_PC     0x78
+#define SF_V9_RETP   0x80
+#define SF_V9_XARG0  0x88
+#define SF_V9_XARG1  0x90
+#define SF_V9_XARG2  0x98
+#define SF_V9_XARG3  0xa0
+#define SF_V9_XARG4  0xa8
+#define SF_V9_XARG5  0xb0
+#define SF_V9_XXARG  0xb8
+
+#define SF_L0     0x00
+#define SF_L1     0x04
+#define SF_L2     0x08
+#define SF_L3     0x0c
+#define SF_L4     0x10
+#define SF_L5     0x14
+#define SF_L6     0x18
+#define SF_L7     0x1c
+#define SF_I0     0x20
+#define SF_I1     0x24
+#define SF_I2     0x28
+#define SF_I3     0x2c
+#define SF_I4     0x30
+#define SF_I5     0x34
+#define SF_FP     0x38
+#define SF_PC     0x3c
+#define SF_RETP   0x40
+#define SF_XARG0  0x44
+#define SF_XARG1  0x48
+#define SF_XARG2  0x4c
+#define SF_XARG3  0x50
+#define SF_XARG4  0x54
+#define SF_XARG5  0x58
+#define SF_XXARG  0x5c
+
+#ifdef __KERNEL__
+
+/* global_reg_snapshot offsets */
+#define GR_SNAP_TSTATE	0x00
+#define GR_SNAP_TPC	0x08
+#define GR_SNAP_TNPC	0x10
+#define GR_SNAP_O7	0x18
+#define GR_SNAP_I7	0x20
+#define GR_SNAP_THREAD	0x28
+#define GR_SNAP_PAD1	0x30
+#define GR_SNAP_PAD2	0x38
+
+#endif  /*  __KERNEL__  */
+
+/* Stuff for the ptrace system call */
+#define PTRACE_SPARC_DETACH       11
+#define PTRACE_GETREGS            12
+#define PTRACE_SETREGS            13
+#define PTRACE_GETFPREGS          14
+#define PTRACE_SETFPREGS          15
+#define PTRACE_READDATA           16
+#define PTRACE_WRITEDATA          17
+#define PTRACE_READTEXT           18
+#define PTRACE_WRITETEXT          19
+#define PTRACE_GETFPAREGS         20
+#define PTRACE_SETFPAREGS         21
+
+/* There are for debugging 64-bit processes, either from a 32 or 64 bit
+ * parent.  Thus their complements are for debugging 32-bit processes only.
+ */
+
+#define PTRACE_GETREGS64	  22
+#define PTRACE_SETREGS64	  23
+/* PTRACE_SYSCALL is 24 */
+#define PTRACE_GETFPREGS64	  25
+#define PTRACE_SETFPREGS64	  26
+
+#endif /* !(_SPARC64_PTRACE_H) */
diff --git a/include/asm-sparc/reboot.h b/include/asm-sparc/reboot.h
new file mode 100644
index 00000000000..3f3f43f5be5
--- /dev/null
+++ b/include/asm-sparc/reboot.h
@@ -0,0 +1,6 @@
+#ifndef _SPARC64_REBOOT_H
+#define _SPARC64_REBOOT_H
+
+extern void machine_alt_power_off(void);
+
+#endif /* _SPARC64_REBOOT_H */
diff --git a/include/asm-sparc/reg.h b/include/asm-sparc/reg.h
index ea0a7e590bb..cb34b0a49aa 100644
--- a/include/asm-sparc/reg.h
+++ b/include/asm-sparc/reg.h
@@ -1,79 +1,8 @@
-/*
- * linux/include/asm-sparc/reg.h
- * Layout of the registers as expected by gdb on the Sparc
- * we should replace the user.h definitions with those in
- * this file, we don't even use the other 
- * -miguel
- *
- * The names of the structures, constants and aliases in this file
- * have the same names as the sunos ones, some programs rely on these
- * names (gdb for example).
- *
- */
-
-#ifndef __SPARC_REG_H
-#define __SPARC_REG_H
-
-struct regs {
-	int     r_psr;
-#define r_ps r_psr
-        int     r_pc; 
-        int     r_npc;
-        int     r_y;  
-        int     r_g1; 
-        int     r_g2;
-        int     r_g3;
-        int     r_g4;
-        int     r_g5;
-        int     r_g6;
-        int     r_g7;
-        int     r_o0;
-        int     r_o1;
-        int     r_o2;
-        int     r_o3;
-        int     r_o4;
-        int     r_o5;
-        int     r_o6;
-        int     r_o7;
-};
-
-struct fpq {
-        unsigned long *addr;
-        unsigned long instr;
-};
-
-struct  fq {
-        union {
-                double  whole;
-                struct  fpq fpq;
-        } FQu;
-};
-
-#define FPU_REGS_TYPE unsigned int
-#define FPU_FSR_TYPE unsigned
-
-struct fp_status {
-        union {
-                FPU_REGS_TYPE Fpu_regs[32];
-                double  Fpu_dregs[16];
-        } fpu_fr;
-        FPU_FSR_TYPE Fpu_fsr;
-        unsigned Fpu_flags;
-        unsigned Fpu_extra;
-        unsigned Fpu_qcnt;
-        struct fq Fpu_q[16];
-};
-
-#define fpu_regs  f_fpstatus.fpu_fr.Fpu_regs
-#define fpu_dregs f_fpstatus.fpu_fr.Fpu_dregs
-#define fpu_fsr   f_fpstatus.Fpu_fsr
-#define fpu_flags f_fpstatus.Fpu_flags
-#define fpu_extra f_fpstatus.Fpu_extra
-#define fpu_q     f_fpstatus.Fpu_q
-#define fpu_qcnt  f_fpstatus.Fpu_qcnt
-
-struct fpu {
-        struct fp_status f_fpstatus;
-};
-
-#endif /* __SPARC_REG_H */
+#ifndef ___ASM_SPARC_REG_H
+#define ___ASM_SPARC_REG_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/reg_64.h>
+#else
+#include <asm-sparc/reg_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/reg_32.h b/include/asm-sparc/reg_32.h
new file mode 100644
index 00000000000..42fecfcd97e
--- /dev/null
+++ b/include/asm-sparc/reg_32.h
@@ -0,0 +1,79 @@
+/*
+ * linux/include/asm-sparc/reg.h
+ * Layout of the registers as expected by gdb on the Sparc
+ * we should replace the user.h definitions with those in
+ * this file, we don't even use the other
+ * -miguel
+ *
+ * The names of the structures, constants and aliases in this file
+ * have the same names as the sunos ones, some programs rely on these
+ * names (gdb for example).
+ *
+ */
+
+#ifndef __SPARC_REG_H
+#define __SPARC_REG_H
+
+struct regs {
+	int     r_psr;
+#define r_ps r_psr
+        int     r_pc;
+        int     r_npc;
+        int     r_y;
+        int     r_g1;
+        int     r_g2;
+        int     r_g3;
+        int     r_g4;
+        int     r_g5;
+        int     r_g6;
+        int     r_g7;
+        int     r_o0;
+        int     r_o1;
+        int     r_o2;
+        int     r_o3;
+        int     r_o4;
+        int     r_o5;
+        int     r_o6;
+        int     r_o7;
+};
+
+struct fpq {
+        unsigned long *addr;
+        unsigned long instr;
+};
+
+struct  fq {
+        union {
+                double  whole;
+                struct  fpq fpq;
+        } FQu;
+};
+
+#define FPU_REGS_TYPE unsigned int
+#define FPU_FSR_TYPE unsigned
+
+struct fp_status {
+        union {
+                FPU_REGS_TYPE Fpu_regs[32];
+                double  Fpu_dregs[16];
+        } fpu_fr;
+        FPU_FSR_TYPE Fpu_fsr;
+        unsigned Fpu_flags;
+        unsigned Fpu_extra;
+        unsigned Fpu_qcnt;
+        struct fq Fpu_q[16];
+};
+
+#define fpu_regs  f_fpstatus.fpu_fr.Fpu_regs
+#define fpu_dregs f_fpstatus.fpu_fr.Fpu_dregs
+#define fpu_fsr   f_fpstatus.Fpu_fsr
+#define fpu_flags f_fpstatus.Fpu_flags
+#define fpu_extra f_fpstatus.Fpu_extra
+#define fpu_q     f_fpstatus.Fpu_q
+#define fpu_qcnt  f_fpstatus.Fpu_qcnt
+
+struct fpu {
+        struct fp_status f_fpstatus;
+};
+
+#endif /* __SPARC_REG_H */
diff --git a/include/asm-sparc/reg_64.h b/include/asm-sparc/reg_64.h
new file mode 100644
index 00000000000..eb24a07ff4d
--- /dev/null
+++ b/include/asm-sparc/reg_64.h
@@ -0,0 +1,56 @@
+/*
+ * linux/asm-sparc64/reg.h
+ * Layout of the registers as expected by gdb on the Sparc
+ * we should replace the user.h definitions with those in
+ * this file, we don't even use the other
+ * -miguel
+ *
+ * The names of the structures, constants and aliases in this file
+ * have the same names as the sunos ones, some programs rely on these
+ * names (gdb for example).
+ *
+ */
+
+#ifndef __SPARC64_REG_H
+#define __SPARC64_REG_H
+
+struct regs {
+        unsigned long r_g1;
+        unsigned long r_g2;
+        unsigned long r_g3;
+        unsigned long r_g4;
+        unsigned long r_g5;
+        unsigned long r_g6;
+        unsigned long r_g7;
+        unsigned long r_o0;
+        unsigned long r_o1;
+        unsigned long r_o2;
+        unsigned long r_o3;
+        unsigned long r_o4;
+        unsigned long r_o5;
+        unsigned long r_o6;
+        unsigned long r_o7;
+        unsigned long __pad;
+        unsigned long r_tstate;
+        unsigned long r_tpc;
+        unsigned long r_tnpc;
+        unsigned int  r_y;
+        unsigned int  r_fprs;
+};
+
+#define FPU_REGS_TYPE unsigned int
+#define FPU_FSR_TYPE unsigned long
+
+struct fp_status {
+        unsigned long fpu_fr[32];
+        unsigned long Fpu_fsr;
+};
+
+struct fpu {
+	struct fp_status f_fpstatus;
+};
+
+#define fpu_regs  f_fpstatus.fpu_fr
+#define fpu_fsr   f_fpstatus.Fpu_fsr
+
+#endif /* __SPARC64_REG_H */
diff --git a/include/asm-sparc/resource.h b/include/asm-sparc/resource.h
index 985948a4129..fe163cafb4c 100644
--- a/include/asm-sparc/resource.h
+++ b/include/asm-sparc/resource.h
@@ -1,7 +1,7 @@
 /*
  * resource.h: Resource definitions.
  *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
  */
 
 #ifndef _SPARC_RESOURCE_H
@@ -14,12 +14,16 @@
 #define RLIMIT_NOFILE		6	/* max number of open files */
 #define RLIMIT_NPROC		7	/* max number of processes */
 
+#if defined(__sparc__) && defined(__arch64__)
+/* Use generic version */
+#else
 /*
  * SuS says limits have to be unsigned.
  * We make this unsigned, but keep the
  * old value for compatibility:
  */
 #define RLIM_INFINITY		0x7fffffff
+#endif
 
 #include <asm-generic/resource.h>
 
diff --git a/include/asm-sparc/rwsem-const.h b/include/asm-sparc/rwsem-const.h
new file mode 100644
index 00000000000..a303c9d64d8
--- /dev/null
+++ b/include/asm-sparc/rwsem-const.h
@@ -0,0 +1,12 @@
+/* rwsem-const.h: RW semaphore counter constants.  */
+#ifndef _SPARC64_RWSEM_CONST_H
+#define _SPARC64_RWSEM_CONST_H
+
+#define RWSEM_UNLOCKED_VALUE		0x00000000
+#define RWSEM_ACTIVE_BIAS		0x00000001
+#define RWSEM_ACTIVE_MASK		0x0000ffff
+#define RWSEM_WAITING_BIAS		0xffff0000
+#define RWSEM_ACTIVE_READ_BIAS		RWSEM_ACTIVE_BIAS
+#define RWSEM_ACTIVE_WRITE_BIAS		(RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS)
+
+#endif /* _SPARC64_RWSEM_CONST_H */
diff --git a/include/asm-sparc/rwsem.h b/include/asm-sparc/rwsem.h
new file mode 100644
index 00000000000..1dc129ac2fe
--- /dev/null
+++ b/include/asm-sparc/rwsem.h
@@ -0,0 +1,84 @@
+/*
+ * rwsem.h: R/W semaphores implemented using CAS
+ *
+ * Written by David S. Miller (davem@redhat.com), 2001.
+ * Derived from asm-i386/rwsem.h
+ */
+#ifndef _SPARC64_RWSEM_H
+#define _SPARC64_RWSEM_H
+
+#ifndef _LINUX_RWSEM_H
+#error "please don't include asm/rwsem.h directly, use linux/rwsem.h instead"
+#endif
+
+#ifdef __KERNEL__
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <asm/rwsem-const.h>
+
+struct rwsem_waiter;
+
+struct rw_semaphore {
+	signed int count;
+	spinlock_t		wait_lock;
+	struct list_head	wait_list;
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	struct lockdep_map	dep_map;
+#endif
+};
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname }
+#else
+# define __RWSEM_DEP_MAP_INIT(lockname)
+#endif
+
+#define __RWSEM_INITIALIZER(name) \
+{ RWSEM_UNLOCKED_VALUE, SPIN_LOCK_UNLOCKED, LIST_HEAD_INIT((name).wait_list) \
+  __RWSEM_DEP_MAP_INIT(name) }
+
+#define DECLARE_RWSEM(name) \
+	struct rw_semaphore name = __RWSEM_INITIALIZER(name)
+
+extern void __init_rwsem(struct rw_semaphore *sem, const char *name,
+			 struct lock_class_key *key);
+
+#define init_rwsem(sem)						\
+do {								\
+	static struct lock_class_key __key;			\
+								\
+	__init_rwsem((sem), #sem, &__key);			\
+} while (0)
+
+extern void __down_read(struct rw_semaphore *sem);
+extern int __down_read_trylock(struct rw_semaphore *sem);
+extern void __down_write(struct rw_semaphore *sem);
+extern int __down_write_trylock(struct rw_semaphore *sem);
+extern void __up_read(struct rw_semaphore *sem);
+extern void __up_write(struct rw_semaphore *sem);
+extern void __downgrade_write(struct rw_semaphore *sem);
+
+static inline void __down_write_nested(struct rw_semaphore *sem, int subclass)
+{
+	__down_write(sem);
+}
+
+static inline int rwsem_atomic_update(int delta, struct rw_semaphore *sem)
+{
+	return atomic_add_return(delta, (atomic_t *)(&sem->count));
+}
+
+static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem)
+{
+	atomic_add(delta, (atomic_t *)(&sem->count));
+}
+
+static inline int rwsem_is_locked(struct rw_semaphore *sem)
+{
+	return (sem->count != 0);
+}
+
+#endif /* __KERNEL__ */
+
+#endif /* _SPARC64_RWSEM_H */
diff --git a/include/asm-sparc/sbus.h b/include/asm-sparc/sbus.h
index f1d2fe1c9a3..8f29a197966 100644
--- a/include/asm-sparc/sbus.h
+++ b/include/asm-sparc/sbus.h
@@ -1,153 +1,8 @@
-/*
- * sbus.h:  Defines for the Sun SBus.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC_SBUS_H
-#define _SPARC_SBUS_H
-
-#include <linux/dma-mapping.h>
-#include <linux/ioport.h>
-
-#include <asm/oplib.h>
-#include <asm/prom.h>
-#include <asm/of_device.h>
-#include <asm/scatterlist.h>
-
-/* We scan which devices are on the SBus using the PROM node device
- * tree.  SBus devices are described in two different ways.  You can
- * either get an absolute address at which to access the device, or
- * you can get a SBus 'slot' number and an offset within that slot.
- */
-
-/* The base address at which to calculate device OBIO addresses. */
-#define SUN_SBUS_BVADDR        0xf8000000
-#define SBUS_OFF_MASK          0x01ffffff
-
-/* These routines are used to calculate device address from slot
- * numbers + offsets, and vice versa.
- */
-
-static inline unsigned long sbus_devaddr(int slotnum, unsigned long offset)
-{
-  return (unsigned long) (SUN_SBUS_BVADDR+((slotnum)<<25)+(offset));
-}
-
-static inline int sbus_dev_slot(unsigned long dev_addr)
-{
-  return (int) (((dev_addr)-SUN_SBUS_BVADDR)>>25);
-}
-
-struct sbus_bus;
-
-/* Linux SBUS device tables */
-struct sbus_dev {
-	struct of_device	ofdev;
-	struct sbus_bus		*bus;
-	struct sbus_dev		*next;
-	struct sbus_dev		*child;
-	struct sbus_dev		*parent;
-	int prom_node;	
-	char prom_name[64];
-	int slot;
-
-	struct resource resource[PROMREG_MAX];
-
-	struct linux_prom_registers reg_addrs[PROMREG_MAX];
-	int num_registers;
-
-	struct linux_prom_ranges device_ranges[PROMREG_MAX];
-	int num_device_ranges;
-
-	unsigned int irqs[4];
-	int num_irqs;
-};
-#define to_sbus_device(d) container_of(d, struct sbus_dev, ofdev.dev)
-
-/* This struct describes the SBus(s) found on this machine. */
-struct sbus_bus {
-	struct of_device	ofdev;
-	struct sbus_dev		*devices;	/* Link to devices on this SBus */
-	struct sbus_bus		*next;		/* next SBus, if more than one SBus */
-	int			prom_node;	/* PROM device tree node for this SBus */
-	char			prom_name[64];  /* Usually "sbus" or "sbi" */
-	int			clock_freq;
-
-	struct linux_prom_ranges sbus_ranges[PROMREG_MAX];
-	int num_sbus_ranges;
-
-	int devid;
-	int board;
-};
-#define to_sbus(d) container_of(d, struct sbus_bus, ofdev.dev)
-
-extern struct sbus_bus *sbus_root;
-
-static inline int
-sbus_is_slave(struct sbus_dev *dev)
-{
-	/* XXX Have to write this for sun4c's */
-	return 0;
-}
-
-/* Device probing routines could find these handy */
-#define for_each_sbus(bus) \
-        for((bus) = sbus_root; (bus); (bus)=(bus)->next)
-
-#define for_each_sbusdev(device, bus) \
-        for((device) = (bus)->devices; (device); (device)=(device)->next)
-        
-#define for_all_sbusdev(device, bus) \
-	for ((bus) = sbus_root; (bus); (bus) = (bus)->next) \
-		for ((device) = (bus)->devices; (device); (device) = (device)->next)
-
-/* Driver DVMA interfaces. */
-#define sbus_can_dma_64bit(sdev)	(0) /* actually, sparc_cpu_model==sun4d */
-#define sbus_can_burst64(sdev)		(0) /* actually, sparc_cpu_model==sun4d */
-extern void sbus_set_sbus64(struct sbus_dev *, int);
-extern void sbus_fill_device_irq(struct sbus_dev *);
-
-/* These yield IOMMU mappings in consistent mode. */
-extern void *sbus_alloc_consistent(struct sbus_dev *, long, u32 *dma_addrp);
-extern void sbus_free_consistent(struct sbus_dev *, long, void *, u32);
-void prom_adjust_ranges(struct linux_prom_ranges *, int,
-			struct linux_prom_ranges *, int);
-
-#define SBUS_DMA_BIDIRECTIONAL	DMA_BIDIRECTIONAL
-#define SBUS_DMA_TODEVICE	DMA_TO_DEVICE
-#define SBUS_DMA_FROMDEVICE	DMA_FROM_DEVICE
-#define	SBUS_DMA_NONE		DMA_NONE
-
-/* All the rest use streaming mode mappings. */
-extern dma_addr_t sbus_map_single(struct sbus_dev *, void *, size_t, int);
-extern void sbus_unmap_single(struct sbus_dev *, dma_addr_t, size_t, int);
-extern int sbus_map_sg(struct sbus_dev *, struct scatterlist *, int, int);
-extern void sbus_unmap_sg(struct sbus_dev *, struct scatterlist *, int, int);
-
-/* Finally, allow explicit synchronization of streamable mappings. */
-extern void sbus_dma_sync_single_for_cpu(struct sbus_dev *, dma_addr_t, size_t, int);
-#define sbus_dma_sync_single sbus_dma_sync_single_for_cpu
-extern void sbus_dma_sync_single_for_device(struct sbus_dev *, dma_addr_t, size_t, int);
-extern void sbus_dma_sync_sg_for_cpu(struct sbus_dev *, struct scatterlist *, int, int);
-#define sbus_dma_sync_sg sbus_dma_sync_sg_for_cpu
-extern void sbus_dma_sync_sg_for_device(struct sbus_dev *, struct scatterlist *, int, int);
-
-/* Eric Brower (ebrower@usa.net)
- * Translate SBus interrupt levels to ino values--
- * this is used when converting sbus "interrupts" OBP 
- * node values to "intr" node values, and is platform 
- * dependent.  If only we could call OBP with 
- * "sbus-intr>cpu (sbint -- ino)" from kernel...
- * See .../drivers/sbus/sbus.c for details.
- */
-BTFIXUPDEF_CALL(unsigned int, sbint_to_irq, struct sbus_dev *sdev, unsigned int)
-#define sbint_to_irq(sdev, sbint) BTFIXUP_CALL(sbint_to_irq)(sdev, sbint)
-
-extern void sbus_arch_bus_ranges_init(struct device_node *, struct sbus_bus *);
-extern void sbus_setup_iommu(struct sbus_bus *, struct device_node *);
-extern void sbus_setup_arch_props(struct sbus_bus *, struct device_node *);
-extern int sbus_arch_preinit(void);
-extern void sbus_arch_postinit(void);
-
-#endif /* !(_SPARC_SBUS_H) */
+#ifndef ___ASM_SPARC_SBUS_H
+#define ___ASM_SPARC_SBUS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/sbus_64.h>
+#else
+#include <asm-sparc/sbus_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/sbus_32.h b/include/asm-sparc/sbus_32.h
new file mode 100644
index 00000000000..77b5d3aadc9
--- /dev/null
+++ b/include/asm-sparc/sbus_32.h
@@ -0,0 +1,153 @@
+/*
+ * sbus.h:  Defines for the Sun SBus.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _SPARC_SBUS_H
+#define _SPARC_SBUS_H
+
+#include <linux/dma-mapping.h>
+#include <linux/ioport.h>
+
+#include <asm/oplib.h>
+#include <asm/prom.h>
+#include <asm/of_device.h>
+#include <asm/scatterlist.h>
+
+/* We scan which devices are on the SBus using the PROM node device
+ * tree.  SBus devices are described in two different ways.  You can
+ * either get an absolute address at which to access the device, or
+ * you can get a SBus 'slot' number and an offset within that slot.
+ */
+
+/* The base address at which to calculate device OBIO addresses. */
+#define SUN_SBUS_BVADDR        0xf8000000
+#define SBUS_OFF_MASK          0x01ffffff
+
+/* These routines are used to calculate device address from slot
+ * numbers + offsets, and vice versa.
+ */
+
+static inline unsigned long sbus_devaddr(int slotnum, unsigned long offset)
+{
+  return (unsigned long) (SUN_SBUS_BVADDR+((slotnum)<<25)+(offset));
+}
+
+static inline int sbus_dev_slot(unsigned long dev_addr)
+{
+  return (int) (((dev_addr)-SUN_SBUS_BVADDR)>>25);
+}
+
+struct sbus_bus;
+
+/* Linux SBUS device tables */
+struct sbus_dev {
+	struct of_device	ofdev;
+	struct sbus_bus		*bus;
+	struct sbus_dev		*next;
+	struct sbus_dev		*child;
+	struct sbus_dev		*parent;
+	int prom_node;
+	char prom_name[64];
+	int slot;
+
+	struct resource resource[PROMREG_MAX];
+
+	struct linux_prom_registers reg_addrs[PROMREG_MAX];
+	int num_registers;
+
+	struct linux_prom_ranges device_ranges[PROMREG_MAX];
+	int num_device_ranges;
+
+	unsigned int irqs[4];
+	int num_irqs;
+};
+#define to_sbus_device(d) container_of(d, struct sbus_dev, ofdev.dev)
+
+/* This struct describes the SBus(s) found on this machine. */
+struct sbus_bus {
+	struct of_device	ofdev;
+	struct sbus_dev		*devices;	/* Link to devices on this SBus */
+	struct sbus_bus		*next;		/* next SBus, if more than one SBus */
+	int			prom_node;	/* PROM device tree node for this SBus */
+	char			prom_name[64];  /* Usually "sbus" or "sbi" */
+	int			clock_freq;
+
+	struct linux_prom_ranges sbus_ranges[PROMREG_MAX];
+	int num_sbus_ranges;
+
+	int devid;
+	int board;
+};
+#define to_sbus(d) container_of(d, struct sbus_bus, ofdev.dev)
+
+extern struct sbus_bus *sbus_root;
+
+static inline int
+sbus_is_slave(struct sbus_dev *dev)
+{
+	/* XXX Have to write this for sun4c's */
+	return 0;
+}
+
+/* Device probing routines could find these handy */
+#define for_each_sbus(bus) \
+        for((bus) = sbus_root; (bus); (bus)=(bus)->next)
+
+#define for_each_sbusdev(device, bus) \
+        for((device) = (bus)->devices; (device); (device)=(device)->next)
+
+#define for_all_sbusdev(device, bus) \
+	for ((bus) = sbus_root; (bus); (bus) = (bus)->next) \
+		for ((device) = (bus)->devices; (device); (device) = (device)->next)
+
+/* Driver DVMA interfaces. */
+#define sbus_can_dma_64bit(sdev)	(0) /* actually, sparc_cpu_model==sun4d */
+#define sbus_can_burst64(sdev)		(0) /* actually, sparc_cpu_model==sun4d */
+extern void sbus_set_sbus64(struct sbus_dev *, int);
+extern void sbus_fill_device_irq(struct sbus_dev *);
+
+/* These yield IOMMU mappings in consistent mode. */
+extern void *sbus_alloc_consistent(struct sbus_dev *, long, u32 *dma_addrp);
+extern void sbus_free_consistent(struct sbus_dev *, long, void *, u32);
+void prom_adjust_ranges(struct linux_prom_ranges *, int,
+			struct linux_prom_ranges *, int);
+
+#define SBUS_DMA_BIDIRECTIONAL	DMA_BIDIRECTIONAL
+#define SBUS_DMA_TODEVICE	DMA_TO_DEVICE
+#define SBUS_DMA_FROMDEVICE	DMA_FROM_DEVICE
+#define	SBUS_DMA_NONE		DMA_NONE
+
+/* All the rest use streaming mode mappings. */
+extern dma_addr_t sbus_map_single(struct sbus_dev *, void *, size_t, int);
+extern void sbus_unmap_single(struct sbus_dev *, dma_addr_t, size_t, int);
+extern int sbus_map_sg(struct sbus_dev *, struct scatterlist *, int, int);
+extern void sbus_unmap_sg(struct sbus_dev *, struct scatterlist *, int, int);
+
+/* Finally, allow explicit synchronization of streamable mappings. */
+extern void sbus_dma_sync_single_for_cpu(struct sbus_dev *, dma_addr_t, size_t, int);
+#define sbus_dma_sync_single sbus_dma_sync_single_for_cpu
+extern void sbus_dma_sync_single_for_device(struct sbus_dev *, dma_addr_t, size_t, int);
+extern void sbus_dma_sync_sg_for_cpu(struct sbus_dev *, struct scatterlist *, int, int);
+#define sbus_dma_sync_sg sbus_dma_sync_sg_for_cpu
+extern void sbus_dma_sync_sg_for_device(struct sbus_dev *, struct scatterlist *, int, int);
+
+/* Eric Brower (ebrower@usa.net)
+ * Translate SBus interrupt levels to ino values--
+ * this is used when converting sbus "interrupts" OBP
+ * node values to "intr" node values, and is platform
+ * dependent.  If only we could call OBP with
+ * "sbus-intr>cpu (sbint -- ino)" from kernel...
+ * See .../drivers/sbus/sbus.c for details.
+ */
+BTFIXUPDEF_CALL(unsigned int, sbint_to_irq, struct sbus_dev *sdev, unsigned int)
+#define sbint_to_irq(sdev, sbint) BTFIXUP_CALL(sbint_to_irq)(sdev, sbint)
+
+extern void sbus_arch_bus_ranges_init(struct device_node *, struct sbus_bus *);
+extern void sbus_setup_iommu(struct sbus_bus *, struct device_node *);
+extern void sbus_setup_arch_props(struct sbus_bus *, struct device_node *);
+extern int sbus_arch_preinit(void);
+extern void sbus_arch_postinit(void);
+
+#endif /* !(_SPARC_SBUS_H) */
diff --git a/include/asm-sparc/sbus_64.h b/include/asm-sparc/sbus_64.h
new file mode 100644
index 00000000000..0e16b6dd7e9
--- /dev/null
+++ b/include/asm-sparc/sbus_64.h
@@ -0,0 +1,190 @@
+/* sbus.h: Defines for the Sun SBus.
+ *
+ * Copyright (C) 1996, 1999, 2007 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC64_SBUS_H
+#define _SPARC64_SBUS_H
+
+#include <linux/dma-mapping.h>
+#include <linux/ioport.h>
+
+#include <asm/oplib.h>
+#include <asm/prom.h>
+#include <asm/of_device.h>
+#include <asm/iommu.h>
+#include <asm/scatterlist.h>
+
+/* We scan which devices are on the SBus using the PROM node device
+ * tree.  SBus devices are described in two different ways.  You can
+ * either get an absolute address at which to access the device, or
+ * you can get a SBus 'slot' number and an offset within that slot.
+ */
+
+/* The base address at which to calculate device OBIO addresses. */
+#define SUN_SBUS_BVADDR        0x00000000
+#define SBUS_OFF_MASK          0x0fffffff
+
+/* These routines are used to calculate device address from slot
+ * numbers + offsets, and vice versa.
+ */
+
+static inline unsigned long sbus_devaddr(int slotnum, unsigned long offset)
+{
+  return (unsigned long) (SUN_SBUS_BVADDR+((slotnum)<<28)+(offset));
+}
+
+static inline int sbus_dev_slot(unsigned long dev_addr)
+{
+  return (int) (((dev_addr)-SUN_SBUS_BVADDR)>>28);
+}
+
+struct sbus_bus;
+
+/* Linux SBUS device tables */
+struct sbus_dev {
+	struct of_device	ofdev;
+	struct sbus_bus		*bus;
+	struct sbus_dev		*next;
+	struct sbus_dev		*child;
+	struct sbus_dev		*parent;
+	int prom_node;
+	char prom_name[64];
+	int slot;
+
+	struct resource resource[PROMREG_MAX];
+
+	struct linux_prom_registers reg_addrs[PROMREG_MAX];
+	int num_registers;
+
+	struct linux_prom_ranges device_ranges[PROMREG_MAX];
+	int num_device_ranges;
+
+	unsigned int irqs[4];
+	int num_irqs;
+};
+#define to_sbus_device(d) container_of(d, struct sbus_dev, ofdev.dev)
+
+/* This struct describes the SBus(s) found on this machine. */
+struct sbus_bus {
+	struct of_device	ofdev;
+	struct sbus_dev		*devices;	/* Tree of SBUS devices	*/
+	struct sbus_bus		*next;		/* Next SBUS in system	*/
+	int			prom_node;      /* OBP node of SBUS	*/
+	char			prom_name[64];	/* Usually "sbus" or "sbi" */
+	int			clock_freq;
+
+	struct linux_prom_ranges sbus_ranges[PROMREG_MAX];
+	int num_sbus_ranges;
+
+	int portid;
+};
+#define to_sbus(d) container_of(d, struct sbus_bus, ofdev.dev)
+
+extern struct sbus_bus *sbus_root;
+
+/* Device probing routines could find these handy */
+#define for_each_sbus(bus) \
+        for((bus) = sbus_root; (bus); (bus)=(bus)->next)
+
+#define for_each_sbusdev(device, bus) \
+        for((device) = (bus)->devices; (device); (device)=(device)->next)
+
+#define for_all_sbusdev(device, bus) \
+	for ((bus) = sbus_root; (bus); (bus) = (bus)->next) \
+		for ((device) = (bus)->devices; (device); (device) = (device)->next)
+
+/* Driver DVMA interfaces. */
+#define sbus_can_dma_64bit(sdev)	(1)
+#define sbus_can_burst64(sdev)		(1)
+extern void sbus_set_sbus64(struct sbus_dev *, int);
+extern void sbus_fill_device_irq(struct sbus_dev *);
+
+static inline void *sbus_alloc_consistent(struct sbus_dev *sdev , size_t size,
+					  dma_addr_t *dma_handle)
+{
+	return dma_alloc_coherent(&sdev->ofdev.dev, size,
+				  dma_handle, GFP_ATOMIC);
+}
+
+static inline void sbus_free_consistent(struct sbus_dev *sdev, size_t size,
+					void *vaddr, dma_addr_t dma_handle)
+{
+	return dma_free_coherent(&sdev->ofdev.dev, size, vaddr, dma_handle);
+}
+
+#define SBUS_DMA_BIDIRECTIONAL	DMA_BIDIRECTIONAL
+#define SBUS_DMA_TODEVICE	DMA_TO_DEVICE
+#define SBUS_DMA_FROMDEVICE	DMA_FROM_DEVICE
+#define	SBUS_DMA_NONE		DMA_NONE
+
+/* All the rest use streaming mode mappings. */
+static inline dma_addr_t sbus_map_single(struct sbus_dev *sdev, void *ptr,
+					 size_t size, int direction)
+{
+	return dma_map_single(&sdev->ofdev.dev, ptr, size,
+			      (enum dma_data_direction) direction);
+}
+
+static inline void sbus_unmap_single(struct sbus_dev *sdev,
+				     dma_addr_t dma_addr, size_t size,
+				     int direction)
+{
+	dma_unmap_single(&sdev->ofdev.dev, dma_addr, size,
+			 (enum dma_data_direction) direction);
+}
+
+static inline int sbus_map_sg(struct sbus_dev *sdev, struct scatterlist *sg,
+			      int nents, int direction)
+{
+	return dma_map_sg(&sdev->ofdev.dev, sg, nents,
+			  (enum dma_data_direction) direction);
+}
+
+static inline void sbus_unmap_sg(struct sbus_dev *sdev, struct scatterlist *sg,
+				 int nents, int direction)
+{
+	dma_unmap_sg(&sdev->ofdev.dev, sg, nents,
+		     (enum dma_data_direction) direction);
+}
+
+/* Finally, allow explicit synchronization of streamable mappings. */
+static inline void sbus_dma_sync_single_for_cpu(struct sbus_dev *sdev,
+						dma_addr_t dma_handle,
+						size_t size, int direction)
+{
+	dma_sync_single_for_cpu(&sdev->ofdev.dev, dma_handle, size,
+				(enum dma_data_direction) direction);
+}
+#define sbus_dma_sync_single sbus_dma_sync_single_for_cpu
+
+static inline void sbus_dma_sync_single_for_device(struct sbus_dev *sdev,
+						   dma_addr_t dma_handle,
+						   size_t size, int direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+static inline void sbus_dma_sync_sg_for_cpu(struct sbus_dev *sdev,
+					    struct scatterlist *sg,
+					    int nents, int direction)
+{
+	dma_sync_sg_for_cpu(&sdev->ofdev.dev, sg, nents,
+			    (enum dma_data_direction) direction);
+}
+#define sbus_dma_sync_sg sbus_dma_sync_sg_for_cpu
+
+static inline void sbus_dma_sync_sg_for_device(struct sbus_dev *sdev,
+					       struct scatterlist *sg,
+					       int nents, int direction)
+{
+	/* No flushing needed to sync cpu writes to the device.  */
+}
+
+extern void sbus_arch_bus_ranges_init(struct device_node *, struct sbus_bus *);
+extern void sbus_setup_iommu(struct sbus_bus *, struct device_node *);
+extern void sbus_setup_arch_props(struct sbus_bus *, struct device_node *);
+extern int sbus_arch_preinit(void);
+extern void sbus_arch_postinit(void);
+
+#endif /* !(_SPARC64_SBUS_H) */
diff --git a/include/asm-sparc/scatterlist.h b/include/asm-sparc/scatterlist.h
index c82609ca1d0..b1a0e316c2b 100644
--- a/include/asm-sparc/scatterlist.h
+++ b/include/asm-sparc/scatterlist.h
@@ -1,26 +1,8 @@
-#ifndef _SPARC_SCATTERLIST_H
-#define _SPARC_SCATTERLIST_H
-
-#include <linux/types.h>
-
-struct scatterlist {
-#ifdef CONFIG_DEBUG_SG
-	unsigned long sg_magic;
+#ifndef ___ASM_SPARC_SCATTERLIST_H
+#define ___ASM_SPARC_SCATTERLIST_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/scatterlist_64.h>
+#else
+#include <asm-sparc/scatterlist_32.h>
+#endif
 #endif
-	unsigned long page_link;
-	unsigned int offset;
-
-	unsigned int length;
-
-	__u32 dvma_address; /* A place to hang host-specific addresses at. */
-	__u32 dvma_length;
-};
-
-#define sg_dma_address(sg) ((sg)->dvma_address)
-#define sg_dma_len(sg)     ((sg)->dvma_length)
-
-#define ISA_DMA_THRESHOLD (~0UL)
-
-#define ARCH_HAS_SG_CHAIN
-
-#endif /* !(_SPARC_SCATTERLIST_H) */
diff --git a/include/asm-sparc/scatterlist_32.h b/include/asm-sparc/scatterlist_32.h
new file mode 100644
index 00000000000..c82609ca1d0
--- /dev/null
+++ b/include/asm-sparc/scatterlist_32.h
@@ -0,0 +1,26 @@
+#ifndef _SPARC_SCATTERLIST_H
+#define _SPARC_SCATTERLIST_H
+
+#include <linux/types.h>
+
+struct scatterlist {
+#ifdef CONFIG_DEBUG_SG
+	unsigned long sg_magic;
+#endif
+	unsigned long page_link;
+	unsigned int offset;
+
+	unsigned int length;
+
+	__u32 dvma_address; /* A place to hang host-specific addresses at. */
+	__u32 dvma_length;
+};
+
+#define sg_dma_address(sg) ((sg)->dvma_address)
+#define sg_dma_len(sg)     ((sg)->dvma_length)
+
+#define ISA_DMA_THRESHOLD (~0UL)
+
+#define ARCH_HAS_SG_CHAIN
+
+#endif /* !(_SPARC_SCATTERLIST_H) */
diff --git a/include/asm-sparc/scatterlist_64.h b/include/asm-sparc/scatterlist_64.h
new file mode 100644
index 00000000000..81bd058f938
--- /dev/null
+++ b/include/asm-sparc/scatterlist_64.h
@@ -0,0 +1,27 @@
+#ifndef _SPARC64_SCATTERLIST_H
+#define _SPARC64_SCATTERLIST_H
+
+#include <asm/page.h>
+#include <asm/types.h>
+
+struct scatterlist {
+#ifdef CONFIG_DEBUG_SG
+	unsigned long	sg_magic;
+#endif
+	unsigned long	page_link;
+	unsigned int	offset;
+
+	unsigned int	length;
+
+	dma_addr_t	dma_address;
+	__u32		dma_length;
+};
+
+#define sg_dma_address(sg)	((sg)->dma_address)
+#define sg_dma_len(sg)     	((sg)->dma_length)
+
+#define ISA_DMA_THRESHOLD	(~0UL)
+
+#define ARCH_HAS_SG_CHAIN
+
+#endif /* !(_SPARC64_SCATTERLIST_H) */
diff --git a/include/asm-sparc/scratchpad.h b/include/asm-sparc/scratchpad.h
new file mode 100644
index 00000000000..5e8b01fb334
--- /dev/null
+++ b/include/asm-sparc/scratchpad.h
@@ -0,0 +1,14 @@
+#ifndef _SPARC64_SCRATCHPAD_H
+#define _SPARC64_SCRATCHPAD_H
+
+/* Sun4v scratchpad registers, accessed via ASI_SCRATCHPAD.  */
+
+#define SCRATCHPAD_MMU_MISS	0x00 /* Shared with OBP - set by OBP	    */
+#define SCRATCHPAD_CPUID	0x08 /* Shared with OBP - set by hypervisor */
+#define SCRATCHPAD_UTSBREG1	0x10
+#define SCRATCHPAD_UTSBREG2	0x18
+	/* 0x20 and 0x28, hypervisor only... */
+#define SCRATCHPAD_UNUSED1	0x30
+#define SCRATCHPAD_UNUSED2	0x38 /* Reserved for OBP		    */
+
+#endif /* !(_SPARC64_SCRATCHPAD_H) */
diff --git a/include/asm-sparc/seccomp.h b/include/asm-sparc/seccomp.h
new file mode 100644
index 00000000000..7fcd9968192
--- /dev/null
+++ b/include/asm-sparc/seccomp.h
@@ -0,0 +1,21 @@
+#ifndef _ASM_SECCOMP_H
+
+#include <linux/thread_info.h> /* already defines TIF_32BIT */
+
+#ifndef TIF_32BIT
+#error "unexpected TIF_32BIT on sparc64"
+#endif
+
+#include <linux/unistd.h>
+
+#define __NR_seccomp_read __NR_read
+#define __NR_seccomp_write __NR_write
+#define __NR_seccomp_exit __NR_exit
+#define __NR_seccomp_sigreturn __NR_rt_sigreturn
+
+#define __NR_seccomp_read_32 __NR_read
+#define __NR_seccomp_write_32 __NR_write
+#define __NR_seccomp_exit_32 __NR_exit
+#define __NR_seccomp_sigreturn_32 __NR_sigreturn
+
+#endif /* _ASM_SECCOMP_H */
diff --git a/include/asm-sparc/sections.h b/include/asm-sparc/sections.h
index 6832841df05..cbd01916242 100644
--- a/include/asm-sparc/sections.h
+++ b/include/asm-sparc/sections.h
@@ -1,6 +1,8 @@
-#ifndef _SPARC_SECTIONS_H
-#define _SPARC_SECTIONS_H
-
-#include <asm-generic/sections.h>
-
+#ifndef ___ASM_SPARC_SECTIONS_H
+#define ___ASM_SPARC_SECTIONS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/sections_64.h>
+#else
+#include <asm-sparc/sections_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/sections_32.h b/include/asm-sparc/sections_32.h
new file mode 100644
index 00000000000..6832841df05
--- /dev/null
+++ b/include/asm-sparc/sections_32.h
@@ -0,0 +1,6 @@
+#ifndef _SPARC_SECTIONS_H
+#define _SPARC_SECTIONS_H
+
+#include <asm-generic/sections.h>
+
+#endif
diff --git a/include/asm-sparc/sections_64.h b/include/asm-sparc/sections_64.h
new file mode 100644
index 00000000000..3f4b9fdc28d
--- /dev/null
+++ b/include/asm-sparc/sections_64.h
@@ -0,0 +1,9 @@
+#ifndef _SPARC64_SECTIONS_H
+#define _SPARC64_SECTIONS_H
+
+/* nothing to see, move along */
+#include <asm-generic/sections.h>
+
+extern char _start[];
+
+#endif
diff --git a/include/asm-sparc/sembuf.h b/include/asm-sparc/sembuf.h
index a79c4bb3c08..faee1be08d6 100644
--- a/include/asm-sparc/sembuf.h
+++ b/include/asm-sparc/sembuf.h
@@ -1,7 +1,7 @@
 #ifndef _SPARC_SEMBUF_H
 #define _SPARC_SEMBUF_H
 
-/* 
+/*
  * The semid64_ds structure for sparc architecture.
  * Note extra padding because this structure is passed back and forth
  * between kernel and user space.
@@ -10,16 +10,22 @@
  * - 64-bit time_t to solve y2038 problem
  * - 2 miscellaneous 32-bit values
  */
+#if defined(__sparc__) && defined(__arch64__)
+# define PADDING(x)
+#else
+# define PADDING(x) unsigned int x;
+#endif
 
 struct semid64_ds {
 	struct ipc64_perm sem_perm;		/* permissions .. see ipc.h */
-	unsigned int	__pad1;
+	PADDING(__pad1)
 	__kernel_time_t	sem_otime;		/* last semop time */
-	unsigned int	__pad2;
+	PADDING(__pad2)
 	__kernel_time_t	sem_ctime;		/* last change time */
 	unsigned long	sem_nsems;		/* no. of semaphores in array */
 	unsigned long	__unused1;
 	unsigned long	__unused2;
 };
+#undef PADDING
 
 #endif /* _SPARC64_SEMBUF_H */
diff --git a/include/asm-sparc/setup.h b/include/asm-sparc/setup.h
index b3af958a2ad..2643c62f4ac 100644
--- a/include/asm-sparc/setup.h
+++ b/include/asm-sparc/setup.h
@@ -5,6 +5,10 @@
 #ifndef _SPARC_SETUP_H
 #define _SPARC_SETUP_H
 
-#define COMMAND_LINE_SIZE	256
+#if defined(__sparc__) && defined(__arch64__)
+# define COMMAND_LINE_SIZE 2048
+#else
+# define COMMAND_LINE_SIZE 256
+#endif
 
 #endif /* _SPARC_SETUP_H */
diff --git a/include/asm-sparc/sfafsr.h b/include/asm-sparc/sfafsr.h
new file mode 100644
index 00000000000..e96137b04a4
--- /dev/null
+++ b/include/asm-sparc/sfafsr.h
@@ -0,0 +1,82 @@
+#ifndef _SPARC64_SFAFSR_H
+#define _SPARC64_SFAFSR_H
+
+#include <linux/const.h>
+
+/* Spitfire Asynchronous Fault Status register, ASI=0x4C VA<63:0>=0x0 */
+
+#define SFAFSR_ME		(_AC(1,UL) << SFAFSR_ME_SHIFT)
+#define SFAFSR_ME_SHIFT		32
+#define SFAFSR_PRIV		(_AC(1,UL) << SFAFSR_PRIV_SHIFT)
+#define SFAFSR_PRIV_SHIFT	31
+#define SFAFSR_ISAP		(_AC(1,UL) << SFAFSR_ISAP_SHIFT)
+#define SFAFSR_ISAP_SHIFT	30
+#define SFAFSR_ETP		(_AC(1,UL) << SFAFSR_ETP_SHIFT)
+#define SFAFSR_ETP_SHIFT	29
+#define SFAFSR_IVUE		(_AC(1,UL) << SFAFSR_IVUE_SHIFT)
+#define SFAFSR_IVUE_SHIFT	28
+#define SFAFSR_TO		(_AC(1,UL) << SFAFSR_TO_SHIFT)
+#define SFAFSR_TO_SHIFT		27
+#define SFAFSR_BERR		(_AC(1,UL) << SFAFSR_BERR_SHIFT)
+#define SFAFSR_BERR_SHIFT	26
+#define SFAFSR_LDP		(_AC(1,UL) << SFAFSR_LDP_SHIFT)
+#define SFAFSR_LDP_SHIFT	25
+#define SFAFSR_CP		(_AC(1,UL) << SFAFSR_CP_SHIFT)
+#define SFAFSR_CP_SHIFT		24
+#define SFAFSR_WP		(_AC(1,UL) << SFAFSR_WP_SHIFT)
+#define SFAFSR_WP_SHIFT		23
+#define SFAFSR_EDP		(_AC(1,UL) << SFAFSR_EDP_SHIFT)
+#define SFAFSR_EDP_SHIFT	22
+#define SFAFSR_UE		(_AC(1,UL) << SFAFSR_UE_SHIFT)
+#define SFAFSR_UE_SHIFT		21
+#define SFAFSR_CE		(_AC(1,UL) << SFAFSR_CE_SHIFT)
+#define SFAFSR_CE_SHIFT		20
+#define SFAFSR_ETS		(_AC(0xf,UL) << SFAFSR_ETS_SHIFT)
+#define SFAFSR_ETS_SHIFT	16
+#define SFAFSR_PSYND		(_AC(0xffff,UL) << SFAFSR_PSYND_SHIFT)
+#define SFAFSR_PSYND_SHIFT	0
+
+/* UDB Error Register, ASI=0x7f VA<63:0>=0x0(High),0x18(Low) for read
+ *                     ASI=0x77 VA<63:0>=0x0(High),0x18(Low) for write
+ */
+
+#define UDBE_UE			(_AC(1,UL) << 9)
+#define UDBE_CE			(_AC(1,UL) << 8)
+#define UDBE_E_SYNDR		(_AC(0xff,UL) << 0)
+
+/* The trap handlers for asynchronous errors encode the AFSR and
+ * other pieces of information into a 64-bit argument for C code
+ * encoded as follows:
+ *
+ * -----------------------------------------------
+ * |  UDB_H  |  UDB_L  | TL>1  |  TT  |   AFSR   |
+ * -----------------------------------------------
+ *  63     54 53     44    42   41  33 32       0
+ *
+ * The AFAR is passed in unchanged.
+ */
+#define SFSTAT_UDBH_MASK	(_AC(0x3ff,UL) << SFSTAT_UDBH_SHIFT)
+#define SFSTAT_UDBH_SHIFT	54
+#define SFSTAT_UDBL_MASK	(_AC(0x3ff,UL) << SFSTAT_UDBH_SHIFT)
+#define SFSTAT_UDBL_SHIFT	44
+#define SFSTAT_TL_GT_ONE	(_AC(1,UL) << SFSTAT_TL_GT_ONE_SHIFT)
+#define SFSTAT_TL_GT_ONE_SHIFT	42
+#define SFSTAT_TRAP_TYPE	(_AC(0x1FF,UL) << SFSTAT_TRAP_TYPE_SHIFT)
+#define SFSTAT_TRAP_TYPE_SHIFT	33
+#define SFSTAT_AFSR_MASK	(_AC(0x1ffffffff,UL) << SFSTAT_AFSR_SHIFT)
+#define SFSTAT_AFSR_SHIFT	0
+
+/* ESTATE Error Enable Register, ASI=0x4b VA<63:0>=0x0 */
+#define ESTATE_ERR_CE		0x1 /* Correctable errors                    */
+#define ESTATE_ERR_NCE		0x2 /* TO, BERR, LDP, ETP, EDP, WP, UE, IVUE */
+#define ESTATE_ERR_ISAP		0x4 /* System address parity error           */
+#define ESTATE_ERR_ALL		(ESTATE_ERR_CE | \
+				 ESTATE_ERR_NCE | \
+				 ESTATE_ERR_ISAP)
+
+/* The various trap types that report using the above state. */
+#define TRAP_TYPE_IAE		0x09 /* Instruction Access Error             */
+#define TRAP_TYPE_DAE		0x32 /* Data Access Error                    */
+#define TRAP_TYPE_CEE		0x63 /* Correctable ECC Error                */
+
+#endif /* _SPARC64_SFAFSR_H */
diff --git a/include/asm-sparc/sfp-machine.h b/include/asm-sparc/sfp-machine.h
index 266a42b8f99..c676fcc2dd2 100644
--- a/include/asm-sparc/sfp-machine.h
+++ b/include/asm-sparc/sfp-machine.h
@@ -1,212 +1,8 @@
-/* Machine-dependent software floating-point definitions.
-   Sparc userland (_Q_*) version.
-   Copyright (C) 1997,1998,1999 Free Software Foundation, Inc.
-   This file is part of the GNU C Library.
-   Contributed by Richard Henderson (rth@cygnus.com),
-		  Jakub Jelinek (jj@ultra.linux.cz),
-		  David S. Miller (davem@redhat.com) and
-		  Peter Maydell (pmaydell@chiark.greenend.org.uk).
-
-   The GNU C Library is free software; you can redistribute it and/or
-   modify it under the terms of the GNU Library General Public License as
-   published by the Free Software Foundation; either version 2 of the
-   License, or (at your option) any later version.
-
-   The GNU C Library is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
-   Library General Public License for more details.
-
-   You should have received a copy of the GNU Library General Public
-   License along with the GNU C Library; see the file COPYING.LIB.  If
-   not, write to the Free Software Foundation, Inc.,
-   59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
-
-#ifndef _SFP_MACHINE_H
-#define _SFP_MACHINE_H
-
-   
-#define _FP_W_TYPE_SIZE		32
-#define _FP_W_TYPE		unsigned long
-#define _FP_WS_TYPE		signed long
-#define _FP_I_TYPE		long
-
-#define _FP_MUL_MEAT_S(R,X,Y)					\
-  _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
-#define _FP_MUL_MEAT_D(R,X,Y)					\
-  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
-#define _FP_MUL_MEAT_Q(R,X,Y)					\
-  _FP_MUL_MEAT_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
-
-#define _FP_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_udiv(S,R,X,Y)
-#define _FP_DIV_MEAT_D(R,X,Y)	_FP_DIV_MEAT_2_udiv(D,R,X,Y)
-#define _FP_DIV_MEAT_Q(R,X,Y)	_FP_DIV_MEAT_4_udiv(Q,R,X,Y)
-
-#define _FP_NANFRAC_S		((_FP_QNANBIT_S << 1) - 1)
-#define _FP_NANFRAC_D		((_FP_QNANBIT_D << 1) - 1), -1
-#define _FP_NANFRAC_Q		((_FP_QNANBIT_Q << 1) - 1), -1, -1, -1
-#define _FP_NANSIGN_S		0
-#define _FP_NANSIGN_D		0
-#define _FP_NANSIGN_Q		0
-
-#define _FP_KEEPNANFRACP 1
-
-/* If one NaN is signaling and the other is not,
- * we choose that one, otherwise we choose X.
- */
-/* For _Qp_* and _Q_*, this should prefer X, for
- * CPU instruction emulation this should prefer Y.
- * (see SPAMv9 B.2.2 section).
- */
-#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
-  do {								\
-    if ((_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs)		\
-	&& !(_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs))	\
-      {								\
-	R##_s = X##_s;						\
-	_FP_FRAC_COPY_##wc(R,X);				\
-      }								\
-    else							\
-      {								\
-	R##_s = Y##_s;						\
-	_FP_FRAC_COPY_##wc(R,Y);				\
-      }								\
-    R##_c = FP_CLS_NAN;						\
-  } while (0)
-
-/* Some assembly to speed things up. */
-#define __FP_FRAC_ADD_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
-  __asm__ ("addcc %r7,%8,%2\n\t"					\
-	   "addxcc %r5,%6,%1\n\t"					\
-	   "addx %r3,%4,%0\n"						\
-	   : "=r" ((USItype)(r2)),					\
-	     "=&r" ((USItype)(r1)),					\
-	     "=&r" ((USItype)(r0))					\
-	   : "%rJ" ((USItype)(x2)),					\
-	     "rI" ((USItype)(y2)),					\
-	     "%rJ" ((USItype)(x1)),					\
-	     "rI" ((USItype)(y1)),					\
-	     "%rJ" ((USItype)(x0)),					\
-	     "rI" ((USItype)(y0))					\
-	   : "cc")
-
-#define __FP_FRAC_SUB_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
-  __asm__ ("subcc %r7,%8,%2\n\t"					\
-	    "subxcc %r5,%6,%1\n\t"					\
-	    "subx %r3,%4,%0\n"						\
-	   : "=r" ((USItype)(r2)),					\
-	     "=&r" ((USItype)(r1)),					\
-	     "=&r" ((USItype)(r0))					\
-	   : "%rJ" ((USItype)(x2)),					\
-	     "rI" ((USItype)(y2)),					\
-	     "%rJ" ((USItype)(x1)),					\
-	     "rI" ((USItype)(y1)),					\
-	     "%rJ" ((USItype)(x0)),					\
-	     "rI" ((USItype)(y0))					\
-	   : "cc")
-
-#define __FP_FRAC_ADD_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
-  do {									\
-    /* We need to fool gcc,  as we need to pass more than 10		\
-       input/outputs.  */						\
-    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
-    __asm__ __volatile__ (						\
-	    "addcc %r8,%9,%1\n\t"					\
-	    "addxcc %r6,%7,%0\n\t"					\
-	    "addxcc %r4,%5,%%g2\n\t"					\
-	    "addx %r2,%3,%%g1\n\t"					\
-	   : "=&r" ((USItype)(r1)),					\
-	     "=&r" ((USItype)(r0))					\
-	   : "%rJ" ((USItype)(x3)),					\
-	     "rI" ((USItype)(y3)),					\
-	     "%rJ" ((USItype)(x2)),					\
-	     "rI" ((USItype)(y2)),					\
-	     "%rJ" ((USItype)(x1)),					\
-	     "rI" ((USItype)(y1)),					\
-	     "%rJ" ((USItype)(x0)),					\
-	     "rI" ((USItype)(y0))					\
-	   : "cc", "g1", "g2");						\
-    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
-    r3 = _t1; r2 = _t2;							\
-  } while (0)
-
-#define __FP_FRAC_SUB_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
-  do {									\
-    /* We need to fool gcc,  as we need to pass more than 10		\
-       input/outputs.  */						\
-    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
-    __asm__ __volatile__ (						\
-	    "subcc %r8,%9,%1\n\t"					\
-	    "subxcc %r6,%7,%0\n\t"					\
-	    "subxcc %r4,%5,%%g2\n\t"					\
-	    "subx %r2,%3,%%g1\n\t"					\
-	   : "=&r" ((USItype)(r1)),					\
-	     "=&r" ((USItype)(r0))					\
-	   : "%rJ" ((USItype)(x3)),					\
-	     "rI" ((USItype)(y3)),					\
-	     "%rJ" ((USItype)(x2)),					\
-	     "rI" ((USItype)(y2)),					\
-	     "%rJ" ((USItype)(x1)),					\
-	     "rI" ((USItype)(y1)),					\
-	     "%rJ" ((USItype)(x0)),					\
-	     "rI" ((USItype)(y0))					\
-	   : "cc", "g1", "g2");						\
-    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
-    r3 = _t1; r2 = _t2;							\
-  } while (0)
-
-#define __FP_FRAC_DEC_3(x2,x1,x0,y2,y1,y0) __FP_FRAC_SUB_3(x2,x1,x0,x2,x1,x0,y2,y1,y0)
-
-#define __FP_FRAC_DEC_4(x3,x2,x1,x0,y3,y2,y1,y0) __FP_FRAC_SUB_4(x3,x2,x1,x0,x3,x2,x1,x0,y3,y2,y1,y0)
-
-#define __FP_FRAC_ADDI_4(x3,x2,x1,x0,i)					\
-  __asm__ ("addcc %3,%4,%3\n\t"						\
-	   "addxcc %2,%%g0,%2\n\t"					\
-	   "addxcc %1,%%g0,%1\n\t"					\
-	   "addx %0,%%g0,%0\n\t"					\
-	   : "=&r" ((USItype)(x3)),					\
-	     "=&r" ((USItype)(x2)),					\
-	     "=&r" ((USItype)(x1)),					\
-	     "=&r" ((USItype)(x0))					\
-	   : "rI" ((USItype)(i)),					\
-	     "0" ((USItype)(x3)),					\
-	     "1" ((USItype)(x2)),					\
-	     "2" ((USItype)(x1)),					\
-	     "3" ((USItype)(x0))					\
-	   : "cc")
-
-#ifndef CONFIG_SMP
-extern struct task_struct *last_task_used_math;
-#endif
-
-/* Obtain the current rounding mode. */
-#ifndef FP_ROUNDMODE
-#ifdef CONFIG_SMP
-#define FP_ROUNDMODE	((current->thread.fsr >> 30) & 0x3)
-#else
-#define FP_ROUNDMODE	((last_task_used_math->thread.fsr >> 30) & 0x3)
-#endif
-#endif
-
-/* Exception flags. */
-#define FP_EX_INVALID		(1 << 4)
-#define FP_EX_OVERFLOW		(1 << 3)
-#define FP_EX_UNDERFLOW		(1 << 2)
-#define FP_EX_DIVZERO		(1 << 1)
-#define FP_EX_INEXACT		(1 << 0)
-
-#define FP_HANDLE_EXCEPTIONS return _fex
-
-#ifdef CONFIG_SMP
-#define FP_INHIBIT_RESULTS ((current->thread.fsr >> 23) & _fex)
-#else
-#define FP_INHIBIT_RESULTS ((last_task_used_math->thread.fsr >> 23) & _fex)
-#endif
-
-#ifdef CONFIG_SMP
-#define FP_TRAPPING_EXCEPTIONS ((current->thread.fsr >> 23) & 0x1f)
+#ifndef ___ASM_SPARC_SFP_MACHINE_H
+#define ___ASM_SPARC_SFP_MACHINE_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/sfp-machine_64.h>
 #else
-#define FP_TRAPPING_EXCEPTIONS ((last_task_used_math->thread.fsr >> 23) & 0x1f)
+#include <asm-sparc/sfp-machine_32.h>
 #endif
-
 #endif
diff --git a/include/asm-sparc/sfp-machine_32.h b/include/asm-sparc/sfp-machine_32.h
new file mode 100644
index 00000000000..01d9c3b5a73
--- /dev/null
+++ b/include/asm-sparc/sfp-machine_32.h
@@ -0,0 +1,212 @@
+/* Machine-dependent software floating-point definitions.
+   Sparc userland (_Q_*) version.
+   Copyright (C) 1997,1998,1999 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Richard Henderson (rth@cygnus.com),
+		  Jakub Jelinek (jj@ultra.linux.cz),
+		  David S. Miller (davem@redhat.com) and
+		  Peter Maydell (pmaydell@chiark.greenend.org.uk).
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If
+   not, write to the Free Software Foundation, Inc.,
+   59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
+
+#ifndef _SFP_MACHINE_H
+#define _SFP_MACHINE_H
+
+
+#define _FP_W_TYPE_SIZE		32
+#define _FP_W_TYPE		unsigned long
+#define _FP_WS_TYPE		signed long
+#define _FP_I_TYPE		long
+
+#define _FP_MUL_MEAT_S(R,X,Y)					\
+  _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
+#define _FP_MUL_MEAT_D(R,X,Y)					\
+  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
+#define _FP_MUL_MEAT_Q(R,X,Y)					\
+  _FP_MUL_MEAT_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
+
+#define _FP_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_udiv(S,R,X,Y)
+#define _FP_DIV_MEAT_D(R,X,Y)	_FP_DIV_MEAT_2_udiv(D,R,X,Y)
+#define _FP_DIV_MEAT_Q(R,X,Y)	_FP_DIV_MEAT_4_udiv(Q,R,X,Y)
+
+#define _FP_NANFRAC_S		((_FP_QNANBIT_S << 1) - 1)
+#define _FP_NANFRAC_D		((_FP_QNANBIT_D << 1) - 1), -1
+#define _FP_NANFRAC_Q		((_FP_QNANBIT_Q << 1) - 1), -1, -1, -1
+#define _FP_NANSIGN_S		0
+#define _FP_NANSIGN_D		0
+#define _FP_NANSIGN_Q		0
+
+#define _FP_KEEPNANFRACP 1
+
+/* If one NaN is signaling and the other is not,
+ * we choose that one, otherwise we choose X.
+ */
+/* For _Qp_* and _Q_*, this should prefer X, for
+ * CPU instruction emulation this should prefer Y.
+ * (see SPAMv9 B.2.2 section).
+ */
+#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
+  do {								\
+    if ((_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs)		\
+	&& !(_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs))	\
+      {								\
+	R##_s = X##_s;						\
+	_FP_FRAC_COPY_##wc(R,X);				\
+      }								\
+    else							\
+      {								\
+	R##_s = Y##_s;						\
+	_FP_FRAC_COPY_##wc(R,Y);				\
+      }								\
+    R##_c = FP_CLS_NAN;						\
+  } while (0)
+
+/* Some assembly to speed things up. */
+#define __FP_FRAC_ADD_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
+  __asm__ ("addcc %r7,%8,%2\n\t"					\
+	   "addxcc %r5,%6,%1\n\t"					\
+	   "addx %r3,%4,%0\n"						\
+	   : "=r" ((USItype)(r2)),					\
+	     "=&r" ((USItype)(r1)),					\
+	     "=&r" ((USItype)(r0))					\
+	   : "%rJ" ((USItype)(x2)),					\
+	     "rI" ((USItype)(y2)),					\
+	     "%rJ" ((USItype)(x1)),					\
+	     "rI" ((USItype)(y1)),					\
+	     "%rJ" ((USItype)(x0)),					\
+	     "rI" ((USItype)(y0))					\
+	   : "cc")
+
+#define __FP_FRAC_SUB_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
+  __asm__ ("subcc %r7,%8,%2\n\t"					\
+	    "subxcc %r5,%6,%1\n\t"					\
+	    "subx %r3,%4,%0\n"						\
+	   : "=r" ((USItype)(r2)),					\
+	     "=&r" ((USItype)(r1)),					\
+	     "=&r" ((USItype)(r0))					\
+	   : "%rJ" ((USItype)(x2)),					\
+	     "rI" ((USItype)(y2)),					\
+	     "%rJ" ((USItype)(x1)),					\
+	     "rI" ((USItype)(y1)),					\
+	     "%rJ" ((USItype)(x0)),					\
+	     "rI" ((USItype)(y0))					\
+	   : "cc")
+
+#define __FP_FRAC_ADD_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
+  do {									\
+    /* We need to fool gcc,  as we need to pass more than 10		\
+       input/outputs.  */						\
+    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
+    __asm__ __volatile__ (						\
+	    "addcc %r8,%9,%1\n\t"					\
+	    "addxcc %r6,%7,%0\n\t"					\
+	    "addxcc %r4,%5,%%g2\n\t"					\
+	    "addx %r2,%3,%%g1\n\t"					\
+	   : "=&r" ((USItype)(r1)),					\
+	     "=&r" ((USItype)(r0))					\
+	   : "%rJ" ((USItype)(x3)),					\
+	     "rI" ((USItype)(y3)),					\
+	     "%rJ" ((USItype)(x2)),					\
+	     "rI" ((USItype)(y2)),					\
+	     "%rJ" ((USItype)(x1)),					\
+	     "rI" ((USItype)(y1)),					\
+	     "%rJ" ((USItype)(x0)),					\
+	     "rI" ((USItype)(y0))					\
+	   : "cc", "g1", "g2");						\
+    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
+    r3 = _t1; r2 = _t2;							\
+  } while (0)
+
+#define __FP_FRAC_SUB_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
+  do {									\
+    /* We need to fool gcc,  as we need to pass more than 10		\
+       input/outputs.  */						\
+    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
+    __asm__ __volatile__ (						\
+	    "subcc %r8,%9,%1\n\t"					\
+	    "subxcc %r6,%7,%0\n\t"					\
+	    "subxcc %r4,%5,%%g2\n\t"					\
+	    "subx %r2,%3,%%g1\n\t"					\
+	   : "=&r" ((USItype)(r1)),					\
+	     "=&r" ((USItype)(r0))					\
+	   : "%rJ" ((USItype)(x3)),					\
+	     "rI" ((USItype)(y3)),					\
+	     "%rJ" ((USItype)(x2)),					\
+	     "rI" ((USItype)(y2)),					\
+	     "%rJ" ((USItype)(x1)),					\
+	     "rI" ((USItype)(y1)),					\
+	     "%rJ" ((USItype)(x0)),					\
+	     "rI" ((USItype)(y0))					\
+	   : "cc", "g1", "g2");						\
+    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
+    r3 = _t1; r2 = _t2;							\
+  } while (0)
+
+#define __FP_FRAC_DEC_3(x2,x1,x0,y2,y1,y0) __FP_FRAC_SUB_3(x2,x1,x0,x2,x1,x0,y2,y1,y0)
+
+#define __FP_FRAC_DEC_4(x3,x2,x1,x0,y3,y2,y1,y0) __FP_FRAC_SUB_4(x3,x2,x1,x0,x3,x2,x1,x0,y3,y2,y1,y0)
+
+#define __FP_FRAC_ADDI_4(x3,x2,x1,x0,i)					\
+  __asm__ ("addcc %3,%4,%3\n\t"						\
+	   "addxcc %2,%%g0,%2\n\t"					\
+	   "addxcc %1,%%g0,%1\n\t"					\
+	   "addx %0,%%g0,%0\n\t"					\
+	   : "=&r" ((USItype)(x3)),					\
+	     "=&r" ((USItype)(x2)),					\
+	     "=&r" ((USItype)(x1)),					\
+	     "=&r" ((USItype)(x0))					\
+	   : "rI" ((USItype)(i)),					\
+	     "0" ((USItype)(x3)),					\
+	     "1" ((USItype)(x2)),					\
+	     "2" ((USItype)(x1)),					\
+	     "3" ((USItype)(x0))					\
+	   : "cc")
+
+#ifndef CONFIG_SMP
+extern struct task_struct *last_task_used_math;
+#endif
+
+/* Obtain the current rounding mode. */
+#ifndef FP_ROUNDMODE
+#ifdef CONFIG_SMP
+#define FP_ROUNDMODE	((current->thread.fsr >> 30) & 0x3)
+#else
+#define FP_ROUNDMODE	((last_task_used_math->thread.fsr >> 30) & 0x3)
+#endif
+#endif
+
+/* Exception flags. */
+#define FP_EX_INVALID		(1 << 4)
+#define FP_EX_OVERFLOW		(1 << 3)
+#define FP_EX_UNDERFLOW		(1 << 2)
+#define FP_EX_DIVZERO		(1 << 1)
+#define FP_EX_INEXACT		(1 << 0)
+
+#define FP_HANDLE_EXCEPTIONS return _fex
+
+#ifdef CONFIG_SMP
+#define FP_INHIBIT_RESULTS ((current->thread.fsr >> 23) & _fex)
+#else
+#define FP_INHIBIT_RESULTS ((last_task_used_math->thread.fsr >> 23) & _fex)
+#endif
+
+#ifdef CONFIG_SMP
+#define FP_TRAPPING_EXCEPTIONS ((current->thread.fsr >> 23) & 0x1f)
+#else
+#define FP_TRAPPING_EXCEPTIONS ((last_task_used_math->thread.fsr >> 23) & 0x1f)
+#endif
+
+#endif
diff --git a/include/asm-sparc/sfp-machine_64.h b/include/asm-sparc/sfp-machine_64.h
new file mode 100644
index 00000000000..ca913ef40bd
--- /dev/null
+++ b/include/asm-sparc/sfp-machine_64.h
@@ -0,0 +1,93 @@
+/* Machine-dependent software floating-point definitions.
+   Sparc64 kernel version.
+   Copyright (C) 1997,1998,1999 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Richard Henderson (rth@cygnus.com),
+		  Jakub Jelinek (jj@ultra.linux.cz) and
+		  David S. Miller (davem@redhat.com).
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If
+   not, write to the Free Software Foundation, Inc.,
+   59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
+
+#ifndef _SFP_MACHINE_H
+#define _SFP_MACHINE_H
+
+#define _FP_W_TYPE_SIZE		64
+#define _FP_W_TYPE		unsigned long
+#define _FP_WS_TYPE		signed long
+#define _FP_I_TYPE		long
+
+#define _FP_MUL_MEAT_S(R,X,Y)					\
+  _FP_MUL_MEAT_1_imm(_FP_WFRACBITS_S,R,X,Y)
+#define _FP_MUL_MEAT_D(R,X,Y)					\
+  _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
+#define _FP_MUL_MEAT_Q(R,X,Y)					\
+  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
+
+#define _FP_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_imm(S,R,X,Y,_FP_DIV_HELP_imm)
+#define _FP_DIV_MEAT_D(R,X,Y)	_FP_DIV_MEAT_1_udiv_norm(D,R,X,Y)
+#define _FP_DIV_MEAT_Q(R,X,Y)	_FP_DIV_MEAT_2_udiv(Q,R,X,Y)
+
+#define _FP_NANFRAC_S		((_FP_QNANBIT_S << 1) - 1)
+#define _FP_NANFRAC_D		((_FP_QNANBIT_D << 1) - 1)
+#define _FP_NANFRAC_Q		((_FP_QNANBIT_Q << 1) - 1), -1
+#define _FP_NANSIGN_S		0
+#define _FP_NANSIGN_D		0
+#define _FP_NANSIGN_Q		0
+
+#define _FP_KEEPNANFRACP 1
+
+/* If one NaN is signaling and the other is not,
+ * we choose that one, otherwise we choose X.
+ */
+/* For _Qp_* and _Q_*, this should prefer X, for
+ * CPU instruction emulation this should prefer Y.
+ * (see SPAMv9 B.2.2 section).
+ */
+#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
+  do {								\
+    if ((_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs)		\
+	&& !(_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs))	\
+      {								\
+	R##_s = X##_s;						\
+	_FP_FRAC_COPY_##wc(R,X);				\
+      }								\
+    else							\
+      {								\
+	R##_s = Y##_s;						\
+	_FP_FRAC_COPY_##wc(R,Y);				\
+      }								\
+    R##_c = FP_CLS_NAN;						\
+  } while (0)
+
+/* Obtain the current rounding mode. */
+#ifndef FP_ROUNDMODE
+#define FP_ROUNDMODE	((current_thread_info()->xfsr[0] >> 30) & 0x3)
+#endif
+
+/* Exception flags. */
+#define FP_EX_INVALID		(1 << 4)
+#define FP_EX_OVERFLOW		(1 << 3)
+#define FP_EX_UNDERFLOW		(1 << 2)
+#define FP_EX_DIVZERO		(1 << 1)
+#define FP_EX_INEXACT		(1 << 0)
+
+#define FP_HANDLE_EXCEPTIONS return _fex
+
+#define FP_INHIBIT_RESULTS ((current_thread_info()->xfsr[0] >> 23) & _fex)
+
+#define FP_TRAPPING_EXCEPTIONS ((current_thread_info()->xfsr[0] >> 23) & 0x1f)
+
+#endif
diff --git a/include/asm-sparc/shmbuf.h b/include/asm-sparc/shmbuf.h
index 1ff9da8bec7..83a16055363 100644
--- a/include/asm-sparc/shmbuf.h
+++ b/include/asm-sparc/shmbuf.h
@@ -11,13 +11,19 @@
  * - 2 miscellaneous 32-bit values
  */
 
+#if defined(__sparc__) && defined(__arch64__)
+# define PADDING(x)
+#else
+# define PADDING(x) unsigned int x;
+#endif
+
 struct shmid64_ds {
 	struct ipc64_perm	shm_perm;	/* operation perms */
-	unsigned int		__pad1;
+	PADDING(__pad1)
 	__kernel_time_t		shm_atime;	/* last attach time */
-	unsigned int		__pad2;
+	PADDING(__pad2)
 	__kernel_time_t		shm_dtime;	/* last detach time */
-	unsigned int		__pad3;
+	PADDING(__pad3)
 	__kernel_time_t		shm_ctime;	/* last change time */
 	size_t			shm_segsz;	/* size of segment (bytes) */
 	__kernel_pid_t		shm_cpid;	/* pid of creator */
@@ -39,4 +45,6 @@ struct shminfo64 {
 	unsigned long	__unused4;
 };
 
+#undef PADDING
+
 #endif /* _SPARC_SHMBUF_H */
diff --git a/include/asm-sparc/shmparam.h b/include/asm-sparc/shmparam.h
index 59a1243c12f..16fda7e9acc 100644
--- a/include/asm-sparc/shmparam.h
+++ b/include/asm-sparc/shmparam.h
@@ -1,11 +1,8 @@
-#ifndef _ASMSPARC_SHMPARAM_H
-#define _ASMSPARC_SHMPARAM_H
-
-#define __ARCH_FORCE_SHMLBA 	1
-
-extern int vac_cache_size;
-#define SHMLBA (vac_cache_size ? vac_cache_size : \
-		(sparc_cpu_model == sun4c ? (64 * 1024) : \
-		 (sparc_cpu_model == sun4 ? (128 * 1024) : PAGE_SIZE)))
-
-#endif /* _ASMSPARC_SHMPARAM_H */
+#ifndef ___ASM_SPARC_SHMPARAM_H
+#define ___ASM_SPARC_SHMPARAM_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/shmparam_64.h>
+#else
+#include <asm-sparc/shmparam_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/shmparam_32.h b/include/asm-sparc/shmparam_32.h
new file mode 100644
index 00000000000..59a1243c12f
--- /dev/null
+++ b/include/asm-sparc/shmparam_32.h
@@ -0,0 +1,11 @@
+#ifndef _ASMSPARC_SHMPARAM_H
+#define _ASMSPARC_SHMPARAM_H
+
+#define __ARCH_FORCE_SHMLBA 	1
+
+extern int vac_cache_size;
+#define SHMLBA (vac_cache_size ? vac_cache_size : \
+		(sparc_cpu_model == sun4c ? (64 * 1024) : \
+		 (sparc_cpu_model == sun4 ? (128 * 1024) : PAGE_SIZE)))
+
+#endif /* _ASMSPARC_SHMPARAM_H */
diff --git a/include/asm-sparc/shmparam_64.h b/include/asm-sparc/shmparam_64.h
new file mode 100644
index 00000000000..1ed0d6701a9
--- /dev/null
+++ b/include/asm-sparc/shmparam_64.h
@@ -0,0 +1,10 @@
+#ifndef _ASMSPARC64_SHMPARAM_H
+#define _ASMSPARC64_SHMPARAM_H
+
+#include <asm/spitfire.h>
+
+#define __ARCH_FORCE_SHMLBA	1
+/* attach addr a multiple of this */
+#define	SHMLBA	((PAGE_SIZE > L1DCACHE_SIZE) ? PAGE_SIZE : L1DCACHE_SIZE)
+
+#endif /* _ASMSPARC64_SHMPARAM_H */
diff --git a/include/asm-sparc/sigcontext.h b/include/asm-sparc/sigcontext.h
index c5fb60dcbd7..82fc7d54a4f 100644
--- a/include/asm-sparc/sigcontext.h
+++ b/include/asm-sparc/sigcontext.h
@@ -1,62 +1,8 @@
-#ifndef __SPARC_SIGCONTEXT_H
-#define __SPARC_SIGCONTEXT_H
-
-#ifdef __KERNEL__
-#include <asm/ptrace.h>
-
-#ifndef __ASSEMBLY__
-
-#define __SUNOS_MAXWIN   31
-
-/* This is what SunOS does, so shall I. */
-struct sigcontext {
-	int sigc_onstack;      /* state to restore */
-	int sigc_mask;         /* sigmask to restore */
-	int sigc_sp;           /* stack pointer */
-	int sigc_pc;           /* program counter */
-	int sigc_npc;          /* next program counter */
-	int sigc_psr;          /* for condition codes etc */
-	int sigc_g1;           /* User uses these two registers */
-	int sigc_o0;           /* within the trampoline code. */
-
-	/* Now comes information regarding the users window set
-	 * at the time of the signal.
-	 */
-	int sigc_oswins;       /* outstanding windows */
-
-	/* stack ptrs for each regwin buf */
-	char *sigc_spbuf[__SUNOS_MAXWIN];
-
-	/* Windows to restore after signal */
-	struct {
-		unsigned long	locals[8];
-		unsigned long	ins[8];
-	} sigc_wbuf[__SUNOS_MAXWIN];
-};
-
-typedef struct {
-	struct {
-		unsigned long psr;
-		unsigned long pc;
-		unsigned long npc;
-		unsigned long y;
-		unsigned long u_regs[16]; /* globals and ins */
-	}		si_regs;
-	int		si_mask;
-} __siginfo_t;
-
-typedef struct {
-	unsigned   long si_float_regs [32];
-	unsigned   long si_fsr;
-	unsigned   long si_fpqdepth;
-	struct {
-		unsigned long *insn_addr;
-		unsigned long insn;
-	} si_fpqueue [16];
-} __siginfo_fpu_t;
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* (__KERNEL__) */
-
-#endif /* !(__SPARC_SIGCONTEXT_H) */
+#ifndef ___ASM_SPARC_SIGCONTEXT_H
+#define ___ASM_SPARC_SIGCONTEXT_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/sigcontext_64.h>
+#else
+#include <asm-sparc/sigcontext_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/sigcontext_32.h b/include/asm-sparc/sigcontext_32.h
new file mode 100644
index 00000000000..c5fb60dcbd7
--- /dev/null
+++ b/include/asm-sparc/sigcontext_32.h
@@ -0,0 +1,62 @@
+#ifndef __SPARC_SIGCONTEXT_H
+#define __SPARC_SIGCONTEXT_H
+
+#ifdef __KERNEL__
+#include <asm/ptrace.h>
+
+#ifndef __ASSEMBLY__
+
+#define __SUNOS_MAXWIN   31
+
+/* This is what SunOS does, so shall I. */
+struct sigcontext {
+	int sigc_onstack;      /* state to restore */
+	int sigc_mask;         /* sigmask to restore */
+	int sigc_sp;           /* stack pointer */
+	int sigc_pc;           /* program counter */
+	int sigc_npc;          /* next program counter */
+	int sigc_psr;          /* for condition codes etc */
+	int sigc_g1;           /* User uses these two registers */
+	int sigc_o0;           /* within the trampoline code. */
+
+	/* Now comes information regarding the users window set
+	 * at the time of the signal.
+	 */
+	int sigc_oswins;       /* outstanding windows */
+
+	/* stack ptrs for each regwin buf */
+	char *sigc_spbuf[__SUNOS_MAXWIN];
+
+	/* Windows to restore after signal */
+	struct {
+		unsigned long	locals[8];
+		unsigned long	ins[8];
+	} sigc_wbuf[__SUNOS_MAXWIN];
+};
+
+typedef struct {
+	struct {
+		unsigned long psr;
+		unsigned long pc;
+		unsigned long npc;
+		unsigned long y;
+		unsigned long u_regs[16]; /* globals and ins */
+	}		si_regs;
+	int		si_mask;
+} __siginfo_t;
+
+typedef struct {
+	unsigned   long si_float_regs [32];
+	unsigned   long si_fsr;
+	unsigned   long si_fpqdepth;
+	struct {
+		unsigned long *insn_addr;
+		unsigned long insn;
+	} si_fpqueue [16];
+} __siginfo_fpu_t;
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* (__KERNEL__) */
+
+#endif /* !(__SPARC_SIGCONTEXT_H) */
diff --git a/include/asm-sparc/sigcontext_64.h b/include/asm-sparc/sigcontext_64.h
new file mode 100644
index 00000000000..1c868d680cf
--- /dev/null
+++ b/include/asm-sparc/sigcontext_64.h
@@ -0,0 +1,87 @@
+#ifndef __SPARC64_SIGCONTEXT_H
+#define __SPARC64_SIGCONTEXT_H
+
+#ifdef __KERNEL__
+#include <asm/ptrace.h>
+#endif
+
+#ifndef __ASSEMBLY__
+
+#ifdef __KERNEL__
+
+#define __SUNOS_MAXWIN   31
+
+/* This is what SunOS does, so shall I unless we use new 32bit signals or rt signals. */
+struct sigcontext32 {
+	int sigc_onstack;      /* state to restore */
+	int sigc_mask;         /* sigmask to restore */
+	int sigc_sp;           /* stack pointer */
+	int sigc_pc;           /* program counter */
+	int sigc_npc;          /* next program counter */
+	int sigc_psr;          /* for condition codes etc */
+	int sigc_g1;           /* User uses these two registers */
+	int sigc_o0;           /* within the trampoline code. */
+
+	/* Now comes information regarding the users window set
+	 * at the time of the signal.
+	 */
+	int sigc_oswins;       /* outstanding windows */
+
+	/* stack ptrs for each regwin buf */
+	unsigned sigc_spbuf[__SUNOS_MAXWIN];
+
+	/* Windows to restore after signal */
+	struct reg_window32 sigc_wbuf[__SUNOS_MAXWIN];
+};
+
+#endif
+
+#ifdef __KERNEL__
+
+/* This is what we use for 32bit new non-rt signals. */
+
+typedef struct {
+	struct {
+		unsigned int psr;
+		unsigned int pc;
+		unsigned int npc;
+		unsigned int y;
+		unsigned int u_regs[16]; /* globals and ins */
+	}			si_regs;
+	int			si_mask;
+} __siginfo32_t;
+
+#endif
+
+typedef struct {
+	unsigned   int si_float_regs [64];
+	unsigned   long si_fsr;
+	unsigned   long si_gsr;
+	unsigned   long si_fprs;
+} __siginfo_fpu_t;
+
+/* This is what SunOS doesn't, so we have to write this alone
+   and do it properly. */
+struct sigcontext {
+	/* The size of this array has to match SI_MAX_SIZE from siginfo.h */
+	char			sigc_info[128];
+	struct {
+		unsigned long	u_regs[16]; /* globals and ins */
+		unsigned long	tstate;
+		unsigned long	tpc;
+		unsigned long	tnpc;
+		unsigned int	y;
+		unsigned int	fprs;
+	}			sigc_regs;
+	__siginfo_fpu_t *	sigc_fpu_save;
+	struct {
+		void	*	ss_sp;
+		int		ss_flags;
+		unsigned long	ss_size;
+	}			sigc_stack;
+	unsigned long		sigc_mask;
+};
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC64_SIGCONTEXT_H) */
diff --git a/include/asm-sparc/siginfo.h b/include/asm-sparc/siginfo.h
index 3c71af135c5..2c9fccf4ce1 100644
--- a/include/asm-sparc/siginfo.h
+++ b/include/asm-sparc/siginfo.h
@@ -1,17 +1,8 @@
-#ifndef _SPARC_SIGINFO_H
-#define _SPARC_SIGINFO_H
-
-#define __ARCH_SI_UID_T		unsigned int
-#define __ARCH_SI_TRAPNO
-
-#include <asm-generic/siginfo.h>
-
-#define SI_NOINFO	32767		/* no information in siginfo_t */
-
-/*
- * SIGEMT si_codes
- */
-#define EMT_TAGOVF	(__SI_FAULT|1)	/* tag overflow */
-#define NSIGEMT		1
-
-#endif /* !(_SPARC_SIGINFO_H) */
+#ifndef ___ASM_SPARC_SIGINFO_H
+#define ___ASM_SPARC_SIGINFO_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/siginfo_64.h>
+#else
+#include <asm-sparc/siginfo_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/siginfo_32.h b/include/asm-sparc/siginfo_32.h
new file mode 100644
index 00000000000..3c71af135c5
--- /dev/null
+++ b/include/asm-sparc/siginfo_32.h
@@ -0,0 +1,17 @@
+#ifndef _SPARC_SIGINFO_H
+#define _SPARC_SIGINFO_H
+
+#define __ARCH_SI_UID_T		unsigned int
+#define __ARCH_SI_TRAPNO
+
+#include <asm-generic/siginfo.h>
+
+#define SI_NOINFO	32767		/* no information in siginfo_t */
+
+/*
+ * SIGEMT si_codes
+ */
+#define EMT_TAGOVF	(__SI_FAULT|1)	/* tag overflow */
+#define NSIGEMT		1
+
+#endif /* !(_SPARC_SIGINFO_H) */
diff --git a/include/asm-sparc/siginfo_64.h b/include/asm-sparc/siginfo_64.h
new file mode 100644
index 00000000000..c96e6c30f8b
--- /dev/null
+++ b/include/asm-sparc/siginfo_64.h
@@ -0,0 +1,32 @@
+#ifndef _SPARC64_SIGINFO_H
+#define _SPARC64_SIGINFO_H
+
+#define SI_PAD_SIZE32	((SI_MAX_SIZE/sizeof(int)) - 3)
+
+#define __ARCH_SI_PREAMBLE_SIZE	(4 * sizeof(int))
+#define __ARCH_SI_TRAPNO
+#define __ARCH_SI_BAND_T int
+
+#include <asm-generic/siginfo.h>
+
+#ifdef __KERNEL__
+
+#include <linux/compat.h>
+
+#ifdef CONFIG_COMPAT
+
+struct compat_siginfo;
+
+#endif /* CONFIG_COMPAT */
+
+#endif /* __KERNEL__ */
+
+#define SI_NOINFO	32767		/* no information in siginfo_t */
+
+/*
+ * SIGEMT si_codes
+ */
+#define EMT_TAGOVF	(__SI_FAULT|1)	/* tag overflow */
+#define NSIGEMT		1
+
+#endif
diff --git a/include/asm-sparc/signal.h b/include/asm-sparc/signal.h
index 683657d6e68..36f5f9e482f 100644
--- a/include/asm-sparc/signal.h
+++ b/include/asm-sparc/signal.h
@@ -1,207 +1,8 @@
-#ifndef _ASMSPARC_SIGNAL_H
-#define _ASMSPARC_SIGNAL_H
-
-#include <asm/sigcontext.h>
-#include <linux/compiler.h>
-
-#ifdef __KERNEL__
-#ifndef __ASSEMBLY__
-#include <linux/personality.h>
-#include <linux/types.h>
-#endif
-#endif
-
-/* On the Sparc the signal handlers get passed a 'sub-signal' code
- * for certain signal types, which we document here.
- */
-#define SIGHUP		 1
-#define SIGINT		 2
-#define SIGQUIT		 3
-#define SIGILL		 4
-#define    SUBSIG_STACK       0
-#define    SUBSIG_ILLINST     2
-#define    SUBSIG_PRIVINST    3
-#define    SUBSIG_BADTRAP(t)  (0x80 + (t))
-
-#define SIGTRAP		 5
-#define SIGABRT		 6
-#define SIGIOT		 6
-
-#define SIGEMT           7
-#define    SUBSIG_TAG    10
-
-#define SIGFPE		 8
-#define    SUBSIG_FPDISABLED     0x400
-#define    SUBSIG_FPERROR        0x404
-#define    SUBSIG_FPINTOVFL      0x001
-#define    SUBSIG_FPSTSIG        0x002
-#define    SUBSIG_IDIVZERO       0x014
-#define    SUBSIG_FPINEXACT      0x0c4
-#define    SUBSIG_FPDIVZERO      0x0c8
-#define    SUBSIG_FPUNFLOW       0x0cc
-#define    SUBSIG_FPOPERROR      0x0d0
-#define    SUBSIG_FPOVFLOW       0x0d4
-
-#define SIGKILL		 9
-#define SIGBUS          10
-#define    SUBSIG_BUSTIMEOUT    1
-#define    SUBSIG_ALIGNMENT     2
-#define    SUBSIG_MISCERROR     5
-
-#define SIGSEGV		11
-#define    SUBSIG_NOMAPPING     3
-#define    SUBSIG_PROTECTION    4
-#define    SUBSIG_SEGERROR      5
-
-#define SIGSYS		12
-
-#define SIGPIPE		13
-#define SIGALRM		14
-#define SIGTERM		15
-#define SIGURG          16
-
-/* SunOS values which deviate from the Linux/i386 ones */
-#define SIGSTOP		17
-#define SIGTSTP		18
-#define SIGCONT		19
-#define SIGCHLD		20
-#define SIGTTIN		21
-#define SIGTTOU		22
-#define SIGIO		23
-#define SIGPOLL		SIGIO   /* SysV name for SIGIO */
-#define SIGXCPU		24
-#define SIGXFSZ		25
-#define SIGVTALRM	26
-#define SIGPROF		27
-#define SIGWINCH	28
-#define SIGLOST		29
-#define SIGPWR		SIGLOST
-#define SIGUSR1		30
-#define SIGUSR2		31
-
-/* Most things should be clean enough to redefine this at will, if care
- * is taken to make libc match.
- */
-
-#define __OLD_NSIG	32
-#define __NEW_NSIG	64
-#define _NSIG_BPW	32
-#define _NSIG_WORDS	(__NEW_NSIG / _NSIG_BPW)
-
-#define SIGRTMIN	32
-#define SIGRTMAX	__NEW_NSIG
-
-#if defined(__KERNEL__) || defined(__WANT_POSIX1B_SIGNALS__)
-#define	_NSIG		__NEW_NSIG
-#define __new_sigset_t	sigset_t
-#define __new_sigaction	sigaction
-#define __old_sigset_t	old_sigset_t
-#define __old_sigaction	old_sigaction
+#ifndef ___ASM_SPARC_SIGNAL_H
+#define ___ASM_SPARC_SIGNAL_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/signal_64.h>
 #else
-#define _NSIG		__OLD_NSIG
-#define __old_sigset_t	sigset_t
-#define __old_sigaction	sigaction
-#endif
-
-#ifndef __ASSEMBLY__
-
-typedef unsigned long __old_sigset_t;
-
-typedef struct {
-	unsigned long	sig[_NSIG_WORDS];
-} __new_sigset_t;
-
-
-#ifdef __KERNEL__
-/* A SunOS sigstack */
-struct sigstack {
-	char *the_stack;
-	int   cur_status;
-};
+#include <asm-sparc/signal_32.h>
 #endif
-
-/* Sigvec flags */
-#define _SV_SSTACK    1u    /* This signal handler should use sig-stack */
-#define _SV_INTR      2u    /* Sig return should not restart system call */
-#define _SV_RESET     4u    /* Set handler to SIG_DFL upon taken signal */
-#define _SV_IGNCHILD  8u    /* Do not send SIGCHLD */
-
-/*
- * sa_flags values: SA_STACK is not currently supported, but will allow the
- * usage of signal stacks by using the (now obsolete) sa_restorer field in
- * the sigaction structure as a stack pointer. This is now possible due to
- * the changes in signal handling. LBT 010493.
- * SA_RESTART flag to get restarting signals (which were the default long ago)
- */
-#define SA_NOCLDSTOP	_SV_IGNCHILD
-#define SA_STACK	_SV_SSTACK
-#define SA_ONSTACK	_SV_SSTACK
-#define SA_RESTART	_SV_INTR
-#define SA_ONESHOT	_SV_RESET
-#define SA_NOMASK	0x20u
-#define SA_NOCLDWAIT	0x100u
-#define SA_SIGINFO	0x200u
-
-#define SIG_BLOCK          0x01	/* for blocking signals */
-#define SIG_UNBLOCK        0x02	/* for unblocking signals */
-#define SIG_SETMASK        0x04	/* for setting the signal mask */
-
-/* 
- * sigaltstack controls
- */
-#define SS_ONSTACK	1
-#define SS_DISABLE	2
-
-#define MINSIGSTKSZ	4096
-#define SIGSTKSZ	16384
-
-#ifdef __KERNEL__
-/*
- * DJHR
- * SA_STATIC_ALLOC is used for the SPARC system to indicate that this
- * interrupt handler's irq structure should be statically allocated
- * by the request_irq routine.
- * The alternative is that arch/sparc/kernel/irq.c has carnal knowledge
- * of interrupt usage and that sucks. Also without a flag like this
- * it may be possible for the free_irq routine to attempt to free
- * statically allocated data.. which is NOT GOOD.
- *
- */
-#define SA_STATIC_ALLOC		0x8000
 #endif
-
-#include <asm-generic/signal.h>
-
-#ifdef __KERNEL__
-struct __new_sigaction {
-	__sighandler_t	sa_handler;
-	unsigned long	sa_flags;
-	void		(*sa_restorer)(void);	/* Not used by Linux/SPARC */
-	__new_sigset_t	sa_mask;
-};
-
-struct k_sigaction {
-	struct __new_sigaction	sa;
-	void			__user *ka_restorer;
-};
-
-struct __old_sigaction {
-	__sighandler_t	sa_handler;
-	__old_sigset_t	sa_mask;
-	unsigned long	sa_flags;
-	void		(*sa_restorer) (void);	/* not used by Linux/SPARC */
-};
-
-typedef struct sigaltstack {
-	void		__user *ss_sp;
-	int		ss_flags;
-	size_t		ss_size;
-} stack_t;
-
-#define ptrace_signal_deliver(regs, cookie) do { } while (0)
-
-#endif /* !(__KERNEL__) */
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(_ASMSPARC_SIGNAL_H) */
diff --git a/include/asm-sparc/signal_32.h b/include/asm-sparc/signal_32.h
new file mode 100644
index 00000000000..96a60ab03ca
--- /dev/null
+++ b/include/asm-sparc/signal_32.h
@@ -0,0 +1,207 @@
+#ifndef _ASMSPARC_SIGNAL_H
+#define _ASMSPARC_SIGNAL_H
+
+#include <asm/sigcontext.h>
+#include <linux/compiler.h>
+
+#ifdef __KERNEL__
+#ifndef __ASSEMBLY__
+#include <linux/personality.h>
+#include <linux/types.h>
+#endif
+#endif
+
+/* On the Sparc the signal handlers get passed a 'sub-signal' code
+ * for certain signal types, which we document here.
+ */
+#define SIGHUP		 1
+#define SIGINT		 2
+#define SIGQUIT		 3
+#define SIGILL		 4
+#define    SUBSIG_STACK       0
+#define    SUBSIG_ILLINST     2
+#define    SUBSIG_PRIVINST    3
+#define    SUBSIG_BADTRAP(t)  (0x80 + (t))
+
+#define SIGTRAP		 5
+#define SIGABRT		 6
+#define SIGIOT		 6
+
+#define SIGEMT           7
+#define    SUBSIG_TAG    10
+
+#define SIGFPE		 8
+#define    SUBSIG_FPDISABLED     0x400
+#define    SUBSIG_FPERROR        0x404
+#define    SUBSIG_FPINTOVFL      0x001
+#define    SUBSIG_FPSTSIG        0x002
+#define    SUBSIG_IDIVZERO       0x014
+#define    SUBSIG_FPINEXACT      0x0c4
+#define    SUBSIG_FPDIVZERO      0x0c8
+#define    SUBSIG_FPUNFLOW       0x0cc
+#define    SUBSIG_FPOPERROR      0x0d0
+#define    SUBSIG_FPOVFLOW       0x0d4
+
+#define SIGKILL		 9
+#define SIGBUS          10
+#define    SUBSIG_BUSTIMEOUT    1
+#define    SUBSIG_ALIGNMENT     2
+#define    SUBSIG_MISCERROR     5
+
+#define SIGSEGV		11
+#define    SUBSIG_NOMAPPING     3
+#define    SUBSIG_PROTECTION    4
+#define    SUBSIG_SEGERROR      5
+
+#define SIGSYS		12
+
+#define SIGPIPE		13
+#define SIGALRM		14
+#define SIGTERM		15
+#define SIGURG          16
+
+/* SunOS values which deviate from the Linux/i386 ones */
+#define SIGSTOP		17
+#define SIGTSTP		18
+#define SIGCONT		19
+#define SIGCHLD		20
+#define SIGTTIN		21
+#define SIGTTOU		22
+#define SIGIO		23
+#define SIGPOLL		SIGIO   /* SysV name for SIGIO */
+#define SIGXCPU		24
+#define SIGXFSZ		25
+#define SIGVTALRM	26
+#define SIGPROF		27
+#define SIGWINCH	28
+#define SIGLOST		29
+#define SIGPWR		SIGLOST
+#define SIGUSR1		30
+#define SIGUSR2		31
+
+/* Most things should be clean enough to redefine this at will, if care
+ * is taken to make libc match.
+ */
+
+#define __OLD_NSIG	32
+#define __NEW_NSIG	64
+#define _NSIG_BPW	32
+#define _NSIG_WORDS	(__NEW_NSIG / _NSIG_BPW)
+
+#define SIGRTMIN	32
+#define SIGRTMAX	__NEW_NSIG
+
+#if defined(__KERNEL__) || defined(__WANT_POSIX1B_SIGNALS__)
+#define	_NSIG		__NEW_NSIG
+#define __new_sigset_t	sigset_t
+#define __new_sigaction	sigaction
+#define __old_sigset_t	old_sigset_t
+#define __old_sigaction	old_sigaction
+#else
+#define _NSIG		__OLD_NSIG
+#define __old_sigset_t	sigset_t
+#define __old_sigaction	sigaction
+#endif
+
+#ifndef __ASSEMBLY__
+
+typedef unsigned long __old_sigset_t;
+
+typedef struct {
+	unsigned long	sig[_NSIG_WORDS];
+} __new_sigset_t;
+
+
+#ifdef __KERNEL__
+/* A SunOS sigstack */
+struct sigstack {
+	char *the_stack;
+	int   cur_status;
+};
+#endif
+
+/* Sigvec flags */
+#define _SV_SSTACK    1u    /* This signal handler should use sig-stack */
+#define _SV_INTR      2u    /* Sig return should not restart system call */
+#define _SV_RESET     4u    /* Set handler to SIG_DFL upon taken signal */
+#define _SV_IGNCHILD  8u    /* Do not send SIGCHLD */
+
+/*
+ * sa_flags values: SA_STACK is not currently supported, but will allow the
+ * usage of signal stacks by using the (now obsolete) sa_restorer field in
+ * the sigaction structure as a stack pointer. This is now possible due to
+ * the changes in signal handling. LBT 010493.
+ * SA_RESTART flag to get restarting signals (which were the default long ago)
+ */
+#define SA_NOCLDSTOP	_SV_IGNCHILD
+#define SA_STACK	_SV_SSTACK
+#define SA_ONSTACK	_SV_SSTACK
+#define SA_RESTART	_SV_INTR
+#define SA_ONESHOT	_SV_RESET
+#define SA_NOMASK	0x20u
+#define SA_NOCLDWAIT	0x100u
+#define SA_SIGINFO	0x200u
+
+#define SIG_BLOCK          0x01	/* for blocking signals */
+#define SIG_UNBLOCK        0x02	/* for unblocking signals */
+#define SIG_SETMASK        0x04	/* for setting the signal mask */
+
+/*
+ * sigaltstack controls
+ */
+#define SS_ONSTACK	1
+#define SS_DISABLE	2
+
+#define MINSIGSTKSZ	4096
+#define SIGSTKSZ	16384
+
+#ifdef __KERNEL__
+/*
+ * DJHR
+ * SA_STATIC_ALLOC is used for the SPARC system to indicate that this
+ * interrupt handler's irq structure should be statically allocated
+ * by the request_irq routine.
+ * The alternative is that arch/sparc/kernel/irq.c has carnal knowledge
+ * of interrupt usage and that sucks. Also without a flag like this
+ * it may be possible for the free_irq routine to attempt to free
+ * statically allocated data.. which is NOT GOOD.
+ *
+ */
+#define SA_STATIC_ALLOC		0x8000
+#endif
+
+#include <asm-generic/signal.h>
+
+#ifdef __KERNEL__
+struct __new_sigaction {
+	__sighandler_t	sa_handler;
+	unsigned long	sa_flags;
+	void		(*sa_restorer)(void);	/* Not used by Linux/SPARC */
+	__new_sigset_t	sa_mask;
+};
+
+struct k_sigaction {
+	struct __new_sigaction	sa;
+	void			__user *ka_restorer;
+};
+
+struct __old_sigaction {
+	__sighandler_t	sa_handler;
+	__old_sigset_t	sa_mask;
+	unsigned long	sa_flags;
+	void		(*sa_restorer) (void);	/* not used by Linux/SPARC */
+};
+
+typedef struct sigaltstack {
+	void		__user *ss_sp;
+	int		ss_flags;
+	size_t		ss_size;
+} stack_t;
+
+#define ptrace_signal_deliver(regs, cookie) do { } while (0)
+
+#endif /* !(__KERNEL__) */
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(_ASMSPARC_SIGNAL_H) */
diff --git a/include/asm-sparc/signal_64.h b/include/asm-sparc/signal_64.h
new file mode 100644
index 00000000000..ab1509a101c
--- /dev/null
+++ b/include/asm-sparc/signal_64.h
@@ -0,0 +1,194 @@
+#ifndef _ASMSPARC64_SIGNAL_H
+#define _ASMSPARC64_SIGNAL_H
+
+#include <asm/sigcontext.h>
+
+#ifdef __KERNEL__
+#ifndef __ASSEMBLY__
+#include <linux/personality.h>
+#include <linux/types.h>
+#endif
+#endif
+
+/* On the Sparc the signal handlers get passed a 'sub-signal' code
+ * for certain signal types, which we document here.
+ */
+#define SIGHUP		 1
+#define SIGINT		 2
+#define SIGQUIT		 3
+#define SIGILL		 4
+#define    SUBSIG_STACK       0
+#define    SUBSIG_ILLINST     2
+#define    SUBSIG_PRIVINST    3
+#define    SUBSIG_BADTRAP(t)  (0x80 + (t))
+
+#define SIGTRAP		 5
+#define SIGABRT		 6
+#define SIGIOT		 6
+
+#define SIGEMT           7
+#define    SUBSIG_TAG    10
+
+#define SIGFPE		 8
+#define    SUBSIG_FPDISABLED     0x400
+#define    SUBSIG_FPERROR        0x404
+#define    SUBSIG_FPINTOVFL      0x001
+#define    SUBSIG_FPSTSIG        0x002
+#define    SUBSIG_IDIVZERO       0x014
+#define    SUBSIG_FPINEXACT      0x0c4
+#define    SUBSIG_FPDIVZERO      0x0c8
+#define    SUBSIG_FPUNFLOW       0x0cc
+#define    SUBSIG_FPOPERROR      0x0d0
+#define    SUBSIG_FPOVFLOW       0x0d4
+
+#define SIGKILL		 9
+#define SIGBUS          10
+#define    SUBSIG_BUSTIMEOUT    1
+#define    SUBSIG_ALIGNMENT     2
+#define    SUBSIG_MISCERROR     5
+
+#define SIGSEGV		11
+#define    SUBSIG_NOMAPPING     3
+#define    SUBSIG_PROTECTION    4
+#define    SUBSIG_SEGERROR      5
+
+#define SIGSYS		12
+
+#define SIGPIPE		13
+#define SIGALRM		14
+#define SIGTERM		15
+#define SIGURG          16
+
+/* SunOS values which deviate from the Linux/i386 ones */
+#define SIGSTOP		17
+#define SIGTSTP		18
+#define SIGCONT		19
+#define SIGCHLD		20
+#define SIGTTIN		21
+#define SIGTTOU		22
+#define SIGIO		23
+#define SIGPOLL		SIGIO   /* SysV name for SIGIO */
+#define SIGXCPU		24
+#define SIGXFSZ		25
+#define SIGVTALRM	26
+#define SIGPROF		27
+#define SIGWINCH	28
+#define SIGLOST		29
+#define SIGPWR		SIGLOST
+#define SIGUSR1		30
+#define SIGUSR2		31
+
+/* Most things should be clean enough to redefine this at will, if care
+   is taken to make libc match.  */
+
+#define __OLD_NSIG	32
+#define __NEW_NSIG      64
+#define _NSIG_BPW     	64
+#define _NSIG_WORDS   	(__NEW_NSIG / _NSIG_BPW)
+
+#define SIGRTMIN       32
+#define SIGRTMAX       __NEW_NSIG
+
+#if defined(__KERNEL__) || defined(__WANT_POSIX1B_SIGNALS__)
+#define _NSIG			__NEW_NSIG
+#define __new_sigset_t		sigset_t
+#define __new_sigaction		sigaction
+#define __new_sigaction32	sigaction32
+#define __old_sigset_t		old_sigset_t
+#define __old_sigaction		old_sigaction
+#define __old_sigaction32	old_sigaction32
+#else
+#define _NSIG			__OLD_NSIG
+#define NSIG			_NSIG
+#define __old_sigset_t		sigset_t
+#define __old_sigaction		sigaction
+#define __old_sigaction32	sigaction32
+#endif
+
+#ifndef __ASSEMBLY__
+
+typedef unsigned long __old_sigset_t;            /* at least 32 bits */
+
+typedef struct {
+       unsigned long sig[_NSIG_WORDS];
+} __new_sigset_t;
+
+/* A SunOS sigstack */
+struct sigstack {
+	/* XXX 32-bit pointers pinhead XXX */
+	char *the_stack;
+	int   cur_status;
+};
+
+/* Sigvec flags */
+#define _SV_SSTACK    1u    /* This signal handler should use sig-stack */
+#define _SV_INTR      2u    /* Sig return should not restart system call */
+#define _SV_RESET     4u    /* Set handler to SIG_DFL upon taken signal */
+#define _SV_IGNCHILD  8u    /* Do not send SIGCHLD */
+
+/*
+ * sa_flags values: SA_STACK is not currently supported, but will allow the
+ * usage of signal stacks by using the (now obsolete) sa_restorer field in
+ * the sigaction structure as a stack pointer. This is now possible due to
+ * the changes in signal handling. LBT 010493.
+ * SA_RESTART flag to get restarting signals (which were the default long ago)
+ */
+#define SA_NOCLDSTOP	_SV_IGNCHILD
+#define SA_STACK	_SV_SSTACK
+#define SA_ONSTACK	_SV_SSTACK
+#define SA_RESTART	_SV_INTR
+#define SA_ONESHOT	_SV_RESET
+#define SA_NOMASK	0x20u
+#define SA_NOCLDWAIT    0x100u
+#define SA_SIGINFO      0x200u
+
+
+#define SIG_BLOCK          0x01	/* for blocking signals */
+#define SIG_UNBLOCK        0x02	/* for unblocking signals */
+#define SIG_SETMASK        0x04	/* for setting the signal mask */
+
+/*
+ * sigaltstack controls
+ */
+#define SS_ONSTACK	1
+#define SS_DISABLE	2
+
+#define MINSIGSTKSZ	4096
+#define SIGSTKSZ	16384
+
+#include <asm-generic/signal.h>
+
+struct __new_sigaction {
+	__sighandler_t		sa_handler;
+	unsigned long		sa_flags;
+	__sigrestore_t 		sa_restorer;  /* not used by Linux/SPARC yet */
+	__new_sigset_t		sa_mask;
+};
+
+struct __old_sigaction {
+	__sighandler_t  	sa_handler;
+	__old_sigset_t  	sa_mask;
+	unsigned long   	sa_flags;
+	void 			(*sa_restorer)(void);     /* not used by Linux/SPARC yet */
+};
+
+typedef struct sigaltstack {
+	void			__user *ss_sp;
+	int			ss_flags;
+	size_t			ss_size;
+} stack_t;
+
+#ifdef __KERNEL__
+
+struct k_sigaction {
+	struct __new_sigaction 	sa;
+	void __user		*ka_restorer;
+};
+
+#define ptrace_signal_deliver(regs, cookie) do { } while (0)
+
+#endif /* !(__KERNEL__) */
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(_ASMSPARC64_SIGNAL_H) */
diff --git a/include/asm-sparc/smp.h b/include/asm-sparc/smp.h
index b61e74bea06..1f9dedfbabd 100644
--- a/include/asm-sparc/smp.h
+++ b/include/asm-sparc/smp.h
@@ -1,173 +1,8 @@
-/* smp.h: Sparc specific SMP stuff.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC_SMP_H
-#define _SPARC_SMP_H
-
-#include <linux/threads.h>
-#include <asm/head.h>
-#include <asm/btfixup.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/cpumask.h>
-
-#endif /* __ASSEMBLY__ */
-
-#ifdef CONFIG_SMP
-
-#ifndef __ASSEMBLY__
-
-#include <asm/ptrace.h>
-#include <asm/asi.h>
-#include <asm/atomic.h>
-
-/*
- *	Private routines/data
- */
- 
-extern unsigned char boot_cpu_id;
-extern cpumask_t phys_cpu_present_map;
-#define cpu_possible_map phys_cpu_present_map
-
-typedef void (*smpfunc_t)(unsigned long, unsigned long, unsigned long,
-		       unsigned long, unsigned long);
-
-/*
- *	General functions that each host system must provide.
- */
- 
-void sun4m_init_smp(void);
-void sun4d_init_smp(void);
-
-void smp_callin(void);
-void smp_boot_cpus(void);
-void smp_store_cpu_info(int);
-
-struct seq_file;
-void smp_bogo(struct seq_file *);
-void smp_info(struct seq_file *);
-
-BTFIXUPDEF_CALL(void, smp_cross_call, smpfunc_t, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(int, __hard_smp_processor_id, void)
-BTFIXUPDEF_BLACKBOX(hard_smp_processor_id)
-BTFIXUPDEF_BLACKBOX(load_current)
-
-#define smp_cross_call(func,arg1,arg2,arg3,arg4,arg5) BTFIXUP_CALL(smp_cross_call)(func,arg1,arg2,arg3,arg4,arg5)
-
-static inline void xc0(smpfunc_t func) { smp_cross_call(func, 0, 0, 0, 0, 0); }
-static inline void xc1(smpfunc_t func, unsigned long arg1)
-{ smp_cross_call(func, arg1, 0, 0, 0, 0); }
-static inline void xc2(smpfunc_t func, unsigned long arg1, unsigned long arg2)
-{ smp_cross_call(func, arg1, arg2, 0, 0, 0); }
-static inline void xc3(smpfunc_t func, unsigned long arg1, unsigned long arg2,
-			   unsigned long arg3)
-{ smp_cross_call(func, arg1, arg2, arg3, 0, 0); }
-static inline void xc4(smpfunc_t func, unsigned long arg1, unsigned long arg2,
-			   unsigned long arg3, unsigned long arg4)
-{ smp_cross_call(func, arg1, arg2, arg3, arg4, 0); }
-static inline void xc5(smpfunc_t func, unsigned long arg1, unsigned long arg2,
-			   unsigned long arg3, unsigned long arg4, unsigned long arg5)
-{ smp_cross_call(func, arg1, arg2, arg3, arg4, arg5); }
-
-static inline int smp_call_function(void (*func)(void *info), void *info, int wait)
-{
-	xc1((smpfunc_t)func, (unsigned long)info);
-	return 0;
-}
-
-static inline int cpu_logical_map(int cpu)
-{
-	return cpu;
-}
-
-static inline int hard_smp4m_processor_id(void)
-{
-	int cpuid;
-
-	__asm__ __volatile__("rd %%tbr, %0\n\t"
-			     "srl %0, 12, %0\n\t"
-			     "and %0, 3, %0\n\t" :
-			     "=&r" (cpuid));
-	return cpuid;
-}
-
-static inline int hard_smp4d_processor_id(void)
-{
-	int cpuid;
-
-	__asm__ __volatile__("lda [%%g0] %1, %0\n\t" :
-			     "=&r" (cpuid) : "i" (ASI_M_VIKING_TMP1));
-	return cpuid;
-}
-
-#ifndef MODULE
-static inline int hard_smp_processor_id(void)
-{
-	int cpuid;
-
-	/* Black box - sun4m
-		__asm__ __volatile__("rd %%tbr, %0\n\t"
-				     "srl %0, 12, %0\n\t"
-				     "and %0, 3, %0\n\t" :
-				     "=&r" (cpuid));
-	             - sun4d
-	   	__asm__ __volatile__("lda [%g0] ASI_M_VIKING_TMP1, %0\n\t"
-	   			     "nop; nop" :
-	   			     "=&r" (cpuid));
-	   See btfixup.h and btfixupprep.c to understand how a blackbox works.
-	 */
-	__asm__ __volatile__("sethi %%hi(___b_hard_smp_processor_id), %0\n\t"
-			     "sethi %%hi(boot_cpu_id), %0\n\t"
-			     "ldub [%0 + %%lo(boot_cpu_id)], %0\n\t" :
-			     "=&r" (cpuid));
-	return cpuid;
-}
+#ifndef ___ASM_SPARC_SMP_H
+#define ___ASM_SPARC_SMP_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/smp_64.h>
 #else
-static inline int hard_smp_processor_id(void)
-{
-	int cpuid;
-	
-	__asm__ __volatile__("mov %%o7, %%g1\n\t"
-			     "call ___f___hard_smp_processor_id\n\t"
-			     " nop\n\t"
-			     "mov %%g2, %0\n\t" : "=r"(cpuid) : : "g1", "g2");
-	return cpuid;
-}
+#include <asm-sparc/smp_32.h>
+#endif
 #endif
-
-#define raw_smp_processor_id()		(current_thread_info()->cpu)
-
-#define prof_multiplier(__cpu)		cpu_data(__cpu).multiplier
-#define prof_counter(__cpu)		cpu_data(__cpu).counter
-
-void smp_setup_cpu_possible_map(void);
-
-#endif /* !(__ASSEMBLY__) */
-
-/* Sparc specific messages. */
-#define MSG_CROSS_CALL         0x0005       /* run func on cpus */
-
-/* Empirical PROM processor mailbox constants.  If the per-cpu mailbox
- * contains something other than one of these then the ipi is from
- * Linux's active_kernel_processor.  This facility exists so that
- * the boot monitor can capture all the other cpus when one catches
- * a watchdog reset or the user enters the monitor using L1-A keys.
- */
-#define MBOX_STOPCPU          0xFB
-#define MBOX_IDLECPU          0xFC
-#define MBOX_IDLECPU2         0xFD
-#define MBOX_STOPCPU2         0xFE
-
-#else /* SMP */
-
-#define hard_smp_processor_id()		0
-#define smp_setup_cpu_possible_map() do { } while (0)
-
-#endif /* !(SMP) */
-
-#define NO_PROC_ID            0xFF
-
-#endif /* !(_SPARC_SMP_H) */
diff --git a/include/asm-sparc/smp_32.h b/include/asm-sparc/smp_32.h
new file mode 100644
index 00000000000..7201752cf93
--- /dev/null
+++ b/include/asm-sparc/smp_32.h
@@ -0,0 +1,173 @@
+/* smp.h: Sparc specific SMP stuff.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef _SPARC_SMP_H
+#define _SPARC_SMP_H
+
+#include <linux/threads.h>
+#include <asm/head.h>
+#include <asm/btfixup.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/cpumask.h>
+
+#endif /* __ASSEMBLY__ */
+
+#ifdef CONFIG_SMP
+
+#ifndef __ASSEMBLY__
+
+#include <asm/ptrace.h>
+#include <asm/asi.h>
+#include <asm/atomic.h>
+
+/*
+ *	Private routines/data
+ */
+
+extern unsigned char boot_cpu_id;
+extern cpumask_t phys_cpu_present_map;
+#define cpu_possible_map phys_cpu_present_map
+
+typedef void (*smpfunc_t)(unsigned long, unsigned long, unsigned long,
+		       unsigned long, unsigned long);
+
+/*
+ *	General functions that each host system must provide.
+ */
+
+void sun4m_init_smp(void);
+void sun4d_init_smp(void);
+
+void smp_callin(void);
+void smp_boot_cpus(void);
+void smp_store_cpu_info(int);
+
+struct seq_file;
+void smp_bogo(struct seq_file *);
+void smp_info(struct seq_file *);
+
+BTFIXUPDEF_CALL(void, smp_cross_call, smpfunc_t, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(int, __hard_smp_processor_id, void)
+BTFIXUPDEF_BLACKBOX(hard_smp_processor_id)
+BTFIXUPDEF_BLACKBOX(load_current)
+
+#define smp_cross_call(func,arg1,arg2,arg3,arg4,arg5) BTFIXUP_CALL(smp_cross_call)(func,arg1,arg2,arg3,arg4,arg5)
+
+static inline void xc0(smpfunc_t func) { smp_cross_call(func, 0, 0, 0, 0, 0); }
+static inline void xc1(smpfunc_t func, unsigned long arg1)
+{ smp_cross_call(func, arg1, 0, 0, 0, 0); }
+static inline void xc2(smpfunc_t func, unsigned long arg1, unsigned long arg2)
+{ smp_cross_call(func, arg1, arg2, 0, 0, 0); }
+static inline void xc3(smpfunc_t func, unsigned long arg1, unsigned long arg2,
+			   unsigned long arg3)
+{ smp_cross_call(func, arg1, arg2, arg3, 0, 0); }
+static inline void xc4(smpfunc_t func, unsigned long arg1, unsigned long arg2,
+			   unsigned long arg3, unsigned long arg4)
+{ smp_cross_call(func, arg1, arg2, arg3, arg4, 0); }
+static inline void xc5(smpfunc_t func, unsigned long arg1, unsigned long arg2,
+			   unsigned long arg3, unsigned long arg4, unsigned long arg5)
+{ smp_cross_call(func, arg1, arg2, arg3, arg4, arg5); }
+
+static inline int smp_call_function(void (*func)(void *info), void *info, int wait)
+{
+	xc1((smpfunc_t)func, (unsigned long)info);
+	return 0;
+}
+
+static inline int cpu_logical_map(int cpu)
+{
+	return cpu;
+}
+
+static inline int hard_smp4m_processor_id(void)
+{
+	int cpuid;
+
+	__asm__ __volatile__("rd %%tbr, %0\n\t"
+			     "srl %0, 12, %0\n\t"
+			     "and %0, 3, %0\n\t" :
+			     "=&r" (cpuid));
+	return cpuid;
+}
+
+static inline int hard_smp4d_processor_id(void)
+{
+	int cpuid;
+
+	__asm__ __volatile__("lda [%%g0] %1, %0\n\t" :
+			     "=&r" (cpuid) : "i" (ASI_M_VIKING_TMP1));
+	return cpuid;
+}
+
+#ifndef MODULE
+static inline int hard_smp_processor_id(void)
+{
+	int cpuid;
+
+	/* Black box - sun4m
+		__asm__ __volatile__("rd %%tbr, %0\n\t"
+				     "srl %0, 12, %0\n\t"
+				     "and %0, 3, %0\n\t" :
+				     "=&r" (cpuid));
+	             - sun4d
+		__asm__ __volatile__("lda [%g0] ASI_M_VIKING_TMP1, %0\n\t"
+				     "nop; nop" :
+				     "=&r" (cpuid));
+	   See btfixup.h and btfixupprep.c to understand how a blackbox works.
+	 */
+	__asm__ __volatile__("sethi %%hi(___b_hard_smp_processor_id), %0\n\t"
+			     "sethi %%hi(boot_cpu_id), %0\n\t"
+			     "ldub [%0 + %%lo(boot_cpu_id)], %0\n\t" :
+			     "=&r" (cpuid));
+	return cpuid;
+}
+#else
+static inline int hard_smp_processor_id(void)
+{
+	int cpuid;
+
+	__asm__ __volatile__("mov %%o7, %%g1\n\t"
+			     "call ___f___hard_smp_processor_id\n\t"
+			     " nop\n\t"
+			     "mov %%g2, %0\n\t" : "=r"(cpuid) : : "g1", "g2");
+	return cpuid;
+}
+#endif
+
+#define raw_smp_processor_id()		(current_thread_info()->cpu)
+
+#define prof_multiplier(__cpu)		cpu_data(__cpu).multiplier
+#define prof_counter(__cpu)		cpu_data(__cpu).counter
+
+void smp_setup_cpu_possible_map(void);
+
+#endif /* !(__ASSEMBLY__) */
+
+/* Sparc specific messages. */
+#define MSG_CROSS_CALL         0x0005       /* run func on cpus */
+
+/* Empirical PROM processor mailbox constants.  If the per-cpu mailbox
+ * contains something other than one of these then the ipi is from
+ * Linux's active_kernel_processor.  This facility exists so that
+ * the boot monitor can capture all the other cpus when one catches
+ * a watchdog reset or the user enters the monitor using L1-A keys.
+ */
+#define MBOX_STOPCPU          0xFB
+#define MBOX_IDLECPU          0xFC
+#define MBOX_IDLECPU2         0xFD
+#define MBOX_STOPCPU2         0xFE
+
+#else /* SMP */
+
+#define hard_smp_processor_id()		0
+#define smp_setup_cpu_possible_map() do { } while (0)
+
+#endif /* !(SMP) */
+
+#define NO_PROC_ID            0xFF
+
+#endif /* !(_SPARC_SMP_H) */
diff --git a/include/asm-sparc/smp_64.h b/include/asm-sparc/smp_64.h
new file mode 100644
index 00000000000..57224dd37b3
--- /dev/null
+++ b/include/asm-sparc/smp_64.h
@@ -0,0 +1,67 @@
+/* smp.h: Sparc64 specific SMP stuff.
+ *
+ * Copyright (C) 1996, 2008 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC64_SMP_H
+#define _SPARC64_SMP_H
+
+#include <linux/threads.h>
+#include <asm/asi.h>
+#include <asm/starfire.h>
+#include <asm/spitfire.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/cpumask.h>
+#include <linux/cache.h>
+
+#endif /* !(__ASSEMBLY__) */
+
+#ifdef CONFIG_SMP
+
+#ifndef __ASSEMBLY__
+
+/*
+ *	Private routines/data
+ */
+
+#include <linux/bitops.h>
+#include <asm/atomic.h>
+#include <asm/percpu.h>
+
+DECLARE_PER_CPU(cpumask_t, cpu_sibling_map);
+extern cpumask_t cpu_core_map[NR_CPUS];
+extern int sparc64_multi_core;
+
+extern void arch_send_call_function_single_ipi(int cpu);
+extern void arch_send_call_function_ipi(cpumask_t mask);
+
+/*
+ *	General functions that each host system must provide.
+ */
+
+extern int hard_smp_processor_id(void);
+#define raw_smp_processor_id() (current_thread_info()->cpu)
+
+extern void smp_fill_in_sib_core_maps(void);
+extern void cpu_play_dead(void);
+
+extern void smp_fetch_global_regs(void);
+
+#ifdef CONFIG_HOTPLUG_CPU
+extern int __cpu_disable(void);
+extern void __cpu_die(unsigned int cpu);
+#endif
+
+#endif /* !(__ASSEMBLY__) */
+
+#else
+
+#define hard_smp_processor_id()		0
+#define smp_fill_in_sib_core_maps() do { } while (0)
+#define smp_fetch_global_regs() do { } while (0)
+
+#endif /* !(CONFIG_SMP) */
+
+#endif /* !(_SPARC64_SMP_H) */
diff --git a/include/asm-sparc/sparsemem.h b/include/asm-sparc/sparsemem.h
new file mode 100644
index 00000000000..b99d4e4b6d2
--- /dev/null
+++ b/include/asm-sparc/sparsemem.h
@@ -0,0 +1,12 @@
+#ifndef _SPARC64_SPARSEMEM_H
+#define _SPARC64_SPARSEMEM_H
+
+#ifdef __KERNEL__
+
+#define SECTION_SIZE_BITS       30
+#define MAX_PHYSADDR_BITS       42
+#define MAX_PHYSMEM_BITS        42
+
+#endif /* !(__KERNEL__) */
+
+#endif /* !(_SPARC64_SPARSEMEM_H) */
diff --git a/include/asm-sparc/spinlock.h b/include/asm-sparc/spinlock.h
index de2249b267c..3b71c50b72e 100644
--- a/include/asm-sparc/spinlock.h
+++ b/include/asm-sparc/spinlock.h
@@ -1,192 +1,8 @@
-/* spinlock.h: 32-bit Sparc spinlock support.
- *
- * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __SPARC_SPINLOCK_H
-#define __SPARC_SPINLOCK_H
-
-#include <linux/threads.h>	/* For NR_CPUS */
-
-#ifndef __ASSEMBLY__
-
-#include <asm/psr.h>
-
-#define __raw_spin_is_locked(lock) (*((volatile unsigned char *)(lock)) != 0)
-
-#define __raw_spin_unlock_wait(lock) \
-	do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
-
-static inline void __raw_spin_lock(raw_spinlock_t *lock)
-{
-	__asm__ __volatile__(
-	"\n1:\n\t"
-	"ldstub	[%0], %%g2\n\t"
-	"orcc	%%g2, 0x0, %%g0\n\t"
-	"bne,a	2f\n\t"
-	" ldub	[%0], %%g2\n\t"
-	".subsection	2\n"
-	"2:\n\t"
-	"orcc	%%g2, 0x0, %%g0\n\t"
-	"bne,a	2b\n\t"
-	" ldub	[%0], %%g2\n\t"
-	"b,a	1b\n\t"
-	".previous\n"
-	: /* no outputs */
-	: "r" (lock)
-	: "g2", "memory", "cc");
-}
-
-static inline int __raw_spin_trylock(raw_spinlock_t *lock)
-{
-	unsigned int result;
-	__asm__ __volatile__("ldstub [%1], %0"
-			     : "=r" (result)
-			     : "r" (lock)
-			     : "memory");
-	return (result == 0);
-}
-
-static inline void __raw_spin_unlock(raw_spinlock_t *lock)
-{
-	__asm__ __volatile__("stb %%g0, [%0]" : : "r" (lock) : "memory");
-}
-
-/* Read-write spinlocks, allowing multiple readers
- * but only one writer.
- *
- * NOTE! it is quite common to have readers in interrupts
- * but no interrupt writers. For those circumstances we
- * can "mix" irq-safe locks - any writer needs to get a
- * irq-safe write-lock, but readers can get non-irqsafe
- * read-locks.
- *
- * XXX This might create some problems with my dual spinlock
- * XXX scheme, deadlocks etc. -DaveM
- *
- * Sort of like atomic_t's on Sparc, but even more clever.
- *
- *	------------------------------------
- *	| 24-bit counter           | wlock |  raw_rwlock_t
- *	------------------------------------
- *	 31                       8 7     0
- *
- * wlock signifies the one writer is in or somebody is updating
- * counter. For a writer, if he successfully acquires the wlock,
- * but counter is non-zero, he has to release the lock and wait,
- * till both counter and wlock are zero.
- *
- * Unfortunately this scheme limits us to ~16,000,000 cpus.
- */
-static inline void __read_lock(raw_rwlock_t *rw)
-{
-	register raw_rwlock_t *lp asm("g1");
-	lp = rw;
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___rw_read_enter\n\t"
-	" ldstub	[%%g1 + 3], %%g2\n"
-	: /* no outputs */
-	: "r" (lp)
-	: "g2", "g4", "memory", "cc");
-}
-
-#define __raw_read_lock(lock) \
-do {	unsigned long flags; \
-	local_irq_save(flags); \
-	__read_lock(lock); \
-	local_irq_restore(flags); \
-} while(0)
-
-static inline void __read_unlock(raw_rwlock_t *rw)
-{
-	register raw_rwlock_t *lp asm("g1");
-	lp = rw;
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___rw_read_exit\n\t"
-	" ldstub	[%%g1 + 3], %%g2\n"
-	: /* no outputs */
-	: "r" (lp)
-	: "g2", "g4", "memory", "cc");
-}
-
-#define __raw_read_unlock(lock) \
-do {	unsigned long flags; \
-	local_irq_save(flags); \
-	__read_unlock(lock); \
-	local_irq_restore(flags); \
-} while(0)
-
-static inline void __raw_write_lock(raw_rwlock_t *rw)
-{
-	register raw_rwlock_t *lp asm("g1");
-	lp = rw;
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___rw_write_enter\n\t"
-	" ldstub	[%%g1 + 3], %%g2\n"
-	: /* no outputs */
-	: "r" (lp)
-	: "g2", "g4", "memory", "cc");
-	*(volatile __u32 *)&lp->lock = ~0U;
-}
-
-static inline int __raw_write_trylock(raw_rwlock_t *rw)
-{
-	unsigned int val;
-
-	__asm__ __volatile__("ldstub [%1 + 3], %0"
-			     : "=r" (val)
-			     : "r" (&rw->lock)
-			     : "memory");
-
-	if (val == 0) {
-		val = rw->lock & ~0xff;
-		if (val)
-			((volatile u8*)&rw->lock)[3] = 0;
-		else
-			*(volatile u32*)&rw->lock = ~0U;
-	}
-
-	return (val == 0);
-}
-
-static inline int __read_trylock(raw_rwlock_t *rw)
-{
-	register raw_rwlock_t *lp asm("g1");
-	register int res asm("o0");
-	lp = rw;
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___rw_read_try\n\t"
-	" ldstub	[%%g1 + 3], %%g2\n"
-	: "=r" (res)
-	: "r" (lp)
-	: "g2", "g4", "memory", "cc");
-	return res;
-}
-
-#define __raw_read_trylock(lock) \
-({	unsigned long flags; \
-	int res; \
-	local_irq_save(flags); \
-	res = __read_trylock(lock); \
-	local_irq_restore(flags); \
-	res; \
-})
-
-#define __raw_write_unlock(rw)	do { (rw)->lock = 0; } while(0)
-
-#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
-
-#define _raw_spin_relax(lock)	cpu_relax()
-#define _raw_read_relax(lock)	cpu_relax()
-#define _raw_write_relax(lock)	cpu_relax()
-
-#define __raw_read_can_lock(rw) (!((rw)->lock & 0xff))
-#define __raw_write_can_lock(rw) (!(rw)->lock)
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* __SPARC_SPINLOCK_H */
+#ifndef ___ASM_SPARC_SPINLOCK_H
+#define ___ASM_SPARC_SPINLOCK_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/spinlock_64.h>
+#else
+#include <asm-sparc/spinlock_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/spinlock_32.h b/include/asm-sparc/spinlock_32.h
new file mode 100644
index 00000000000..de2249b267c
--- /dev/null
+++ b/include/asm-sparc/spinlock_32.h
@@ -0,0 +1,192 @@
+/* spinlock.h: 32-bit Sparc spinlock support.
+ *
+ * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __SPARC_SPINLOCK_H
+#define __SPARC_SPINLOCK_H
+
+#include <linux/threads.h>	/* For NR_CPUS */
+
+#ifndef __ASSEMBLY__
+
+#include <asm/psr.h>
+
+#define __raw_spin_is_locked(lock) (*((volatile unsigned char *)(lock)) != 0)
+
+#define __raw_spin_unlock_wait(lock) \
+	do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
+
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
+{
+	__asm__ __volatile__(
+	"\n1:\n\t"
+	"ldstub	[%0], %%g2\n\t"
+	"orcc	%%g2, 0x0, %%g0\n\t"
+	"bne,a	2f\n\t"
+	" ldub	[%0], %%g2\n\t"
+	".subsection	2\n"
+	"2:\n\t"
+	"orcc	%%g2, 0x0, %%g0\n\t"
+	"bne,a	2b\n\t"
+	" ldub	[%0], %%g2\n\t"
+	"b,a	1b\n\t"
+	".previous\n"
+	: /* no outputs */
+	: "r" (lock)
+	: "g2", "memory", "cc");
+}
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+{
+	unsigned int result;
+	__asm__ __volatile__("ldstub [%1], %0"
+			     : "=r" (result)
+			     : "r" (lock)
+			     : "memory");
+	return (result == 0);
+}
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+	__asm__ __volatile__("stb %%g0, [%0]" : : "r" (lock) : "memory");
+}
+
+/* Read-write spinlocks, allowing multiple readers
+ * but only one writer.
+ *
+ * NOTE! it is quite common to have readers in interrupts
+ * but no interrupt writers. For those circumstances we
+ * can "mix" irq-safe locks - any writer needs to get a
+ * irq-safe write-lock, but readers can get non-irqsafe
+ * read-locks.
+ *
+ * XXX This might create some problems with my dual spinlock
+ * XXX scheme, deadlocks etc. -DaveM
+ *
+ * Sort of like atomic_t's on Sparc, but even more clever.
+ *
+ *	------------------------------------
+ *	| 24-bit counter           | wlock |  raw_rwlock_t
+ *	------------------------------------
+ *	 31                       8 7     0
+ *
+ * wlock signifies the one writer is in or somebody is updating
+ * counter. For a writer, if he successfully acquires the wlock,
+ * but counter is non-zero, he has to release the lock and wait,
+ * till both counter and wlock are zero.
+ *
+ * Unfortunately this scheme limits us to ~16,000,000 cpus.
+ */
+static inline void __read_lock(raw_rwlock_t *rw)
+{
+	register raw_rwlock_t *lp asm("g1");
+	lp = rw;
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___rw_read_enter\n\t"
+	" ldstub	[%%g1 + 3], %%g2\n"
+	: /* no outputs */
+	: "r" (lp)
+	: "g2", "g4", "memory", "cc");
+}
+
+#define __raw_read_lock(lock) \
+do {	unsigned long flags; \
+	local_irq_save(flags); \
+	__read_lock(lock); \
+	local_irq_restore(flags); \
+} while(0)
+
+static inline void __read_unlock(raw_rwlock_t *rw)
+{
+	register raw_rwlock_t *lp asm("g1");
+	lp = rw;
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___rw_read_exit\n\t"
+	" ldstub	[%%g1 + 3], %%g2\n"
+	: /* no outputs */
+	: "r" (lp)
+	: "g2", "g4", "memory", "cc");
+}
+
+#define __raw_read_unlock(lock) \
+do {	unsigned long flags; \
+	local_irq_save(flags); \
+	__read_unlock(lock); \
+	local_irq_restore(flags); \
+} while(0)
+
+static inline void __raw_write_lock(raw_rwlock_t *rw)
+{
+	register raw_rwlock_t *lp asm("g1");
+	lp = rw;
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___rw_write_enter\n\t"
+	" ldstub	[%%g1 + 3], %%g2\n"
+	: /* no outputs */
+	: "r" (lp)
+	: "g2", "g4", "memory", "cc");
+	*(volatile __u32 *)&lp->lock = ~0U;
+}
+
+static inline int __raw_write_trylock(raw_rwlock_t *rw)
+{
+	unsigned int val;
+
+	__asm__ __volatile__("ldstub [%1 + 3], %0"
+			     : "=r" (val)
+			     : "r" (&rw->lock)
+			     : "memory");
+
+	if (val == 0) {
+		val = rw->lock & ~0xff;
+		if (val)
+			((volatile u8*)&rw->lock)[3] = 0;
+		else
+			*(volatile u32*)&rw->lock = ~0U;
+	}
+
+	return (val == 0);
+}
+
+static inline int __read_trylock(raw_rwlock_t *rw)
+{
+	register raw_rwlock_t *lp asm("g1");
+	register int res asm("o0");
+	lp = rw;
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___rw_read_try\n\t"
+	" ldstub	[%%g1 + 3], %%g2\n"
+	: "=r" (res)
+	: "r" (lp)
+	: "g2", "g4", "memory", "cc");
+	return res;
+}
+
+#define __raw_read_trylock(lock) \
+({	unsigned long flags; \
+	int res; \
+	local_irq_save(flags); \
+	res = __read_trylock(lock); \
+	local_irq_restore(flags); \
+	res; \
+})
+
+#define __raw_write_unlock(rw)	do { (rw)->lock = 0; } while(0)
+
+#define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
+
+#define _raw_spin_relax(lock)	cpu_relax()
+#define _raw_read_relax(lock)	cpu_relax()
+#define _raw_write_relax(lock)	cpu_relax()
+
+#define __raw_read_can_lock(rw) (!((rw)->lock & 0xff))
+#define __raw_write_can_lock(rw) (!(rw)->lock)
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* __SPARC_SPINLOCK_H */
diff --git a/include/asm-sparc/spinlock_64.h b/include/asm-sparc/spinlock_64.h
new file mode 100644
index 00000000000..0006fe9f8c7
--- /dev/null
+++ b/include/asm-sparc/spinlock_64.h
@@ -0,0 +1,250 @@
+/* spinlock.h: 64-bit Sparc spinlock support.
+ *
+ * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __SPARC64_SPINLOCK_H
+#define __SPARC64_SPINLOCK_H
+
+#include <linux/threads.h>	/* For NR_CPUS */
+
+#ifndef __ASSEMBLY__
+
+/* To get debugging spinlocks which detect and catch
+ * deadlock situations, set CONFIG_DEBUG_SPINLOCK
+ * and rebuild your kernel.
+ */
+
+/* All of these locking primitives are expected to work properly
+ * even in an RMO memory model, which currently is what the kernel
+ * runs in.
+ *
+ * There is another issue.  Because we play games to save cycles
+ * in the non-contention case, we need to be extra careful about
+ * branch targets into the "spinning" code.  They live in their
+ * own section, but the newer V9 branches have a shorter range
+ * than the traditional 32-bit sparc branch variants.  The rule
+ * is that the branches that go into and out of the spinner sections
+ * must be pre-V9 branches.
+ */
+
+#define __raw_spin_is_locked(lp)	((lp)->lock != 0)
+
+#define __raw_spin_unlock_wait(lp)	\
+	do {	rmb();			\
+	} while((lp)->lock)
+
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
+{
+	unsigned long tmp;
+
+	__asm__ __volatile__(
+"1:	ldstub		[%1], %0\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	brnz,pn		%0, 2f\n"
+"	 nop\n"
+"	.subsection	2\n"
+"2:	ldub		[%1], %0\n"
+"	membar		#LoadLoad\n"
+"	brnz,pt		%0, 2b\n"
+"	 nop\n"
+"	ba,a,pt		%%xcc, 1b\n"
+"	.previous"
+	: "=&r" (tmp)
+	: "r" (lock)
+	: "memory");
+}
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+{
+	unsigned long result;
+
+	__asm__ __volatile__(
+"	ldstub		[%1], %0\n"
+"	membar		#StoreLoad | #StoreStore"
+	: "=r" (result)
+	: "r" (lock)
+	: "memory");
+
+	return (result == 0UL);
+}
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+	__asm__ __volatile__(
+"	membar		#StoreStore | #LoadStore\n"
+"	stb		%%g0, [%0]"
+	: /* No outputs */
+	: "r" (lock)
+	: "memory");
+}
+
+static inline void __raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__(
+"1:	ldstub		[%2], %0\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	brnz,pn		%0, 2f\n"
+"	 nop\n"
+"	.subsection	2\n"
+"2:	rdpr		%%pil, %1\n"
+"	wrpr		%3, %%pil\n"
+"3:	ldub		[%2], %0\n"
+"	membar		#LoadLoad\n"
+"	brnz,pt		%0, 3b\n"
+"	 nop\n"
+"	ba,pt		%%xcc, 1b\n"
+"	 wrpr		%1, %%pil\n"
+"	.previous"
+	: "=&r" (tmp1), "=&r" (tmp2)
+	: "r"(lock), "r"(flags)
+	: "memory");
+}
+
+/* Multi-reader locks, these are much saner than the 32-bit Sparc ones... */
+
+static void inline __read_lock(raw_rwlock_t *lock)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__ (
+"1:	ldsw		[%2], %0\n"
+"	brlz,pn		%0, 2f\n"
+"4:	 add		%0, 1, %1\n"
+"	cas		[%2], %0, %1\n"
+"	cmp		%0, %1\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	bne,pn		%%icc, 1b\n"
+"	 nop\n"
+"	.subsection	2\n"
+"2:	ldsw		[%2], %0\n"
+"	membar		#LoadLoad\n"
+"	brlz,pt		%0, 2b\n"
+"	 nop\n"
+"	ba,a,pt		%%xcc, 4b\n"
+"	.previous"
+	: "=&r" (tmp1), "=&r" (tmp2)
+	: "r" (lock)
+	: "memory");
+}
+
+static int inline __read_trylock(raw_rwlock_t *lock)
+{
+	int tmp1, tmp2;
+
+	__asm__ __volatile__ (
+"1:	ldsw		[%2], %0\n"
+"	brlz,a,pn	%0, 2f\n"
+"	 mov		0, %0\n"
+"	add		%0, 1, %1\n"
+"	cas		[%2], %0, %1\n"
+"	cmp		%0, %1\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	bne,pn		%%icc, 1b\n"
+"	 mov		1, %0\n"
+"2:"
+	: "=&r" (tmp1), "=&r" (tmp2)
+	: "r" (lock)
+	: "memory");
+
+	return tmp1;
+}
+
+static void inline __read_unlock(raw_rwlock_t *lock)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__(
+"	membar	#StoreLoad | #LoadLoad\n"
+"1:	lduw	[%2], %0\n"
+"	sub	%0, 1, %1\n"
+"	cas	[%2], %0, %1\n"
+"	cmp	%0, %1\n"
+"	bne,pn	%%xcc, 1b\n"
+"	 nop"
+	: "=&r" (tmp1), "=&r" (tmp2)
+	: "r" (lock)
+	: "memory");
+}
+
+static void inline __write_lock(raw_rwlock_t *lock)
+{
+	unsigned long mask, tmp1, tmp2;
+
+	mask = 0x80000000UL;
+
+	__asm__ __volatile__(
+"1:	lduw		[%2], %0\n"
+"	brnz,pn		%0, 2f\n"
+"4:	 or		%0, %3, %1\n"
+"	cas		[%2], %0, %1\n"
+"	cmp		%0, %1\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	bne,pn		%%icc, 1b\n"
+"	 nop\n"
+"	.subsection	2\n"
+"2:	lduw		[%2], %0\n"
+"	membar		#LoadLoad\n"
+"	brnz,pt		%0, 2b\n"
+"	 nop\n"
+"	ba,a,pt		%%xcc, 4b\n"
+"	.previous"
+	: "=&r" (tmp1), "=&r" (tmp2)
+	: "r" (lock), "r" (mask)
+	: "memory");
+}
+
+static void inline __write_unlock(raw_rwlock_t *lock)
+{
+	__asm__ __volatile__(
+"	membar		#LoadStore | #StoreStore\n"
+"	stw		%%g0, [%0]"
+	: /* no outputs */
+	: "r" (lock)
+	: "memory");
+}
+
+static int inline __write_trylock(raw_rwlock_t *lock)
+{
+	unsigned long mask, tmp1, tmp2, result;
+
+	mask = 0x80000000UL;
+
+	__asm__ __volatile__(
+"	mov		0, %2\n"
+"1:	lduw		[%3], %0\n"
+"	brnz,pn		%0, 2f\n"
+"	 or		%0, %4, %1\n"
+"	cas		[%3], %0, %1\n"
+"	cmp		%0, %1\n"
+"	membar		#StoreLoad | #StoreStore\n"
+"	bne,pn		%%icc, 1b\n"
+"	 nop\n"
+"	mov		1, %2\n"
+"2:"
+	: "=&r" (tmp1), "=&r" (tmp2), "=&r" (result)
+	: "r" (lock), "r" (mask)
+	: "memory");
+
+	return result;
+}
+
+#define __raw_read_lock(p)	__read_lock(p)
+#define __raw_read_trylock(p)	__read_trylock(p)
+#define __raw_read_unlock(p)	__read_unlock(p)
+#define __raw_write_lock(p)	__write_lock(p)
+#define __raw_write_unlock(p)	__write_unlock(p)
+#define __raw_write_trylock(p)	__write_trylock(p)
+
+#define __raw_read_can_lock(rw)		(!((rw)->lock & 0x80000000UL))
+#define __raw_write_can_lock(rw)	(!(rw)->lock)
+
+#define _raw_spin_relax(lock)	cpu_relax()
+#define _raw_read_relax(lock)	cpu_relax()
+#define _raw_write_relax(lock)	cpu_relax()
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(__SPARC64_SPINLOCK_H) */
diff --git a/include/asm-sparc/spinlock_types.h b/include/asm-sparc/spinlock_types.h
index 0a0fb116c4e..37cbe01c585 100644
--- a/include/asm-sparc/spinlock_types.h
+++ b/include/asm-sparc/spinlock_types.h
@@ -6,7 +6,7 @@
 #endif
 
 typedef struct {
-	unsigned char lock;
+	volatile unsigned char lock;
 } raw_spinlock_t;
 
 #define __RAW_SPIN_LOCK_UNLOCKED	{ 0 }
diff --git a/include/asm-sparc/spitfire.h b/include/asm-sparc/spitfire.h
new file mode 100644
index 00000000000..985ea7e3199
--- /dev/null
+++ b/include/asm-sparc/spitfire.h
@@ -0,0 +1,342 @@
+/* spitfire.h: SpitFire/BlackBird/Cheetah inline MMU operations.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC64_SPITFIRE_H
+#define _SPARC64_SPITFIRE_H
+
+#include <asm/asi.h>
+
+/* The following register addresses are accessible via ASI_DMMU
+ * and ASI_IMMU, that is there is a distinct and unique copy of
+ * each these registers for each TLB.
+ */
+#define TSB_TAG_TARGET		0x0000000000000000 /* All chips				*/
+#define TLB_SFSR		0x0000000000000018 /* All chips				*/
+#define TSB_REG			0x0000000000000028 /* All chips				*/
+#define TLB_TAG_ACCESS		0x0000000000000030 /* All chips				*/
+#define VIRT_WATCHPOINT		0x0000000000000038 /* All chips				*/
+#define PHYS_WATCHPOINT		0x0000000000000040 /* All chips				*/
+#define TSB_EXTENSION_P		0x0000000000000048 /* Ultra-III and later		*/
+#define TSB_EXTENSION_S		0x0000000000000050 /* Ultra-III and later, D-TLB only	*/
+#define TSB_EXTENSION_N		0x0000000000000058 /* Ultra-III and later		*/
+#define TLB_TAG_ACCESS_EXT	0x0000000000000060 /* Ultra-III+ and later		*/
+
+/* These registers only exist as one entity, and are accessed
+ * via ASI_DMMU only.
+ */
+#define PRIMARY_CONTEXT		0x0000000000000008
+#define SECONDARY_CONTEXT	0x0000000000000010
+#define DMMU_SFAR		0x0000000000000020
+#define VIRT_WATCHPOINT		0x0000000000000038
+#define PHYS_WATCHPOINT		0x0000000000000040
+
+#define SPITFIRE_HIGHEST_LOCKED_TLBENT	(64 - 1)
+#define CHEETAH_HIGHEST_LOCKED_TLBENT	(16 - 1)
+
+#define L1DCACHE_SIZE		0x4000
+
+#define SUN4V_CHIP_INVALID	0x00
+#define SUN4V_CHIP_NIAGARA1	0x01
+#define SUN4V_CHIP_NIAGARA2	0x02
+#define SUN4V_CHIP_UNKNOWN	0xff
+
+#ifndef __ASSEMBLY__
+
+enum ultra_tlb_layout {
+	spitfire = 0,
+	cheetah = 1,
+	cheetah_plus = 2,
+	hypervisor = 3,
+};
+
+extern enum ultra_tlb_layout tlb_type;
+
+extern int sun4v_chip_type;
+
+extern int cheetah_pcache_forced_on;
+extern void cheetah_enable_pcache(void);
+
+#define sparc64_highest_locked_tlbent()	\
+	(tlb_type == spitfire ? \
+	 SPITFIRE_HIGHEST_LOCKED_TLBENT : \
+	 CHEETAH_HIGHEST_LOCKED_TLBENT)
+
+extern int num_kernel_image_mappings;
+
+/* The data cache is write through, so this just invalidates the
+ * specified line.
+ */
+static inline void spitfire_put_dcache_tag(unsigned long addr, unsigned long tag)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (tag), "r" (addr), "i" (ASI_DCACHE_TAG));
+}
+
+/* The instruction cache lines are flushed with this, but note that
+ * this does not flush the pipeline.  It is possible for a line to
+ * get flushed but stale instructions to still be in the pipeline,
+ * a flush instruction (to any address) is sufficient to handle
+ * this issue after the line is invalidated.
+ */
+static inline void spitfire_put_icache_tag(unsigned long addr, unsigned long tag)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (tag), "r" (addr), "i" (ASI_IC_TAG));
+}
+
+static inline unsigned long spitfire_get_dtlb_data(int entry)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" (entry << 3), "i" (ASI_DTLB_DATA_ACCESS));
+
+	/* Clear TTE diag bits. */
+	data &= ~0x0003fe0000000000UL;
+
+	return data;
+}
+
+static inline unsigned long spitfire_get_dtlb_tag(int entry)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" (entry << 3), "i" (ASI_DTLB_TAG_READ));
+	return tag;
+}
+
+static inline void spitfire_put_dtlb_data(int entry, unsigned long data)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data), "r" (entry << 3),
+			       "i" (ASI_DTLB_DATA_ACCESS));
+}
+
+static inline unsigned long spitfire_get_itlb_data(int entry)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" (entry << 3), "i" (ASI_ITLB_DATA_ACCESS));
+
+	/* Clear TTE diag bits. */
+	data &= ~0x0003fe0000000000UL;
+
+	return data;
+}
+
+static inline unsigned long spitfire_get_itlb_tag(int entry)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" (entry << 3), "i" (ASI_ITLB_TAG_READ));
+	return tag;
+}
+
+static inline void spitfire_put_itlb_data(int entry, unsigned long data)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data), "r" (entry << 3),
+			       "i" (ASI_ITLB_DATA_ACCESS));
+}
+
+static inline void spitfire_flush_dtlb_nucleus_page(unsigned long page)
+{
+	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (page | 0x20), "i" (ASI_DMMU_DEMAP));
+}
+
+static inline void spitfire_flush_itlb_nucleus_page(unsigned long page)
+{
+	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (page | 0x20), "i" (ASI_IMMU_DEMAP));
+}
+
+/* Cheetah has "all non-locked" tlb flushes. */
+static inline void cheetah_flush_dtlb_all(void)
+{
+	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (0x80), "i" (ASI_DMMU_DEMAP));
+}
+
+static inline void cheetah_flush_itlb_all(void)
+{
+	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (0x80), "i" (ASI_IMMU_DEMAP));
+}
+
+/* Cheetah has a 4-tlb layout so direct access is a bit different.
+ * The first two TLBs are fully assosciative, hold 16 entries, and are
+ * used only for locked and >8K sized translations.  One exists for
+ * data accesses and one for instruction accesses.
+ *
+ * The third TLB is for data accesses to 8K non-locked translations, is
+ * 2 way assosciative, and holds 512 entries.  The fourth TLB is for
+ * instruction accesses to 8K non-locked translations, is 2 way
+ * assosciative, and holds 128 entries.
+ *
+ * Cheetah has some bug where bogus data can be returned from
+ * ASI_{D,I}TLB_DATA_ACCESS loads, doing the load twice fixes
+ * the problem for me. -DaveM
+ */
+static inline unsigned long cheetah_get_ldtlb_data(int entry)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
+			     "ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" ((0 << 16) | (entry << 3)),
+			     "i" (ASI_DTLB_DATA_ACCESS));
+
+	return data;
+}
+
+static inline unsigned long cheetah_get_litlb_data(int entry)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
+			     "ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" ((0 << 16) | (entry << 3)),
+			     "i" (ASI_ITLB_DATA_ACCESS));
+
+	return data;
+}
+
+static inline unsigned long cheetah_get_ldtlb_tag(int entry)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" ((0 << 16) | (entry << 3)),
+			     "i" (ASI_DTLB_TAG_READ));
+
+	return tag;
+}
+
+static inline unsigned long cheetah_get_litlb_tag(int entry)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" ((0 << 16) | (entry << 3)),
+			     "i" (ASI_ITLB_TAG_READ));
+
+	return tag;
+}
+
+static inline void cheetah_put_ldtlb_data(int entry, unsigned long data)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data),
+			       "r" ((0 << 16) | (entry << 3)),
+			       "i" (ASI_DTLB_DATA_ACCESS));
+}
+
+static inline void cheetah_put_litlb_data(int entry, unsigned long data)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data),
+			       "r" ((0 << 16) | (entry << 3)),
+			       "i" (ASI_ITLB_DATA_ACCESS));
+}
+
+static inline unsigned long cheetah_get_dtlb_data(int entry, int tlb)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
+			     "ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" ((tlb << 16) | (entry << 3)), "i" (ASI_DTLB_DATA_ACCESS));
+
+	return data;
+}
+
+static inline unsigned long cheetah_get_dtlb_tag(int entry, int tlb)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" ((tlb << 16) | (entry << 3)), "i" (ASI_DTLB_TAG_READ));
+	return tag;
+}
+
+static inline void cheetah_put_dtlb_data(int entry, unsigned long data, int tlb)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data),
+			       "r" ((tlb << 16) | (entry << 3)),
+			       "i" (ASI_DTLB_DATA_ACCESS));
+}
+
+static inline unsigned long cheetah_get_itlb_data(int entry)
+{
+	unsigned long data;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
+			     "ldxa	[%1] %2, %0"
+			     : "=r" (data)
+			     : "r" ((2 << 16) | (entry << 3)),
+                               "i" (ASI_ITLB_DATA_ACCESS));
+
+	return data;
+}
+
+static inline unsigned long cheetah_get_itlb_tag(int entry)
+{
+	unsigned long tag;
+
+	__asm__ __volatile__("ldxa	[%1] %2, %0"
+			     : "=r" (tag)
+			     : "r" ((2 << 16) | (entry << 3)), "i" (ASI_ITLB_TAG_READ));
+	return tag;
+}
+
+static inline void cheetah_put_itlb_data(int entry, unsigned long data)
+{
+	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
+			     "membar	#Sync"
+			     : /* No outputs */
+			     : "r" (data), "r" ((2 << 16) | (entry << 3)),
+			       "i" (ASI_ITLB_DATA_ACCESS));
+}
+
+#endif /* !(__ASSEMBLY__) */
+
+#endif /* !(_SPARC64_SPITFIRE_H) */
diff --git a/include/asm-sparc/sstate.h b/include/asm-sparc/sstate.h
new file mode 100644
index 00000000000..a7c35dbcb28
--- /dev/null
+++ b/include/asm-sparc/sstate.h
@@ -0,0 +1,13 @@
+#ifndef _SPARC64_SSTATE_H
+#define _SPARC64_SSTATE_H
+
+extern void sstate_booting(void);
+extern void sstate_running(void);
+extern void sstate_halt(void);
+extern void sstate_poweroff(void);
+extern void sstate_panic(void);
+extern void sstate_reboot(void);
+
+extern void sun4v_sstate_init(void);
+
+#endif /* _SPARC64_SSTATE_H */
diff --git a/include/asm-sparc/stacktrace.h b/include/asm-sparc/stacktrace.h
new file mode 100644
index 00000000000..6cee39adf6d
--- /dev/null
+++ b/include/asm-sparc/stacktrace.h
@@ -0,0 +1,6 @@
+#ifndef _SPARC64_STACKTRACE_H
+#define _SPARC64_STACKTRACE_H
+
+extern void stack_trace_flush(void);
+
+#endif /* _SPARC64_STACKTRACE_H */
diff --git a/include/asm-sparc/starfire.h b/include/asm-sparc/starfire.h
new file mode 100644
index 00000000000..07bafd31e33
--- /dev/null
+++ b/include/asm-sparc/starfire.h
@@ -0,0 +1,21 @@
+/*
+ * starfire.h: Group all starfire specific code together.
+ *
+ * Copyright (C) 2000 Anton Blanchard (anton@samba.org)
+ */
+
+#ifndef _SPARC64_STARFIRE_H
+#define _SPARC64_STARFIRE_H
+
+#ifndef __ASSEMBLY__
+
+extern int this_is_starfire;
+
+extern void check_if_starfire(void);
+extern void starfire_cpu_setup(void);
+extern int starfire_hard_smp_processor_id(void);
+extern void starfire_hookup(int);
+extern unsigned int starfire_translate(unsigned long imap, unsigned int upaid);
+
+#endif
+#endif
diff --git a/include/asm-sparc/stat.h b/include/asm-sparc/stat.h
index 2299e1d5d94..9fdcaf8c9cd 100644
--- a/include/asm-sparc/stat.h
+++ b/include/asm-sparc/stat.h
@@ -1,76 +1,8 @@
-#ifndef _SPARC_STAT_H
-#define _SPARC_STAT_H
-
-#include <linux/types.h>
-
-struct __old_kernel_stat {
-	unsigned short st_dev;
-	unsigned short st_ino;
-	unsigned short st_mode;
-	unsigned short st_nlink;
-	unsigned short st_uid;
-	unsigned short st_gid;
-	unsigned short st_rdev;
-	unsigned long  st_size;
-	unsigned long  st_atime;
-	unsigned long  st_mtime;
-	unsigned long  st_ctime;
-};
-
-struct stat {
-	unsigned short	st_dev;
-	unsigned long	st_ino;
-	unsigned short	st_mode;
-	short		st_nlink;
-	unsigned short	st_uid;
-	unsigned short	st_gid;
-	unsigned short	st_rdev;
-	long		st_size;
-	long		st_atime;
-	unsigned long	st_atime_nsec;
-	long		st_mtime;
-	unsigned long	st_mtime_nsec;
-	long		st_ctime;
-	unsigned long	st_ctime_nsec;
-	long		st_blksize;
-	long		st_blocks;
-	unsigned long	__unused4[2];
-};
-
-#define STAT_HAVE_NSEC 1
-
-struct stat64 {
-	unsigned long long st_dev;
-
-	unsigned long long st_ino;
-
-	unsigned int	st_mode;
-	unsigned int	st_nlink;
-
-	unsigned int	st_uid;
-	unsigned int	st_gid;
-
-	unsigned long long st_rdev;
-
-	unsigned char	__pad3[8];
-
-	long long	st_size;
-	unsigned int	st_blksize;
-
-	unsigned char	__pad4[8];
-	unsigned int	st_blocks;
-
-	unsigned int	st_atime;
-	unsigned int	st_atime_nsec;
-
-	unsigned int	st_mtime;
-	unsigned int	st_mtime_nsec;
-
-	unsigned int	st_ctime;
-	unsigned int	st_ctime_nsec;
-
-	unsigned int	__unused4;
-	unsigned int	__unused5;
-};
-
+#ifndef ___ASM_SPARC_STAT_H
+#define ___ASM_SPARC_STAT_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/stat_64.h>
+#else
+#include <asm-sparc/stat_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/stat_32.h b/include/asm-sparc/stat_32.h
new file mode 100644
index 00000000000..2299e1d5d94
--- /dev/null
+++ b/include/asm-sparc/stat_32.h
@@ -0,0 +1,76 @@
+#ifndef _SPARC_STAT_H
+#define _SPARC_STAT_H
+
+#include <linux/types.h>
+
+struct __old_kernel_stat {
+	unsigned short st_dev;
+	unsigned short st_ino;
+	unsigned short st_mode;
+	unsigned short st_nlink;
+	unsigned short st_uid;
+	unsigned short st_gid;
+	unsigned short st_rdev;
+	unsigned long  st_size;
+	unsigned long  st_atime;
+	unsigned long  st_mtime;
+	unsigned long  st_ctime;
+};
+
+struct stat {
+	unsigned short	st_dev;
+	unsigned long	st_ino;
+	unsigned short	st_mode;
+	short		st_nlink;
+	unsigned short	st_uid;
+	unsigned short	st_gid;
+	unsigned short	st_rdev;
+	long		st_size;
+	long		st_atime;
+	unsigned long	st_atime_nsec;
+	long		st_mtime;
+	unsigned long	st_mtime_nsec;
+	long		st_ctime;
+	unsigned long	st_ctime_nsec;
+	long		st_blksize;
+	long		st_blocks;
+	unsigned long	__unused4[2];
+};
+
+#define STAT_HAVE_NSEC 1
+
+struct stat64 {
+	unsigned long long st_dev;
+
+	unsigned long long st_ino;
+
+	unsigned int	st_mode;
+	unsigned int	st_nlink;
+
+	unsigned int	st_uid;
+	unsigned int	st_gid;
+
+	unsigned long long st_rdev;
+
+	unsigned char	__pad3[8];
+
+	long long	st_size;
+	unsigned int	st_blksize;
+
+	unsigned char	__pad4[8];
+	unsigned int	st_blocks;
+
+	unsigned int	st_atime;
+	unsigned int	st_atime_nsec;
+
+	unsigned int	st_mtime;
+	unsigned int	st_mtime_nsec;
+
+	unsigned int	st_ctime;
+	unsigned int	st_ctime_nsec;
+
+	unsigned int	__unused4;
+	unsigned int	__unused5;
+};
+
+#endif
diff --git a/include/asm-sparc/stat_64.h b/include/asm-sparc/stat_64.h
new file mode 100644
index 00000000000..9650fdea847
--- /dev/null
+++ b/include/asm-sparc/stat_64.h
@@ -0,0 +1,47 @@
+#ifndef _SPARC64_STAT_H
+#define _SPARC64_STAT_H
+
+#include <linux/types.h>
+
+struct stat {
+	unsigned   st_dev;
+	ino_t   st_ino;
+	mode_t  st_mode;
+	short   st_nlink;
+	uid_t   st_uid;
+	gid_t   st_gid;
+	unsigned   st_rdev;
+	off_t   st_size;
+	time_t  st_atime;
+	time_t  st_mtime;
+	time_t  st_ctime;
+	off_t   st_blksize;
+	off_t   st_blocks;
+	unsigned long  __unused4[2];
+};
+
+struct stat64 {
+	unsigned long	st_dev;
+	unsigned long	st_ino;
+	unsigned long	st_nlink;
+
+	unsigned int	st_mode;
+	unsigned int	st_uid;
+	unsigned int	st_gid;
+	unsigned int	__pad0;
+
+	unsigned long	st_rdev;
+	long		st_size;
+	long		st_blksize;
+	long		st_blocks;
+
+	unsigned long	st_atime;
+	unsigned long	st_atime_nsec;
+	unsigned long	st_mtime;
+	unsigned long	st_mtime_nsec;
+	unsigned long	st_ctime;
+	unsigned long	st_ctime_nsec;
+	long		__unused[3];
+};
+
+#endif
diff --git a/include/asm-sparc/statfs.h b/include/asm-sparc/statfs.h
index 304520fa886..a70cc52e701 100644
--- a/include/asm-sparc/statfs.h
+++ b/include/asm-sparc/statfs.h
@@ -1,6 +1,8 @@
-#ifndef _SPARC_STATFS_H
-#define _SPARC_STATFS_H
-
-#include <asm-generic/statfs.h>
-
+#ifndef ___ASM_SPARC_STATFS_H
+#define ___ASM_SPARC_STATFS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/statfs_64.h>
+#else
+#include <asm-sparc/statfs_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/statfs_32.h b/include/asm-sparc/statfs_32.h
new file mode 100644
index 00000000000..304520fa886
--- /dev/null
+++ b/include/asm-sparc/statfs_32.h
@@ -0,0 +1,6 @@
+#ifndef _SPARC_STATFS_H
+#define _SPARC_STATFS_H
+
+#include <asm-generic/statfs.h>
+
+#endif
diff --git a/include/asm-sparc/statfs_64.h b/include/asm-sparc/statfs_64.h
new file mode 100644
index 00000000000..79b3c890a5f
--- /dev/null
+++ b/include/asm-sparc/statfs_64.h
@@ -0,0 +1,54 @@
+#ifndef _SPARC64_STATFS_H
+#define _SPARC64_STATFS_H
+
+#ifndef __KERNEL_STRICT_NAMES
+
+#include <linux/types.h>
+
+typedef __kernel_fsid_t	fsid_t;
+
+#endif
+
+struct statfs {
+	long f_type;
+	long f_bsize;
+	long f_blocks;
+	long f_bfree;
+	long f_bavail;
+	long f_files;
+	long f_ffree;
+	__kernel_fsid_t f_fsid;
+	long f_namelen;
+	long f_frsize;
+	long f_spare[5];
+};
+
+struct statfs64 {
+	long f_type;
+	long f_bsize;
+	long f_blocks;
+	long f_bfree;
+	long f_bavail;
+	long f_files;
+	long f_ffree;
+	__kernel_fsid_t f_fsid;
+	long f_namelen;
+	long f_frsize;
+	long f_spare[5];
+};
+
+struct compat_statfs64 {
+	__u32 f_type;
+	__u32 f_bsize;
+	__u64 f_blocks;
+	__u64 f_bfree;
+	__u64 f_bavail;
+	__u64 f_files;
+	__u64 f_ffree;
+	__kernel_fsid_t f_fsid;
+	__u32 f_namelen;
+	__u32 f_frsize;
+	__u32 f_spare[5];
+};
+
+#endif
diff --git a/include/asm-sparc/string.h b/include/asm-sparc/string.h
index 8d7c0dd4f29..14c04c7697a 100644
--- a/include/asm-sparc/string.h
+++ b/include/asm-sparc/string.h
@@ -1,205 +1,8 @@
-/*
- * string.h: External definitions for optimized assembly string
- *           routines for the Linux Kernel.
- *
- * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1996,1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-
-#ifndef __SPARC_STRING_H__
-#define __SPARC_STRING_H__
-
-#include <asm/page.h>
-
-/* Really, userland/ksyms should not see any of this stuff. */
-
-#ifdef __KERNEL__
-
-extern void __memmove(void *,const void *,__kernel_size_t);
-extern __kernel_size_t __memcpy(void *,const void *,__kernel_size_t);
-extern __kernel_size_t __memset(void *,int,__kernel_size_t);
-
-#ifndef EXPORT_SYMTAB_STROPS
-
-/* First the mem*() things. */
-#define __HAVE_ARCH_MEMMOVE
-#undef memmove
-#define memmove(_to, _from, _n) \
-({ \
-	void *_t = (_to); \
-	__memmove(_t, (_from), (_n)); \
-	_t; \
-})
-
-#define __HAVE_ARCH_MEMCPY
-
-static inline void *__constant_memcpy(void *to, const void *from, __kernel_size_t n)
-{
-	extern void __copy_1page(void *, const void *);
-
-	if(n <= 32) {
-		__builtin_memcpy(to, from, n);
-	} else if (((unsigned int) to & 7) != 0) {
-		/* Destination is not aligned on the double-word boundary */
-		__memcpy(to, from, n);
-	} else {
-		switch(n) {
-		case PAGE_SIZE:
-			__copy_1page(to, from);
-			break;
-		default:
-			__memcpy(to, from, n);
-			break;
-		}
-	}
-	return to;
-}
-
-static inline void *__nonconstant_memcpy(void *to, const void *from, __kernel_size_t n)
-{
-	__memcpy(to, from, n);
-	return to;
-}
-
-#undef memcpy
-#define memcpy(t, f, n) \
-(__builtin_constant_p(n) ? \
- __constant_memcpy((t),(f),(n)) : \
- __nonconstant_memcpy((t),(f),(n)))
-
-#define __HAVE_ARCH_MEMSET
-
-static inline void *__constant_c_and_count_memset(void *s, char c, __kernel_size_t count)
-{
-	extern void bzero_1page(void *);
-	extern __kernel_size_t __bzero(void *, __kernel_size_t);
-
-	if(!c) {
-		if(count == PAGE_SIZE)
-			bzero_1page(s);
-		else
-			__bzero(s, count);
-	} else {
-		__memset(s, c, count);
-	}
-	return s;
-}
-
-static inline void *__constant_c_memset(void *s, char c, __kernel_size_t count)
-{
-	extern __kernel_size_t __bzero(void *, __kernel_size_t);
-
-	if(!c)
-		__bzero(s, count);
-	else
-		__memset(s, c, count);
-	return s;
-}
-
-static inline void *__nonconstant_memset(void *s, char c, __kernel_size_t count)
-{
-	__memset(s, c, count);
-	return s;
-}
-
-#undef memset
-#define memset(s, c, count) \
-(__builtin_constant_p(c) ? (__builtin_constant_p(count) ? \
-                            __constant_c_and_count_memset((s), (c), (count)) : \
-                            __constant_c_memset((s), (c), (count))) \
-                          : __nonconstant_memset((s), (c), (count)))
-
-#define __HAVE_ARCH_MEMSCAN
-
-#undef memscan
-#define memscan(__arg0, __char, __arg2)						\
-({										\
-	extern void *__memscan_zero(void *, size_t);				\
-	extern void *__memscan_generic(void *, int, size_t);			\
-	void *__retval, *__addr = (__arg0);					\
-	size_t __size = (__arg2);						\
-										\
-	if(__builtin_constant_p(__char) && !(__char))				\
-		__retval = __memscan_zero(__addr, __size);			\
-	else									\
-		__retval = __memscan_generic(__addr, (__char), __size);		\
-										\
-	__retval;								\
-})
-
-#define __HAVE_ARCH_MEMCMP
-extern int memcmp(const void *,const void *,__kernel_size_t);
-
-/* Now the str*() stuff... */
-#define __HAVE_ARCH_STRLEN
-extern __kernel_size_t strlen(const char *);
-
-#define __HAVE_ARCH_STRNCMP
-
-extern int __strncmp(const char *, const char *, __kernel_size_t);
-
-static inline int __constant_strncmp(const char *src, const char *dest, __kernel_size_t count)
-{
-	register int retval;
-	switch(count) {
-	case 0: return 0;
-	case 1: return (src[0] - dest[0]);
-	case 2: retval = (src[0] - dest[0]);
-		if(!retval && src[0])
-		  retval = (src[1] - dest[1]);
-		return retval;
-	case 3: retval = (src[0] - dest[0]);
-		if(!retval && src[0]) {
-		  retval = (src[1] - dest[1]);
-		  if(!retval && src[1])
-		    retval = (src[2] - dest[2]);
-		}
-		return retval;
-	case 4: retval = (src[0] - dest[0]);
-		if(!retval && src[0]) {
-		  retval = (src[1] - dest[1]);
-		  if(!retval && src[1]) {
-		    retval = (src[2] - dest[2]);
-		    if (!retval && src[2])
-		      retval = (src[3] - dest[3]);
-		  }
-		}
-		return retval;
-	case 5: retval = (src[0] - dest[0]);
-		if(!retval && src[0]) {
-		  retval = (src[1] - dest[1]);
-		  if(!retval && src[1]) {
-		    retval = (src[2] - dest[2]);
-		    if (!retval && src[2]) {
-		      retval = (src[3] - dest[3]);
-		      if (!retval && src[3])
-		        retval = (src[4] - dest[4]);
-		    }
-		  }
-		}
-		return retval;
-	default:
-		retval = (src[0] - dest[0]);
-		if(!retval && src[0]) {
-		  retval = (src[1] - dest[1]);
-		  if(!retval && src[1]) {
-		    retval = (src[2] - dest[2]);
-		    if(!retval && src[2])
-		      retval = __strncmp(src+3,dest+3,count-3);
-		  }
-		}
-		return retval;
-	}
-}
-
-#undef strncmp
-#define strncmp(__arg0, __arg1, __arg2)	\
-(__builtin_constant_p(__arg2) ?	\
- __constant_strncmp(__arg0, __arg1, __arg2) : \
- __strncmp(__arg0, __arg1, __arg2))
- 
-#endif /* !EXPORT_SYMTAB_STROPS */
-
-#endif /* __KERNEL__ */
-
-#endif /* !(__SPARC_STRING_H__) */
+#ifndef ___ASM_SPARC_STRING_H
+#define ___ASM_SPARC_STRING_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/string_64.h>
+#else
+#include <asm-sparc/string_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/string_32.h b/include/asm-sparc/string_32.h
new file mode 100644
index 00000000000..6c5fddb7e6b
--- /dev/null
+++ b/include/asm-sparc/string_32.h
@@ -0,0 +1,205 @@
+/*
+ * string.h: External definitions for optimized assembly string
+ *           routines for the Linux Kernel.
+ *
+ * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996,1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+
+#ifndef __SPARC_STRING_H__
+#define __SPARC_STRING_H__
+
+#include <asm/page.h>
+
+/* Really, userland/ksyms should not see any of this stuff. */
+
+#ifdef __KERNEL__
+
+extern void __memmove(void *,const void *,__kernel_size_t);
+extern __kernel_size_t __memcpy(void *,const void *,__kernel_size_t);
+extern __kernel_size_t __memset(void *,int,__kernel_size_t);
+
+#ifndef EXPORT_SYMTAB_STROPS
+
+/* First the mem*() things. */
+#define __HAVE_ARCH_MEMMOVE
+#undef memmove
+#define memmove(_to, _from, _n) \
+({ \
+	void *_t = (_to); \
+	__memmove(_t, (_from), (_n)); \
+	_t; \
+})
+
+#define __HAVE_ARCH_MEMCPY
+
+static inline void *__constant_memcpy(void *to, const void *from, __kernel_size_t n)
+{
+	extern void __copy_1page(void *, const void *);
+
+	if(n <= 32) {
+		__builtin_memcpy(to, from, n);
+	} else if (((unsigned int) to & 7) != 0) {
+		/* Destination is not aligned on the double-word boundary */
+		__memcpy(to, from, n);
+	} else {
+		switch(n) {
+		case PAGE_SIZE:
+			__copy_1page(to, from);
+			break;
+		default:
+			__memcpy(to, from, n);
+			break;
+		}
+	}
+	return to;
+}
+
+static inline void *__nonconstant_memcpy(void *to, const void *from, __kernel_size_t n)
+{
+	__memcpy(to, from, n);
+	return to;
+}
+
+#undef memcpy
+#define memcpy(t, f, n) \
+(__builtin_constant_p(n) ? \
+ __constant_memcpy((t),(f),(n)) : \
+ __nonconstant_memcpy((t),(f),(n)))
+
+#define __HAVE_ARCH_MEMSET
+
+static inline void *__constant_c_and_count_memset(void *s, char c, __kernel_size_t count)
+{
+	extern void bzero_1page(void *);
+	extern __kernel_size_t __bzero(void *, __kernel_size_t);
+
+	if(!c) {
+		if(count == PAGE_SIZE)
+			bzero_1page(s);
+		else
+			__bzero(s, count);
+	} else {
+		__memset(s, c, count);
+	}
+	return s;
+}
+
+static inline void *__constant_c_memset(void *s, char c, __kernel_size_t count)
+{
+	extern __kernel_size_t __bzero(void *, __kernel_size_t);
+
+	if(!c)
+		__bzero(s, count);
+	else
+		__memset(s, c, count);
+	return s;
+}
+
+static inline void *__nonconstant_memset(void *s, char c, __kernel_size_t count)
+{
+	__memset(s, c, count);
+	return s;
+}
+
+#undef memset
+#define memset(s, c, count) \
+(__builtin_constant_p(c) ? (__builtin_constant_p(count) ? \
+                            __constant_c_and_count_memset((s), (c), (count)) : \
+                            __constant_c_memset((s), (c), (count))) \
+                          : __nonconstant_memset((s), (c), (count)))
+
+#define __HAVE_ARCH_MEMSCAN
+
+#undef memscan
+#define memscan(__arg0, __char, __arg2)						\
+({										\
+	extern void *__memscan_zero(void *, size_t);				\
+	extern void *__memscan_generic(void *, int, size_t);			\
+	void *__retval, *__addr = (__arg0);					\
+	size_t __size = (__arg2);						\
+										\
+	if(__builtin_constant_p(__char) && !(__char))				\
+		__retval = __memscan_zero(__addr, __size);			\
+	else									\
+		__retval = __memscan_generic(__addr, (__char), __size);		\
+										\
+	__retval;								\
+})
+
+#define __HAVE_ARCH_MEMCMP
+extern int memcmp(const void *,const void *,__kernel_size_t);
+
+/* Now the str*() stuff... */
+#define __HAVE_ARCH_STRLEN
+extern __kernel_size_t strlen(const char *);
+
+#define __HAVE_ARCH_STRNCMP
+
+extern int __strncmp(const char *, const char *, __kernel_size_t);
+
+static inline int __constant_strncmp(const char *src, const char *dest, __kernel_size_t count)
+{
+	register int retval;
+	switch(count) {
+	case 0: return 0;
+	case 1: return (src[0] - dest[0]);
+	case 2: retval = (src[0] - dest[0]);
+		if(!retval && src[0])
+		  retval = (src[1] - dest[1]);
+		return retval;
+	case 3: retval = (src[0] - dest[0]);
+		if(!retval && src[0]) {
+		  retval = (src[1] - dest[1]);
+		  if(!retval && src[1])
+		    retval = (src[2] - dest[2]);
+		}
+		return retval;
+	case 4: retval = (src[0] - dest[0]);
+		if(!retval && src[0]) {
+		  retval = (src[1] - dest[1]);
+		  if(!retval && src[1]) {
+		    retval = (src[2] - dest[2]);
+		    if (!retval && src[2])
+		      retval = (src[3] - dest[3]);
+		  }
+		}
+		return retval;
+	case 5: retval = (src[0] - dest[0]);
+		if(!retval && src[0]) {
+		  retval = (src[1] - dest[1]);
+		  if(!retval && src[1]) {
+		    retval = (src[2] - dest[2]);
+		    if (!retval && src[2]) {
+		      retval = (src[3] - dest[3]);
+		      if (!retval && src[3])
+		        retval = (src[4] - dest[4]);
+		    }
+		  }
+		}
+		return retval;
+	default:
+		retval = (src[0] - dest[0]);
+		if(!retval && src[0]) {
+		  retval = (src[1] - dest[1]);
+		  if(!retval && src[1]) {
+		    retval = (src[2] - dest[2]);
+		    if(!retval && src[2])
+		      retval = __strncmp(src+3,dest+3,count-3);
+		  }
+		}
+		return retval;
+	}
+}
+
+#undef strncmp
+#define strncmp(__arg0, __arg1, __arg2)	\
+(__builtin_constant_p(__arg2) ?	\
+ __constant_strncmp(__arg0, __arg1, __arg2) : \
+ __strncmp(__arg0, __arg1, __arg2))
+
+#endif /* !EXPORT_SYMTAB_STROPS */
+
+#endif /* __KERNEL__ */
+
+#endif /* !(__SPARC_STRING_H__) */
diff --git a/include/asm-sparc/string_64.h b/include/asm-sparc/string_64.h
new file mode 100644
index 00000000000..43161f2d17e
--- /dev/null
+++ b/include/asm-sparc/string_64.h
@@ -0,0 +1,83 @@
+/*
+ * string.h: External definitions for optimized assembly string
+ *           routines for the Linux Kernel.
+ *
+ * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996,1997,1999 Jakub Jelinek (jakub@redhat.com)
+ */
+
+#ifndef __SPARC64_STRING_H__
+#define __SPARC64_STRING_H__
+
+/* Really, userland/ksyms should not see any of this stuff. */
+
+#ifdef __KERNEL__
+
+#include <asm/asi.h>
+
+extern void *__memset(void *,int,__kernel_size_t);
+
+#ifndef EXPORT_SYMTAB_STROPS
+
+/* First the mem*() things. */
+#define __HAVE_ARCH_MEMMOVE
+extern void *memmove(void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMCPY
+extern void *memcpy(void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMSET
+extern void *__builtin_memset(void *,int,__kernel_size_t);
+
+static inline void *__constant_memset(void *s, int c, __kernel_size_t count)
+{
+	extern __kernel_size_t __bzero(void *, __kernel_size_t);
+
+	if (!c) {
+		__bzero(s, count);
+		return s;
+	} else
+		return __memset(s, c, count);
+}
+
+#undef memset
+#define memset(s, c, count) \
+((__builtin_constant_p(count) && (count) <= 32) ? \
+ __builtin_memset((s), (c), (count)) : \
+ (__builtin_constant_p(c) ? \
+  __constant_memset((s), (c), (count)) : \
+  __memset((s), (c), (count))))
+
+#define __HAVE_ARCH_MEMSCAN
+
+#undef memscan
+#define memscan(__arg0, __char, __arg2)					\
+({									\
+	extern void *__memscan_zero(void *, size_t);			\
+	extern void *__memscan_generic(void *, int, size_t);		\
+	void *__retval, *__addr = (__arg0);				\
+	size_t __size = (__arg2);					\
+									\
+	if(__builtin_constant_p(__char) && !(__char))			\
+		__retval = __memscan_zero(__addr, __size);		\
+	else								\
+		__retval = __memscan_generic(__addr, (__char), __size);	\
+									\
+	__retval;							\
+})
+
+#define __HAVE_ARCH_MEMCMP
+extern int memcmp(const void *,const void *,__kernel_size_t);
+
+/* Now the str*() stuff... */
+#define __HAVE_ARCH_STRLEN
+extern __kernel_size_t strlen(const char *);
+
+#define __HAVE_ARCH_STRNCMP
+extern int strncmp(const char *, const char *, __kernel_size_t);
+
+#endif /* !EXPORT_SYMTAB_STROPS */
+
+#endif /* __KERNEL__ */
+
+#endif /* !(__SPARC64_STRING_H__) */
diff --git a/include/asm-sparc/syscalls.h b/include/asm-sparc/syscalls.h
new file mode 100644
index 00000000000..45a43f637a1
--- /dev/null
+++ b/include/asm-sparc/syscalls.h
@@ -0,0 +1,13 @@
+#ifndef _SPARC64_SYSCALLS_H
+#define _SPARC64_SYSCALLS_H
+
+struct pt_regs;
+
+extern asmlinkage long sparc_do_fork(unsigned long clone_flags,
+				     unsigned long stack_start,
+				     struct pt_regs *regs,
+				     unsigned long stack_size);
+
+extern asmlinkage int sparc_execve(struct pt_regs *regs);
+
+#endif /* _SPARC64_SYSCALLS_H */
diff --git a/include/asm-sparc/system.h b/include/asm-sparc/system.h
index b4b024445fc..15e2a3bc4f6 100644
--- a/include/asm-sparc/system.h
+++ b/include/asm-sparc/system.h
@@ -1,288 +1,8 @@
-#ifndef __SPARC_SYSTEM_H
-#define __SPARC_SYSTEM_H
-
-#include <linux/kernel.h>
-#include <linux/threads.h>	/* NR_CPUS */
-#include <linux/thread_info.h>
-
-#include <asm/page.h>
-#include <asm/psr.h>
-#include <asm/ptrace.h>
-#include <asm/btfixup.h>
-#include <asm/smp.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/irqflags.h>
-
-/*
- * Sparc (general) CPU types
- */
-enum sparc_cpu {
-  sun4        = 0x00,
-  sun4c       = 0x01,
-  sun4m       = 0x02,
-  sun4d       = 0x03,
-  sun4e       = 0x04,
-  sun4u       = 0x05, /* V8 ploos ploos */
-  sun_unknown = 0x06,
-  ap1000      = 0x07, /* almost a sun4m */
-};
-
-/* Really, userland should not be looking at any of this... */
-#ifdef __KERNEL__
-
-extern enum sparc_cpu sparc_cpu_model;
-
-#ifndef CONFIG_SUN4
-#define ARCH_SUN4C_SUN4 (sparc_cpu_model==sun4c)
-#define ARCH_SUN4 0
+#ifndef ___ASM_SPARC_SYSTEM_H
+#define ___ASM_SPARC_SYSTEM_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/system_64.h>
 #else
-#define ARCH_SUN4C_SUN4 1
-#define ARCH_SUN4 1
+#include <asm-sparc/system_32.h>
 #endif
-
-#define SUN4M_NCPUS            4              /* Architectural limit of sun4m. */
-
-extern char reboot_command[];
-
-extern struct thread_info *current_set[NR_CPUS];
-
-extern unsigned long empty_bad_page;
-extern unsigned long empty_bad_page_table;
-extern unsigned long empty_zero_page;
-
-extern void sun_do_break(void);
-extern int serial_console;
-extern int stop_a_enabled;
-
-static inline int con_is_present(void)
-{
-	return serial_console ? 0 : 1;
-}
-
-/* When a context switch happens we must flush all user windows so that
- * the windows of the current process are flushed onto its stack. This
- * way the windows are all clean for the next process and the stack
- * frames are up to date.
- */
-extern void flush_user_windows(void);
-extern void kill_user_windows(void);
-extern void synchronize_user_stack(void);
-extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
-		   void *fpqueue, unsigned long *fpqdepth);
-
-#ifdef CONFIG_SMP
-#define SWITCH_ENTER(prv) \
-	do {			\
-	if (test_tsk_thread_flag(prv, TIF_USEDFPU)) { \
-		put_psr(get_psr() | PSR_EF); \
-		fpsave(&(prv)->thread.float_regs[0], &(prv)->thread.fsr, \
-		       &(prv)->thread.fpqueue[0], &(prv)->thread.fpqdepth); \
-		clear_tsk_thread_flag(prv, TIF_USEDFPU); \
-		(prv)->thread.kregs->psr &= ~PSR_EF; \
-	} \
-	} while(0)
-
-#define SWITCH_DO_LAZY_FPU(next)	/* */
-#else
-#define SWITCH_ENTER(prv)		/* */
-#define SWITCH_DO_LAZY_FPU(nxt)	\
-	do {			\
-	if (last_task_used_math != (nxt))		\
-		(nxt)->thread.kregs->psr&=~PSR_EF;	\
-	} while(0)
-#endif
-
-extern void flushw_all(void);
-
-/*
- * Flush windows so that the VM switch which follows
- * would not pull the stack from under us.
- *
- * SWITCH_ENTER and SWITH_DO_LAZY_FPU do not work yet (e.g. SMP does not work)
- * XXX WTF is the above comment? Found in late teen 2.4.x.
- */
-#define prepare_arch_switch(next) do { \
-	__asm__ __volatile__( \
-	".globl\tflush_patch_switch\nflush_patch_switch:\n\t" \
-	"save %sp, -0x40, %sp; save %sp, -0x40, %sp; save %sp, -0x40, %sp\n\t" \
-	"save %sp, -0x40, %sp; save %sp, -0x40, %sp; save %sp, -0x40, %sp\n\t" \
-	"save %sp, -0x40, %sp\n\t" \
-	"restore; restore; restore; restore; restore; restore; restore"); \
-} while(0)
-
-	/* Much care has gone into this code, do not touch it.
-	 *
-	 * We need to loadup regs l0/l1 for the newly forked child
-	 * case because the trap return path relies on those registers
-	 * holding certain values, gcc is told that they are clobbered.
-	 * Gcc needs registers for 3 values in and 1 value out, so we
-	 * clobber every non-fixed-usage register besides l2/l3/o4/o5.  -DaveM
-	 *
-	 * Hey Dave, that do not touch sign is too much of an incentive
-	 * - Anton & Pete
-	 */
-#define switch_to(prev, next, last) do {						\
-	SWITCH_ENTER(prev);								\
-	SWITCH_DO_LAZY_FPU(next);							\
-	cpu_set(smp_processor_id(), next->active_mm->cpu_vm_mask);			\
-	__asm__ __volatile__(								\
-	"sethi	%%hi(here - 0x8), %%o7\n\t"						\
-	"mov	%%g6, %%g3\n\t"								\
-	"or	%%o7, %%lo(here - 0x8), %%o7\n\t"					\
-	"rd	%%psr, %%g4\n\t"							\
-	"std	%%sp, [%%g6 + %4]\n\t"							\
-	"rd	%%wim, %%g5\n\t"							\
-	"wr	%%g4, 0x20, %%psr\n\t"							\
-	"nop\n\t"									\
-	"std	%%g4, [%%g6 + %3]\n\t"							\
-	"ldd	[%2 + %3], %%g4\n\t"							\
-	"mov	%2, %%g6\n\t"								\
-	".globl	patchme_store_new_current\n"						\
-"patchme_store_new_current:\n\t"							\
-	"st	%2, [%1]\n\t"								\
-	"wr	%%g4, 0x20, %%psr\n\t"							\
-	"nop\n\t"									\
-	"nop\n\t"									\
-	"nop\n\t"	/* LEON needs all 3 nops: load to %sp depends on CWP. */		\
-	"ldd	[%%g6 + %4], %%sp\n\t"							\
-	"wr	%%g5, 0x0, %%wim\n\t"							\
-	"ldd	[%%sp + 0x00], %%l0\n\t"						\
-	"ldd	[%%sp + 0x38], %%i6\n\t"						\
-	"wr	%%g4, 0x0, %%psr\n\t"							\
-	"nop\n\t"									\
-	"nop\n\t"									\
-	"jmpl	%%o7 + 0x8, %%g0\n\t"							\
-	" ld	[%%g3 + %5], %0\n\t"							\
-	"here:\n"									\
-        : "=&r" (last)									\
-        : "r" (&(current_set[hard_smp_processor_id()])),	\
-	  "r" (task_thread_info(next)),				\
-	  "i" (TI_KPSR),					\
-	  "i" (TI_KSP),						\
-	  "i" (TI_TASK)						\
-	:       "g1", "g2", "g3", "g4", "g5",       "g7",	\
-	  "l0", "l1",       "l3", "l4", "l5", "l6", "l7",	\
-	  "i0", "i1", "i2", "i3", "i4", "i5",			\
-	  "o0", "o1", "o2", "o3",                   "o7");	\
-	} while(0)
-
-/* XXX Change this if we ever use a PSO mode kernel. */
-#define mb()	__asm__ __volatile__ ("" : : : "memory")
-#define rmb()	mb()
-#define wmb()	mb()
-#define read_barrier_depends()	do { } while(0)
-#define set_mb(__var, __value)  do { __var = __value; mb(); } while(0)
-#define smp_mb()	__asm__ __volatile__("":::"memory")
-#define smp_rmb()	__asm__ __volatile__("":::"memory")
-#define smp_wmb()	__asm__ __volatile__("":::"memory")
-#define smp_read_barrier_depends()	do { } while(0)
-
-#define nop() __asm__ __volatile__ ("nop")
-
-/* This has special calling conventions */
-#ifndef CONFIG_SMP
-BTFIXUPDEF_CALL(void, ___xchg32, void)
-#endif
-
-static inline unsigned long xchg_u32(__volatile__ unsigned long *m, unsigned long val)
-{
-#ifdef CONFIG_SMP
-	__asm__ __volatile__("swap [%2], %0"
-			     : "=&r" (val)
-			     : "0" (val), "r" (m)
-			     : "memory");
-	return val;
-#else
-	register unsigned long *ptr asm("g1");
-	register unsigned long ret asm("g2");
-
-	ptr = (unsigned long *) m;
-	ret = val;
-
-	/* Note: this is magic and the nop there is
-	   really needed. */
-	__asm__ __volatile__(
-	"mov	%%o7, %%g4\n\t"
-	"call	___f____xchg32\n\t"
-	" nop\n\t"
-	: "=&r" (ret)
-	: "0" (ret), "r" (ptr)
-	: "g3", "g4", "g7", "memory", "cc");
-
-	return ret;
 #endif
-}
-
-#define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
-
-extern void __xchg_called_with_bad_pointer(void);
-
-static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr, int size)
-{
-	switch (size) {
-	case 4:
-		return xchg_u32(ptr, x);
-	};
-	__xchg_called_with_bad_pointer();
-	return x;
-}
-
-/* Emulate cmpxchg() the same way we emulate atomics,
- * by hashing the object address and indexing into an array
- * of spinlocks to get a bit of performance...
- *
- * See arch/sparc/lib/atomic32.c for implementation.
- *
- * Cribbed from <asm-parisc/atomic.h>
- */
-#define __HAVE_ARCH_CMPXCHG	1
-
-/* bug catcher for when unsupported size is used - won't link */
-extern void __cmpxchg_called_with_bad_pointer(void);
-/* we only need to support cmpxchg of a u32 on sparc */
-extern unsigned long __cmpxchg_u32(volatile u32 *m, u32 old, u32 new_);
-
-/* don't worry...optimizer will get rid of most of this */
-static inline unsigned long
-__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
-{
-	switch (size) {
-	case 4:
-		return __cmpxchg_u32((u32 *)ptr, (u32)old, (u32)new_);
-	default:
-		__cmpxchg_called_with_bad_pointer();
-		break;
-	}
-	return old;
-}
-
-#define cmpxchg(ptr, o, n)						\
-({									\
-	__typeof__(*(ptr)) _o_ = (o);					\
-	__typeof__(*(ptr)) _n_ = (n);					\
-	(__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,	\
-			(unsigned long)_n_, sizeof(*(ptr)));		\
-})
-
-#include <asm-generic/cmpxchg-local.h>
-
-/*
- * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
- * them available.
- */
-#define cmpxchg_local(ptr, o, n)				  	       \
-	((__typeof__(*(ptr)))__cmpxchg_local_generic((ptr), (unsigned long)(o),\
-			(unsigned long)(n), sizeof(*(ptr))))
-#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
-
-extern void die_if_kernel(char *str, struct pt_regs *regs) __attribute__ ((noreturn));
-
-#endif /* __KERNEL__ */
-
-#endif /* __ASSEMBLY__ */
-
-#define arch_align_stack(x) (x)
-
-#endif /* !(__SPARC_SYSTEM_H) */
diff --git a/include/asm-sparc/system_32.h b/include/asm-sparc/system_32.h
new file mode 100644
index 00000000000..b4b024445fc
--- /dev/null
+++ b/include/asm-sparc/system_32.h
@@ -0,0 +1,288 @@
+#ifndef __SPARC_SYSTEM_H
+#define __SPARC_SYSTEM_H
+
+#include <linux/kernel.h>
+#include <linux/threads.h>	/* NR_CPUS */
+#include <linux/thread_info.h>
+
+#include <asm/page.h>
+#include <asm/psr.h>
+#include <asm/ptrace.h>
+#include <asm/btfixup.h>
+#include <asm/smp.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/irqflags.h>
+
+/*
+ * Sparc (general) CPU types
+ */
+enum sparc_cpu {
+  sun4        = 0x00,
+  sun4c       = 0x01,
+  sun4m       = 0x02,
+  sun4d       = 0x03,
+  sun4e       = 0x04,
+  sun4u       = 0x05, /* V8 ploos ploos */
+  sun_unknown = 0x06,
+  ap1000      = 0x07, /* almost a sun4m */
+};
+
+/* Really, userland should not be looking at any of this... */
+#ifdef __KERNEL__
+
+extern enum sparc_cpu sparc_cpu_model;
+
+#ifndef CONFIG_SUN4
+#define ARCH_SUN4C_SUN4 (sparc_cpu_model==sun4c)
+#define ARCH_SUN4 0
+#else
+#define ARCH_SUN4C_SUN4 1
+#define ARCH_SUN4 1
+#endif
+
+#define SUN4M_NCPUS            4              /* Architectural limit of sun4m. */
+
+extern char reboot_command[];
+
+extern struct thread_info *current_set[NR_CPUS];
+
+extern unsigned long empty_bad_page;
+extern unsigned long empty_bad_page_table;
+extern unsigned long empty_zero_page;
+
+extern void sun_do_break(void);
+extern int serial_console;
+extern int stop_a_enabled;
+
+static inline int con_is_present(void)
+{
+	return serial_console ? 0 : 1;
+}
+
+/* When a context switch happens we must flush all user windows so that
+ * the windows of the current process are flushed onto its stack. This
+ * way the windows are all clean for the next process and the stack
+ * frames are up to date.
+ */
+extern void flush_user_windows(void);
+extern void kill_user_windows(void);
+extern void synchronize_user_stack(void);
+extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
+		   void *fpqueue, unsigned long *fpqdepth);
+
+#ifdef CONFIG_SMP
+#define SWITCH_ENTER(prv) \
+	do {			\
+	if (test_tsk_thread_flag(prv, TIF_USEDFPU)) { \
+		put_psr(get_psr() | PSR_EF); \
+		fpsave(&(prv)->thread.float_regs[0], &(prv)->thread.fsr, \
+		       &(prv)->thread.fpqueue[0], &(prv)->thread.fpqdepth); \
+		clear_tsk_thread_flag(prv, TIF_USEDFPU); \
+		(prv)->thread.kregs->psr &= ~PSR_EF; \
+	} \
+	} while(0)
+
+#define SWITCH_DO_LAZY_FPU(next)	/* */
+#else
+#define SWITCH_ENTER(prv)		/* */
+#define SWITCH_DO_LAZY_FPU(nxt)	\
+	do {			\
+	if (last_task_used_math != (nxt))		\
+		(nxt)->thread.kregs->psr&=~PSR_EF;	\
+	} while(0)
+#endif
+
+extern void flushw_all(void);
+
+/*
+ * Flush windows so that the VM switch which follows
+ * would not pull the stack from under us.
+ *
+ * SWITCH_ENTER and SWITH_DO_LAZY_FPU do not work yet (e.g. SMP does not work)
+ * XXX WTF is the above comment? Found in late teen 2.4.x.
+ */
+#define prepare_arch_switch(next) do { \
+	__asm__ __volatile__( \
+	".globl\tflush_patch_switch\nflush_patch_switch:\n\t" \
+	"save %sp, -0x40, %sp; save %sp, -0x40, %sp; save %sp, -0x40, %sp\n\t" \
+	"save %sp, -0x40, %sp; save %sp, -0x40, %sp; save %sp, -0x40, %sp\n\t" \
+	"save %sp, -0x40, %sp\n\t" \
+	"restore; restore; restore; restore; restore; restore; restore"); \
+} while(0)
+
+	/* Much care has gone into this code, do not touch it.
+	 *
+	 * We need to loadup regs l0/l1 for the newly forked child
+	 * case because the trap return path relies on those registers
+	 * holding certain values, gcc is told that they are clobbered.
+	 * Gcc needs registers for 3 values in and 1 value out, so we
+	 * clobber every non-fixed-usage register besides l2/l3/o4/o5.  -DaveM
+	 *
+	 * Hey Dave, that do not touch sign is too much of an incentive
+	 * - Anton & Pete
+	 */
+#define switch_to(prev, next, last) do {						\
+	SWITCH_ENTER(prev);								\
+	SWITCH_DO_LAZY_FPU(next);							\
+	cpu_set(smp_processor_id(), next->active_mm->cpu_vm_mask);			\
+	__asm__ __volatile__(								\
+	"sethi	%%hi(here - 0x8), %%o7\n\t"						\
+	"mov	%%g6, %%g3\n\t"								\
+	"or	%%o7, %%lo(here - 0x8), %%o7\n\t"					\
+	"rd	%%psr, %%g4\n\t"							\
+	"std	%%sp, [%%g6 + %4]\n\t"							\
+	"rd	%%wim, %%g5\n\t"							\
+	"wr	%%g4, 0x20, %%psr\n\t"							\
+	"nop\n\t"									\
+	"std	%%g4, [%%g6 + %3]\n\t"							\
+	"ldd	[%2 + %3], %%g4\n\t"							\
+	"mov	%2, %%g6\n\t"								\
+	".globl	patchme_store_new_current\n"						\
+"patchme_store_new_current:\n\t"							\
+	"st	%2, [%1]\n\t"								\
+	"wr	%%g4, 0x20, %%psr\n\t"							\
+	"nop\n\t"									\
+	"nop\n\t"									\
+	"nop\n\t"	/* LEON needs all 3 nops: load to %sp depends on CWP. */		\
+	"ldd	[%%g6 + %4], %%sp\n\t"							\
+	"wr	%%g5, 0x0, %%wim\n\t"							\
+	"ldd	[%%sp + 0x00], %%l0\n\t"						\
+	"ldd	[%%sp + 0x38], %%i6\n\t"						\
+	"wr	%%g4, 0x0, %%psr\n\t"							\
+	"nop\n\t"									\
+	"nop\n\t"									\
+	"jmpl	%%o7 + 0x8, %%g0\n\t"							\
+	" ld	[%%g3 + %5], %0\n\t"							\
+	"here:\n"									\
+        : "=&r" (last)									\
+        : "r" (&(current_set[hard_smp_processor_id()])),	\
+	  "r" (task_thread_info(next)),				\
+	  "i" (TI_KPSR),					\
+	  "i" (TI_KSP),						\
+	  "i" (TI_TASK)						\
+	:       "g1", "g2", "g3", "g4", "g5",       "g7",	\
+	  "l0", "l1",       "l3", "l4", "l5", "l6", "l7",	\
+	  "i0", "i1", "i2", "i3", "i4", "i5",			\
+	  "o0", "o1", "o2", "o3",                   "o7");	\
+	} while(0)
+
+/* XXX Change this if we ever use a PSO mode kernel. */
+#define mb()	__asm__ __volatile__ ("" : : : "memory")
+#define rmb()	mb()
+#define wmb()	mb()
+#define read_barrier_depends()	do { } while(0)
+#define set_mb(__var, __value)  do { __var = __value; mb(); } while(0)
+#define smp_mb()	__asm__ __volatile__("":::"memory")
+#define smp_rmb()	__asm__ __volatile__("":::"memory")
+#define smp_wmb()	__asm__ __volatile__("":::"memory")
+#define smp_read_barrier_depends()	do { } while(0)
+
+#define nop() __asm__ __volatile__ ("nop")
+
+/* This has special calling conventions */
+#ifndef CONFIG_SMP
+BTFIXUPDEF_CALL(void, ___xchg32, void)
+#endif
+
+static inline unsigned long xchg_u32(__volatile__ unsigned long *m, unsigned long val)
+{
+#ifdef CONFIG_SMP
+	__asm__ __volatile__("swap [%2], %0"
+			     : "=&r" (val)
+			     : "0" (val), "r" (m)
+			     : "memory");
+	return val;
+#else
+	register unsigned long *ptr asm("g1");
+	register unsigned long ret asm("g2");
+
+	ptr = (unsigned long *) m;
+	ret = val;
+
+	/* Note: this is magic and the nop there is
+	   really needed. */
+	__asm__ __volatile__(
+	"mov	%%o7, %%g4\n\t"
+	"call	___f____xchg32\n\t"
+	" nop\n\t"
+	: "=&r" (ret)
+	: "0" (ret), "r" (ptr)
+	: "g3", "g4", "g7", "memory", "cc");
+
+	return ret;
+#endif
+}
+
+#define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
+
+extern void __xchg_called_with_bad_pointer(void);
+
+static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr, int size)
+{
+	switch (size) {
+	case 4:
+		return xchg_u32(ptr, x);
+	};
+	__xchg_called_with_bad_pointer();
+	return x;
+}
+
+/* Emulate cmpxchg() the same way we emulate atomics,
+ * by hashing the object address and indexing into an array
+ * of spinlocks to get a bit of performance...
+ *
+ * See arch/sparc/lib/atomic32.c for implementation.
+ *
+ * Cribbed from <asm-parisc/atomic.h>
+ */
+#define __HAVE_ARCH_CMPXCHG	1
+
+/* bug catcher for when unsupported size is used - won't link */
+extern void __cmpxchg_called_with_bad_pointer(void);
+/* we only need to support cmpxchg of a u32 on sparc */
+extern unsigned long __cmpxchg_u32(volatile u32 *m, u32 old, u32 new_);
+
+/* don't worry...optimizer will get rid of most of this */
+static inline unsigned long
+__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
+{
+	switch (size) {
+	case 4:
+		return __cmpxchg_u32((u32 *)ptr, (u32)old, (u32)new_);
+	default:
+		__cmpxchg_called_with_bad_pointer();
+		break;
+	}
+	return old;
+}
+
+#define cmpxchg(ptr, o, n)						\
+({									\
+	__typeof__(*(ptr)) _o_ = (o);					\
+	__typeof__(*(ptr)) _n_ = (n);					\
+	(__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,	\
+			(unsigned long)_n_, sizeof(*(ptr)));		\
+})
+
+#include <asm-generic/cmpxchg-local.h>
+
+/*
+ * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
+ * them available.
+ */
+#define cmpxchg_local(ptr, o, n)				  	       \
+	((__typeof__(*(ptr)))__cmpxchg_local_generic((ptr), (unsigned long)(o),\
+			(unsigned long)(n), sizeof(*(ptr))))
+#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
+
+extern void die_if_kernel(char *str, struct pt_regs *regs) __attribute__ ((noreturn));
+
+#endif /* __KERNEL__ */
+
+#endif /* __ASSEMBLY__ */
+
+#define arch_align_stack(x) (x)
+
+#endif /* !(__SPARC_SYSTEM_H) */
diff --git a/include/asm-sparc/system_64.h b/include/asm-sparc/system_64.h
new file mode 100644
index 00000000000..db9e742a406
--- /dev/null
+++ b/include/asm-sparc/system_64.h
@@ -0,0 +1,355 @@
+#ifndef __SPARC64_SYSTEM_H
+#define __SPARC64_SYSTEM_H
+
+#include <asm/ptrace.h>
+#include <asm/processor.h>
+#include <asm/visasm.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/irqflags.h>
+#include <asm-generic/cmpxchg-local.h>
+
+/*
+ * Sparc (general) CPU types
+ */
+enum sparc_cpu {
+  sun4        = 0x00,
+  sun4c       = 0x01,
+  sun4m       = 0x02,
+  sun4d       = 0x03,
+  sun4e       = 0x04,
+  sun4u       = 0x05, /* V8 ploos ploos */
+  sun_unknown = 0x06,
+  ap1000      = 0x07, /* almost a sun4m */
+};
+
+#define sparc_cpu_model sun4u
+
+/* This cannot ever be a sun4c nor sun4 :) That's just history. */
+#define ARCH_SUN4C_SUN4 0
+#define ARCH_SUN4 0
+
+extern char reboot_command[];
+
+/* These are here in an effort to more fully work around Spitfire Errata
+ * #51.  Essentially, if a memory barrier occurs soon after a mispredicted
+ * branch, the chip can stop executing instructions until a trap occurs.
+ * Therefore, if interrupts are disabled, the chip can hang forever.
+ *
+ * It used to be believed that the memory barrier had to be right in the
+ * delay slot, but a case has been traced recently wherein the memory barrier
+ * was one instruction after the branch delay slot and the chip still hung.
+ * The offending sequence was the following in sym_wakeup_done() of the
+ * sym53c8xx_2 driver:
+ *
+ *	call	sym_ccb_from_dsa, 0
+ *	 movge	%icc, 0, %l0
+ *	brz,pn	%o0, .LL1303
+ *	 mov	%o0, %l2
+ *	membar	#LoadLoad
+ *
+ * The branch has to be mispredicted for the bug to occur.  Therefore, we put
+ * the memory barrier explicitly into a "branch always, predicted taken"
+ * delay slot to avoid the problem case.
+ */
+#define membar_safe(type) \
+do {	__asm__ __volatile__("ba,pt	%%xcc, 1f\n\t" \
+			     " membar	" type "\n" \
+			     "1:\n" \
+			     : : : "memory"); \
+} while (0)
+
+#define mb()	\
+	membar_safe("#LoadLoad | #LoadStore | #StoreStore | #StoreLoad")
+#define rmb()	\
+	membar_safe("#LoadLoad")
+#define wmb()	\
+	membar_safe("#StoreStore")
+#define membar_storeload() \
+	membar_safe("#StoreLoad")
+#define membar_storeload_storestore() \
+	membar_safe("#StoreLoad | #StoreStore")
+#define membar_storeload_loadload() \
+	membar_safe("#StoreLoad | #LoadLoad")
+#define membar_storestore_loadstore() \
+	membar_safe("#StoreStore | #LoadStore")
+
+#endif
+
+#define nop() 		__asm__ __volatile__ ("nop")
+
+#define read_barrier_depends()		do { } while(0)
+#define set_mb(__var, __value) \
+	do { __var = __value; membar_storeload_storestore(); } while(0)
+
+#ifdef CONFIG_SMP
+#define smp_mb()	mb()
+#define smp_rmb()	rmb()
+#define smp_wmb()	wmb()
+#define smp_read_barrier_depends()	read_barrier_depends()
+#else
+#define smp_mb()	__asm__ __volatile__("":::"memory")
+#define smp_rmb()	__asm__ __volatile__("":::"memory")
+#define smp_wmb()	__asm__ __volatile__("":::"memory")
+#define smp_read_barrier_depends()	do { } while(0)
+#endif
+
+#define flushi(addr)	__asm__ __volatile__ ("flush %0" : : "r" (addr) : "memory")
+
+#define flushw_all()	__asm__ __volatile__("flushw")
+
+/* Performance counter register access. */
+#define read_pcr(__p)  __asm__ __volatile__("rd	%%pcr, %0" : "=r" (__p))
+#define write_pcr(__p) __asm__ __volatile__("wr	%0, 0x0, %%pcr" : : "r" (__p))
+#define read_pic(__p)  __asm__ __volatile__("rd %%pic, %0" : "=r" (__p))
+
+/* Blackbird errata workaround.  See commentary in
+ * arch/sparc64/kernel/smp.c:smp_percpu_timer_interrupt()
+ * for more information.
+ */
+#define reset_pic()    						\
+	__asm__ __volatile__("ba,pt	%xcc, 99f\n\t"		\
+			     ".align	64\n"			\
+			  "99:wr	%g0, 0x0, %pic\n\t"	\
+			     "rd	%pic, %g0")
+
+#ifndef __ASSEMBLY__
+
+extern void sun_do_break(void);
+extern int stop_a_enabled;
+
+extern void fault_in_user_windows(void);
+extern void synchronize_user_stack(void);
+
+extern void __flushw_user(void);
+#define flushw_user() __flushw_user()
+
+#define flush_user_windows flushw_user
+#define flush_register_windows flushw_all
+
+/* Don't hold the runqueue lock over context switch */
+#define __ARCH_WANT_UNLOCKED_CTXSW
+#define prepare_arch_switch(next)		\
+do {						\
+	flushw_all();				\
+} while (0)
+
+	/* See what happens when you design the chip correctly?
+	 *
+	 * We tell gcc we clobber all non-fixed-usage registers except
+	 * for l0/l1.  It will use one for 'next' and the other to hold
+	 * the output value of 'last'.  'next' is not referenced again
+	 * past the invocation of switch_to in the scheduler, so we need
+	 * not preserve it's value.  Hairy, but it lets us remove 2 loads
+	 * and 2 stores in this critical code path.  -DaveM
+	 */
+#define switch_to(prev, next, last)					\
+do {	if (test_thread_flag(TIF_PERFCTR)) {				\
+		unsigned long __tmp;					\
+		read_pcr(__tmp);					\
+		current_thread_info()->pcr_reg = __tmp;			\
+		read_pic(__tmp);					\
+		current_thread_info()->kernel_cntd0 += (unsigned int)(__tmp);\
+		current_thread_info()->kernel_cntd1 += ((__tmp) >> 32);	\
+	}								\
+	flush_tlb_pending();						\
+	save_and_clear_fpu();						\
+	/* If you are tempted to conditionalize the following */	\
+	/* so that ASI is only written if it changes, think again. */	\
+	__asm__ __volatile__("wr %%g0, %0, %%asi"			\
+	: : "r" (__thread_flag_byte_ptr(task_thread_info(next))[TI_FLAG_BYTE_CURRENT_DS]));\
+	trap_block[current_thread_info()->cpu].thread =			\
+		task_thread_info(next);					\
+	__asm__ __volatile__(						\
+	"mov	%%g4, %%g7\n\t"						\
+	"stx	%%i6, [%%sp + 2047 + 0x70]\n\t"				\
+	"stx	%%i7, [%%sp + 2047 + 0x78]\n\t"				\
+	"rdpr	%%wstate, %%o5\n\t"					\
+	"stx	%%o6, [%%g6 + %6]\n\t"					\
+	"stb	%%o5, [%%g6 + %5]\n\t"					\
+	"rdpr	%%cwp, %%o5\n\t"					\
+	"stb	%%o5, [%%g6 + %8]\n\t"					\
+	"mov	%4, %%g6\n\t"						\
+	"ldub	[%4 + %8], %%g1\n\t"					\
+	"wrpr	%%g1, %%cwp\n\t"					\
+	"ldx	[%%g6 + %6], %%o6\n\t"					\
+	"ldub	[%%g6 + %5], %%o5\n\t"					\
+	"ldub	[%%g6 + %7], %%o7\n\t"					\
+	"wrpr	%%o5, 0x0, %%wstate\n\t"				\
+	"ldx	[%%sp + 2047 + 0x70], %%i6\n\t"				\
+	"ldx	[%%sp + 2047 + 0x78], %%i7\n\t"				\
+	"ldx	[%%g6 + %9], %%g4\n\t"					\
+	"brz,pt %%o7, switch_to_pc\n\t"					\
+	" mov	%%g7, %0\n\t"						\
+	"sethi	%%hi(ret_from_syscall), %%g1\n\t"			\
+	"jmpl	%%g1 + %%lo(ret_from_syscall), %%g0\n\t"		\
+	" nop\n\t"							\
+	".globl switch_to_pc\n\t"					\
+	"switch_to_pc:\n\t"						\
+	: "=&r" (last), "=r" (current), "=r" (current_thread_info_reg),	\
+	  "=r" (__local_per_cpu_offset)					\
+	: "0" (task_thread_info(next)),					\
+	  "i" (TI_WSTATE), "i" (TI_KSP), "i" (TI_NEW_CHILD),            \
+	  "i" (TI_CWP), "i" (TI_TASK)					\
+	: "cc",								\
+	        "g1", "g2", "g3",                   "g7",		\
+	        "l1", "l2", "l3", "l4", "l5", "l6", "l7",		\
+	  "i0", "i1", "i2", "i3", "i4", "i5",				\
+	  "o0", "o1", "o2", "o3", "o4", "o5",       "o7");		\
+	/* If you fuck with this, update ret_from_syscall code too. */	\
+	if (test_thread_flag(TIF_PERFCTR)) {				\
+		write_pcr(current_thread_info()->pcr_reg);		\
+		reset_pic();						\
+	}								\
+} while(0)
+
+static inline unsigned long xchg32(__volatile__ unsigned int *m, unsigned int val)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__(
+"	membar		#StoreLoad | #LoadLoad\n"
+"	mov		%0, %1\n"
+"1:	lduw		[%4], %2\n"
+"	cas		[%4], %2, %0\n"
+"	cmp		%2, %0\n"
+"	bne,a,pn	%%icc, 1b\n"
+"	 mov		%1, %0\n"
+"	membar		#StoreLoad | #StoreStore\n"
+	: "=&r" (val), "=&r" (tmp1), "=&r" (tmp2)
+	: "0" (val), "r" (m)
+	: "cc", "memory");
+	return val;
+}
+
+static inline unsigned long xchg64(__volatile__ unsigned long *m, unsigned long val)
+{
+	unsigned long tmp1, tmp2;
+
+	__asm__ __volatile__(
+"	membar		#StoreLoad | #LoadLoad\n"
+"	mov		%0, %1\n"
+"1:	ldx		[%4], %2\n"
+"	casx		[%4], %2, %0\n"
+"	cmp		%2, %0\n"
+"	bne,a,pn	%%xcc, 1b\n"
+"	 mov		%1, %0\n"
+"	membar		#StoreLoad | #StoreStore\n"
+	: "=&r" (val), "=&r" (tmp1), "=&r" (tmp2)
+	: "0" (val), "r" (m)
+	: "cc", "memory");
+	return val;
+}
+
+#define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
+
+extern void __xchg_called_with_bad_pointer(void);
+
+static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr,
+				       int size)
+{
+	switch (size) {
+	case 4:
+		return xchg32(ptr, x);
+	case 8:
+		return xchg64(ptr, x);
+	};
+	__xchg_called_with_bad_pointer();
+	return x;
+}
+
+extern void die_if_kernel(char *str, struct pt_regs *regs) __attribute__ ((noreturn));
+
+/*
+ * Atomic compare and exchange.  Compare OLD with MEM, if identical,
+ * store NEW in MEM.  Return the initial value in MEM.  Success is
+ * indicated by comparing RETURN with OLD.
+ */
+
+#define __HAVE_ARCH_CMPXCHG 1
+
+static inline unsigned long
+__cmpxchg_u32(volatile int *m, int old, int new)
+{
+	__asm__ __volatile__("membar #StoreLoad | #LoadLoad\n"
+			     "cas [%2], %3, %0\n\t"
+			     "membar #StoreLoad | #StoreStore"
+			     : "=&r" (new)
+			     : "0" (new), "r" (m), "r" (old)
+			     : "memory");
+
+	return new;
+}
+
+static inline unsigned long
+__cmpxchg_u64(volatile long *m, unsigned long old, unsigned long new)
+{
+	__asm__ __volatile__("membar #StoreLoad | #LoadLoad\n"
+			     "casx [%2], %3, %0\n\t"
+			     "membar #StoreLoad | #StoreStore"
+			     : "=&r" (new)
+			     : "0" (new), "r" (m), "r" (old)
+			     : "memory");
+
+	return new;
+}
+
+/* This function doesn't exist, so you'll get a linker error
+   if something tries to do an invalid cmpxchg().  */
+extern void __cmpxchg_called_with_bad_pointer(void);
+
+static inline unsigned long
+__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
+{
+	switch (size) {
+		case 4:
+			return __cmpxchg_u32(ptr, old, new);
+		case 8:
+			return __cmpxchg_u64(ptr, old, new);
+	}
+	__cmpxchg_called_with_bad_pointer();
+	return old;
+}
+
+#define cmpxchg(ptr,o,n)						 \
+  ({									 \
+     __typeof__(*(ptr)) _o_ = (o);					 \
+     __typeof__(*(ptr)) _n_ = (n);					 \
+     (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,		 \
+				    (unsigned long)_n_, sizeof(*(ptr))); \
+  })
+
+/*
+ * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
+ * them available.
+ */
+
+static inline unsigned long __cmpxchg_local(volatile void *ptr,
+				      unsigned long old,
+				      unsigned long new, int size)
+{
+	switch (size) {
+	case 4:
+	case 8:	return __cmpxchg(ptr, old, new, size);
+	default:
+		return __cmpxchg_local_generic(ptr, old, new, size);
+	}
+
+	return old;
+}
+
+#define cmpxchg_local(ptr, o, n)				  	\
+	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
+			(unsigned long)(n), sizeof(*(ptr))))
+#define cmpxchg64_local(ptr, o, n)					\
+  ({									\
+	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
+	cmpxchg_local((ptr), (o), (n));					\
+  })
+
+#endif /* !(__ASSEMBLY__) */
+
+#define arch_align_stack(x) (x)
+
+#endif /* !(__SPARC64_SYSTEM_H) */
diff --git a/include/asm-sparc/termbits.h b/include/asm-sparc/termbits.h
index 90cf2210118..d6ca3e2754f 100644
--- a/include/asm-sparc/termbits.h
+++ b/include/asm-sparc/termbits.h
@@ -5,7 +5,12 @@
 
 typedef unsigned char   cc_t;
 typedef unsigned int    speed_t;
+
+#if defined(__sparc__) && defined(__arch64__)
+typedef unsigned int    tcflag_t;
+#else
 typedef unsigned long   tcflag_t;
+#endif
 
 #define NCC 8
 struct termio {
diff --git a/include/asm-sparc/termios.h b/include/asm-sparc/termios.h
index f7b4409c35f..e8ba9539964 100644
--- a/include/asm-sparc/termios.h
+++ b/include/asm-sparc/termios.h
@@ -53,7 +53,6 @@ struct winsize {
 #define _VMIN	4
 #define _VTIME	5
 
-
 /*	intr=^C		quit=^\		erase=del	kill=^U
 	eof=^D		eol=\0		eol2=\0		sxtc=\0
 	start=^Q	stop=^S		susp=^Z		dsusp=^Y
@@ -68,16 +67,17 @@ struct winsize {
 #define user_termio_to_kernel_termios(termios, termio) \
 ({ \
 	unsigned short tmp; \
-	get_user(tmp, &(termio)->c_iflag); \
+	int err; \
+	err = get_user(tmp, &(termio)->c_iflag); \
 	(termios)->c_iflag = (0xffff0000 & ((termios)->c_iflag)) | tmp; \
-	get_user(tmp, &(termio)->c_oflag); \
+	err |= get_user(tmp, &(termio)->c_oflag); \
 	(termios)->c_oflag = (0xffff0000 & ((termios)->c_oflag)) | tmp; \
-	get_user(tmp, &(termio)->c_cflag); \
+	err |= get_user(tmp, &(termio)->c_cflag); \
 	(termios)->c_cflag = (0xffff0000 & ((termios)->c_cflag)) | tmp; \
-	get_user(tmp, &(termio)->c_lflag); \
+	err |= get_user(tmp, &(termio)->c_lflag); \
 	(termios)->c_lflag = (0xffff0000 & ((termios)->c_lflag)) | tmp; \
-	copy_from_user((termios)->c_cc, (termio)->c_cc, NCC); \
-	0; \
+	err |= copy_from_user((termios)->c_cc, (termio)->c_cc, NCC); \
+	err; \
 })
 
 /*
@@ -87,17 +87,18 @@ struct winsize {
  */
 #define kernel_termios_to_user_termio(termio, termios) \
 ({ \
-	put_user((termios)->c_iflag, &(termio)->c_iflag); \
-	put_user((termios)->c_oflag, &(termio)->c_oflag); \
-	put_user((termios)->c_cflag, &(termio)->c_cflag); \
-	put_user((termios)->c_lflag, &(termio)->c_lflag); \
-	put_user((termios)->c_line,  &(termio)->c_line); \
-	copy_to_user((termio)->c_cc, (termios)->c_cc, NCC); \
+	int err; \
+	err  = put_user((termios)->c_iflag, &(termio)->c_iflag); \
+	err |= put_user((termios)->c_oflag, &(termio)->c_oflag); \
+	err |= put_user((termios)->c_cflag, &(termio)->c_cflag); \
+	err |= put_user((termios)->c_lflag, &(termio)->c_lflag); \
+	err |= put_user((termios)->c_line,  &(termio)->c_line); \
+	err |= copy_to_user((termio)->c_cc, (termios)->c_cc, NCC); \
 	if (!((termios)->c_lflag & ICANON)) { \
-		put_user((termios)->c_cc[VMIN], &(termio)->c_cc[_VMIN]); \
-		put_user((termios)->c_cc[VTIME], &(termio)->c_cc[_VTIME]); \
+		err |= put_user((termios)->c_cc[VMIN], &(termio)->c_cc[_VMIN]); \
+		err |= put_user((termios)->c_cc[VTIME], &(termio)->c_cc[_VTIME]); \
 	} \
-	0; \
+	err; \
 })
 
 #define user_termios_to_kernel_termios(k, u) \
@@ -144,38 +145,40 @@ struct winsize {
 
 #define user_termios_to_kernel_termios_1(k, u) \
 ({ \
-	get_user((k)->c_iflag, &(u)->c_iflag); \
-	get_user((k)->c_oflag, &(u)->c_oflag); \
-	get_user((k)->c_cflag, &(u)->c_cflag); \
-	get_user((k)->c_lflag, &(u)->c_lflag); \
-	get_user((k)->c_line,  &(u)->c_line); \
-	copy_from_user((k)->c_cc, (u)->c_cc, NCCS); \
+	int err; \
+	err  = get_user((k)->c_iflag, &(u)->c_iflag); \
+	err |= get_user((k)->c_oflag, &(u)->c_oflag); \
+	err |= get_user((k)->c_cflag, &(u)->c_cflag); \
+	err |= get_user((k)->c_lflag, &(u)->c_lflag); \
+	err |= get_user((k)->c_line,  &(u)->c_line); \
+	err |= copy_from_user((k)->c_cc, (u)->c_cc, NCCS); \
 	if ((k)->c_lflag & ICANON) { \
-		get_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		get_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
+		err |= get_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
+		err |= get_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
 	} else { \
-		get_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		get_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
+		err |= get_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
+		err |= get_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
 	} \
-	0; \
+	err; \
 })
 
 #define kernel_termios_to_user_termios_1(u, k) \
 ({ \
-	put_user((k)->c_iflag, &(u)->c_iflag); \
-	put_user((k)->c_oflag, &(u)->c_oflag); \
-	put_user((k)->c_cflag, &(u)->c_cflag); \
-	put_user((k)->c_lflag, &(u)->c_lflag); \
-	put_user((k)->c_line, &(u)->c_line); \
-	copy_to_user((u)->c_cc, (k)->c_cc, NCCS); \
+	int err; \
+	err  = put_user((k)->c_iflag, &(u)->c_iflag); \
+	err |= put_user((k)->c_oflag, &(u)->c_oflag); \
+	err |= put_user((k)->c_cflag, &(u)->c_cflag); \
+	err |= put_user((k)->c_lflag, &(u)->c_lflag); \
+	err |= put_user((k)->c_line, &(u)->c_line); \
+	err |= copy_to_user((u)->c_cc, (k)->c_cc, NCCS); \
 	if (!((k)->c_lflag & ICANON)) { \
-		put_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		put_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
+		err |= put_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
+		err |= put_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
 	} else { \
-		put_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		put_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
+		err |= put_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
+		err |= put_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
 	} \
-	0; \
+	err; \
 })
 
 #endif	/* __KERNEL__ */
diff --git a/include/asm-sparc/thread_info.h b/include/asm-sparc/thread_info.h
index 91b9f5888c8..64155cf89f3 100644
--- a/include/asm-sparc/thread_info.h
+++ b/include/asm-sparc/thread_info.h
@@ -1,151 +1,8 @@
-/*
- * thread_info.h: sparc low-level thread information
- * adapted from the ppc version by Pete Zaitcev, which was
- * adapted from the i386 version by Paul Mackerras
- *
- * Copyright (C) 2002  David Howells (dhowells@redhat.com)
- * Copyright (c) 2002  Pete Zaitcev (zaitcev@yahoo.com)
- * - Incorporating suggestions made by Linus Torvalds and Dave Miller
- */
-
-#ifndef _ASM_THREAD_INFO_H
-#define _ASM_THREAD_INFO_H
-
-#ifdef __KERNEL__
-
-#ifndef __ASSEMBLY__
-
-#include <asm/btfixup.h>
-#include <asm/ptrace.h>
-#include <asm/page.h>
-
-/*
- * Low level task data.
- *
- * If you change this, change the TI_* offsets below to match.
- */
-#define NSWINS 8
-struct thread_info {
-	unsigned long		uwinmask;
-	struct task_struct	*task;		/* main task structure */
-	struct exec_domain	*exec_domain;	/* execution domain */
-	unsigned long		flags;		/* low level flags */
-	int			cpu;		/* cpu we're on */
-	int			preempt_count;	/* 0 => preemptable,
-						   <0 => BUG */
-	int			softirq_count;
-	int			hardirq_count;
-
-	/* Context switch saved kernel state. */
-	unsigned long ksp;	/* ... ksp __attribute__ ((aligned (8))); */
-	unsigned long kpc;
-	unsigned long kpsr;
-	unsigned long kwim;
-
-	/* A place to store user windows and stack pointers
-	 * when the stack needs inspection.
-	 */
-	struct reg_window	reg_window[NSWINS];	/* align for ldd! */
-	unsigned long		rwbuf_stkptrs[NSWINS];
-	unsigned long		w_saved;
-
-	struct restart_block	restart_block;
-};
-
-/*
- * macros/functions for gaining access to the thread information structure
- *
- * preempt_count needs to be 1 initially, until the scheduler is functional.
- */
-#define INIT_THREAD_INFO(tsk)				\
-{							\
-	.uwinmask	=	0,			\
-	.task		=	&tsk,			\
-	.exec_domain	=	&default_exec_domain,	\
-	.flags		=	0,			\
-	.cpu		=	0,			\
-	.preempt_count	=	1,			\
-	.restart_block	= {				\
-		.fn	=	do_no_restart_syscall,	\
-	},						\
-}
-
-#define init_thread_info	(init_thread_union.thread_info)
-#define init_stack		(init_thread_union.stack)
-
-/* how to get the thread information struct from C */
-register struct thread_info *current_thread_info_reg asm("g6");
-#define current_thread_info()   (current_thread_info_reg)
-
-/*
- * thread information allocation
- */
-#if PAGE_SHIFT == 13
-#define THREAD_INFO_ORDER  0
-#else /* PAGE_SHIFT */
-#define THREAD_INFO_ORDER  1
+#ifndef ___ASM_SPARC_THREAD_INFO_H
+#define ___ASM_SPARC_THREAD_INFO_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/thread_info_64.h>
+#else
+#include <asm-sparc/thread_info_32.h>
+#endif
 #endif
-
-BTFIXUPDEF_CALL(struct thread_info *, alloc_thread_info, void)
-#define alloc_thread_info(tsk) BTFIXUP_CALL(alloc_thread_info)()
-
-BTFIXUPDEF_CALL(void, free_thread_info, struct thread_info *)
-#define free_thread_info(ti) BTFIXUP_CALL(free_thread_info)(ti)
-
-#endif /* __ASSEMBLY__ */
-
-/*
- * Size of kernel stack for each process.
- * Observe the order of get_free_pages() in alloc_thread_info().
- * The sun4 has 8K stack too, because it's short on memory, and 16K is a waste.
- */
-#define THREAD_SIZE		8192
-
-/*
- * Offsets in thread_info structure, used in assembly code
- * The "#define REGWIN_SZ 0x40" was abolished, so no multiplications.
- */
-#define TI_UWINMASK	0x00	/* uwinmask */
-#define TI_TASK		0x04
-#define TI_EXECDOMAIN	0x08	/* exec_domain */
-#define TI_FLAGS	0x0c
-#define TI_CPU		0x10
-#define TI_PREEMPT	0x14	/* preempt_count */
-#define TI_SOFTIRQ	0x18	/* softirq_count */
-#define TI_HARDIRQ	0x1c	/* hardirq_count */
-#define TI_KSP		0x20	/* ksp */
-#define TI_KPC		0x24	/* kpc (ldd'ed with kpc) */
-#define TI_KPSR		0x28	/* kpsr */
-#define TI_KWIM		0x2c	/* kwim (ldd'ed with kpsr) */
-#define TI_REG_WINDOW	0x30
-#define TI_RWIN_SPTRS	0x230
-#define TI_W_SAVED	0x250
-/* #define TI_RESTART_BLOCK 0x25n */ /* Nobody cares */
-
-#define PREEMPT_ACTIVE		0x4000000
-
-/*
- * thread information flag bit numbers
- */
-#define TIF_SYSCALL_TRACE	0	/* syscall trace active */
-/* flag bit 1 is available */
-#define TIF_SIGPENDING		2	/* signal pending */
-#define TIF_NEED_RESCHED	3	/* rescheduling necessary */
-#define TIF_RESTORE_SIGMASK	4	/* restore signal mask in do_signal() */
-#define TIF_USEDFPU		8	/* FPU was used by this task
-					 * this quantum (SMP) */
-#define TIF_POLLING_NRFLAG	9	/* true if poll_idle() is polling
-					 * TIF_NEED_RESCHED */
-#define TIF_MEMDIE		10
-
-/* as above, but as bit values */
-#define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
-#define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
-#define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
-#define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
-#define _TIF_USEDFPU		(1<<TIF_USEDFPU)
-#define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
-
-#endif /* __KERNEL__ */
-
-#endif /* _ASM_THREAD_INFO_H */
diff --git a/include/asm-sparc/thread_info_32.h b/include/asm-sparc/thread_info_32.h
new file mode 100644
index 00000000000..91b9f5888c8
--- /dev/null
+++ b/include/asm-sparc/thread_info_32.h
@@ -0,0 +1,151 @@
+/*
+ * thread_info.h: sparc low-level thread information
+ * adapted from the ppc version by Pete Zaitcev, which was
+ * adapted from the i386 version by Paul Mackerras
+ *
+ * Copyright (C) 2002  David Howells (dhowells@redhat.com)
+ * Copyright (c) 2002  Pete Zaitcev (zaitcev@yahoo.com)
+ * - Incorporating suggestions made by Linus Torvalds and Dave Miller
+ */
+
+#ifndef _ASM_THREAD_INFO_H
+#define _ASM_THREAD_INFO_H
+
+#ifdef __KERNEL__
+
+#ifndef __ASSEMBLY__
+
+#include <asm/btfixup.h>
+#include <asm/ptrace.h>
+#include <asm/page.h>
+
+/*
+ * Low level task data.
+ *
+ * If you change this, change the TI_* offsets below to match.
+ */
+#define NSWINS 8
+struct thread_info {
+	unsigned long		uwinmask;
+	struct task_struct	*task;		/* main task structure */
+	struct exec_domain	*exec_domain;	/* execution domain */
+	unsigned long		flags;		/* low level flags */
+	int			cpu;		/* cpu we're on */
+	int			preempt_count;	/* 0 => preemptable,
+						   <0 => BUG */
+	int			softirq_count;
+	int			hardirq_count;
+
+	/* Context switch saved kernel state. */
+	unsigned long ksp;	/* ... ksp __attribute__ ((aligned (8))); */
+	unsigned long kpc;
+	unsigned long kpsr;
+	unsigned long kwim;
+
+	/* A place to store user windows and stack pointers
+	 * when the stack needs inspection.
+	 */
+	struct reg_window	reg_window[NSWINS];	/* align for ldd! */
+	unsigned long		rwbuf_stkptrs[NSWINS];
+	unsigned long		w_saved;
+
+	struct restart_block	restart_block;
+};
+
+/*
+ * macros/functions for gaining access to the thread information structure
+ *
+ * preempt_count needs to be 1 initially, until the scheduler is functional.
+ */
+#define INIT_THREAD_INFO(tsk)				\
+{							\
+	.uwinmask	=	0,			\
+	.task		=	&tsk,			\
+	.exec_domain	=	&default_exec_domain,	\
+	.flags		=	0,			\
+	.cpu		=	0,			\
+	.preempt_count	=	1,			\
+	.restart_block	= {				\
+		.fn	=	do_no_restart_syscall,	\
+	},						\
+}
+
+#define init_thread_info	(init_thread_union.thread_info)
+#define init_stack		(init_thread_union.stack)
+
+/* how to get the thread information struct from C */
+register struct thread_info *current_thread_info_reg asm("g6");
+#define current_thread_info()   (current_thread_info_reg)
+
+/*
+ * thread information allocation
+ */
+#if PAGE_SHIFT == 13
+#define THREAD_INFO_ORDER  0
+#else /* PAGE_SHIFT */
+#define THREAD_INFO_ORDER  1
+#endif
+
+BTFIXUPDEF_CALL(struct thread_info *, alloc_thread_info, void)
+#define alloc_thread_info(tsk) BTFIXUP_CALL(alloc_thread_info)()
+
+BTFIXUPDEF_CALL(void, free_thread_info, struct thread_info *)
+#define free_thread_info(ti) BTFIXUP_CALL(free_thread_info)(ti)
+
+#endif /* __ASSEMBLY__ */
+
+/*
+ * Size of kernel stack for each process.
+ * Observe the order of get_free_pages() in alloc_thread_info().
+ * The sun4 has 8K stack too, because it's short on memory, and 16K is a waste.
+ */
+#define THREAD_SIZE		8192
+
+/*
+ * Offsets in thread_info structure, used in assembly code
+ * The "#define REGWIN_SZ 0x40" was abolished, so no multiplications.
+ */
+#define TI_UWINMASK	0x00	/* uwinmask */
+#define TI_TASK		0x04
+#define TI_EXECDOMAIN	0x08	/* exec_domain */
+#define TI_FLAGS	0x0c
+#define TI_CPU		0x10
+#define TI_PREEMPT	0x14	/* preempt_count */
+#define TI_SOFTIRQ	0x18	/* softirq_count */
+#define TI_HARDIRQ	0x1c	/* hardirq_count */
+#define TI_KSP		0x20	/* ksp */
+#define TI_KPC		0x24	/* kpc (ldd'ed with kpc) */
+#define TI_KPSR		0x28	/* kpsr */
+#define TI_KWIM		0x2c	/* kwim (ldd'ed with kpsr) */
+#define TI_REG_WINDOW	0x30
+#define TI_RWIN_SPTRS	0x230
+#define TI_W_SAVED	0x250
+/* #define TI_RESTART_BLOCK 0x25n */ /* Nobody cares */
+
+#define PREEMPT_ACTIVE		0x4000000
+
+/*
+ * thread information flag bit numbers
+ */
+#define TIF_SYSCALL_TRACE	0	/* syscall trace active */
+/* flag bit 1 is available */
+#define TIF_SIGPENDING		2	/* signal pending */
+#define TIF_NEED_RESCHED	3	/* rescheduling necessary */
+#define TIF_RESTORE_SIGMASK	4	/* restore signal mask in do_signal() */
+#define TIF_USEDFPU		8	/* FPU was used by this task
+					 * this quantum (SMP) */
+#define TIF_POLLING_NRFLAG	9	/* true if poll_idle() is polling
+					 * TIF_NEED_RESCHED */
+#define TIF_MEMDIE		10
+
+/* as above, but as bit values */
+#define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
+#define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
+#define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
+#define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
+#define _TIF_USEDFPU		(1<<TIF_USEDFPU)
+#define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_THREAD_INFO_H */
diff --git a/include/asm-sparc/thread_info_64.h b/include/asm-sparc/thread_info_64.h
new file mode 100644
index 00000000000..c6d2e6c7f84
--- /dev/null
+++ b/include/asm-sparc/thread_info_64.h
@@ -0,0 +1,277 @@
+/* thread_info.h: sparc64 low-level thread information
+ *
+ * Copyright (C) 2002  David S. Miller (davem@redhat.com)
+ */
+
+#ifndef _ASM_THREAD_INFO_H
+#define _ASM_THREAD_INFO_H
+
+#ifdef __KERNEL__
+
+#define NSWINS		7
+
+#define TI_FLAG_BYTE_FAULT_CODE		0
+#define TI_FLAG_FAULT_CODE_SHIFT	56
+#define TI_FLAG_BYTE_WSTATE		1
+#define TI_FLAG_WSTATE_SHIFT		48
+#define TI_FLAG_BYTE_CWP		2
+#define TI_FLAG_CWP_SHIFT		40
+#define TI_FLAG_BYTE_CURRENT_DS		3
+#define TI_FLAG_CURRENT_DS_SHIFT	32
+#define TI_FLAG_BYTE_FPDEPTH		4
+#define TI_FLAG_FPDEPTH_SHIFT		24
+#define TI_FLAG_BYTE_WSAVED		5
+#define TI_FLAG_WSAVED_SHIFT		16
+
+#include <asm/page.h>
+
+#ifndef __ASSEMBLY__
+
+#include <asm/ptrace.h>
+#include <asm/types.h>
+
+struct task_struct;
+struct exec_domain;
+
+struct thread_info {
+	/* D$ line 1 */
+	struct task_struct	*task;
+	unsigned long		flags;
+	__u8			fpsaved[7];
+	__u8			status;
+	unsigned long		ksp;
+
+	/* D$ line 2 */
+	unsigned long		fault_address;
+	struct pt_regs		*kregs;
+	struct exec_domain	*exec_domain;
+	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
+	__u8			new_child;
+	__u8			syscall_noerror;
+	__u16			cpu;
+
+	unsigned long		*utraps;
+
+	struct reg_window 	reg_window[NSWINS];
+	unsigned long 		rwbuf_stkptrs[NSWINS];
+
+	unsigned long		gsr[7];
+	unsigned long		xfsr[7];
+
+	__u64			__user *user_cntd0;
+	__u64			__user *user_cntd1;
+	__u64			kernel_cntd0, kernel_cntd1;
+	__u64			pcr_reg;
+
+	struct restart_block	restart_block;
+
+	struct pt_regs		*kern_una_regs;
+	unsigned int		kern_una_insn;
+
+	unsigned long		fpregs[0] __attribute__ ((aligned(64)));
+};
+
+#endif /* !(__ASSEMBLY__) */
+
+/* offsets into the thread_info struct for assembly code access */
+#define TI_TASK		0x00000000
+#define TI_FLAGS	0x00000008
+#define TI_FAULT_CODE	(TI_FLAGS + TI_FLAG_BYTE_FAULT_CODE)
+#define TI_WSTATE	(TI_FLAGS + TI_FLAG_BYTE_WSTATE)
+#define TI_CWP		(TI_FLAGS + TI_FLAG_BYTE_CWP)
+#define TI_CURRENT_DS	(TI_FLAGS + TI_FLAG_BYTE_CURRENT_DS)
+#define TI_FPDEPTH	(TI_FLAGS + TI_FLAG_BYTE_FPDEPTH)
+#define TI_WSAVED	(TI_FLAGS + TI_FLAG_BYTE_WSAVED)
+#define TI_FPSAVED	0x00000010
+#define TI_KSP		0x00000018
+#define TI_FAULT_ADDR	0x00000020
+#define TI_KREGS	0x00000028
+#define TI_EXEC_DOMAIN	0x00000030
+#define TI_PRE_COUNT	0x00000038
+#define TI_NEW_CHILD	0x0000003c
+#define TI_SYS_NOERROR	0x0000003d
+#define TI_CPU		0x0000003e
+#define TI_UTRAPS	0x00000040
+#define TI_REG_WINDOW	0x00000048
+#define TI_RWIN_SPTRS	0x000003c8
+#define TI_GSR		0x00000400
+#define TI_XFSR		0x00000438
+#define TI_USER_CNTD0	0x00000470
+#define TI_USER_CNTD1	0x00000478
+#define TI_KERN_CNTD0	0x00000480
+#define TI_KERN_CNTD1	0x00000488
+#define TI_PCR		0x00000490
+#define TI_RESTART_BLOCK 0x00000498
+#define TI_KUNA_REGS	0x000004c0
+#define TI_KUNA_INSN	0x000004c8
+#define TI_FPREGS	0x00000500
+
+/* We embed this in the uppermost byte of thread_info->flags */
+#define FAULT_CODE_WRITE	0x01	/* Write access, implies D-TLB	   */
+#define FAULT_CODE_DTLB		0x02	/* Miss happened in D-TLB	   */
+#define FAULT_CODE_ITLB		0x04	/* Miss happened in I-TLB	   */
+#define FAULT_CODE_WINFIXUP	0x08	/* Miss happened during spill/fill */
+#define FAULT_CODE_BLKCOMMIT	0x10	/* Use blk-commit ASI in copy_page */
+
+#if PAGE_SHIFT == 13
+#define THREAD_SIZE (2*PAGE_SIZE)
+#define THREAD_SHIFT (PAGE_SHIFT + 1)
+#else /* PAGE_SHIFT == 13 */
+#define THREAD_SIZE PAGE_SIZE
+#define THREAD_SHIFT PAGE_SHIFT
+#endif /* PAGE_SHIFT == 13 */
+
+#define PREEMPT_ACTIVE		0x4000000
+
+/*
+ * macros/functions for gaining access to the thread information structure
+ *
+ * preempt_count needs to be 1 initially, until the scheduler is functional.
+ */
+#ifndef __ASSEMBLY__
+
+#define INIT_THREAD_INFO(tsk)				\
+{							\
+	.task		=	&tsk,			\
+	.flags		= ((unsigned long)ASI_P) << TI_FLAG_CURRENT_DS_SHIFT,	\
+	.exec_domain	=	&default_exec_domain,	\
+	.preempt_count	=	1,			\
+	.restart_block	= {				\
+		.fn	=	do_no_restart_syscall,	\
+	},						\
+}
+
+#define init_thread_info	(init_thread_union.thread_info)
+#define init_stack		(init_thread_union.stack)
+
+/* how to get the thread information struct from C */
+register struct thread_info *current_thread_info_reg asm("g6");
+#define current_thread_info()	(current_thread_info_reg)
+
+/* thread information allocation */
+#if PAGE_SHIFT == 13
+#define __THREAD_INFO_ORDER	1
+#else /* PAGE_SHIFT == 13 */
+#define __THREAD_INFO_ORDER	0
+#endif /* PAGE_SHIFT == 13 */
+
+#ifdef CONFIG_DEBUG_STACK_USAGE
+#define alloc_thread_info(tsk)					\
+({								\
+	struct thread_info *ret;				\
+								\
+	ret = (struct thread_info *)				\
+	  __get_free_pages(GFP_KERNEL, __THREAD_INFO_ORDER);	\
+	if (ret)						\
+		memset(ret, 0, PAGE_SIZE<<__THREAD_INFO_ORDER);	\
+	ret;							\
+})
+#else
+#define alloc_thread_info(tsk) \
+	((struct thread_info *)__get_free_pages(GFP_KERNEL, __THREAD_INFO_ORDER))
+#endif
+
+#define free_thread_info(ti) \
+	free_pages((unsigned long)(ti),__THREAD_INFO_ORDER)
+
+#define __thread_flag_byte_ptr(ti)	\
+	((unsigned char *)(&((ti)->flags)))
+#define __cur_thread_flag_byte_ptr	__thread_flag_byte_ptr(current_thread_info())
+
+#define get_thread_fault_code()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FAULT_CODE])
+#define set_thread_fault_code(val)	(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FAULT_CODE] = (val))
+#define get_thread_wstate()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSTATE])
+#define set_thread_wstate(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSTATE] = (val))
+#define get_thread_cwp()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CWP])
+#define set_thread_cwp(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CWP] = (val))
+#define get_thread_current_ds()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CURRENT_DS])
+#define set_thread_current_ds(val)	(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CURRENT_DS] = (val))
+#define get_thread_fpdepth()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FPDEPTH])
+#define set_thread_fpdepth(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FPDEPTH] = (val))
+#define get_thread_wsaved()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSAVED])
+#define set_thread_wsaved(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSAVED] = (val))
+
+#endif /* !(__ASSEMBLY__) */
+
+/*
+ * Thread information flags, only 16 bits are available as we encode
+ * other values into the upper 6 bytes.
+ *
+ * On trap return we need to test several values:
+ *
+ * user:	need_resched, notify_resume, sigpending, wsaved, perfctr
+ * kernel:	fpdepth
+ *
+ * So to check for work in the kernel case we simply load the fpdepth
+ * byte out of the flags and test it.  For the user case we encode the
+ * lower 3 bytes of flags as follows:
+ *	----------------------------------------
+ *	| wsaved | flags byte 1 | flags byte 2 |
+ *	----------------------------------------
+ * This optimizes the user test into:
+ *	ldx		[%g6 + TI_FLAGS], REG1
+ *	sethi		%hi(_TIF_USER_WORK_MASK), REG2
+ *	or		REG2, %lo(_TIF_USER_WORK_MASK), REG2
+ *	andcc		REG1, REG2, %g0
+ *	be,pt		no_work_to_do
+ *	 nop
+ */
+#define TIF_SYSCALL_TRACE	0	/* syscall trace active */
+/* flags bit 1 is available */
+#define TIF_SIGPENDING		2	/* signal pending */
+#define TIF_NEED_RESCHED	3	/* rescheduling necessary */
+#define TIF_PERFCTR		4	/* performance counters active */
+#define TIF_UNALIGNED		5	/* allowed to do unaligned accesses */
+/* flag bit 6 is available */
+#define TIF_32BIT		7	/* 32-bit binary */
+/* flag bit 8 is available */
+#define TIF_SECCOMP		9	/* secure computing */
+#define TIF_SYSCALL_AUDIT	10	/* syscall auditing active */
+/* flag bit 11 is available */
+/* NOTE: Thread flags >= 12 should be ones we have no interest
+ *       in using in assembly, else we can't use the mask as
+ *       an immediate value in instructions such as andcc.
+ */
+#define TIF_ABI_PENDING		12
+#define TIF_MEMDIE		13
+#define TIF_POLLING_NRFLAG	14
+
+#define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
+#define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
+#define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
+#define _TIF_PERFCTR		(1<<TIF_PERFCTR)
+#define _TIF_UNALIGNED		(1<<TIF_UNALIGNED)
+#define _TIF_32BIT		(1<<TIF_32BIT)
+#define _TIF_SECCOMP		(1<<TIF_SECCOMP)
+#define _TIF_SYSCALL_AUDIT	(1<<TIF_SYSCALL_AUDIT)
+#define _TIF_ABI_PENDING	(1<<TIF_ABI_PENDING)
+#define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
+
+#define _TIF_USER_WORK_MASK	((0xff << TI_FLAG_WSAVED_SHIFT) | \
+				 (_TIF_SIGPENDING | \
+				  _TIF_NEED_RESCHED | _TIF_PERFCTR))
+
+/*
+ * Thread-synchronous status.
+ *
+ * This is different from the flags in that nobody else
+ * ever touches our thread-synchronous status, so we don't
+ * have to worry about atomic accesses.
+ *
+ * Note that there are only 8 bits available.
+ */
+#define TS_RESTORE_SIGMASK	0x0001	/* restore signal mask in do_signal() */
+
+#ifndef __ASSEMBLY__
+#define HAVE_SET_RESTORE_SIGMASK	1
+static inline void set_restore_sigmask(void)
+{
+	struct thread_info *ti = current_thread_info();
+	ti->status |= TS_RESTORE_SIGMASK;
+	set_bit(TIF_SIGPENDING, &ti->flags);
+}
+#endif	/* !__ASSEMBLY__ */
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_THREAD_INFO_H */
diff --git a/include/asm-sparc/timer.h b/include/asm-sparc/timer.h
index d909565f941..475baa05a96 100644
--- a/include/asm-sparc/timer.h
+++ b/include/asm-sparc/timer.h
@@ -1,109 +1,8 @@
-/*
- * timer.h:  Definitions for the timer chips on the Sparc.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- */
-
-
-#ifndef _SPARC_TIMER_H
-#define _SPARC_TIMER_H
-
-#include <asm/system.h>  /* For SUN4M_NCPUS */
-#include <asm/sun4paddr.h>
-#include <asm/btfixup.h>
-
-/* Timer structures. The interrupt timer has two properties which
- * are the counter (which is handled in do_timer in sched.c) and the limit.
- * This limit is where the timer's counter 'wraps' around. Oddly enough,
- * the sun4c timer when it hits the limit wraps back to 1 and not zero
- * thus when calculating the value at which it will fire a microsecond you
- * must adjust by one.  Thanks SUN for designing such great hardware ;(
- */
-
-/* Note that I am only going to use the timer that interrupts at
- * Sparc IRQ 10.  There is another one available that can fire at
- * IRQ 14. Currently it is left untouched, we keep the PROM's limit
- * register value and let the prom take these interrupts.  This allows
- * L1-A to work.
- */
-
-struct sun4c_timer_info {
-  __volatile__ unsigned int cur_count10;
-  __volatile__ unsigned int timer_limit10;
-  __volatile__ unsigned int cur_count14;
-  __volatile__ unsigned int timer_limit14;
-};
-
-#define SUN4C_TIMER_PHYSADDR   0xf3000000
-#ifdef CONFIG_SUN4
-#define SUN_TIMER_PHYSADDR SUN4_300_TIMER_PHYSADDR
+#ifndef ___ASM_SPARC_TIMER_H
+#define ___ASM_SPARC_TIMER_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/timer_64.h>
 #else
-#define SUN_TIMER_PHYSADDR SUN4C_TIMER_PHYSADDR
+#include <asm-sparc/timer_32.h>
+#endif
 #endif
-
-/* A sun4m has two blocks of registers which are probably of the same
- * structure. LSI Logic's L64851 is told to _decrement_ from the limit
- * value. Aurora behaves similarly but its limit value is compacted in
- * other fashion (it's wider). Documented fields are defined here.
- */
-
-/* As with the interrupt register, we have two classes of timer registers
- * which are per-cpu and master.  Per-cpu timers only hit that cpu and are
- * only level 14 ticks, master timer hits all cpus and is level 10.
- */
-
-#define SUN4M_PRM_CNT_L       0x80000000
-#define SUN4M_PRM_CNT_LVALUE  0x7FFFFC00
-
-struct sun4m_timer_percpu_info {
-  __volatile__ unsigned int l14_timer_limit;    /* Initial value is 0x009c4000 */
-  __volatile__ unsigned int l14_cur_count;
-
-  /* This register appears to be write only and/or inaccessible
-   * on Uni-Processor sun4m machines.
-   */
-  __volatile__ unsigned int l14_limit_noclear;  /* Data access error is here */
-
-  __volatile__ unsigned int cntrl;            /* =1 after POST on Aurora */
-  __volatile__ unsigned char space[PAGE_SIZE - 16];
-};
-
-struct sun4m_timer_regs {
-	struct sun4m_timer_percpu_info cpu_timers[SUN4M_NCPUS];
-	volatile unsigned int l10_timer_limit;
-	volatile unsigned int l10_cur_count;
-
-	/* Again, this appears to be write only and/or inaccessible
-	 * on uni-processor sun4m machines.
-	 */
-	volatile unsigned int l10_limit_noclear;
-
-	/* This register too, it must be magic. */
-	volatile unsigned int foobar;
-
-	volatile unsigned int cfg;     /* equals zero at boot time... */
-};
-
-extern struct sun4m_timer_regs *sun4m_timers;
-
-#define SUN4D_PRM_CNT_L       0x80000000
-#define SUN4D_PRM_CNT_LVALUE  0x7FFFFC00
-
-struct sun4d_timer_regs {
-	volatile unsigned int l10_timer_limit;
-	volatile unsigned int l10_cur_countx;
-	volatile unsigned int l10_limit_noclear;
-	volatile unsigned int ctrl;
-	volatile unsigned int l10_cur_count;
-};
-
-extern struct sun4d_timer_regs *sun4d_timers;
-
-extern __volatile__ unsigned int *master_l10_counter;
-extern __volatile__ unsigned int *master_l10_limit;
-
-/* FIXME: Make do_[gs]ettimeofday btfixup calls */
-BTFIXUPDEF_CALL(int, bus_do_settimeofday, struct timespec *tv)
-#define bus_do_settimeofday(tv) BTFIXUP_CALL(bus_do_settimeofday)(tv)
-
-#endif /* !(_SPARC_TIMER_H) */
diff --git a/include/asm-sparc/timer_32.h b/include/asm-sparc/timer_32.h
new file mode 100644
index 00000000000..361e53898dd
--- /dev/null
+++ b/include/asm-sparc/timer_32.h
@@ -0,0 +1,107 @@
+/*
+ * timer.h:  Definitions for the timer chips on the Sparc.
+ *
+ * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+
+#ifndef _SPARC_TIMER_H
+#define _SPARC_TIMER_H
+
+#include <asm/system.h>  /* For SUN4M_NCPUS */
+#include <asm/sun4paddr.h>
+#include <asm/btfixup.h>
+
+/* Timer structures. The interrupt timer has two properties which
+ * are the counter (which is handled in do_timer in sched.c) and the limit.
+ * This limit is where the timer's counter 'wraps' around. Oddly enough,
+ * the sun4c timer when it hits the limit wraps back to 1 and not zero
+ * thus when calculating the value at which it will fire a microsecond you
+ * must adjust by one.  Thanks SUN for designing such great hardware ;(
+ */
+
+/* Note that I am only going to use the timer that interrupts at
+ * Sparc IRQ 10.  There is another one available that can fire at
+ * IRQ 14. Currently it is left untouched, we keep the PROM's limit
+ * register value and let the prom take these interrupts.  This allows
+ * L1-A to work.
+ */
+
+struct sun4c_timer_info {
+  __volatile__ unsigned int cur_count10;
+  __volatile__ unsigned int timer_limit10;
+  __volatile__ unsigned int cur_count14;
+  __volatile__ unsigned int timer_limit14;
+};
+
+#define SUN4C_TIMER_PHYSADDR   0xf3000000
+#ifdef CONFIG_SUN4
+#define SUN_TIMER_PHYSADDR SUN4_300_TIMER_PHYSADDR
+#else
+#define SUN_TIMER_PHYSADDR SUN4C_TIMER_PHYSADDR
+#endif
+
+/* A sun4m has two blocks of registers which are probably of the same
+ * structure. LSI Logic's L64851 is told to _decrement_ from the limit
+ * value. Aurora behaves similarly but its limit value is compacted in
+ * other fashion (it's wider). Documented fields are defined here.
+ */
+
+/* As with the interrupt register, we have two classes of timer registers
+ * which are per-cpu and master.  Per-cpu timers only hit that cpu and are
+ * only level 14 ticks, master timer hits all cpus and is level 10.
+ */
+
+#define SUN4M_PRM_CNT_L       0x80000000
+#define SUN4M_PRM_CNT_LVALUE  0x7FFFFC00
+
+struct sun4m_timer_percpu_info {
+  __volatile__ unsigned int l14_timer_limit;    /* Initial value is 0x009c4000 */
+  __volatile__ unsigned int l14_cur_count;
+
+  /* This register appears to be write only and/or inaccessible
+   * on Uni-Processor sun4m machines.
+   */
+  __volatile__ unsigned int l14_limit_noclear;  /* Data access error is here */
+
+  __volatile__ unsigned int cntrl;            /* =1 after POST on Aurora */
+  __volatile__ unsigned char space[PAGE_SIZE - 16];
+};
+
+struct sun4m_timer_regs {
+	struct sun4m_timer_percpu_info cpu_timers[SUN4M_NCPUS];
+	volatile unsigned int l10_timer_limit;
+	volatile unsigned int l10_cur_count;
+
+	/* Again, this appears to be write only and/or inaccessible
+	 * on uni-processor sun4m machines.
+	 */
+	volatile unsigned int l10_limit_noclear;
+
+	/* This register too, it must be magic. */
+	volatile unsigned int foobar;
+
+	volatile unsigned int cfg;     /* equals zero at boot time... */
+};
+
+#define SUN4D_PRM_CNT_L       0x80000000
+#define SUN4D_PRM_CNT_LVALUE  0x7FFFFC00
+
+struct sun4d_timer_regs {
+	volatile unsigned int l10_timer_limit;
+	volatile unsigned int l10_cur_countx;
+	volatile unsigned int l10_limit_noclear;
+	volatile unsigned int ctrl;
+	volatile unsigned int l10_cur_count;
+};
+
+extern struct sun4d_timer_regs *sun4d_timers;
+
+extern __volatile__ unsigned int *master_l10_counter;
+extern __volatile__ unsigned int *master_l10_limit;
+
+/* FIXME: Make do_[gs]ettimeofday btfixup calls */
+BTFIXUPDEF_CALL(int, bus_do_settimeofday, struct timespec *tv)
+#define bus_do_settimeofday(tv) BTFIXUP_CALL(bus_do_settimeofday)(tv)
+
+#endif /* !(_SPARC_TIMER_H) */
diff --git a/include/asm-sparc/timer_64.h b/include/asm-sparc/timer_64.h
new file mode 100644
index 00000000000..5b779fd1f78
--- /dev/null
+++ b/include/asm-sparc/timer_64.h
@@ -0,0 +1,30 @@
+/* timer.h: System timer definitions for sun5.
+ *
+ * Copyright (C) 1997, 2008 David S. Miller (davem@davemloft.net)
+ */
+
+#ifndef _SPARC64_TIMER_H
+#define _SPARC64_TIMER_H
+
+#include <linux/types.h>
+#include <linux/init.h>
+
+struct sparc64_tick_ops {
+	unsigned long (*get_tick)(void);
+	int (*add_compare)(unsigned long);
+	unsigned long softint_mask;
+	void (*disable_irq)(void);
+
+	void (*init_tick)(void);
+	unsigned long (*add_tick)(unsigned long);
+
+	char *name;
+};
+
+extern struct sparc64_tick_ops *tick_ops;
+
+extern unsigned long sparc64_get_clock_tick(unsigned int cpu);
+extern void __devinit setup_sparc64_timer(void);
+extern void __init time_init(void);
+
+#endif /* _SPARC64_TIMER_H */
diff --git a/include/asm-sparc/timex.h b/include/asm-sparc/timex.h
index 71b45c90cca..01d9f199d45 100644
--- a/include/asm-sparc/timex.h
+++ b/include/asm-sparc/timex.h
@@ -1,15 +1,8 @@
-/*
- * linux/include/asm-sparc/timex.h
- *
- * sparc architecture timex specifications
- */
-#ifndef _ASMsparc_TIMEX_H
-#define _ASMsparc_TIMEX_H
-
-#define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
-
-/* XXX Maybe do something better at some point... -DaveM */
-typedef unsigned long cycles_t;
-#define get_cycles()	(0)
-
+#ifndef ___ASM_SPARC_TIMEX_H
+#define ___ASM_SPARC_TIMEX_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/timex_64.h>
+#else
+#include <asm-sparc/timex_32.h>
+#endif
 #endif
diff --git a/include/asm-sparc/timex_32.h b/include/asm-sparc/timex_32.h
new file mode 100644
index 00000000000..71b45c90cca
--- /dev/null
+++ b/include/asm-sparc/timex_32.h
@@ -0,0 +1,15 @@
+/*
+ * linux/include/asm-sparc/timex.h
+ *
+ * sparc architecture timex specifications
+ */
+#ifndef _ASMsparc_TIMEX_H
+#define _ASMsparc_TIMEX_H
+
+#define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
+
+/* XXX Maybe do something better at some point... -DaveM */
+typedef unsigned long cycles_t;
+#define get_cycles()	(0)
+
+#endif
diff --git a/include/asm-sparc/timex_64.h b/include/asm-sparc/timex_64.h
new file mode 100644
index 00000000000..c622535c456
--- /dev/null
+++ b/include/asm-sparc/timex_64.h
@@ -0,0 +1,19 @@
+/*
+ * linux/include/asm-sparc64/timex.h
+ *
+ * sparc64 architecture timex specifications
+ */
+#ifndef _ASMsparc64_TIMEX_H
+#define _ASMsparc64_TIMEX_H
+
+#include <asm/timer.h>
+
+#define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
+
+/* Getting on the cycle counter on sparc64. */
+typedef unsigned long cycles_t;
+#define get_cycles()	tick_ops->get_tick()
+
+#define ARCH_HAS_READ_CURRENT_TIMER
+
+#endif
diff --git a/include/asm-sparc/tlb.h b/include/asm-sparc/tlb.h
index 6d02d1ce53f..a821057327c 100644
--- a/include/asm-sparc/tlb.h
+++ b/include/asm-sparc/tlb.h
@@ -1,24 +1,8 @@
-#ifndef _SPARC_TLB_H
-#define _SPARC_TLB_H
-
-#define tlb_start_vma(tlb, vma) \
-do {								\
-	flush_cache_range(vma, vma->vm_start, vma->vm_end);	\
-} while (0)
-
-#define tlb_end_vma(tlb, vma) \
-do {								\
-	flush_tlb_range(vma, vma->vm_start, vma->vm_end);	\
-} while (0)
-
-#define __tlb_remove_tlb_entry(tlb, pte, address) \
-	do { } while (0)
-
-#define tlb_flush(tlb) \
-do {								\
-	flush_tlb_mm((tlb)->mm);				\
-} while (0)
-
-#include <asm-generic/tlb.h>
-
-#endif /* _SPARC_TLB_H */
+#ifndef ___ASM_SPARC_TLB_H
+#define ___ASM_SPARC_TLB_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/tlb_64.h>
+#else
+#include <asm-sparc/tlb_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/tlb_32.h b/include/asm-sparc/tlb_32.h
new file mode 100644
index 00000000000..6d02d1ce53f
--- /dev/null
+++ b/include/asm-sparc/tlb_32.h
@@ -0,0 +1,24 @@
+#ifndef _SPARC_TLB_H
+#define _SPARC_TLB_H
+
+#define tlb_start_vma(tlb, vma) \
+do {								\
+	flush_cache_range(vma, vma->vm_start, vma->vm_end);	\
+} while (0)
+
+#define tlb_end_vma(tlb, vma) \
+do {								\
+	flush_tlb_range(vma, vma->vm_start, vma->vm_end);	\
+} while (0)
+
+#define __tlb_remove_tlb_entry(tlb, pte, address) \
+	do { } while (0)
+
+#define tlb_flush(tlb) \
+do {								\
+	flush_tlb_mm((tlb)->mm);				\
+} while (0)
+
+#include <asm-generic/tlb.h>
+
+#endif /* _SPARC_TLB_H */
diff --git a/include/asm-sparc/tlb_64.h b/include/asm-sparc/tlb_64.h
new file mode 100644
index 00000000000..ec81cdedef2
--- /dev/null
+++ b/include/asm-sparc/tlb_64.h
@@ -0,0 +1,111 @@
+#ifndef _SPARC64_TLB_H
+#define _SPARC64_TLB_H
+
+#include <linux/swap.h>
+#include <linux/pagemap.h>
+#include <asm/pgalloc.h>
+#include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
+
+#define TLB_BATCH_NR	192
+
+/*
+ * For UP we don't need to worry about TLB flush
+ * and page free order so much..
+ */
+#ifdef CONFIG_SMP
+  #define FREE_PTE_NR	506
+  #define tlb_fast_mode(bp) ((bp)->pages_nr == ~0U)
+#else
+  #define FREE_PTE_NR	1
+  #define tlb_fast_mode(bp) 1
+#endif
+
+struct mmu_gather {
+	struct mm_struct *mm;
+	unsigned int pages_nr;
+	unsigned int need_flush;
+	unsigned int fullmm;
+	unsigned int tlb_nr;
+	unsigned long vaddrs[TLB_BATCH_NR];
+	struct page *pages[FREE_PTE_NR];
+};
+
+DECLARE_PER_CPU(struct mmu_gather, mmu_gathers);
+
+#ifdef CONFIG_SMP
+extern void smp_flush_tlb_pending(struct mm_struct *,
+				  unsigned long, unsigned long *);
+#endif
+
+extern void __flush_tlb_pending(unsigned long, unsigned long, unsigned long *);
+extern void flush_tlb_pending(void);
+
+static inline struct mmu_gather *tlb_gather_mmu(struct mm_struct *mm, unsigned int full_mm_flush)
+{
+	struct mmu_gather *mp = &get_cpu_var(mmu_gathers);
+
+	BUG_ON(mp->tlb_nr);
+
+	mp->mm = mm;
+	mp->pages_nr = num_online_cpus() > 1 ? 0U : ~0U;
+	mp->fullmm = full_mm_flush;
+
+	return mp;
+}
+
+
+static inline void tlb_flush_mmu(struct mmu_gather *mp)
+{
+	if (mp->need_flush) {
+		free_pages_and_swap_cache(mp->pages, mp->pages_nr);
+		mp->pages_nr = 0;
+		mp->need_flush = 0;
+	}
+
+}
+
+#ifdef CONFIG_SMP
+extern void smp_flush_tlb_mm(struct mm_struct *mm);
+#define do_flush_tlb_mm(mm) smp_flush_tlb_mm(mm)
+#else
+#define do_flush_tlb_mm(mm) __flush_tlb_mm(CTX_HWBITS(mm->context), SECONDARY_CONTEXT)
+#endif
+
+static inline void tlb_finish_mmu(struct mmu_gather *mp, unsigned long start, unsigned long end)
+{
+	tlb_flush_mmu(mp);
+
+	if (mp->fullmm)
+		mp->fullmm = 0;
+	else
+		flush_tlb_pending();
+
+	/* keep the page table cache within bounds */
+	check_pgt_cache();
+
+	put_cpu_var(mmu_gathers);
+}
+
+static inline void tlb_remove_page(struct mmu_gather *mp, struct page *page)
+{
+	if (tlb_fast_mode(mp)) {
+		free_page_and_swap_cache(page);
+		return;
+	}
+	mp->need_flush = 1;
+	mp->pages[mp->pages_nr++] = page;
+	if (mp->pages_nr >= FREE_PTE_NR)
+		tlb_flush_mmu(mp);
+}
+
+#define tlb_remove_tlb_entry(mp,ptep,addr) do { } while (0)
+#define pte_free_tlb(mp, ptepage) pte_free((mp)->mm, ptepage)
+#define pmd_free_tlb(mp, pmdp) pmd_free((mp)->mm, pmdp)
+#define pud_free_tlb(tlb,pudp) __pud_free_tlb(tlb,pudp)
+
+#define tlb_migrate_finish(mm)	do { } while (0)
+#define tlb_start_vma(tlb, vma) do { } while (0)
+#define tlb_end_vma(tlb, vma)	do { } while (0)
+
+#endif /* _SPARC64_TLB_H */
diff --git a/include/asm-sparc/tlbflush.h b/include/asm-sparc/tlbflush.h
index b957e29d2ae..6e6bc12227b 100644
--- a/include/asm-sparc/tlbflush.h
+++ b/include/asm-sparc/tlbflush.h
@@ -1,60 +1,8 @@
-#ifndef _SPARC_TLBFLUSH_H
-#define _SPARC_TLBFLUSH_H
-
-#include <linux/mm.h>
-// #include <asm/processor.h>
-
-/*
- * TLB flushing:
- *
- *  - flush_tlb() flushes the current mm struct TLBs	XXX Exists?
- *  - flush_tlb_all() flushes all processes TLBs 
- *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
- *  - flush_tlb_page(vma, vmaddr) flushes one page
- *  - flush_tlb_range(vma, start, end) flushes a range of pages
- *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
- */
-
-#ifdef CONFIG_SMP
-
-BTFIXUPDEF_CALL(void, local_flush_tlb_all, void)
-BTFIXUPDEF_CALL(void, local_flush_tlb_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, local_flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_tlb_page, struct vm_area_struct *, unsigned long)
-
-#define local_flush_tlb_all() BTFIXUP_CALL(local_flush_tlb_all)()
-#define local_flush_tlb_mm(mm) BTFIXUP_CALL(local_flush_tlb_mm)(mm)
-#define local_flush_tlb_range(vma,start,end) BTFIXUP_CALL(local_flush_tlb_range)(vma,start,end)
-#define local_flush_tlb_page(vma,addr) BTFIXUP_CALL(local_flush_tlb_page)(vma,addr)
-
-extern void smp_flush_tlb_all(void);
-extern void smp_flush_tlb_mm(struct mm_struct *mm);
-extern void smp_flush_tlb_range(struct vm_area_struct *vma,
-				  unsigned long start,
-				  unsigned long end);
-extern void smp_flush_tlb_page(struct vm_area_struct *mm, unsigned long page);
-
-#endif /* CONFIG_SMP */
-
-BTFIXUPDEF_CALL(void, flush_tlb_all, void)
-BTFIXUPDEF_CALL(void, flush_tlb_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, flush_tlb_page, struct vm_area_struct *, unsigned long)
-
-#define flush_tlb_all() BTFIXUP_CALL(flush_tlb_all)()
-#define flush_tlb_mm(mm) BTFIXUP_CALL(flush_tlb_mm)(mm)
-#define flush_tlb_range(vma,start,end) BTFIXUP_CALL(flush_tlb_range)(vma,start,end)
-#define flush_tlb_page(vma,addr) BTFIXUP_CALL(flush_tlb_page)(vma,addr)
-
-// #define flush_tlb() flush_tlb_mm(current->active_mm)	/* XXX Sure? */
-
-/*
- * This is a kludge, until I know better. --zaitcev XXX
- */
-static inline void flush_tlb_kernel_range(unsigned long start,
-					  unsigned long end)
-{
-	flush_tlb_all();
-}
-
-#endif /* _SPARC_TLBFLUSH_H */
+#ifndef ___ASM_SPARC_TLBFLUSH_H
+#define ___ASM_SPARC_TLBFLUSH_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/tlbflush_64.h>
+#else
+#include <asm-sparc/tlbflush_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/tlbflush_32.h b/include/asm-sparc/tlbflush_32.h
new file mode 100644
index 00000000000..fe0a71abc9b
--- /dev/null
+++ b/include/asm-sparc/tlbflush_32.h
@@ -0,0 +1,60 @@
+#ifndef _SPARC_TLBFLUSH_H
+#define _SPARC_TLBFLUSH_H
+
+#include <linux/mm.h>
+// #include <asm/processor.h>
+
+/*
+ * TLB flushing:
+ *
+ *  - flush_tlb() flushes the current mm struct TLBs	XXX Exists?
+ *  - flush_tlb_all() flushes all processes TLBs
+ *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
+ *  - flush_tlb_page(vma, vmaddr) flushes one page
+ *  - flush_tlb_range(vma, start, end) flushes a range of pages
+ *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ */
+
+#ifdef CONFIG_SMP
+
+BTFIXUPDEF_CALL(void, local_flush_tlb_all, void)
+BTFIXUPDEF_CALL(void, local_flush_tlb_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, local_flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_tlb_page, struct vm_area_struct *, unsigned long)
+
+#define local_flush_tlb_all() BTFIXUP_CALL(local_flush_tlb_all)()
+#define local_flush_tlb_mm(mm) BTFIXUP_CALL(local_flush_tlb_mm)(mm)
+#define local_flush_tlb_range(vma,start,end) BTFIXUP_CALL(local_flush_tlb_range)(vma,start,end)
+#define local_flush_tlb_page(vma,addr) BTFIXUP_CALL(local_flush_tlb_page)(vma,addr)
+
+extern void smp_flush_tlb_all(void);
+extern void smp_flush_tlb_mm(struct mm_struct *mm);
+extern void smp_flush_tlb_range(struct vm_area_struct *vma,
+				  unsigned long start,
+				  unsigned long end);
+extern void smp_flush_tlb_page(struct vm_area_struct *mm, unsigned long page);
+
+#endif /* CONFIG_SMP */
+
+BTFIXUPDEF_CALL(void, flush_tlb_all, void)
+BTFIXUPDEF_CALL(void, flush_tlb_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, flush_tlb_page, struct vm_area_struct *, unsigned long)
+
+#define flush_tlb_all() BTFIXUP_CALL(flush_tlb_all)()
+#define flush_tlb_mm(mm) BTFIXUP_CALL(flush_tlb_mm)(mm)
+#define flush_tlb_range(vma,start,end) BTFIXUP_CALL(flush_tlb_range)(vma,start,end)
+#define flush_tlb_page(vma,addr) BTFIXUP_CALL(flush_tlb_page)(vma,addr)
+
+// #define flush_tlb() flush_tlb_mm(current->active_mm)	/* XXX Sure? */
+
+/*
+ * This is a kludge, until I know better. --zaitcev XXX
+ */
+static inline void flush_tlb_kernel_range(unsigned long start,
+					  unsigned long end)
+{
+	flush_tlb_all();
+}
+
+#endif /* _SPARC_TLBFLUSH_H */
diff --git a/include/asm-sparc/tlbflush_64.h b/include/asm-sparc/tlbflush_64.h
new file mode 100644
index 00000000000..fbb675dbe0c
--- /dev/null
+++ b/include/asm-sparc/tlbflush_64.h
@@ -0,0 +1,44 @@
+#ifndef _SPARC64_TLBFLUSH_H
+#define _SPARC64_TLBFLUSH_H
+
+#include <linux/mm.h>
+#include <asm/mmu_context.h>
+
+/* TSB flush operations. */
+struct mmu_gather;
+extern void flush_tsb_kernel_range(unsigned long start, unsigned long end);
+extern void flush_tsb_user(struct mmu_gather *mp);
+
+/* TLB flush operations. */
+
+extern void flush_tlb_pending(void);
+
+#define flush_tlb_range(vma,start,end)	\
+	do { (void)(start); flush_tlb_pending(); } while (0)
+#define flush_tlb_page(vma,addr)	flush_tlb_pending()
+#define flush_tlb_mm(mm)		flush_tlb_pending()
+
+/* Local cpu only.  */
+extern void __flush_tlb_all(void);
+
+extern void __flush_tlb_kernel_range(unsigned long start, unsigned long end);
+
+#ifndef CONFIG_SMP
+
+#define flush_tlb_kernel_range(start,end) \
+do {	flush_tsb_kernel_range(start,end); \
+	__flush_tlb_kernel_range(start,end); \
+} while (0)
+
+#else /* CONFIG_SMP */
+
+extern void smp_flush_tlb_kernel_range(unsigned long start, unsigned long end);
+
+#define flush_tlb_kernel_range(start, end) \
+do {	flush_tsb_kernel_range(start,end); \
+	smp_flush_tlb_kernel_range(start, end); \
+} while (0)
+
+#endif /* ! CONFIG_SMP */
+
+#endif /* _SPARC64_TLBFLUSH_H */
diff --git a/include/asm-sparc/topology.h b/include/asm-sparc/topology.h
index ee5ac9c9da2..ed13630f32e 100644
--- a/include/asm-sparc/topology.h
+++ b/include/asm-sparc/topology.h
@@ -1,6 +1,8 @@
-#ifndef _ASM_SPARC_TOPOLOGY_H
-#define _ASM_SPARC_TOPOLOGY_H
-
-#include <asm-generic/topology.h>
-
-#endif /* _ASM_SPARC_TOPOLOGY_H */
+#ifndef ___ASM_SPARC_TOPOLOGY_H
+#define ___ASM_SPARC_TOPOLOGY_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/topology_64.h>
+#else
+#include <asm-sparc/topology_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/topology_32.h b/include/asm-sparc/topology_32.h
new file mode 100644
index 00000000000..ee5ac9c9da2
--- /dev/null
+++ b/include/asm-sparc/topology_32.h
@@ -0,0 +1,6 @@
+#ifndef _ASM_SPARC_TOPOLOGY_H
+#define _ASM_SPARC_TOPOLOGY_H
+
+#include <asm-generic/topology.h>
+
+#endif /* _ASM_SPARC_TOPOLOGY_H */
diff --git a/include/asm-sparc/topology_64.h b/include/asm-sparc/topology_64.h
new file mode 100644
index 00000000000..001c04027c8
--- /dev/null
+++ b/include/asm-sparc/topology_64.h
@@ -0,0 +1,86 @@
+#ifndef _ASM_SPARC64_TOPOLOGY_H
+#define _ASM_SPARC64_TOPOLOGY_H
+
+#ifdef CONFIG_NUMA
+
+#include <asm/mmzone.h>
+
+static inline int cpu_to_node(int cpu)
+{
+	return numa_cpu_lookup_table[cpu];
+}
+
+#define parent_node(node)	(node)
+
+static inline cpumask_t node_to_cpumask(int node)
+{
+	return numa_cpumask_lookup_table[node];
+}
+
+/* Returns a pointer to the cpumask of CPUs on Node 'node'. */
+#define node_to_cpumask_ptr(v, node)		\
+		cpumask_t *v = &(numa_cpumask_lookup_table[node])
+
+#define node_to_cpumask_ptr_next(v, node)	\
+			   v = &(numa_cpumask_lookup_table[node])
+
+static inline int node_to_first_cpu(int node)
+{
+	cpumask_t tmp;
+	tmp = node_to_cpumask(node);
+	return first_cpu(tmp);
+}
+
+struct pci_bus;
+#ifdef CONFIG_PCI
+extern int pcibus_to_node(struct pci_bus *pbus);
+#else
+static inline int pcibus_to_node(struct pci_bus *pbus)
+{
+	return -1;
+}
+#endif
+
+#define pcibus_to_cpumask(bus)	\
+	(pcibus_to_node(bus) == -1 ? \
+	 CPU_MASK_ALL : \
+	 node_to_cpumask(pcibus_to_node(bus)))
+
+#define SD_NODE_INIT (struct sched_domain) {		\
+	.min_interval		= 8,			\
+	.max_interval		= 32,			\
+	.busy_factor		= 32,			\
+	.imbalance_pct		= 125,			\
+	.cache_nice_tries	= 2,			\
+	.busy_idx		= 3,			\
+	.idle_idx		= 2,			\
+	.newidle_idx		= 0, 			\
+	.wake_idx		= 1,			\
+	.forkexec_idx		= 1,			\
+	.flags			= SD_LOAD_BALANCE	\
+				| SD_BALANCE_FORK	\
+				| SD_BALANCE_EXEC	\
+				| SD_SERIALIZE		\
+				| SD_WAKE_BALANCE,	\
+	.last_balance		= jiffies,		\
+	.balance_interval	= 1,			\
+}
+
+#else /* CONFIG_NUMA */
+
+#include <asm-generic/topology.h>
+
+#endif /* !(CONFIG_NUMA) */
+
+#ifdef CONFIG_SMP
+#define topology_physical_package_id(cpu)	(cpu_data(cpu).proc_id)
+#define topology_core_id(cpu)			(cpu_data(cpu).core_id)
+#define topology_core_siblings(cpu)		(cpu_core_map[cpu])
+#define topology_thread_siblings(cpu)		(per_cpu(cpu_sibling_map, cpu))
+#define mc_capable()				(sparc64_multi_core)
+#define smt_capable()				(sparc64_multi_core)
+#endif /* CONFIG_SMP */
+
+#define cpu_coregroup_map(cpu)			(cpu_core_map[cpu])
+
+#endif /* _ASM_SPARC64_TOPOLOGY_H */
diff --git a/include/asm-sparc/tsb.h b/include/asm-sparc/tsb.h
new file mode 100644
index 00000000000..76e4299dd9b
--- /dev/null
+++ b/include/asm-sparc/tsb.h
@@ -0,0 +1,283 @@
+#ifndef _SPARC64_TSB_H
+#define _SPARC64_TSB_H
+
+/* The sparc64 TSB is similar to the powerpc hashtables.  It's a
+ * power-of-2 sized table of TAG/PTE pairs.  The cpu precomputes
+ * pointers into this table for 8K and 64K page sizes, and also a
+ * comparison TAG based upon the virtual address and context which
+ * faults.
+ *
+ * TLB miss trap handler software does the actual lookup via something
+ * of the form:
+ *
+ * 	ldxa		[%g0] ASI_{D,I}MMU_TSB_8KB_PTR, %g1
+ * 	ldxa		[%g0] ASI_{D,I}MMU, %g6
+ *	sllx		%g6, 22, %g6
+ *	srlx		%g6, 22, %g6
+ * 	ldda		[%g1] ASI_NUCLEUS_QUAD_LDD, %g4
+ * 	cmp		%g4, %g6
+ * 	bne,pn	%xcc, tsb_miss_{d,i}tlb
+ * 	 mov		FAULT_CODE_{D,I}TLB, %g3
+ * 	stxa		%g5, [%g0] ASI_{D,I}TLB_DATA_IN
+ * 	retry
+ *
+ *
+ * Each 16-byte slot of the TSB is the 8-byte tag and then the 8-byte
+ * PTE.  The TAG is of the same layout as the TLB TAG TARGET mmu
+ * register which is:
+ *
+ * -------------------------------------------------
+ * |  -  |  CONTEXT |  -  |    VADDR bits 63:22    |
+ * -------------------------------------------------
+ *  63 61 60      48 47 42 41                     0
+ *
+ * But actually, since we use per-mm TSB's, we zero out the CONTEXT
+ * field.
+ *
+ * Like the powerpc hashtables we need to use locking in order to
+ * synchronize while we update the entries.  PTE updates need locking
+ * as well.
+ *
+ * We need to carefully choose a lock bits for the TSB entry.  We
+ * choose to use bit 47 in the tag.  Also, since we never map anything
+ * at page zero in context zero, we use zero as an invalid tag entry.
+ * When the lock bit is set, this forces a tag comparison failure.
+ */
+
+#define TSB_TAG_LOCK_BIT	47
+#define TSB_TAG_LOCK_HIGH	(1 << (TSB_TAG_LOCK_BIT - 32))
+
+#define TSB_TAG_INVALID_BIT	46
+#define TSB_TAG_INVALID_HIGH	(1 << (TSB_TAG_INVALID_BIT - 32))
+
+#define TSB_MEMBAR	membar	#StoreStore
+
+/* Some cpus support physical address quad loads.  We want to use
+ * those if possible so we don't need to hard-lock the TSB mapping
+ * into the TLB.  We encode some instruction patching in order to
+ * support this.
+ *
+ * The kernel TSB is locked into the TLB by virtue of being in the
+ * kernel image, so we don't play these games for swapper_tsb access.
+ */
+#ifndef __ASSEMBLY__
+struct tsb_ldquad_phys_patch_entry {
+	unsigned int	addr;
+	unsigned int	sun4u_insn;
+	unsigned int	sun4v_insn;
+};
+extern struct tsb_ldquad_phys_patch_entry __tsb_ldquad_phys_patch,
+	__tsb_ldquad_phys_patch_end;
+
+struct tsb_phys_patch_entry {
+	unsigned int	addr;
+	unsigned int	insn;
+};
+extern struct tsb_phys_patch_entry __tsb_phys_patch, __tsb_phys_patch_end;
+#endif
+#define TSB_LOAD_QUAD(TSB, REG)	\
+661:	ldda		[TSB] ASI_NUCLEUS_QUAD_LDD, REG; \
+	.section	.tsb_ldquad_phys_patch, "ax"; \
+	.word		661b; \
+	ldda		[TSB] ASI_QUAD_LDD_PHYS, REG; \
+	ldda		[TSB] ASI_QUAD_LDD_PHYS_4V, REG; \
+	.previous
+
+#define TSB_LOAD_TAG_HIGH(TSB, REG) \
+661:	lduwa		[TSB] ASI_N, REG; \
+	.section	.tsb_phys_patch, "ax"; \
+	.word		661b; \
+	lduwa		[TSB] ASI_PHYS_USE_EC, REG; \
+	.previous
+
+#define TSB_LOAD_TAG(TSB, REG) \
+661:	ldxa		[TSB] ASI_N, REG; \
+	.section	.tsb_phys_patch, "ax"; \
+	.word		661b; \
+	ldxa		[TSB] ASI_PHYS_USE_EC, REG; \
+	.previous
+
+#define TSB_CAS_TAG_HIGH(TSB, REG1, REG2) \
+661:	casa		[TSB] ASI_N, REG1, REG2; \
+	.section	.tsb_phys_patch, "ax"; \
+	.word		661b; \
+	casa		[TSB] ASI_PHYS_USE_EC, REG1, REG2; \
+	.previous
+
+#define TSB_CAS_TAG(TSB, REG1, REG2) \
+661:	casxa		[TSB] ASI_N, REG1, REG2; \
+	.section	.tsb_phys_patch, "ax"; \
+	.word		661b; \
+	casxa		[TSB] ASI_PHYS_USE_EC, REG1, REG2; \
+	.previous
+
+#define TSB_STORE(ADDR, VAL) \
+661:	stxa		VAL, [ADDR] ASI_N; \
+	.section	.tsb_phys_patch, "ax"; \
+	.word		661b; \
+	stxa		VAL, [ADDR] ASI_PHYS_USE_EC; \
+	.previous
+
+#define TSB_LOCK_TAG(TSB, REG1, REG2)	\
+99:	TSB_LOAD_TAG_HIGH(TSB, REG1);	\
+	sethi	%hi(TSB_TAG_LOCK_HIGH), REG2;\
+	andcc	REG1, REG2, %g0;	\
+	bne,pn	%icc, 99b;		\
+	 nop;				\
+	TSB_CAS_TAG_HIGH(TSB, REG1, REG2);	\
+	cmp	REG1, REG2;		\
+	bne,pn	%icc, 99b;		\
+	 nop;				\
+	TSB_MEMBAR
+
+#define TSB_WRITE(TSB, TTE, TAG) \
+	add	TSB, 0x8, TSB;   \
+	TSB_STORE(TSB, TTE);     \
+	sub	TSB, 0x8, TSB;   \
+	TSB_MEMBAR;              \
+	TSB_STORE(TSB, TAG);
+
+#define KTSB_LOAD_QUAD(TSB, REG) \
+	ldda		[TSB] ASI_NUCLEUS_QUAD_LDD, REG;
+
+#define KTSB_STORE(ADDR, VAL) \
+	stxa		VAL, [ADDR] ASI_N;
+
+#define KTSB_LOCK_TAG(TSB, REG1, REG2)	\
+99:	lduwa	[TSB] ASI_N, REG1;	\
+	sethi	%hi(TSB_TAG_LOCK_HIGH), REG2;\
+	andcc	REG1, REG2, %g0;	\
+	bne,pn	%icc, 99b;		\
+	 nop;				\
+	casa	[TSB] ASI_N, REG1, REG2;\
+	cmp	REG1, REG2;		\
+	bne,pn	%icc, 99b;		\
+	 nop;				\
+	TSB_MEMBAR
+
+#define KTSB_WRITE(TSB, TTE, TAG) \
+	add	TSB, 0x8, TSB;   \
+	stxa	TTE, [TSB] ASI_N;     \
+	sub	TSB, 0x8, TSB;   \
+	TSB_MEMBAR;              \
+	stxa	TAG, [TSB] ASI_N;
+
+	/* Do a kernel page table walk.  Leaves physical PTE pointer in
+	 * REG1.  Jumps to FAIL_LABEL on early page table walk termination.
+	 * VADDR will not be clobbered, but REG2 will.
+	 */
+#define KERN_PGTABLE_WALK(VADDR, REG1, REG2, FAIL_LABEL)	\
+	sethi		%hi(swapper_pg_dir), REG1; \
+	or		REG1, %lo(swapper_pg_dir), REG1; \
+	sllx		VADDR, 64 - (PGDIR_SHIFT + PGDIR_BITS), REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	andn		REG2, 0x3, REG2; \
+	lduw		[REG1 + REG2], REG1; \
+	brz,pn		REG1, FAIL_LABEL; \
+	 sllx		VADDR, 64 - (PMD_SHIFT + PMD_BITS), REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	sllx		REG1, 11, REG1; \
+	andn		REG2, 0x3, REG2; \
+	lduwa		[REG1 + REG2] ASI_PHYS_USE_EC, REG1; \
+	brz,pn		REG1, FAIL_LABEL; \
+	 sllx		VADDR, 64 - PMD_SHIFT, REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	sllx		REG1, 11, REG1; \
+	andn		REG2, 0x7, REG2; \
+	add		REG1, REG2, REG1;
+
+	/* Do a user page table walk in MMU globals.  Leaves physical PTE
+	 * pointer in REG1.  Jumps to FAIL_LABEL on early page table walk
+	 * termination.  Physical base of page tables is in PHYS_PGD which
+	 * will not be modified.
+	 *
+	 * VADDR will not be clobbered, but REG1 and REG2 will.
+	 */
+#define USER_PGTABLE_WALK_TL1(VADDR, PHYS_PGD, REG1, REG2, FAIL_LABEL)	\
+	sllx		VADDR, 64 - (PGDIR_SHIFT + PGDIR_BITS), REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	andn		REG2, 0x3, REG2; \
+	lduwa		[PHYS_PGD + REG2] ASI_PHYS_USE_EC, REG1; \
+	brz,pn		REG1, FAIL_LABEL; \
+	 sllx		VADDR, 64 - (PMD_SHIFT + PMD_BITS), REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	sllx		REG1, 11, REG1; \
+	andn		REG2, 0x3, REG2; \
+	lduwa		[REG1 + REG2] ASI_PHYS_USE_EC, REG1; \
+	brz,pn		REG1, FAIL_LABEL; \
+	 sllx		VADDR, 64 - PMD_SHIFT, REG2; \
+	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
+	sllx		REG1, 11, REG1; \
+	andn		REG2, 0x7, REG2; \
+	add		REG1, REG2, REG1;
+
+/* Lookup a OBP mapping on VADDR in the prom_trans[] table at TL>0.
+ * If no entry is found, FAIL_LABEL will be branched to.  On success
+ * the resulting PTE value will be left in REG1.  VADDR is preserved
+ * by this routine.
+ */
+#define OBP_TRANS_LOOKUP(VADDR, REG1, REG2, REG3, FAIL_LABEL) \
+	sethi		%hi(prom_trans), REG1; \
+	or		REG1, %lo(prom_trans), REG1; \
+97:	ldx		[REG1 + 0x00], REG2; \
+	brz,pn		REG2, FAIL_LABEL; \
+	 nop; \
+	ldx		[REG1 + 0x08], REG3; \
+	add		REG2, REG3, REG3; \
+	cmp		REG2, VADDR; \
+	bgu,pt		%xcc, 98f; \
+	 cmp		VADDR, REG3; \
+	bgeu,pt		%xcc, 98f; \
+	 ldx		[REG1 + 0x10], REG3; \
+	sub		VADDR, REG2, REG2; \
+	ba,pt		%xcc, 99f; \
+	 add		REG3, REG2, REG1; \
+98:	ba,pt		%xcc, 97b; \
+	 add		REG1, (3 * 8), REG1; \
+99:
+
+	/* We use a 32K TSB for the whole kernel, this allows to
+	 * handle about 16MB of modules and vmalloc mappings without
+	 * incurring many hash conflicts.
+	 */
+#define KERNEL_TSB_SIZE_BYTES	(32 * 1024)
+#define KERNEL_TSB_NENTRIES	\
+	(KERNEL_TSB_SIZE_BYTES / 16)
+#define KERNEL_TSB4M_NENTRIES	4096
+
+	/* Do a kernel TSB lookup at tl>0 on VADDR+TAG, branch to OK_LABEL
+	 * on TSB hit.  REG1, REG2, REG3, and REG4 are used as temporaries
+	 * and the found TTE will be left in REG1.  REG3 and REG4 must
+	 * be an even/odd pair of registers.
+	 *
+	 * VADDR and TAG will be preserved and not clobbered by this macro.
+	 */
+#define KERN_TSB_LOOKUP_TL1(VADDR, TAG, REG1, REG2, REG3, REG4, OK_LABEL) \
+	sethi		%hi(swapper_tsb), REG1; \
+	or		REG1, %lo(swapper_tsb), REG1; \
+	srlx		VADDR, PAGE_SHIFT, REG2; \
+	and		REG2, (KERNEL_TSB_NENTRIES - 1), REG2; \
+	sllx		REG2, 4, REG2; \
+	add		REG1, REG2, REG2; \
+	KTSB_LOAD_QUAD(REG2, REG3); \
+	cmp		REG3, TAG; \
+	be,a,pt		%xcc, OK_LABEL; \
+	 mov		REG4, REG1;
+
+#ifndef CONFIG_DEBUG_PAGEALLOC
+	/* This version uses a trick, the TAG is already (VADDR >> 22) so
+	 * we can make use of that for the index computation.
+	 */
+#define KERN_TSB4M_LOOKUP_TL1(TAG, REG1, REG2, REG3, REG4, OK_LABEL) \
+	sethi		%hi(swapper_4m_tsb), REG1; \
+	or		REG1, %lo(swapper_4m_tsb), REG1; \
+	and		TAG, (KERNEL_TSB4M_NENTRIES - 1), REG2; \
+	sllx		REG2, 4, REG2; \
+	add		REG1, REG2, REG2; \
+	KTSB_LOAD_QUAD(REG2, REG3); \
+	cmp		REG3, TAG; \
+	be,a,pt		%xcc, OK_LABEL; \
+	 mov		REG4, REG1;
+#endif
+
+#endif /* !(_SPARC64_TSB_H) */
diff --git a/include/asm-sparc/ttable.h b/include/asm-sparc/ttable.h
new file mode 100644
index 00000000000..5708ba2719f
--- /dev/null
+++ b/include/asm-sparc/ttable.h
@@ -0,0 +1,658 @@
+#ifndef _SPARC64_TTABLE_H
+#define _SPARC64_TTABLE_H
+
+#include <asm/utrap.h>
+
+#ifdef __ASSEMBLY__
+#include <asm/thread_info.h>
+#endif
+
+#define BOOT_KERNEL b sparc64_boot; nop; nop; nop; nop; nop; nop; nop;
+
+/* We need a "cleaned" instruction... */
+#define CLEAN_WINDOW							\
+	rdpr	%cleanwin, %l0;		add	%l0, 1, %l0;		\
+	wrpr	%l0, 0x0, %cleanwin;					\
+	clr	%o0;	clr	%o1;	clr	%o2;	clr	%o3;	\
+	clr	%o4;	clr	%o5;	clr	%o6;	clr	%o7;	\
+	clr	%l0;	clr	%l1;	clr	%l2;	clr	%l3;	\
+	clr	%l4;	clr	%l5;	clr	%l6;	clr	%l7;	\
+	retry;								\
+	nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;
+
+#define TRAP(routine)					\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etrap;				\
+109:	 or	%g7, %lo(109b), %g7;			\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o0;			\
+	ba,pt	%xcc, rtrap;				\
+	 nop;						\
+	nop;
+
+#define TRAP_7INSNS(routine)				\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etrap;				\
+109:	 or	%g7, %lo(109b), %g7;			\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o0;			\
+	ba,pt	%xcc, rtrap;				\
+	 nop;
+
+#define TRAP_SAVEFPU(routine)				\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, do_fptrap;			\
+109:	 or	%g7, %lo(109b), %g7;			\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o0;			\
+	ba,pt	%xcc, rtrap;				\
+	 nop;						\
+	nop;
+
+#define TRAP_NOSAVE(routine)				\
+	ba,pt	%xcc, routine;				\
+	 nop;						\
+	nop; nop; nop; nop; nop; nop;
+
+#define TRAP_NOSAVE_7INSNS(routine)			\
+	ba,pt	%xcc, routine;				\
+	 nop;						\
+	nop; nop; nop; nop; nop;
+
+#define TRAPTL1(routine)				\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etraptl1;				\
+109:	 or	%g7, %lo(109b), %g7;			\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o0;			\
+	ba,pt	%xcc, rtrap;				\
+	 nop;						\
+	nop;
+
+#define TRAP_ARG(routine, arg)				\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etrap;				\
+109:	 or	%g7, %lo(109b), %g7;			\
+	add	%sp, PTREGS_OFF, %o0;			\
+	call	routine;				\
+	 mov	arg, %o1;				\
+	ba,pt	%xcc, rtrap;				\
+	 nop;
+
+#define TRAPTL1_ARG(routine, arg)			\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etraptl1;				\
+109:	 or	%g7, %lo(109b), %g7;			\
+	add	%sp, PTREGS_OFF, %o0;			\
+	call	routine;				\
+	 mov	arg, %o1;				\
+	ba,pt	%xcc, rtrap;				\
+	 nop;
+
+#define SYSCALL_TRAP(routine, systbl)			\
+	rdpr	%pil, %g2;				\
+	mov	TSTATE_SYSCALL, %g3;			\
+	sethi	%hi(109f), %g7;				\
+	ba,pt	%xcc, etrap_syscall;			\
+109:	 or	%g7, %lo(109b), %g7;			\
+	sethi	%hi(systbl), %l7;			\
+	ba,pt	%xcc, routine;				\
+	 or	%l7, %lo(systbl), %l7;
+
+#define TRAP_UTRAP(handler,lvl)				\
+	mov	handler, %g3;				\
+	ba,pt	%xcc, utrap_trap;			\
+	 mov	lvl, %g4;				\
+	nop;						\
+	nop;						\
+	nop;						\
+	nop;						\
+	nop;
+
+#ifdef CONFIG_COMPAT
+#define	LINUX_32BIT_SYSCALL_TRAP SYSCALL_TRAP(linux_sparc_syscall32, sys_call_table32)
+#else
+#define	LINUX_32BIT_SYSCALL_TRAP BTRAP(0x110)
+#endif
+#define LINUX_64BIT_SYSCALL_TRAP SYSCALL_TRAP(linux_sparc_syscall, sys_call_table64)
+#define GETCC_TRAP TRAP(getcc)
+#define SETCC_TRAP TRAP(setcc)
+#define BREAKPOINT_TRAP TRAP(breakpoint_trap)
+
+#ifdef CONFIG_TRACE_IRQFLAGS
+
+#define TRAP_IRQ(routine, level)			\
+	rdpr	%pil, %g2;				\
+	wrpr	%g0, 15, %pil;				\
+	sethi	%hi(1f-4), %g7;				\
+	ba,pt	%xcc, etrap_irq;			\
+	 or	%g7, %lo(1f-4), %g7;			\
+	nop;						\
+	nop;						\
+	nop;						\
+	.subsection	2;				\
+1:	call	trace_hardirqs_off;			\
+	 nop;						\
+	mov	level, %o0;				\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o1;			\
+	ba,a,pt	%xcc, rtrap_irq;			\
+	.previous;
+
+#else
+
+#define TRAP_IRQ(routine, level)			\
+	rdpr	%pil, %g2;				\
+	wrpr	%g0, 15, %pil;				\
+	ba,pt	%xcc, etrap_irq;			\
+	 rd	%pc, %g7;				\
+	mov	level, %o0;				\
+	call	routine;				\
+	 add	%sp, PTREGS_OFF, %o1;			\
+	ba,a,pt	%xcc, rtrap_irq;
+
+#endif
+
+#define TRAP_IVEC TRAP_NOSAVE(do_ivec)
+
+#define BTRAP(lvl) TRAP_ARG(bad_trap, lvl)
+
+#define BTRAPTL1(lvl) TRAPTL1_ARG(bad_trap_tl1, lvl)
+
+#define FLUSH_WINDOW_TRAP						\
+	ba,pt	%xcc, etrap;						\
+	 rd	%pc, %g7;						\
+	flushw;								\
+	ldx	[%sp + PTREGS_OFF + PT_V9_TNPC], %l1;			\
+	add	%l1, 4, %l2;						\
+	stx	%l1, [%sp + PTREGS_OFF + PT_V9_TPC];			\
+	ba,pt	%xcc, rtrap;						\
+	 stx	%l2, [%sp + PTREGS_OFF + PT_V9_TNPC];
+
+#ifdef CONFIG_KPROBES
+#define KPROBES_TRAP(lvl) TRAP_IRQ(kprobe_trap, lvl)
+#else
+#define KPROBES_TRAP(lvl) TRAP_ARG(bad_trap, lvl)
+#endif
+
+#ifdef CONFIG_KGDB
+#define KGDB_TRAP(lvl) TRAP_IRQ(kgdb_trap, lvl)
+#else
+#define KGDB_TRAP(lvl) TRAP_ARG(bad_trap, lvl)
+#endif
+
+#define SUN4V_ITSB_MISS					\
+	ldxa	[%g0] ASI_SCRATCHPAD, %g2;		\
+	ldx	[%g2 + HV_FAULT_I_ADDR_OFFSET], %g4;	\
+	ldx	[%g2 + HV_FAULT_I_CTX_OFFSET], %g5;	\
+	srlx	%g4, 22, %g6;				\
+	ba,pt	%xcc, sun4v_itsb_miss;			\
+	 nop;						\
+	nop;						\
+	nop;
+
+#define SUN4V_DTSB_MISS					\
+	ldxa	[%g0] ASI_SCRATCHPAD, %g2;		\
+	ldx	[%g2 + HV_FAULT_D_ADDR_OFFSET], %g4;	\
+	ldx	[%g2 + HV_FAULT_D_CTX_OFFSET], %g5;	\
+	srlx	%g4, 22, %g6;				\
+	ba,pt	%xcc, sun4v_dtsb_miss;			\
+	 nop;						\
+	nop;						\
+	nop;
+
+/* Before touching these macros, you owe it to yourself to go and
+ * see how arch/sparc64/kernel/winfixup.S works... -DaveM
+ *
+ * For the user cases we used to use the %asi register, but
+ * it turns out that the "wr xxx, %asi" costs ~5 cycles, so
+ * now we use immediate ASI loads and stores instead.  Kudos
+ * to Greg Onufer for pointing out this performance anomaly.
+ *
+ * Further note that we cannot use the g2, g4, g5, and g7 alternate
+ * globals in the spill routines, check out the save instruction in
+ * arch/sparc64/kernel/etrap.S to see what I mean about g2, and
+ * g4/g5 are the globals which are preserved by etrap processing
+ * for the caller of it.  The g7 register is the return pc for
+ * etrap.  Finally, g6 is the current thread register so we cannot
+ * us it in the spill handlers either.  Most of these rules do not
+ * apply to fill processing, only g6 is not usable.
+ */
+
+/* Normal kernel spill */
+#define SPILL_0_NORMAL					\
+	stx	%l0, [%sp + STACK_BIAS + 0x00];		\
+	stx	%l1, [%sp + STACK_BIAS + 0x08];		\
+	stx	%l2, [%sp + STACK_BIAS + 0x10];		\
+	stx	%l3, [%sp + STACK_BIAS + 0x18];		\
+	stx	%l4, [%sp + STACK_BIAS + 0x20];		\
+	stx	%l5, [%sp + STACK_BIAS + 0x28];		\
+	stx	%l6, [%sp + STACK_BIAS + 0x30];		\
+	stx	%l7, [%sp + STACK_BIAS + 0x38];		\
+	stx	%i0, [%sp + STACK_BIAS + 0x40];		\
+	stx	%i1, [%sp + STACK_BIAS + 0x48];		\
+	stx	%i2, [%sp + STACK_BIAS + 0x50];		\
+	stx	%i3, [%sp + STACK_BIAS + 0x58];		\
+	stx	%i4, [%sp + STACK_BIAS + 0x60];		\
+	stx	%i5, [%sp + STACK_BIAS + 0x68];		\
+	stx	%i6, [%sp + STACK_BIAS + 0x70];		\
+	stx	%i7, [%sp + STACK_BIAS + 0x78];		\
+	saved; retry; nop; nop; nop; nop; nop; nop;	\
+	nop; nop; nop; nop; nop; nop; nop; nop;
+
+#define SPILL_0_NORMAL_ETRAP				\
+etrap_kernel_spill:					\
+	stx	%l0, [%sp + STACK_BIAS + 0x00];		\
+	stx	%l1, [%sp + STACK_BIAS + 0x08];		\
+	stx	%l2, [%sp + STACK_BIAS + 0x10];		\
+	stx	%l3, [%sp + STACK_BIAS + 0x18];		\
+	stx	%l4, [%sp + STACK_BIAS + 0x20];		\
+	stx	%l5, [%sp + STACK_BIAS + 0x28];		\
+	stx	%l6, [%sp + STACK_BIAS + 0x30];		\
+	stx	%l7, [%sp + STACK_BIAS + 0x38];		\
+	stx	%i0, [%sp + STACK_BIAS + 0x40];		\
+	stx	%i1, [%sp + STACK_BIAS + 0x48];		\
+	stx	%i2, [%sp + STACK_BIAS + 0x50];		\
+	stx	%i3, [%sp + STACK_BIAS + 0x58];		\
+	stx	%i4, [%sp + STACK_BIAS + 0x60];		\
+	stx	%i5, [%sp + STACK_BIAS + 0x68];		\
+	stx	%i6, [%sp + STACK_BIAS + 0x70];		\
+	stx	%i7, [%sp + STACK_BIAS + 0x78];		\
+	saved;						\
+	sub	%g1, 2, %g1;				\
+	ba,pt	%xcc, etrap_save;			\
+	wrpr	%g1, %cwp;				\
+	nop; nop; nop; nop; nop; nop; nop; nop;		\
+	nop; nop; nop; nop;
+
+/* Normal 64bit spill */
+#define SPILL_1_GENERIC(ASI)				\
+	add	%sp, STACK_BIAS + 0x00, %g1;		\
+	stxa	%l0, [%g1 + %g0] ASI;			\
+	mov	0x08, %g3;				\
+	stxa	%l1, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%l2, [%g1 + %g0] ASI;			\
+	stxa	%l3, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%l4, [%g1 + %g0] ASI;			\
+	stxa	%l5, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%l6, [%g1 + %g0] ASI;			\
+	stxa	%l7, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%i0, [%g1 + %g0] ASI;			\
+	stxa	%i1, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%i2, [%g1 + %g0] ASI;			\
+	stxa	%i3, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%i4, [%g1 + %g0] ASI;			\
+	stxa	%i5, [%g1 + %g3] ASI;			\
+	add	%g1, 0x10, %g1;				\
+	stxa	%i6, [%g1 + %g0] ASI;			\
+	stxa	%i7, [%g1 + %g3] ASI;			\
+	saved;						\
+	retry; nop; nop;				\
+	b,a,pt	%xcc, spill_fixup_dax;			\
+	b,a,pt	%xcc, spill_fixup_mna;			\
+	b,a,pt	%xcc, spill_fixup;
+
+#define SPILL_1_GENERIC_ETRAP				\
+etrap_user_spill_64bit:					\
+	stxa	%l0, [%sp + STACK_BIAS + 0x00] %asi;	\
+	stxa	%l1, [%sp + STACK_BIAS + 0x08] %asi;	\
+	stxa	%l2, [%sp + STACK_BIAS + 0x10] %asi;	\
+	stxa	%l3, [%sp + STACK_BIAS + 0x18] %asi;	\
+	stxa	%l4, [%sp + STACK_BIAS + 0x20] %asi;	\
+	stxa	%l5, [%sp + STACK_BIAS + 0x28] %asi;	\
+	stxa	%l6, [%sp + STACK_BIAS + 0x30] %asi;	\
+	stxa	%l7, [%sp + STACK_BIAS + 0x38] %asi;	\
+	stxa	%i0, [%sp + STACK_BIAS + 0x40] %asi;	\
+	stxa	%i1, [%sp + STACK_BIAS + 0x48] %asi;	\
+	stxa	%i2, [%sp + STACK_BIAS + 0x50] %asi;	\
+	stxa	%i3, [%sp + STACK_BIAS + 0x58] %asi;	\
+	stxa	%i4, [%sp + STACK_BIAS + 0x60] %asi;	\
+	stxa	%i5, [%sp + STACK_BIAS + 0x68] %asi;	\
+	stxa	%i6, [%sp + STACK_BIAS + 0x70] %asi;	\
+	stxa	%i7, [%sp + STACK_BIAS + 0x78] %asi;	\
+	saved;						\
+	sub	%g1, 2, %g1;				\
+	ba,pt	%xcc, etrap_save;			\
+	 wrpr	%g1, %cwp;				\
+	nop; nop; nop; nop; nop;			\
+	nop; nop; nop; nop;				\
+	ba,a,pt	%xcc, etrap_spill_fixup_64bit;		\
+	ba,a,pt	%xcc, etrap_spill_fixup_64bit;		\
+	ba,a,pt	%xcc, etrap_spill_fixup_64bit;
+
+#define SPILL_1_GENERIC_ETRAP_FIXUP			\
+etrap_spill_fixup_64bit:				\
+	ldub	[%g6 + TI_WSAVED], %g1;			\
+	sll	%g1, 3, %g3;				\
+	add	%g6, %g3, %g3;				\
+	stx	%sp, [%g3 + TI_RWIN_SPTRS];		\
+	sll	%g1, 7, %g3;				\
+	add	%g6, %g3, %g3;				\
+	stx	%l0, [%g3 + TI_REG_WINDOW + 0x00];	\
+	stx	%l1, [%g3 + TI_REG_WINDOW + 0x08];	\
+	stx	%l2, [%g3 + TI_REG_WINDOW + 0x10];	\
+	stx	%l3, [%g3 + TI_REG_WINDOW + 0x18];	\
+	stx	%l4, [%g3 + TI_REG_WINDOW + 0x20];	\
+	stx	%l5, [%g3 + TI_REG_WINDOW + 0x28];	\
+	stx	%l6, [%g3 + TI_REG_WINDOW + 0x30];	\
+	stx	%l7, [%g3 + TI_REG_WINDOW + 0x38];	\
+	stx	%i0, [%g3 + TI_REG_WINDOW + 0x40];	\
+	stx	%i1, [%g3 + TI_REG_WINDOW + 0x48];	\
+	stx	%i2, [%g3 + TI_REG_WINDOW + 0x50];	\
+	stx	%i3, [%g3 + TI_REG_WINDOW + 0x58];	\
+	stx	%i4, [%g3 + TI_REG_WINDOW + 0x60];	\
+	stx	%i5, [%g3 + TI_REG_WINDOW + 0x68];	\
+	stx	%i6, [%g3 + TI_REG_WINDOW + 0x70];	\
+	stx	%i7, [%g3 + TI_REG_WINDOW + 0x78];	\
+	add	%g1, 1, %g1;				\
+	stb	%g1, [%g6 + TI_WSAVED];			\
+	saved;						\
+	rdpr	%cwp, %g1;				\
+	sub	%g1, 2, %g1;				\
+	ba,pt	%xcc, etrap_save;			\
+	 wrpr	%g1, %cwp;				\
+	nop; nop; nop
+
+/* Normal 32bit spill */
+#define SPILL_2_GENERIC(ASI)				\
+	srl	%sp, 0, %sp;				\
+	stwa	%l0, [%sp + %g0] ASI;			\
+	mov	0x04, %g3;				\
+	stwa	%l1, [%sp + %g3] ASI;			\
+	add	%sp, 0x08, %g1;				\
+	stwa	%l2, [%g1 + %g0] ASI;			\
+	stwa	%l3, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%l4, [%g1 + %g0] ASI;			\
+	stwa	%l5, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%l6, [%g1 + %g0] ASI;			\
+	stwa	%l7, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%i0, [%g1 + %g0] ASI;			\
+	stwa	%i1, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%i2, [%g1 + %g0] ASI;			\
+	stwa	%i3, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%i4, [%g1 + %g0] ASI;			\
+	stwa	%i5, [%g1 + %g3] ASI;			\
+	add	%g1, 0x08, %g1;				\
+	stwa	%i6, [%g1 + %g0] ASI;			\
+	stwa	%i7, [%g1 + %g3] ASI;			\
+	saved;						\
+        retry; nop; nop;				\
+	b,a,pt	%xcc, spill_fixup_dax;			\
+	b,a,pt	%xcc, spill_fixup_mna;			\
+	b,a,pt	%xcc, spill_fixup;
+
+#define SPILL_2_GENERIC_ETRAP		\
+etrap_user_spill_32bit:			\
+	srl	%sp, 0, %sp;		\
+	stwa	%l0, [%sp + 0x00] %asi;	\
+	stwa	%l1, [%sp + 0x04] %asi;	\
+	stwa	%l2, [%sp + 0x08] %asi;	\
+	stwa	%l3, [%sp + 0x0c] %asi;	\
+	stwa	%l4, [%sp + 0x10] %asi;	\
+	stwa	%l5, [%sp + 0x14] %asi;	\
+	stwa	%l6, [%sp + 0x18] %asi;	\
+	stwa	%l7, [%sp + 0x1c] %asi;	\
+	stwa	%i0, [%sp + 0x20] %asi;	\
+	stwa	%i1, [%sp + 0x24] %asi;	\
+	stwa	%i2, [%sp + 0x28] %asi;	\
+	stwa	%i3, [%sp + 0x2c] %asi;	\
+	stwa	%i4, [%sp + 0x30] %asi;	\
+	stwa	%i5, [%sp + 0x34] %asi;	\
+	stwa	%i6, [%sp + 0x38] %asi;	\
+	stwa	%i7, [%sp + 0x3c] %asi;	\
+	saved;				\
+	sub	%g1, 2, %g1;		\
+	ba,pt	%xcc, etrap_save;	\
+	 wrpr	%g1, %cwp;		\
+	nop; nop; nop; nop;		\
+	nop; nop; nop; nop;		\
+	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
+	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
+	ba,a,pt	%xcc, etrap_spill_fixup_32bit;
+
+#define SPILL_2_GENERIC_ETRAP_FIXUP			\
+etrap_spill_fixup_32bit:				\
+	ldub	[%g6 + TI_WSAVED], %g1;			\
+	sll	%g1, 3, %g3;				\
+	add	%g6, %g3, %g3;				\
+	stx	%sp, [%g3 + TI_RWIN_SPTRS];		\
+	sll	%g1, 7, %g3;				\
+	add	%g6, %g3, %g3;				\
+	stw	%l0, [%g3 + TI_REG_WINDOW + 0x00];	\
+	stw	%l1, [%g3 + TI_REG_WINDOW + 0x04];	\
+	stw	%l2, [%g3 + TI_REG_WINDOW + 0x08];	\
+	stw	%l3, [%g3 + TI_REG_WINDOW + 0x0c];	\
+	stw	%l4, [%g3 + TI_REG_WINDOW + 0x10];	\
+	stw	%l5, [%g3 + TI_REG_WINDOW + 0x14];	\
+	stw	%l6, [%g3 + TI_REG_WINDOW + 0x18];	\
+	stw	%l7, [%g3 + TI_REG_WINDOW + 0x1c];	\
+	stw	%i0, [%g3 + TI_REG_WINDOW + 0x20];	\
+	stw	%i1, [%g3 + TI_REG_WINDOW + 0x24];	\
+	stw	%i2, [%g3 + TI_REG_WINDOW + 0x28];	\
+	stw	%i3, [%g3 + TI_REG_WINDOW + 0x2c];	\
+	stw	%i4, [%g3 + TI_REG_WINDOW + 0x30];	\
+	stw	%i5, [%g3 + TI_REG_WINDOW + 0x34];	\
+	stw	%i6, [%g3 + TI_REG_WINDOW + 0x38];	\
+	stw	%i7, [%g3 + TI_REG_WINDOW + 0x3c];	\
+	add	%g1, 1, %g1;				\
+	stb	%g1, [%g6 + TI_WSAVED];			\
+	saved;						\
+	rdpr	%cwp, %g1;				\
+	sub	%g1, 2, %g1;				\
+	ba,pt	%xcc, etrap_save;			\
+	 wrpr	%g1, %cwp;				\
+	nop; nop; nop
+
+#define SPILL_1_NORMAL SPILL_1_GENERIC(ASI_AIUP)
+#define SPILL_2_NORMAL SPILL_2_GENERIC(ASI_AIUP)
+#define SPILL_3_NORMAL SPILL_0_NORMAL
+#define SPILL_4_NORMAL SPILL_0_NORMAL
+#define SPILL_5_NORMAL SPILL_0_NORMAL
+#define SPILL_6_NORMAL SPILL_0_NORMAL
+#define SPILL_7_NORMAL SPILL_0_NORMAL
+
+#define SPILL_0_OTHER SPILL_0_NORMAL
+#define SPILL_1_OTHER SPILL_1_GENERIC(ASI_AIUS)
+#define SPILL_2_OTHER SPILL_2_GENERIC(ASI_AIUS)
+#define SPILL_3_OTHER SPILL_3_NORMAL
+#define SPILL_4_OTHER SPILL_4_NORMAL
+#define SPILL_5_OTHER SPILL_5_NORMAL
+#define SPILL_6_OTHER SPILL_6_NORMAL
+#define SPILL_7_OTHER SPILL_7_NORMAL
+
+/* Normal kernel fill */
+#define FILL_0_NORMAL					\
+	ldx	[%sp + STACK_BIAS + 0x00], %l0;		\
+	ldx	[%sp + STACK_BIAS + 0x08], %l1;		\
+	ldx	[%sp + STACK_BIAS + 0x10], %l2;		\
+	ldx	[%sp + STACK_BIAS + 0x18], %l3;		\
+	ldx	[%sp + STACK_BIAS + 0x20], %l4;		\
+	ldx	[%sp + STACK_BIAS + 0x28], %l5;		\
+	ldx	[%sp + STACK_BIAS + 0x30], %l6;		\
+	ldx	[%sp + STACK_BIAS + 0x38], %l7;		\
+	ldx	[%sp + STACK_BIAS + 0x40], %i0;		\
+	ldx	[%sp + STACK_BIAS + 0x48], %i1;		\
+	ldx	[%sp + STACK_BIAS + 0x50], %i2;		\
+	ldx	[%sp + STACK_BIAS + 0x58], %i3;		\
+	ldx	[%sp + STACK_BIAS + 0x60], %i4;		\
+	ldx	[%sp + STACK_BIAS + 0x68], %i5;		\
+	ldx	[%sp + STACK_BIAS + 0x70], %i6;		\
+	ldx	[%sp + STACK_BIAS + 0x78], %i7;		\
+	restored; retry; nop; nop; nop; nop; nop; nop;	\
+	nop; nop; nop; nop; nop; nop; nop; nop;
+
+#define FILL_0_NORMAL_RTRAP				\
+kern_rtt_fill:						\
+	rdpr	%cwp, %g1;				\
+	sub	%g1, 1, %g1;				\
+	wrpr	%g1, %cwp;				\
+	ldx	[%sp + STACK_BIAS + 0x00], %l0;		\
+	ldx	[%sp + STACK_BIAS + 0x08], %l1;		\
+	ldx	[%sp + STACK_BIAS + 0x10], %l2;		\
+	ldx	[%sp + STACK_BIAS + 0x18], %l3;		\
+	ldx	[%sp + STACK_BIAS + 0x20], %l4;		\
+	ldx	[%sp + STACK_BIAS + 0x28], %l5;		\
+	ldx	[%sp + STACK_BIAS + 0x30], %l6;		\
+	ldx	[%sp + STACK_BIAS + 0x38], %l7;		\
+	ldx	[%sp + STACK_BIAS + 0x40], %i0;		\
+	ldx	[%sp + STACK_BIAS + 0x48], %i1;		\
+	ldx	[%sp + STACK_BIAS + 0x50], %i2;		\
+	ldx	[%sp + STACK_BIAS + 0x58], %i3;		\
+	ldx	[%sp + STACK_BIAS + 0x60], %i4;		\
+	ldx	[%sp + STACK_BIAS + 0x68], %i5;		\
+	ldx	[%sp + STACK_BIAS + 0x70], %i6;		\
+	ldx	[%sp + STACK_BIAS + 0x78], %i7;		\
+	restored;					\
+	add	%g1, 1, %g1;				\
+	ba,pt	%xcc, kern_rtt_restore;			\
+	 wrpr	%g1, %cwp;				\
+	nop; nop; nop; nop; nop;			\
+	nop; nop; nop; nop;
+
+
+/* Normal 64bit fill */
+#define FILL_1_GENERIC(ASI)				\
+	add	%sp, STACK_BIAS + 0x00, %g1;		\
+	ldxa	[%g1 + %g0] ASI, %l0;			\
+	mov	0x08, %g2;				\
+	mov	0x10, %g3;				\
+	ldxa	[%g1 + %g2] ASI, %l1;			\
+	mov	0x18, %g5;				\
+	ldxa	[%g1 + %g3] ASI, %l2;			\
+	ldxa	[%g1 + %g5] ASI, %l3;			\
+	add	%g1, 0x20, %g1;				\
+	ldxa	[%g1 + %g0] ASI, %l4;			\
+	ldxa	[%g1 + %g2] ASI, %l5;			\
+	ldxa	[%g1 + %g3] ASI, %l6;			\
+	ldxa	[%g1 + %g5] ASI, %l7;			\
+	add	%g1, 0x20, %g1;				\
+	ldxa	[%g1 + %g0] ASI, %i0;			\
+	ldxa	[%g1 + %g2] ASI, %i1;			\
+	ldxa	[%g1 + %g3] ASI, %i2;			\
+	ldxa	[%g1 + %g5] ASI, %i3;			\
+	add	%g1, 0x20, %g1;				\
+	ldxa	[%g1 + %g0] ASI, %i4;			\
+	ldxa	[%g1 + %g2] ASI, %i5;			\
+	ldxa	[%g1 + %g3] ASI, %i6;			\
+	ldxa	[%g1 + %g5] ASI, %i7;			\
+	restored;					\
+	retry; nop; nop; nop; nop;			\
+	b,a,pt	%xcc, fill_fixup_dax;			\
+	b,a,pt	%xcc, fill_fixup_mna;			\
+	b,a,pt	%xcc, fill_fixup;
+
+#define FILL_1_GENERIC_RTRAP				\
+user_rtt_fill_64bit:					\
+	ldxa	[%sp + STACK_BIAS + 0x00] %asi, %l0;	\
+	ldxa	[%sp + STACK_BIAS + 0x08] %asi, %l1;	\
+	ldxa	[%sp + STACK_BIAS + 0x10] %asi, %l2;	\
+	ldxa	[%sp + STACK_BIAS + 0x18] %asi, %l3;	\
+	ldxa	[%sp + STACK_BIAS + 0x20] %asi, %l4;	\
+	ldxa	[%sp + STACK_BIAS + 0x28] %asi, %l5;	\
+	ldxa	[%sp + STACK_BIAS + 0x30] %asi, %l6;	\
+	ldxa	[%sp + STACK_BIAS + 0x38] %asi, %l7;	\
+	ldxa	[%sp + STACK_BIAS + 0x40] %asi, %i0;	\
+	ldxa	[%sp + STACK_BIAS + 0x48] %asi, %i1;	\
+	ldxa	[%sp + STACK_BIAS + 0x50] %asi, %i2;	\
+	ldxa	[%sp + STACK_BIAS + 0x58] %asi, %i3;	\
+	ldxa	[%sp + STACK_BIAS + 0x60] %asi, %i4;	\
+	ldxa	[%sp + STACK_BIAS + 0x68] %asi, %i5;	\
+	ldxa	[%sp + STACK_BIAS + 0x70] %asi, %i6;	\
+	ldxa	[%sp + STACK_BIAS + 0x78] %asi, %i7;	\
+	ba,pt	%xcc, user_rtt_pre_restore;		\
+	 restored;					\
+	nop; nop; nop; nop; nop; nop;			\
+	nop; nop; nop; nop; nop;			\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;
+
+
+/* Normal 32bit fill */
+#define FILL_2_GENERIC(ASI)				\
+	srl	%sp, 0, %sp;				\
+	lduwa	[%sp + %g0] ASI, %l0;			\
+	mov	0x04, %g2;				\
+	mov	0x08, %g3;				\
+	lduwa	[%sp + %g2] ASI, %l1;			\
+	mov	0x0c, %g5;				\
+	lduwa	[%sp + %g3] ASI, %l2;			\
+	lduwa	[%sp + %g5] ASI, %l3;			\
+	add	%sp, 0x10, %g1;				\
+	lduwa	[%g1 + %g0] ASI, %l4;			\
+	lduwa	[%g1 + %g2] ASI, %l5;			\
+	lduwa	[%g1 + %g3] ASI, %l6;			\
+	lduwa	[%g1 + %g5] ASI, %l7;			\
+	add	%g1, 0x10, %g1;				\
+	lduwa	[%g1 + %g0] ASI, %i0;			\
+	lduwa	[%g1 + %g2] ASI, %i1;			\
+	lduwa	[%g1 + %g3] ASI, %i2;			\
+	lduwa	[%g1 + %g5] ASI, %i3;			\
+	add	%g1, 0x10, %g1;				\
+	lduwa	[%g1 + %g0] ASI, %i4;			\
+	lduwa	[%g1 + %g2] ASI, %i5;			\
+	lduwa	[%g1 + %g3] ASI, %i6;			\
+	lduwa	[%g1 + %g5] ASI, %i7;			\
+	restored;					\
+	retry; nop; nop; nop; nop;			\
+	b,a,pt	%xcc, fill_fixup_dax;			\
+	b,a,pt	%xcc, fill_fixup_mna;			\
+	b,a,pt	%xcc, fill_fixup;
+
+#define FILL_2_GENERIC_RTRAP				\
+user_rtt_fill_32bit:					\
+	srl	%sp, 0, %sp;				\
+	lduwa	[%sp + 0x00] %asi, %l0;			\
+	lduwa	[%sp + 0x04] %asi, %l1;			\
+	lduwa	[%sp + 0x08] %asi, %l2;			\
+	lduwa	[%sp + 0x0c] %asi, %l3;			\
+	lduwa	[%sp + 0x10] %asi, %l4;			\
+	lduwa	[%sp + 0x14] %asi, %l5;			\
+	lduwa	[%sp + 0x18] %asi, %l6;			\
+	lduwa	[%sp + 0x1c] %asi, %l7;			\
+	lduwa	[%sp + 0x20] %asi, %i0;			\
+	lduwa	[%sp + 0x24] %asi, %i1;			\
+	lduwa	[%sp + 0x28] %asi, %i2;			\
+	lduwa	[%sp + 0x2c] %asi, %i3;			\
+	lduwa	[%sp + 0x30] %asi, %i4;			\
+	lduwa	[%sp + 0x34] %asi, %i5;			\
+	lduwa	[%sp + 0x38] %asi, %i6;			\
+	lduwa	[%sp + 0x3c] %asi, %i7;			\
+	ba,pt	%xcc, user_rtt_pre_restore;		\
+	 restored;					\
+	nop; nop; nop; nop; nop;			\
+	nop; nop; nop; nop; nop;			\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
+	ba,a,pt	%xcc, user_rtt_fill_fixup;
+
+
+#define FILL_1_NORMAL FILL_1_GENERIC(ASI_AIUP)
+#define FILL_2_NORMAL FILL_2_GENERIC(ASI_AIUP)
+#define FILL_3_NORMAL FILL_0_NORMAL
+#define FILL_4_NORMAL FILL_0_NORMAL
+#define FILL_5_NORMAL FILL_0_NORMAL
+#define FILL_6_NORMAL FILL_0_NORMAL
+#define FILL_7_NORMAL FILL_0_NORMAL
+
+#define FILL_0_OTHER FILL_0_NORMAL
+#define FILL_1_OTHER FILL_1_GENERIC(ASI_AIUS)
+#define FILL_2_OTHER FILL_2_GENERIC(ASI_AIUS)
+#define FILL_3_OTHER FILL_3_NORMAL
+#define FILL_4_OTHER FILL_4_NORMAL
+#define FILL_5_OTHER FILL_5_NORMAL
+#define FILL_6_OTHER FILL_6_NORMAL
+#define FILL_7_OTHER FILL_7_NORMAL
+
+#endif /* !(_SPARC64_TTABLE_H) */
diff --git a/include/asm-sparc/types.h b/include/asm-sparc/types.h
index 07734f94240..8c28fde5eaa 100644
--- a/include/asm-sparc/types.h
+++ b/include/asm-sparc/types.h
@@ -1,6 +1,5 @@
 #ifndef _SPARC_TYPES_H
 #define _SPARC_TYPES_H
-
 /*
  * This file is never included by application software unless
  * explicitly requested (e.g., via linux/types.h) in which case the
@@ -8,6 +7,35 @@
  * not a major issue.  However, for interoperability, libraries still
  * need to be careful to avoid a name clashes.
  */
+
+#if defined(__sparc__) && defined(__arch64__)
+
+/*** SPARC 64 bit ***/
+#include <asm-generic/int-l64.h>
+
+#ifndef __ASSEMBLY__
+
+typedef unsigned short umode_t;
+
+#endif /* __ASSEMBLY__ */
+
+#ifdef __KERNEL__
+
+#define BITS_PER_LONG 64
+
+#ifndef __ASSEMBLY__
+
+/* Dma addresses come in generic and 64-bit flavours.  */
+
+typedef u32 dma_addr_t;
+typedef u64 dma64_addr_t;
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* __KERNEL__ */
+#else
+
+/*** SPARC 32 bit ***/
 #include <asm-generic/int-ll64.h>
 
 #ifndef __ASSEMBLY__
@@ -29,4 +57,6 @@ typedef u32 dma64_addr_t;
 
 #endif /* __KERNEL__ */
 
+#endif /* defined(__sparc__) && defined(__arch64__) */
+
 #endif /* defined(_SPARC_TYPES_H) */
diff --git a/include/asm-sparc/uaccess.h b/include/asm-sparc/uaccess.h
index 47d5619d43f..424facce523 100644
--- a/include/asm-sparc/uaccess.h
+++ b/include/asm-sparc/uaccess.h
@@ -1,336 +1,8 @@
-/*
- * uaccess.h: User space memore access functions.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1996,1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-#ifndef _ASM_UACCESS_H
-#define _ASM_UACCESS_H
-
-#ifdef __KERNEL__
-#include <linux/compiler.h>
-#include <linux/sched.h>
-#include <linux/string.h>
-#include <linux/errno.h>
-#include <asm/vac-ops.h>
+#ifndef ___ASM_SPARC_UACCESS_H
+#define ___ASM_SPARC_UACCESS_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/uaccess_64.h>
+#else
+#include <asm-sparc/uaccess_32.h>
+#endif
 #endif
-
-#ifndef __ASSEMBLY__
-
-/* Sparc is not segmented, however we need to be able to fool access_ok()
- * when doing system calls from kernel mode legitimately.
- *
- * "For historical reasons, these macros are grossly misnamed." -Linus
- */
-
-#define KERNEL_DS   ((mm_segment_t) { 0 })
-#define USER_DS     ((mm_segment_t) { -1 })
-
-#define VERIFY_READ	0
-#define VERIFY_WRITE	1
-
-#define get_ds()	(KERNEL_DS)
-#define get_fs()	(current->thread.current_ds)
-#define set_fs(val)	((current->thread.current_ds) = (val))
-
-#define segment_eq(a,b)	((a).seg == (b).seg)
-
-/* We have there a nice not-mapped page at PAGE_OFFSET - PAGE_SIZE, so that this test
- * can be fairly lightweight.
- * No one can read/write anything from userland in the kernel space by setting
- * large size and address near to PAGE_OFFSET - a fault will break his intentions.
- */
-#define __user_ok(addr, size) ({ (void)(size); (addr) < STACK_TOP; })
-#define __kernel_ok (segment_eq(get_fs(), KERNEL_DS))
-#define __access_ok(addr,size) (__user_ok((addr) & get_fs().seg,(size)))
-#define access_ok(type, addr, size)					\
-	({ (void)(type); __access_ok((unsigned long)(addr), size); })
-
-/*
- * The exception table consists of pairs of addresses: the first is the
- * address of an instruction that is allowed to fault, and the second is
- * the address at which the program should continue.  No registers are
- * modified, so it is entirely up to the continuation code to figure out
- * what to do.
- *
- * All the routines below use bits of fixup code that are out of line
- * with the main instruction path.  This means when everything is well,
- * we don't even have to jump over them.  Further, they do not intrude
- * on our cache or tlb entries.
- *
- * There is a special way how to put a range of potentially faulting
- * insns (like twenty ldd/std's with now intervening other instructions)
- * You specify address of first in insn and 0 in fixup and in the next
- * exception_table_entry you specify last potentially faulting insn + 1
- * and in fixup the routine which should handle the fault.
- * That fixup code will get
- * (faulting_insn_address - first_insn_in_the_range_address)/4
- * in %g2 (ie. index of the faulting instruction in the range).
- */
-
-struct exception_table_entry
-{
-        unsigned long insn, fixup;
-};
-
-/* Returns 0 if exception not found and fixup otherwise.  */
-extern unsigned long search_extables_range(unsigned long addr, unsigned long *g2);
-
-extern void __ret_efault(void);
-
-/* Uh, these should become the main single-value transfer routines..
- * They automatically use the right size if we just have the right
- * pointer type..
- *
- * This gets kind of ugly. We want to return _two_ values in "get_user()"
- * and yet we don't want to do any pointers, because that is too much
- * of a performance impact. Thus we have a few rather ugly macros here,
- * and hide all the ugliness from the user.
- */
-#define put_user(x,ptr) ({ \
-unsigned long __pu_addr = (unsigned long)(ptr); \
-__chk_user_ptr(ptr); \
-__put_user_check((__typeof__(*(ptr)))(x),__pu_addr,sizeof(*(ptr))); })
-
-#define get_user(x,ptr) ({ \
-unsigned long __gu_addr = (unsigned long)(ptr); \
-__chk_user_ptr(ptr); \
-__get_user_check((x),__gu_addr,sizeof(*(ptr)),__typeof__(*(ptr))); })
-
-/*
- * The "__xxx" versions do not do address space checking, useful when
- * doing multiple accesses to the same area (the user has to do the
- * checks by hand with "access_ok()")
- */
-#define __put_user(x,ptr) __put_user_nocheck((__typeof__(*(ptr)))(x),(ptr),sizeof(*(ptr)))
-#define __get_user(x,ptr) __get_user_nocheck((x),(ptr),sizeof(*(ptr)),__typeof__(*(ptr)))
-
-struct __large_struct { unsigned long buf[100]; };
-#define __m(x) ((struct __large_struct __user *)(x))
-
-#define __put_user_check(x,addr,size) ({ \
-register int __pu_ret; \
-if (__access_ok(addr,size)) { \
-switch (size) { \
-case 1: __put_user_asm(x,b,addr,__pu_ret); break; \
-case 2: __put_user_asm(x,h,addr,__pu_ret); break; \
-case 4: __put_user_asm(x,,addr,__pu_ret); break; \
-case 8: __put_user_asm(x,d,addr,__pu_ret); break; \
-default: __pu_ret = __put_user_bad(); break; \
-} } else { __pu_ret = -EFAULT; } __pu_ret; })
-
-#define __put_user_nocheck(x,addr,size) ({ \
-register int __pu_ret; \
-switch (size) { \
-case 1: __put_user_asm(x,b,addr,__pu_ret); break; \
-case 2: __put_user_asm(x,h,addr,__pu_ret); break; \
-case 4: __put_user_asm(x,,addr,__pu_ret); break; \
-case 8: __put_user_asm(x,d,addr,__pu_ret); break; \
-default: __pu_ret = __put_user_bad(); break; \
-} __pu_ret; })
-
-#define __put_user_asm(x,size,addr,ret)					\
-__asm__ __volatile__(							\
-	"/* Put user asm, inline. */\n"					\
-"1:\t"	"st"#size " %1, %2\n\t"						\
-	"clr	%0\n"							\
-"2:\n\n\t"								\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"b	2b\n\t"							\
-	" mov	%3, %0\n\t"						\
-        ".previous\n\n\t"						\
-	".section __ex_table,#alloc\n\t"				\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\t"						\
-	".previous\n\n\t"						\
-       : "=&r" (ret) : "r" (x), "m" (*__m(addr)),			\
-	 "i" (-EFAULT))
-
-extern int __put_user_bad(void);
-
-#define __get_user_check(x,addr,size,type) ({ \
-register int __gu_ret; \
-register unsigned long __gu_val; \
-if (__access_ok(addr,size)) { \
-switch (size) { \
-case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
-case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
-case 4: __get_user_asm(__gu_val,,addr,__gu_ret); break; \
-case 8: __get_user_asm(__gu_val,d,addr,__gu_ret); break; \
-default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
-} } else { __gu_val = 0; __gu_ret = -EFAULT; } x = (type) __gu_val; __gu_ret; })
-
-#define __get_user_check_ret(x,addr,size,type,retval) ({ \
-register unsigned long __gu_val __asm__ ("l1"); \
-if (__access_ok(addr,size)) { \
-switch (size) { \
-case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
-case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
-case 4: __get_user_asm_ret(__gu_val,,addr,retval); break; \
-case 8: __get_user_asm_ret(__gu_val,d,addr,retval); break; \
-default: if (__get_user_bad()) return retval; \
-} x = (type) __gu_val; } else return retval; })
-
-#define __get_user_nocheck(x,addr,size,type) ({ \
-register int __gu_ret; \
-register unsigned long __gu_val; \
-switch (size) { \
-case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
-case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
-case 4: __get_user_asm(__gu_val,,addr,__gu_ret); break; \
-case 8: __get_user_asm(__gu_val,d,addr,__gu_ret); break; \
-default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
-} x = (type) __gu_val; __gu_ret; })
-
-#define __get_user_nocheck_ret(x,addr,size,type,retval) ({ \
-register unsigned long __gu_val __asm__ ("l1"); \
-switch (size) { \
-case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
-case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
-case 4: __get_user_asm_ret(__gu_val,,addr,retval); break; \
-case 8: __get_user_asm_ret(__gu_val,d,addr,retval); break; \
-default: if (__get_user_bad()) return retval; \
-} x = (type) __gu_val; })
-
-#define __get_user_asm(x,size,addr,ret)					\
-__asm__ __volatile__(							\
-	"/* Get user asm, inline. */\n"					\
-"1:\t"	"ld"#size " %2, %1\n\t"						\
-	"clr	%0\n"							\
-"2:\n\n\t"								\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"clr	%1\n\t"							\
-	"b	2b\n\t"							\
-	" mov	%3, %0\n\n\t"						\
-	".previous\n\t"							\
-	".section __ex_table,#alloc\n\t"				\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\n\t"						\
-	".previous\n\t"							\
-       : "=&r" (ret), "=&r" (x) : "m" (*__m(addr)),			\
-	 "i" (-EFAULT))
-
-#define __get_user_asm_ret(x,size,addr,retval)				\
-if (__builtin_constant_p(retval) && retval == -EFAULT)			\
-__asm__ __volatile__(							\
-	"/* Get user asm ret, inline. */\n"				\
-"1:\t"	"ld"#size " %1, %0\n\n\t"					\
-	".section __ex_table,#alloc\n\t"				\
-	".align	4\n\t"							\
-	".word	1b,__ret_efault\n\n\t"					\
-	".previous\n\t"							\
-       : "=&r" (x) : "m" (*__m(addr)));					\
-else									\
-__asm__ __volatile__(							\
-	"/* Get user asm ret, inline. */\n"				\
-"1:\t"	"ld"#size " %1, %0\n\n\t"					\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"ret\n\t"							\
-	" restore %%g0, %2, %%o0\n\n\t"					\
-	".previous\n\t"							\
-	".section __ex_table,#alloc\n\t"				\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\n\t"						\
-	".previous\n\t"							\
-       : "=&r" (x) : "m" (*__m(addr)), "i" (retval))
-
-extern int __get_user_bad(void);
-
-extern unsigned long __copy_user(void __user *to, const void __user *from, unsigned long size);
-
-static inline unsigned long copy_to_user(void __user *to, const void *from, unsigned long n)
-{
-	if (n && __access_ok((unsigned long) to, n))
-		return __copy_user(to, (__force void __user *) from, n);
-	else
-		return n;
-}
-
-static inline unsigned long __copy_to_user(void __user *to, const void *from, unsigned long n)
-{
-	return __copy_user(to, (__force void __user *) from, n);
-}
-
-static inline unsigned long copy_from_user(void *to, const void __user *from, unsigned long n)
-{
-	if (n && __access_ok((unsigned long) from, n))
-		return __copy_user((__force void __user *) to, from, n);
-	else
-		return n;
-}
-
-static inline unsigned long __copy_from_user(void *to, const void __user *from, unsigned long n)
-{
-	return __copy_user((__force void __user *) to, from, n);
-}
-
-#define __copy_to_user_inatomic __copy_to_user
-#define __copy_from_user_inatomic __copy_from_user
-
-static inline unsigned long __clear_user(void __user *addr, unsigned long size)
-{
-	unsigned long ret;
-
-	__asm__ __volatile__ (
-		".section __ex_table,#alloc\n\t"
-		".align 4\n\t"
-		".word 1f,3\n\t"
-		".previous\n\t"
-		"mov %2, %%o1\n"
-		"1:\n\t"
-		"call __bzero\n\t"
-		" mov %1, %%o0\n\t"
-		"mov %%o0, %0\n"
-		: "=r" (ret) : "r" (addr), "r" (size) :
-		"o0", "o1", "o2", "o3", "o4", "o5", "o7",
-		"g1", "g2", "g3", "g4", "g5", "g7", "cc");
-
-	return ret;
-}
-
-static inline unsigned long clear_user(void __user *addr, unsigned long n)
-{
-	if (n && __access_ok((unsigned long) addr, n))
-		return __clear_user(addr, n);
-	else
-		return n;
-}
-
-extern long __strncpy_from_user(char *dest, const char __user *src, long count);
-
-static inline long strncpy_from_user(char *dest, const char __user *src, long count)
-{
-	if (__access_ok((unsigned long) src, count))
-		return __strncpy_from_user(dest, src, count);
-	else
-		return -EFAULT;
-}
-
-extern long __strlen_user(const char __user *);
-extern long __strnlen_user(const char __user *, long len);
-
-static inline long strlen_user(const char __user *str)
-{
-	if (!access_ok(VERIFY_READ, str, 0))
-		return 0;
-	else
-		return __strlen_user(str);
-}
-
-static inline long strnlen_user(const char __user *str, long len)
-{
-	if (!access_ok(VERIFY_READ, str, 0))
-		return 0;
-	else
-		return __strnlen_user(str, len);
-}
-
-#endif  /* __ASSEMBLY__ */
-
-#endif /* _ASM_UACCESS_H */
diff --git a/include/asm-sparc/uaccess_32.h b/include/asm-sparc/uaccess_32.h
new file mode 100644
index 00000000000..47d5619d43f
--- /dev/null
+++ b/include/asm-sparc/uaccess_32.h
@@ -0,0 +1,336 @@
+/*
+ * uaccess.h: User space memore access functions.
+ *
+ * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
+ * Copyright (C) 1996,1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+#ifndef _ASM_UACCESS_H
+#define _ASM_UACCESS_H
+
+#ifdef __KERNEL__
+#include <linux/compiler.h>
+#include <linux/sched.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <asm/vac-ops.h>
+#endif
+
+#ifndef __ASSEMBLY__
+
+/* Sparc is not segmented, however we need to be able to fool access_ok()
+ * when doing system calls from kernel mode legitimately.
+ *
+ * "For historical reasons, these macros are grossly misnamed." -Linus
+ */
+
+#define KERNEL_DS   ((mm_segment_t) { 0 })
+#define USER_DS     ((mm_segment_t) { -1 })
+
+#define VERIFY_READ	0
+#define VERIFY_WRITE	1
+
+#define get_ds()	(KERNEL_DS)
+#define get_fs()	(current->thread.current_ds)
+#define set_fs(val)	((current->thread.current_ds) = (val))
+
+#define segment_eq(a,b)	((a).seg == (b).seg)
+
+/* We have there a nice not-mapped page at PAGE_OFFSET - PAGE_SIZE, so that this test
+ * can be fairly lightweight.
+ * No one can read/write anything from userland in the kernel space by setting
+ * large size and address near to PAGE_OFFSET - a fault will break his intentions.
+ */
+#define __user_ok(addr, size) ({ (void)(size); (addr) < STACK_TOP; })
+#define __kernel_ok (segment_eq(get_fs(), KERNEL_DS))
+#define __access_ok(addr,size) (__user_ok((addr) & get_fs().seg,(size)))
+#define access_ok(type, addr, size)					\
+	({ (void)(type); __access_ok((unsigned long)(addr), size); })
+
+/*
+ * The exception table consists of pairs of addresses: the first is the
+ * address of an instruction that is allowed to fault, and the second is
+ * the address at which the program should continue.  No registers are
+ * modified, so it is entirely up to the continuation code to figure out
+ * what to do.
+ *
+ * All the routines below use bits of fixup code that are out of line
+ * with the main instruction path.  This means when everything is well,
+ * we don't even have to jump over them.  Further, they do not intrude
+ * on our cache or tlb entries.
+ *
+ * There is a special way how to put a range of potentially faulting
+ * insns (like twenty ldd/std's with now intervening other instructions)
+ * You specify address of first in insn and 0 in fixup and in the next
+ * exception_table_entry you specify last potentially faulting insn + 1
+ * and in fixup the routine which should handle the fault.
+ * That fixup code will get
+ * (faulting_insn_address - first_insn_in_the_range_address)/4
+ * in %g2 (ie. index of the faulting instruction in the range).
+ */
+
+struct exception_table_entry
+{
+        unsigned long insn, fixup;
+};
+
+/* Returns 0 if exception not found and fixup otherwise.  */
+extern unsigned long search_extables_range(unsigned long addr, unsigned long *g2);
+
+extern void __ret_efault(void);
+
+/* Uh, these should become the main single-value transfer routines..
+ * They automatically use the right size if we just have the right
+ * pointer type..
+ *
+ * This gets kind of ugly. We want to return _two_ values in "get_user()"
+ * and yet we don't want to do any pointers, because that is too much
+ * of a performance impact. Thus we have a few rather ugly macros here,
+ * and hide all the ugliness from the user.
+ */
+#define put_user(x,ptr) ({ \
+unsigned long __pu_addr = (unsigned long)(ptr); \
+__chk_user_ptr(ptr); \
+__put_user_check((__typeof__(*(ptr)))(x),__pu_addr,sizeof(*(ptr))); })
+
+#define get_user(x,ptr) ({ \
+unsigned long __gu_addr = (unsigned long)(ptr); \
+__chk_user_ptr(ptr); \
+__get_user_check((x),__gu_addr,sizeof(*(ptr)),__typeof__(*(ptr))); })
+
+/*
+ * The "__xxx" versions do not do address space checking, useful when
+ * doing multiple accesses to the same area (the user has to do the
+ * checks by hand with "access_ok()")
+ */
+#define __put_user(x,ptr) __put_user_nocheck((__typeof__(*(ptr)))(x),(ptr),sizeof(*(ptr)))
+#define __get_user(x,ptr) __get_user_nocheck((x),(ptr),sizeof(*(ptr)),__typeof__(*(ptr)))
+
+struct __large_struct { unsigned long buf[100]; };
+#define __m(x) ((struct __large_struct __user *)(x))
+
+#define __put_user_check(x,addr,size) ({ \
+register int __pu_ret; \
+if (__access_ok(addr,size)) { \
+switch (size) { \
+case 1: __put_user_asm(x,b,addr,__pu_ret); break; \
+case 2: __put_user_asm(x,h,addr,__pu_ret); break; \
+case 4: __put_user_asm(x,,addr,__pu_ret); break; \
+case 8: __put_user_asm(x,d,addr,__pu_ret); break; \
+default: __pu_ret = __put_user_bad(); break; \
+} } else { __pu_ret = -EFAULT; } __pu_ret; })
+
+#define __put_user_nocheck(x,addr,size) ({ \
+register int __pu_ret; \
+switch (size) { \
+case 1: __put_user_asm(x,b,addr,__pu_ret); break; \
+case 2: __put_user_asm(x,h,addr,__pu_ret); break; \
+case 4: __put_user_asm(x,,addr,__pu_ret); break; \
+case 8: __put_user_asm(x,d,addr,__pu_ret); break; \
+default: __pu_ret = __put_user_bad(); break; \
+} __pu_ret; })
+
+#define __put_user_asm(x,size,addr,ret)					\
+__asm__ __volatile__(							\
+	"/* Put user asm, inline. */\n"					\
+"1:\t"	"st"#size " %1, %2\n\t"						\
+	"clr	%0\n"							\
+"2:\n\n\t"								\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"b	2b\n\t"							\
+	" mov	%3, %0\n\t"						\
+        ".previous\n\n\t"						\
+	".section __ex_table,#alloc\n\t"				\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\t"						\
+	".previous\n\n\t"						\
+       : "=&r" (ret) : "r" (x), "m" (*__m(addr)),			\
+	 "i" (-EFAULT))
+
+extern int __put_user_bad(void);
+
+#define __get_user_check(x,addr,size,type) ({ \
+register int __gu_ret; \
+register unsigned long __gu_val; \
+if (__access_ok(addr,size)) { \
+switch (size) { \
+case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
+case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
+case 4: __get_user_asm(__gu_val,,addr,__gu_ret); break; \
+case 8: __get_user_asm(__gu_val,d,addr,__gu_ret); break; \
+default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
+} } else { __gu_val = 0; __gu_ret = -EFAULT; } x = (type) __gu_val; __gu_ret; })
+
+#define __get_user_check_ret(x,addr,size,type,retval) ({ \
+register unsigned long __gu_val __asm__ ("l1"); \
+if (__access_ok(addr,size)) { \
+switch (size) { \
+case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
+case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
+case 4: __get_user_asm_ret(__gu_val,,addr,retval); break; \
+case 8: __get_user_asm_ret(__gu_val,d,addr,retval); break; \
+default: if (__get_user_bad()) return retval; \
+} x = (type) __gu_val; } else return retval; })
+
+#define __get_user_nocheck(x,addr,size,type) ({ \
+register int __gu_ret; \
+register unsigned long __gu_val; \
+switch (size) { \
+case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
+case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
+case 4: __get_user_asm(__gu_val,,addr,__gu_ret); break; \
+case 8: __get_user_asm(__gu_val,d,addr,__gu_ret); break; \
+default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
+} x = (type) __gu_val; __gu_ret; })
+
+#define __get_user_nocheck_ret(x,addr,size,type,retval) ({ \
+register unsigned long __gu_val __asm__ ("l1"); \
+switch (size) { \
+case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
+case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
+case 4: __get_user_asm_ret(__gu_val,,addr,retval); break; \
+case 8: __get_user_asm_ret(__gu_val,d,addr,retval); break; \
+default: if (__get_user_bad()) return retval; \
+} x = (type) __gu_val; })
+
+#define __get_user_asm(x,size,addr,ret)					\
+__asm__ __volatile__(							\
+	"/* Get user asm, inline. */\n"					\
+"1:\t"	"ld"#size " %2, %1\n\t"						\
+	"clr	%0\n"							\
+"2:\n\n\t"								\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"clr	%1\n\t"							\
+	"b	2b\n\t"							\
+	" mov	%3, %0\n\n\t"						\
+	".previous\n\t"							\
+	".section __ex_table,#alloc\n\t"				\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\n\t"						\
+	".previous\n\t"							\
+       : "=&r" (ret), "=&r" (x) : "m" (*__m(addr)),			\
+	 "i" (-EFAULT))
+
+#define __get_user_asm_ret(x,size,addr,retval)				\
+if (__builtin_constant_p(retval) && retval == -EFAULT)			\
+__asm__ __volatile__(							\
+	"/* Get user asm ret, inline. */\n"				\
+"1:\t"	"ld"#size " %1, %0\n\n\t"					\
+	".section __ex_table,#alloc\n\t"				\
+	".align	4\n\t"							\
+	".word	1b,__ret_efault\n\n\t"					\
+	".previous\n\t"							\
+       : "=&r" (x) : "m" (*__m(addr)));					\
+else									\
+__asm__ __volatile__(							\
+	"/* Get user asm ret, inline. */\n"				\
+"1:\t"	"ld"#size " %1, %0\n\n\t"					\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"ret\n\t"							\
+	" restore %%g0, %2, %%o0\n\n\t"					\
+	".previous\n\t"							\
+	".section __ex_table,#alloc\n\t"				\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\n\t"						\
+	".previous\n\t"							\
+       : "=&r" (x) : "m" (*__m(addr)), "i" (retval))
+
+extern int __get_user_bad(void);
+
+extern unsigned long __copy_user(void __user *to, const void __user *from, unsigned long size);
+
+static inline unsigned long copy_to_user(void __user *to, const void *from, unsigned long n)
+{
+	if (n && __access_ok((unsigned long) to, n))
+		return __copy_user(to, (__force void __user *) from, n);
+	else
+		return n;
+}
+
+static inline unsigned long __copy_to_user(void __user *to, const void *from, unsigned long n)
+{
+	return __copy_user(to, (__force void __user *) from, n);
+}
+
+static inline unsigned long copy_from_user(void *to, const void __user *from, unsigned long n)
+{
+	if (n && __access_ok((unsigned long) from, n))
+		return __copy_user((__force void __user *) to, from, n);
+	else
+		return n;
+}
+
+static inline unsigned long __copy_from_user(void *to, const void __user *from, unsigned long n)
+{
+	return __copy_user((__force void __user *) to, from, n);
+}
+
+#define __copy_to_user_inatomic __copy_to_user
+#define __copy_from_user_inatomic __copy_from_user
+
+static inline unsigned long __clear_user(void __user *addr, unsigned long size)
+{
+	unsigned long ret;
+
+	__asm__ __volatile__ (
+		".section __ex_table,#alloc\n\t"
+		".align 4\n\t"
+		".word 1f,3\n\t"
+		".previous\n\t"
+		"mov %2, %%o1\n"
+		"1:\n\t"
+		"call __bzero\n\t"
+		" mov %1, %%o0\n\t"
+		"mov %%o0, %0\n"
+		: "=r" (ret) : "r" (addr), "r" (size) :
+		"o0", "o1", "o2", "o3", "o4", "o5", "o7",
+		"g1", "g2", "g3", "g4", "g5", "g7", "cc");
+
+	return ret;
+}
+
+static inline unsigned long clear_user(void __user *addr, unsigned long n)
+{
+	if (n && __access_ok((unsigned long) addr, n))
+		return __clear_user(addr, n);
+	else
+		return n;
+}
+
+extern long __strncpy_from_user(char *dest, const char __user *src, long count);
+
+static inline long strncpy_from_user(char *dest, const char __user *src, long count)
+{
+	if (__access_ok((unsigned long) src, count))
+		return __strncpy_from_user(dest, src, count);
+	else
+		return -EFAULT;
+}
+
+extern long __strlen_user(const char __user *);
+extern long __strnlen_user(const char __user *, long len);
+
+static inline long strlen_user(const char __user *str)
+{
+	if (!access_ok(VERIFY_READ, str, 0))
+		return 0;
+	else
+		return __strlen_user(str);
+}
+
+static inline long strnlen_user(const char __user *str, long len)
+{
+	if (!access_ok(VERIFY_READ, str, 0))
+		return 0;
+	else
+		return __strnlen_user(str, len);
+}
+
+#endif  /* __ASSEMBLY__ */
+
+#endif /* _ASM_UACCESS_H */
diff --git a/include/asm-sparc/uaccess_64.h b/include/asm-sparc/uaccess_64.h
new file mode 100644
index 00000000000..296ef30e05c
--- /dev/null
+++ b/include/asm-sparc/uaccess_64.h
@@ -0,0 +1,273 @@
+#ifndef _ASM_UACCESS_H
+#define _ASM_UACCESS_H
+
+/*
+ * User space memory access functions
+ */
+
+#ifdef __KERNEL__
+#include <linux/compiler.h>
+#include <linux/sched.h>
+#include <linux/string.h>
+#include <asm/asi.h>
+#include <asm/system.h>
+#include <asm/spitfire.h>
+#include <asm-generic/uaccess.h>
+#endif
+
+#ifndef __ASSEMBLY__
+
+/*
+ * Sparc64 is segmented, though more like the M68K than the I386.
+ * We use the secondary ASI to address user memory, which references a
+ * completely different VM map, thus there is zero chance of the user
+ * doing something queer and tricking us into poking kernel memory.
+ *
+ * What is left here is basically what is needed for the other parts of
+ * the kernel that expect to be able to manipulate, erum, "segments".
+ * Or perhaps more properly, permissions.
+ *
+ * "For historical reasons, these macros are grossly misnamed." -Linus
+ */
+
+#define KERNEL_DS   ((mm_segment_t) { ASI_P })
+#define USER_DS     ((mm_segment_t) { ASI_AIUS })	/* har har har */
+
+#define VERIFY_READ	0
+#define VERIFY_WRITE	1
+
+#define get_fs() ((mm_segment_t) { get_thread_current_ds() })
+#define get_ds() (KERNEL_DS)
+
+#define segment_eq(a,b)  ((a).seg == (b).seg)
+
+#define set_fs(val)								\
+do {										\
+	set_thread_current_ds((val).seg);					\
+	__asm__ __volatile__ ("wr %%g0, %0, %%asi" : : "r" ((val).seg));	\
+} while(0)
+
+static inline int __access_ok(const void __user * addr, unsigned long size)
+{
+	return 1;
+}
+
+static inline int access_ok(int type, const void __user * addr, unsigned long size)
+{
+	return 1;
+}
+
+/*
+ * The exception table consists of pairs of addresses: the first is the
+ * address of an instruction that is allowed to fault, and the second is
+ * the address at which the program should continue.  No registers are
+ * modified, so it is entirely up to the continuation code to figure out
+ * what to do.
+ *
+ * All the routines below use bits of fixup code that are out of line
+ * with the main instruction path.  This means when everything is well,
+ * we don't even have to jump over them.  Further, they do not intrude
+ * on our cache or tlb entries.
+ */
+
+struct exception_table_entry {
+        unsigned int insn, fixup;
+};
+
+extern void __ret_efault(void);
+extern void __retl_efault(void);
+
+/* Uh, these should become the main single-value transfer routines..
+ * They automatically use the right size if we just have the right
+ * pointer type..
+ *
+ * This gets kind of ugly. We want to return _two_ values in "get_user()"
+ * and yet we don't want to do any pointers, because that is too much
+ * of a performance impact. Thus we have a few rather ugly macros here,
+ * and hide all the ugliness from the user.
+ */
+#define put_user(x,ptr) ({ \
+unsigned long __pu_addr = (unsigned long)(ptr); \
+__chk_user_ptr(ptr); \
+__put_user_nocheck((__typeof__(*(ptr)))(x),__pu_addr,sizeof(*(ptr))); })
+
+#define get_user(x,ptr) ({ \
+unsigned long __gu_addr = (unsigned long)(ptr); \
+__chk_user_ptr(ptr); \
+__get_user_nocheck((x),__gu_addr,sizeof(*(ptr)),__typeof__(*(ptr))); })
+
+#define __put_user(x,ptr) put_user(x,ptr)
+#define __get_user(x,ptr) get_user(x,ptr)
+
+struct __large_struct { unsigned long buf[100]; };
+#define __m(x) ((struct __large_struct *)(x))
+
+#define __put_user_nocheck(data,addr,size) ({ \
+register int __pu_ret; \
+switch (size) { \
+case 1: __put_user_asm(data,b,addr,__pu_ret); break; \
+case 2: __put_user_asm(data,h,addr,__pu_ret); break; \
+case 4: __put_user_asm(data,w,addr,__pu_ret); break; \
+case 8: __put_user_asm(data,x,addr,__pu_ret); break; \
+default: __pu_ret = __put_user_bad(); break; \
+} __pu_ret; })
+
+#define __put_user_asm(x,size,addr,ret)					\
+__asm__ __volatile__(							\
+	"/* Put user asm, inline. */\n"					\
+"1:\t"	"st"#size "a %1, [%2] %%asi\n\t"				\
+	"clr	%0\n"							\
+"2:\n\n\t"								\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"sethi	%%hi(2b), %0\n\t"					\
+	"jmpl	%0 + %%lo(2b), %%g0\n\t"				\
+	" mov	%3, %0\n\n\t"						\
+	".previous\n\t"							\
+	".section __ex_table,\"a\"\n\t"					\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\t"						\
+	".previous\n\n\t"						\
+       : "=r" (ret) : "r" (x), "r" (__m(addr)),				\
+	 "i" (-EFAULT))
+
+extern int __put_user_bad(void);
+
+#define __get_user_nocheck(data,addr,size,type) ({ \
+register int __gu_ret; \
+register unsigned long __gu_val; \
+switch (size) { \
+case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
+case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
+case 4: __get_user_asm(__gu_val,uw,addr,__gu_ret); break; \
+case 8: __get_user_asm(__gu_val,x,addr,__gu_ret); break; \
+default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
+} data = (type) __gu_val; __gu_ret; })
+
+#define __get_user_nocheck_ret(data,addr,size,type,retval) ({ \
+register unsigned long __gu_val __asm__ ("l1"); \
+switch (size) { \
+case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
+case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
+case 4: __get_user_asm_ret(__gu_val,uw,addr,retval); break; \
+case 8: __get_user_asm_ret(__gu_val,x,addr,retval); break; \
+default: if (__get_user_bad()) return retval; \
+} data = (type) __gu_val; })
+
+#define __get_user_asm(x,size,addr,ret)					\
+__asm__ __volatile__(							\
+	"/* Get user asm, inline. */\n"					\
+"1:\t"	"ld"#size "a [%2] %%asi, %1\n\t"				\
+	"clr	%0\n"							\
+"2:\n\n\t"								\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"sethi	%%hi(2b), %0\n\t"					\
+	"clr	%1\n\t"							\
+	"jmpl	%0 + %%lo(2b), %%g0\n\t"				\
+	" mov	%3, %0\n\n\t"						\
+	".previous\n\t"							\
+	".section __ex_table,\"a\"\n\t"					\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\n\t"						\
+	".previous\n\t"							\
+       : "=r" (ret), "=r" (x) : "r" (__m(addr)),			\
+	 "i" (-EFAULT))
+
+#define __get_user_asm_ret(x,size,addr,retval)				\
+if (__builtin_constant_p(retval) && retval == -EFAULT)			\
+__asm__ __volatile__(							\
+	"/* Get user asm ret, inline. */\n"				\
+"1:\t"	"ld"#size "a [%1] %%asi, %0\n\n\t"				\
+	".section __ex_table,\"a\"\n\t"					\
+	".align	4\n\t"							\
+	".word	1b,__ret_efault\n\n\t"					\
+	".previous\n\t"							\
+       : "=r" (x) : "r" (__m(addr)));					\
+else									\
+__asm__ __volatile__(							\
+	"/* Get user asm ret, inline. */\n"				\
+"1:\t"	"ld"#size "a [%1] %%asi, %0\n\n\t"				\
+	".section .fixup,#alloc,#execinstr\n\t"				\
+	".align	4\n"							\
+"3:\n\t"								\
+	"ret\n\t"							\
+	" restore %%g0, %2, %%o0\n\n\t"					\
+	".previous\n\t"							\
+	".section __ex_table,\"a\"\n\t"					\
+	".align	4\n\t"							\
+	".word	1b, 3b\n\n\t"						\
+	".previous\n\t"							\
+       : "=r" (x) : "r" (__m(addr)), "i" (retval))
+
+extern int __get_user_bad(void);
+
+extern unsigned long __must_check ___copy_from_user(void *to,
+						    const void __user *from,
+						    unsigned long size);
+extern unsigned long copy_from_user_fixup(void *to, const void __user *from,
+					  unsigned long size);
+static inline unsigned long __must_check
+copy_from_user(void *to, const void __user *from, unsigned long size)
+{
+	unsigned long ret = ___copy_from_user(to, from, size);
+
+	if (unlikely(ret))
+		ret = copy_from_user_fixup(to, from, size);
+	return ret;
+}
+#define __copy_from_user copy_from_user
+
+extern unsigned long __must_check ___copy_to_user(void __user *to,
+						  const void *from,
+						  unsigned long size);
+extern unsigned long copy_to_user_fixup(void __user *to, const void *from,
+					unsigned long size);
+static inline unsigned long __must_check
+copy_to_user(void __user *to, const void *from, unsigned long size)
+{
+	unsigned long ret = ___copy_to_user(to, from, size);
+
+	if (unlikely(ret))
+		ret = copy_to_user_fixup(to, from, size);
+	return ret;
+}
+#define __copy_to_user copy_to_user
+
+extern unsigned long __must_check ___copy_in_user(void __user *to,
+						  const void __user *from,
+						  unsigned long size);
+extern unsigned long copy_in_user_fixup(void __user *to, void __user *from,
+					unsigned long size);
+static inline unsigned long __must_check
+copy_in_user(void __user *to, void __user *from, unsigned long size)
+{
+	unsigned long ret = ___copy_in_user(to, from, size);
+
+	if (unlikely(ret))
+		ret = copy_in_user_fixup(to, from, size);
+	return ret;
+}
+#define __copy_in_user copy_in_user
+
+extern unsigned long __must_check __clear_user(void __user *, unsigned long);
+
+#define clear_user __clear_user
+
+extern long __must_check __strncpy_from_user(char *dest, const char __user *src, long count);
+
+#define strncpy_from_user __strncpy_from_user
+
+extern long __strlen_user(const char __user *);
+extern long __strnlen_user(const char __user *, long len);
+
+#define strlen_user __strlen_user
+#define strnlen_user __strnlen_user
+#define __copy_to_user_inatomic __copy_to_user
+#define __copy_from_user_inatomic __copy_from_user
+
+#endif  /* __ASSEMBLY__ */
+
+#endif /* _ASM_UACCESS_H */
diff --git a/include/asm-sparc/uctx.h b/include/asm-sparc/uctx.h
new file mode 100644
index 00000000000..dc937c75ffd
--- /dev/null
+++ b/include/asm-sparc/uctx.h
@@ -0,0 +1,71 @@
+/*
+ * uctx.h: Sparc64 {set,get}context() register state layouts.
+ *
+ * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
+ */
+
+#ifndef __SPARC64_UCTX_H
+#define __SPARC64_UCTX_H
+
+#define MC_TSTATE	0
+#define MC_PC		1
+#define MC_NPC		2
+#define MC_Y		3
+#define MC_G1		4
+#define MC_G2		5
+#define MC_G3		6
+#define MC_G4		7
+#define MC_G5		8
+#define MC_G6		9
+#define MC_G7		10
+#define MC_O0		11
+#define MC_O1		12
+#define MC_O2		13
+#define MC_O3		14
+#define MC_O4		15
+#define MC_O5		16
+#define MC_O6		17
+#define MC_O7		18
+#define MC_NGREG	19
+
+typedef unsigned long mc_greg_t;
+typedef mc_greg_t mc_gregset_t[MC_NGREG];
+
+#define MC_MAXFPQ	16
+struct mc_fq {
+	unsigned long	*mcfq_addr;
+	unsigned int	mcfq_insn;
+};
+
+struct mc_fpu {
+	union {
+		unsigned int	sregs[32];
+		unsigned long	dregs[32];
+		long double	qregs[16];
+	} mcfpu_fregs;
+	unsigned long	mcfpu_fsr;
+	unsigned long	mcfpu_fprs;
+	unsigned long	mcfpu_gsr;
+	struct mc_fq	*mcfpu_fq;
+	unsigned char	mcfpu_qcnt;
+	unsigned char	mcfpu_qentsz;
+	unsigned char	mcfpu_enab;
+};
+typedef struct mc_fpu mc_fpu_t;
+
+typedef struct {
+	mc_gregset_t	mc_gregs;
+	mc_greg_t	mc_fp;
+	mc_greg_t	mc_i7;
+	mc_fpu_t	mc_fpregs;
+} mcontext_t;
+
+struct ucontext {
+	struct ucontext		*uc_link;
+	unsigned long		uc_flags;
+	sigset_t		uc_sigmask;
+	mcontext_t		uc_mcontext;
+};
+typedef struct ucontext ucontext_t;
+
+#endif /* __SPARC64_UCTX_H */
diff --git a/include/asm-sparc/unistd.h b/include/asm-sparc/unistd.h
index 2338a027637..3c2609618a0 100644
--- a/include/asm-sparc/unistd.h
+++ b/include/asm-sparc/unistd.h
@@ -1,378 +1,8 @@
-#ifndef _SPARC_UNISTD_H
-#define _SPARC_UNISTD_H
-
-/*
- * System calls under the Sparc.
- *
- * Don't be scared by the ugly clobbers, it is the only way I can
- * think of right now to force the arguments into fixed registers
- * before the trap into the system call with gcc 'asm' statements.
- *
- * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
- *
- * SunOS compatibility based upon preliminary work which is:
- *
- * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu)
- */
-
-#define __NR_restart_syscall      0 /* Linux Specific				   */
-#define __NR_exit                 1 /* Common                                      */
-#define __NR_fork                 2 /* Common                                      */
-#define __NR_read                 3 /* Common                                      */
-#define __NR_write                4 /* Common                                      */
-#define __NR_open                 5 /* Common                                      */
-#define __NR_close                6 /* Common                                      */
-#define __NR_wait4                7 /* Common                                      */
-#define __NR_creat                8 /* Common                                      */
-#define __NR_link                 9 /* Common                                      */
-#define __NR_unlink              10 /* Common                                      */
-#define __NR_execv               11 /* SunOS Specific                              */
-#define __NR_chdir               12 /* Common                                      */
-#define __NR_chown		 13 /* Common					   */
-#define __NR_mknod               14 /* Common                                      */
-#define __NR_chmod               15 /* Common                                      */
-#define __NR_lchown              16 /* Common                                      */
-#define __NR_brk                 17 /* Common                                      */
-#define __NR_perfctr             18 /* Performance counter operations              */
-#define __NR_lseek               19 /* Common                                      */
-#define __NR_getpid              20 /* Common                                      */
-#define __NR_capget		 21 /* Linux Specific				   */
-#define __NR_capset		 22 /* Linux Specific				   */
-#define __NR_setuid              23 /* Implemented via setreuid in SunOS           */
-#define __NR_getuid              24 /* Common                                      */
-#define __NR_vmsplice	         25 /* ENOSYS under SunOS			   */
-#define __NR_ptrace              26 /* Common                                      */
-#define __NR_alarm               27 /* Implemented via setitimer in SunOS          */
-#define __NR_sigaltstack	 28 /* Common					   */
-#define __NR_pause               29 /* Is sigblock(0)->sigpause() in SunOS         */
-#define __NR_utime               30 /* Implemented via utimes() under SunOS        */
-#define __NR_lchown32            31 /* Linux sparc32 specific                      */
-#define __NR_fchown32            32 /* Linux sparc32 specific                      */
-#define __NR_access              33 /* Common                                      */
-#define __NR_nice                34 /* Implemented via get/setpriority() in SunOS  */
-#define __NR_chown32             35 /* Linux sparc32 specific                      */
-#define __NR_sync                36 /* Common                                      */
-#define __NR_kill                37 /* Common                                      */
-#define __NR_stat                38 /* Common                                      */
-#define __NR_sendfile		 39 /* Linux Specific				   */
-#define __NR_lstat               40 /* Common                                      */
-#define __NR_dup                 41 /* Common                                      */
-#define __NR_pipe                42 /* Common                                      */
-#define __NR_times               43 /* Implemented via getrusage() in SunOS        */
-#define __NR_getuid32            44 /* Linux sparc32 specific                      */
-#define __NR_umount2             45 /* Linux Specific                              */
-#define __NR_setgid              46 /* Implemented via setregid() in SunOS         */
-#define __NR_getgid              47 /* Common                                      */
-#define __NR_signal              48 /* Implemented via sigvec() in SunOS           */
-#define __NR_geteuid             49 /* SunOS calls getuid()                        */
-#define __NR_getegid             50 /* SunOS calls getgid()                        */
-#define __NR_acct                51 /* Common                                      */
-/* #define __NR_memory_ordering  52    Linux sparc64 specific			   */
-#define __NR_getgid32            53 /* Linux sparc32 specific                      */
-#define __NR_ioctl               54 /* Common                                      */
-#define __NR_reboot              55 /* Common                                      */
-#define __NR_mmap2		 56 /* Linux sparc32 Specific			   */
-#define __NR_symlink             57 /* Common                                      */
-#define __NR_readlink            58 /* Common                                      */
-#define __NR_execve              59 /* Common                                      */
-#define __NR_umask               60 /* Common                                      */
-#define __NR_chroot              61 /* Common                                      */
-#define __NR_fstat               62 /* Common                                      */
-#define __NR_fstat64		 63 /* Linux Specific			           */
-#define __NR_getpagesize         64 /* Common                                      */
-#define __NR_msync               65 /* Common in newer 1.3.x revs...               */
-#define __NR_vfork               66 /* Common                                      */
-#define __NR_pread64             67 /* Linux Specific                              */
-#define __NR_pwrite64            68 /* Linux Specific                              */
-#define __NR_geteuid32           69 /* Linux sparc32, sbrk under SunOS             */
-#define __NR_getegid32           70 /* Linux sparc32, sstk under SunOS             */
-#define __NR_mmap                71 /* Common                                      */
-#define __NR_setreuid32          72 /* Linux sparc32, vadvise under SunOS          */
-#define __NR_munmap              73 /* Common                                      */
-#define __NR_mprotect            74 /* Common                                      */
-#define __NR_madvise             75 /* Common                                      */
-#define __NR_vhangup             76 /* Common                                      */
-#define __NR_truncate64		 77 /* Linux sparc32 Specific			   */
-#define __NR_mincore             78 /* Common                                      */
-#define __NR_getgroups           79 /* Common                                      */
-#define __NR_setgroups           80 /* Common                                      */
-#define __NR_getpgrp             81 /* Common                                      */
-#define __NR_setgroups32         82 /* Linux sparc32, setpgrp under SunOS          */
-#define __NR_setitimer           83 /* Common                                      */
-#define __NR_ftruncate64	 84 /* Linux sparc32 Specific			   */
-#define __NR_swapon              85 /* Common                                      */
-#define __NR_getitimer           86 /* Common                                      */
-#define __NR_setuid32            87 /* Linux sparc32, gethostname under SunOS      */
-#define __NR_sethostname         88 /* Common                                      */
-#define __NR_setgid32            89 /* Linux sparc32, getdtablesize under SunOS    */
-#define __NR_dup2                90 /* Common                                      */
-#define __NR_setfsuid32          91 /* Linux sparc32, getdopt under SunOS          */
-#define __NR_fcntl               92 /* Common                                      */
-#define __NR_select              93 /* Common                                      */
-#define __NR_setfsgid32          94 /* Linux sparc32, setdopt under SunOS          */
-#define __NR_fsync               95 /* Common                                      */
-#define __NR_setpriority         96 /* Common                                      */
-#define __NR_socket              97 /* Common                                      */
-#define __NR_connect             98 /* Common                                      */
-#define __NR_accept              99 /* Common                                      */
-#define __NR_getpriority        100 /* Common                                      */
-#define __NR_rt_sigreturn       101 /* Linux Specific                              */
-#define __NR_rt_sigaction       102 /* Linux Specific                              */
-#define __NR_rt_sigprocmask     103 /* Linux Specific                              */
-#define __NR_rt_sigpending      104 /* Linux Specific                              */
-#define __NR_rt_sigtimedwait    105 /* Linux Specific                              */
-#define __NR_rt_sigqueueinfo    106 /* Linux Specific                              */
-#define __NR_rt_sigsuspend      107 /* Linux Specific                              */
-#define __NR_setresuid32        108 /* Linux Specific, sigvec under SunOS	   */
-#define __NR_getresuid32        109 /* Linux Specific, sigblock under SunOS	   */
-#define __NR_setresgid32        110 /* Linux Specific, sigsetmask under SunOS	   */
-#define __NR_getresgid32        111 /* Linux Specific, sigpause under SunOS	   */
-#define __NR_setregid32         112 /* Linux sparc32, sigstack under SunOS         */
-#define __NR_recvmsg            113 /* Common                                      */
-#define __NR_sendmsg            114 /* Common                                      */
-#define __NR_getgroups32        115 /* Linux sparc32, vtrace under SunOS           */
-#define __NR_gettimeofday       116 /* Common                                      */
-#define __NR_getrusage          117 /* Common                                      */
-#define __NR_getsockopt         118 /* Common                                      */
-#define __NR_getcwd		119 /* Linux Specific				   */
-#define __NR_readv              120 /* Common                                      */
-#define __NR_writev             121 /* Common                                      */
-#define __NR_settimeofday       122 /* Common                                      */
-#define __NR_fchown             123 /* Common                                      */
-#define __NR_fchmod             124 /* Common                                      */
-#define __NR_recvfrom           125 /* Common                                      */
-#define __NR_setreuid           126 /* Common                                      */
-#define __NR_setregid           127 /* Common                                      */
-#define __NR_rename             128 /* Common                                      */
-#define __NR_truncate           129 /* Common                                      */
-#define __NR_ftruncate          130 /* Common                                      */
-#define __NR_flock              131 /* Common                                      */
-#define __NR_lstat64		132 /* Linux Specific			           */
-#define __NR_sendto             133 /* Common                                      */
-#define __NR_shutdown           134 /* Common                                      */
-#define __NR_socketpair         135 /* Common                                      */
-#define __NR_mkdir              136 /* Common                                      */
-#define __NR_rmdir              137 /* Common                                      */
-#define __NR_utimes             138 /* SunOS Specific                              */
-#define __NR_stat64		139 /* Linux Specific			           */
-#define __NR_sendfile64         140 /* adjtime under SunOS                         */
-#define __NR_getpeername        141 /* Common                                      */
-#define __NR_futex              142 /* gethostid under SunOS                       */
-#define __NR_gettid             143 /* ENOSYS under SunOS                          */
-#define __NR_getrlimit          144 /* Common                                      */
-#define __NR_setrlimit          145 /* Common                                      */
-#define __NR_pivot_root		146 /* Linux Specific, killpg under SunOS          */
-#define __NR_prctl		147 /* ENOSYS under SunOS                          */
-#define __NR_pciconfig_read	148 /* ENOSYS under SunOS                          */
-#define __NR_pciconfig_write	149 /* ENOSYS under SunOS                          */
-#define __NR_getsockname        150 /* Common                                      */
-#define __NR_inotify_init       151 /* Linux specific                              */
-#define __NR_inotify_add_watch  152 /* Linux specific                              */
-#define __NR_poll               153 /* Common                                      */
-#define __NR_getdents64		154 /* Linux specific				   */
-#define __NR_fcntl64		155 /* Linux sparc32 Specific                      */
-#define __NR_inotify_rm_watch   156 /* Linux specific				   */
-#define __NR_statfs             157 /* Common                                      */
-#define __NR_fstatfs            158 /* Common                                      */
-#define __NR_umount             159 /* Common                                      */
-#define __NR_sched_set_affinity 160 /* Linux specific, async_daemon under SunOS    */
-#define __NR_sched_get_affinity 161 /* Linux specific, getfh under SunOS           */
-#define __NR_getdomainname      162 /* SunOS Specific                              */
-#define __NR_setdomainname      163 /* Common                                      */
-/* #define __NR_utrap_install   164    Linux sparc64 specific			   */
-#define __NR_quotactl           165 /* Common                                      */
-#define __NR_set_tid_address    166 /* Linux specific, exportfs under SunOS        */
-#define __NR_mount              167 /* Common                                      */
-#define __NR_ustat              168 /* Common                                      */
-#define __NR_setxattr           169 /* SunOS: semsys                               */
-#define __NR_lsetxattr          170 /* SunOS: msgsys                               */
-#define __NR_fsetxattr          171 /* SunOS: shmsys                               */
-#define __NR_getxattr           172 /* SunOS: auditsys                             */
-#define __NR_lgetxattr          173 /* SunOS: rfssys                               */
-#define __NR_getdents           174 /* Common                                      */
-#define __NR_setsid             175 /* Common                                      */
-#define __NR_fchdir             176 /* Common                                      */
-#define __NR_fgetxattr          177 /* SunOS: fchroot                              */
-#define __NR_listxattr          178 /* SunOS: vpixsys                              */
-#define __NR_llistxattr         179 /* SunOS: aioread                              */
-#define __NR_flistxattr         180 /* SunOS: aiowrite                             */
-#define __NR_removexattr        181 /* SunOS: aiowait                              */
-#define __NR_lremovexattr       182 /* SunOS: aiocancel                            */
-#define __NR_sigpending         183 /* Common                                      */
-#define __NR_query_module	184 /* Linux Specific				   */
-#define __NR_setpgid            185 /* Common                                      */
-#define __NR_fremovexattr       186 /* SunOS: pathconf                             */
-#define __NR_tkill              187 /* SunOS: fpathconf                            */
-#define __NR_exit_group		188 /* Linux specific, sysconf undef SunOS         */
-#define __NR_uname              189 /* Linux Specific                              */
-#define __NR_init_module        190 /* Linux Specific                              */
-#define __NR_personality        191 /* Linux Specific                              */
-#define __NR_remap_file_pages   192 /* Linux Specific                              */
-#define __NR_epoll_create       193 /* Linux Specific                              */
-#define __NR_epoll_ctl          194 /* Linux Specific                              */
-#define __NR_epoll_wait         195 /* Linux Specific                              */
-#define __NR_ioprio_set         196 /* Linux Specific                              */
-#define __NR_getppid            197 /* Linux Specific                              */
-#define __NR_sigaction          198 /* Linux Specific                              */
-#define __NR_sgetmask           199 /* Linux Specific                              */
-#define __NR_ssetmask           200 /* Linux Specific                              */
-#define __NR_sigsuspend         201 /* Linux Specific                              */
-#define __NR_oldlstat           202 /* Linux Specific                              */
-#define __NR_uselib             203 /* Linux Specific                              */
-#define __NR_readdir            204 /* Linux Specific                              */
-#define __NR_readahead          205 /* Linux Specific                              */
-#define __NR_socketcall         206 /* Linux Specific                              */
-#define __NR_syslog             207 /* Linux Specific                              */
-#define __NR_lookup_dcookie     208 /* Linux Specific                              */
-#define __NR_fadvise64          209 /* Linux Specific                              */
-#define __NR_fadvise64_64       210 /* Linux Specific                              */
-#define __NR_tgkill             211 /* Linux Specific                              */
-#define __NR_waitpid            212 /* Linux Specific                              */
-#define __NR_swapoff            213 /* Linux Specific                              */
-#define __NR_sysinfo            214 /* Linux Specific                              */
-#define __NR_ipc                215 /* Linux Specific                              */
-#define __NR_sigreturn          216 /* Linux Specific                              */
-#define __NR_clone              217 /* Linux Specific                              */
-#define __NR_ioprio_get         218 /* Linux Specific                              */
-#define __NR_adjtimex           219 /* Linux Specific                              */
-#define __NR_sigprocmask        220 /* Linux Specific                              */
-#define __NR_create_module      221 /* Linux Specific                              */
-#define __NR_delete_module      222 /* Linux Specific                              */
-#define __NR_get_kernel_syms    223 /* Linux Specific                              */
-#define __NR_getpgid            224 /* Linux Specific                              */
-#define __NR_bdflush            225 /* Linux Specific                              */
-#define __NR_sysfs              226 /* Linux Specific                              */
-#define __NR_afs_syscall        227 /* Linux Specific                              */
-#define __NR_setfsuid           228 /* Linux Specific                              */
-#define __NR_setfsgid           229 /* Linux Specific                              */
-#define __NR__newselect         230 /* Linux Specific                              */
-#define __NR_time               231 /* Linux Specific                              */
-#define __NR_splice             232 /* Linux Specific                              */
-#define __NR_stime              233 /* Linux Specific                              */
-#define __NR_statfs64           234 /* Linux Specific                              */
-#define __NR_fstatfs64          235 /* Linux Specific                              */
-#define __NR__llseek            236 /* Linux Specific                              */
-#define __NR_mlock              237
-#define __NR_munlock            238
-#define __NR_mlockall           239
-#define __NR_munlockall         240
-#define __NR_sched_setparam     241
-#define __NR_sched_getparam     242
-#define __NR_sched_setscheduler 243
-#define __NR_sched_getscheduler 244
-#define __NR_sched_yield        245
-#define __NR_sched_get_priority_max 246
-#define __NR_sched_get_priority_min 247
-#define __NR_sched_rr_get_interval  248
-#define __NR_nanosleep          249
-#define __NR_mremap             250
-#define __NR__sysctl            251
-#define __NR_getsid             252
-#define __NR_fdatasync          253
-#define __NR_nfsservctl         254
-#define __NR_sync_file_range	255
-#define __NR_clock_settime	256
-#define __NR_clock_gettime	257
-#define __NR_clock_getres	258
-#define __NR_clock_nanosleep	259
-#define __NR_sched_getaffinity	260
-#define __NR_sched_setaffinity	261
-#define __NR_timer_settime	262
-#define __NR_timer_gettime	263
-#define __NR_timer_getoverrun	264
-#define __NR_timer_delete	265
-#define __NR_timer_create	266
-/* #define __NR_vserver		267 Reserved for VSERVER */
-#define __NR_io_setup		268
-#define __NR_io_destroy		269
-#define __NR_io_submit		270
-#define __NR_io_cancel		271
-#define __NR_io_getevents	272
-#define __NR_mq_open		273
-#define __NR_mq_unlink		274
-#define __NR_mq_timedsend	275
-#define __NR_mq_timedreceive	276
-#define __NR_mq_notify		277
-#define __NR_mq_getsetattr	278
-#define __NR_waitid		279
-#define __NR_tee		280
-#define __NR_add_key		281
-#define __NR_request_key	282
-#define __NR_keyctl		283
-#define __NR_openat		284
-#define __NR_mkdirat		285
-#define __NR_mknodat		286
-#define __NR_fchownat		287
-#define __NR_futimesat		288
-#define __NR_fstatat64		289
-#define __NR_unlinkat		290
-#define __NR_renameat		291
-#define __NR_linkat		292
-#define __NR_symlinkat		293
-#define __NR_readlinkat		294
-#define __NR_fchmodat		295
-#define __NR_faccessat		296
-#define __NR_pselect6		297
-#define __NR_ppoll		298
-#define __NR_unshare		299
-#define __NR_set_robust_list	300
-#define __NR_get_robust_list	301
-#define __NR_migrate_pages	302
-#define __NR_mbind		303
-#define __NR_get_mempolicy	304
-#define __NR_set_mempolicy	305
-#define __NR_kexec_load		306
-#define __NR_move_pages		307
-#define __NR_getcpu		308
-#define __NR_epoll_pwait	309
-#define __NR_utimensat		310
-#define __NR_signalfd		311
-#define __NR_timerfd_create	312
-#define __NR_eventfd		313
-#define __NR_fallocate		314
-#define __NR_timerfd_settime	315
-#define __NR_timerfd_gettime	316
-
-#define NR_SYSCALLS		317
-
-/* Sparc 32-bit only has the "setresuid32", "getresuid32" variants,
- * it never had the plain ones and there is no value to adding those
- * old versions into the syscall table.
- */
-#define __IGNORE_setresuid
-#define __IGNORE_getresuid
-#define __IGNORE_setresgid
-#define __IGNORE_getresgid
-
-#ifdef __KERNEL__
-#define __ARCH_WANT_IPC_PARSE_VERSION
-#define __ARCH_WANT_OLD_READDIR
-#define __ARCH_WANT_STAT64
-#define __ARCH_WANT_SYS_ALARM
-#define __ARCH_WANT_SYS_GETHOSTNAME
-#define __ARCH_WANT_SYS_PAUSE
-#define __ARCH_WANT_SYS_SGETMASK
-#define __ARCH_WANT_SYS_SIGNAL
-#define __ARCH_WANT_SYS_TIME
-#define __ARCH_WANT_SYS_UTIME
-#define __ARCH_WANT_SYS_WAITPID
-#define __ARCH_WANT_SYS_SOCKETCALL
-#define __ARCH_WANT_SYS_FADVISE64
-#define __ARCH_WANT_SYS_GETPGRP
-#define __ARCH_WANT_SYS_LLSEEK
-#define __ARCH_WANT_SYS_NICE
-#define __ARCH_WANT_SYS_OLDUMOUNT
-#define __ARCH_WANT_SYS_SIGPENDING
-#define __ARCH_WANT_SYS_SIGPROCMASK
-#define __ARCH_WANT_SYS_RT_SIGSUSPEND
-
-/*
- * "Conditional" syscalls
- *
- * What we want is __attribute__((weak,alias("sys_ni_syscall"))),
- * but it doesn't work on all toolchains, so we just do it by hand
- */
-#define cond_syscall(x) asm(".weak\t" #x "\n\t.set\t" #x ",sys_ni_syscall")
-
-#endif /* __KERNEL__ */
-#endif /* _SPARC_UNISTD_H */
+#ifndef ___ASM_SPARC_UNISTD_H
+#define ___ASM_SPARC_UNISTD_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/unistd_64.h>
+#else
+#include <asm-sparc/unistd_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/unistd_32.h b/include/asm-sparc/unistd_32.h
new file mode 100644
index 00000000000..2338a027637
--- /dev/null
+++ b/include/asm-sparc/unistd_32.h
@@ -0,0 +1,378 @@
+#ifndef _SPARC_UNISTD_H
+#define _SPARC_UNISTD_H
+
+/*
+ * System calls under the Sparc.
+ *
+ * Don't be scared by the ugly clobbers, it is the only way I can
+ * think of right now to force the arguments into fixed registers
+ * before the trap into the system call with gcc 'asm' statements.
+ *
+ * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
+ *
+ * SunOS compatibility based upon preliminary work which is:
+ *
+ * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu)
+ */
+
+#define __NR_restart_syscall      0 /* Linux Specific				   */
+#define __NR_exit                 1 /* Common                                      */
+#define __NR_fork                 2 /* Common                                      */
+#define __NR_read                 3 /* Common                                      */
+#define __NR_write                4 /* Common                                      */
+#define __NR_open                 5 /* Common                                      */
+#define __NR_close                6 /* Common                                      */
+#define __NR_wait4                7 /* Common                                      */
+#define __NR_creat                8 /* Common                                      */
+#define __NR_link                 9 /* Common                                      */
+#define __NR_unlink              10 /* Common                                      */
+#define __NR_execv               11 /* SunOS Specific                              */
+#define __NR_chdir               12 /* Common                                      */
+#define __NR_chown		 13 /* Common					   */
+#define __NR_mknod               14 /* Common                                      */
+#define __NR_chmod               15 /* Common                                      */
+#define __NR_lchown              16 /* Common                                      */
+#define __NR_brk                 17 /* Common                                      */
+#define __NR_perfctr             18 /* Performance counter operations              */
+#define __NR_lseek               19 /* Common                                      */
+#define __NR_getpid              20 /* Common                                      */
+#define __NR_capget		 21 /* Linux Specific				   */
+#define __NR_capset		 22 /* Linux Specific				   */
+#define __NR_setuid              23 /* Implemented via setreuid in SunOS           */
+#define __NR_getuid              24 /* Common                                      */
+#define __NR_vmsplice	         25 /* ENOSYS under SunOS			   */
+#define __NR_ptrace              26 /* Common                                      */
+#define __NR_alarm               27 /* Implemented via setitimer in SunOS          */
+#define __NR_sigaltstack	 28 /* Common					   */
+#define __NR_pause               29 /* Is sigblock(0)->sigpause() in SunOS         */
+#define __NR_utime               30 /* Implemented via utimes() under SunOS        */
+#define __NR_lchown32            31 /* Linux sparc32 specific                      */
+#define __NR_fchown32            32 /* Linux sparc32 specific                      */
+#define __NR_access              33 /* Common                                      */
+#define __NR_nice                34 /* Implemented via get/setpriority() in SunOS  */
+#define __NR_chown32             35 /* Linux sparc32 specific                      */
+#define __NR_sync                36 /* Common                                      */
+#define __NR_kill                37 /* Common                                      */
+#define __NR_stat                38 /* Common                                      */
+#define __NR_sendfile		 39 /* Linux Specific				   */
+#define __NR_lstat               40 /* Common                                      */
+#define __NR_dup                 41 /* Common                                      */
+#define __NR_pipe                42 /* Common                                      */
+#define __NR_times               43 /* Implemented via getrusage() in SunOS        */
+#define __NR_getuid32            44 /* Linux sparc32 specific                      */
+#define __NR_umount2             45 /* Linux Specific                              */
+#define __NR_setgid              46 /* Implemented via setregid() in SunOS         */
+#define __NR_getgid              47 /* Common                                      */
+#define __NR_signal              48 /* Implemented via sigvec() in SunOS           */
+#define __NR_geteuid             49 /* SunOS calls getuid()                        */
+#define __NR_getegid             50 /* SunOS calls getgid()                        */
+#define __NR_acct                51 /* Common                                      */
+/* #define __NR_memory_ordering  52    Linux sparc64 specific			   */
+#define __NR_getgid32            53 /* Linux sparc32 specific                      */
+#define __NR_ioctl               54 /* Common                                      */
+#define __NR_reboot              55 /* Common                                      */
+#define __NR_mmap2		 56 /* Linux sparc32 Specific			   */
+#define __NR_symlink             57 /* Common                                      */
+#define __NR_readlink            58 /* Common                                      */
+#define __NR_execve              59 /* Common                                      */
+#define __NR_umask               60 /* Common                                      */
+#define __NR_chroot              61 /* Common                                      */
+#define __NR_fstat               62 /* Common                                      */
+#define __NR_fstat64		 63 /* Linux Specific			           */
+#define __NR_getpagesize         64 /* Common                                      */
+#define __NR_msync               65 /* Common in newer 1.3.x revs...               */
+#define __NR_vfork               66 /* Common                                      */
+#define __NR_pread64             67 /* Linux Specific                              */
+#define __NR_pwrite64            68 /* Linux Specific                              */
+#define __NR_geteuid32           69 /* Linux sparc32, sbrk under SunOS             */
+#define __NR_getegid32           70 /* Linux sparc32, sstk under SunOS             */
+#define __NR_mmap                71 /* Common                                      */
+#define __NR_setreuid32          72 /* Linux sparc32, vadvise under SunOS          */
+#define __NR_munmap              73 /* Common                                      */
+#define __NR_mprotect            74 /* Common                                      */
+#define __NR_madvise             75 /* Common                                      */
+#define __NR_vhangup             76 /* Common                                      */
+#define __NR_truncate64		 77 /* Linux sparc32 Specific			   */
+#define __NR_mincore             78 /* Common                                      */
+#define __NR_getgroups           79 /* Common                                      */
+#define __NR_setgroups           80 /* Common                                      */
+#define __NR_getpgrp             81 /* Common                                      */
+#define __NR_setgroups32         82 /* Linux sparc32, setpgrp under SunOS          */
+#define __NR_setitimer           83 /* Common                                      */
+#define __NR_ftruncate64	 84 /* Linux sparc32 Specific			   */
+#define __NR_swapon              85 /* Common                                      */
+#define __NR_getitimer           86 /* Common                                      */
+#define __NR_setuid32            87 /* Linux sparc32, gethostname under SunOS      */
+#define __NR_sethostname         88 /* Common                                      */
+#define __NR_setgid32            89 /* Linux sparc32, getdtablesize under SunOS    */
+#define __NR_dup2                90 /* Common                                      */
+#define __NR_setfsuid32          91 /* Linux sparc32, getdopt under SunOS          */
+#define __NR_fcntl               92 /* Common                                      */
+#define __NR_select              93 /* Common                                      */
+#define __NR_setfsgid32          94 /* Linux sparc32, setdopt under SunOS          */
+#define __NR_fsync               95 /* Common                                      */
+#define __NR_setpriority         96 /* Common                                      */
+#define __NR_socket              97 /* Common                                      */
+#define __NR_connect             98 /* Common                                      */
+#define __NR_accept              99 /* Common                                      */
+#define __NR_getpriority        100 /* Common                                      */
+#define __NR_rt_sigreturn       101 /* Linux Specific                              */
+#define __NR_rt_sigaction       102 /* Linux Specific                              */
+#define __NR_rt_sigprocmask     103 /* Linux Specific                              */
+#define __NR_rt_sigpending      104 /* Linux Specific                              */
+#define __NR_rt_sigtimedwait    105 /* Linux Specific                              */
+#define __NR_rt_sigqueueinfo    106 /* Linux Specific                              */
+#define __NR_rt_sigsuspend      107 /* Linux Specific                              */
+#define __NR_setresuid32        108 /* Linux Specific, sigvec under SunOS	   */
+#define __NR_getresuid32        109 /* Linux Specific, sigblock under SunOS	   */
+#define __NR_setresgid32        110 /* Linux Specific, sigsetmask under SunOS	   */
+#define __NR_getresgid32        111 /* Linux Specific, sigpause under SunOS	   */
+#define __NR_setregid32         112 /* Linux sparc32, sigstack under SunOS         */
+#define __NR_recvmsg            113 /* Common                                      */
+#define __NR_sendmsg            114 /* Common                                      */
+#define __NR_getgroups32        115 /* Linux sparc32, vtrace under SunOS           */
+#define __NR_gettimeofday       116 /* Common                                      */
+#define __NR_getrusage          117 /* Common                                      */
+#define __NR_getsockopt         118 /* Common                                      */
+#define __NR_getcwd		119 /* Linux Specific				   */
+#define __NR_readv              120 /* Common                                      */
+#define __NR_writev             121 /* Common                                      */
+#define __NR_settimeofday       122 /* Common                                      */
+#define __NR_fchown             123 /* Common                                      */
+#define __NR_fchmod             124 /* Common                                      */
+#define __NR_recvfrom           125 /* Common                                      */
+#define __NR_setreuid           126 /* Common                                      */
+#define __NR_setregid           127 /* Common                                      */
+#define __NR_rename             128 /* Common                                      */
+#define __NR_truncate           129 /* Common                                      */
+#define __NR_ftruncate          130 /* Common                                      */
+#define __NR_flock              131 /* Common                                      */
+#define __NR_lstat64		132 /* Linux Specific			           */
+#define __NR_sendto             133 /* Common                                      */
+#define __NR_shutdown           134 /* Common                                      */
+#define __NR_socketpair         135 /* Common                                      */
+#define __NR_mkdir              136 /* Common                                      */
+#define __NR_rmdir              137 /* Common                                      */
+#define __NR_utimes             138 /* SunOS Specific                              */
+#define __NR_stat64		139 /* Linux Specific			           */
+#define __NR_sendfile64         140 /* adjtime under SunOS                         */
+#define __NR_getpeername        141 /* Common                                      */
+#define __NR_futex              142 /* gethostid under SunOS                       */
+#define __NR_gettid             143 /* ENOSYS under SunOS                          */
+#define __NR_getrlimit          144 /* Common                                      */
+#define __NR_setrlimit          145 /* Common                                      */
+#define __NR_pivot_root		146 /* Linux Specific, killpg under SunOS          */
+#define __NR_prctl		147 /* ENOSYS under SunOS                          */
+#define __NR_pciconfig_read	148 /* ENOSYS under SunOS                          */
+#define __NR_pciconfig_write	149 /* ENOSYS under SunOS                          */
+#define __NR_getsockname        150 /* Common                                      */
+#define __NR_inotify_init       151 /* Linux specific                              */
+#define __NR_inotify_add_watch  152 /* Linux specific                              */
+#define __NR_poll               153 /* Common                                      */
+#define __NR_getdents64		154 /* Linux specific				   */
+#define __NR_fcntl64		155 /* Linux sparc32 Specific                      */
+#define __NR_inotify_rm_watch   156 /* Linux specific				   */
+#define __NR_statfs             157 /* Common                                      */
+#define __NR_fstatfs            158 /* Common                                      */
+#define __NR_umount             159 /* Common                                      */
+#define __NR_sched_set_affinity 160 /* Linux specific, async_daemon under SunOS    */
+#define __NR_sched_get_affinity 161 /* Linux specific, getfh under SunOS           */
+#define __NR_getdomainname      162 /* SunOS Specific                              */
+#define __NR_setdomainname      163 /* Common                                      */
+/* #define __NR_utrap_install   164    Linux sparc64 specific			   */
+#define __NR_quotactl           165 /* Common                                      */
+#define __NR_set_tid_address    166 /* Linux specific, exportfs under SunOS        */
+#define __NR_mount              167 /* Common                                      */
+#define __NR_ustat              168 /* Common                                      */
+#define __NR_setxattr           169 /* SunOS: semsys                               */
+#define __NR_lsetxattr          170 /* SunOS: msgsys                               */
+#define __NR_fsetxattr          171 /* SunOS: shmsys                               */
+#define __NR_getxattr           172 /* SunOS: auditsys                             */
+#define __NR_lgetxattr          173 /* SunOS: rfssys                               */
+#define __NR_getdents           174 /* Common                                      */
+#define __NR_setsid             175 /* Common                                      */
+#define __NR_fchdir             176 /* Common                                      */
+#define __NR_fgetxattr          177 /* SunOS: fchroot                              */
+#define __NR_listxattr          178 /* SunOS: vpixsys                              */
+#define __NR_llistxattr         179 /* SunOS: aioread                              */
+#define __NR_flistxattr         180 /* SunOS: aiowrite                             */
+#define __NR_removexattr        181 /* SunOS: aiowait                              */
+#define __NR_lremovexattr       182 /* SunOS: aiocancel                            */
+#define __NR_sigpending         183 /* Common                                      */
+#define __NR_query_module	184 /* Linux Specific				   */
+#define __NR_setpgid            185 /* Common                                      */
+#define __NR_fremovexattr       186 /* SunOS: pathconf                             */
+#define __NR_tkill              187 /* SunOS: fpathconf                            */
+#define __NR_exit_group		188 /* Linux specific, sysconf undef SunOS         */
+#define __NR_uname              189 /* Linux Specific                              */
+#define __NR_init_module        190 /* Linux Specific                              */
+#define __NR_personality        191 /* Linux Specific                              */
+#define __NR_remap_file_pages   192 /* Linux Specific                              */
+#define __NR_epoll_create       193 /* Linux Specific                              */
+#define __NR_epoll_ctl          194 /* Linux Specific                              */
+#define __NR_epoll_wait         195 /* Linux Specific                              */
+#define __NR_ioprio_set         196 /* Linux Specific                              */
+#define __NR_getppid            197 /* Linux Specific                              */
+#define __NR_sigaction          198 /* Linux Specific                              */
+#define __NR_sgetmask           199 /* Linux Specific                              */
+#define __NR_ssetmask           200 /* Linux Specific                              */
+#define __NR_sigsuspend         201 /* Linux Specific                              */
+#define __NR_oldlstat           202 /* Linux Specific                              */
+#define __NR_uselib             203 /* Linux Specific                              */
+#define __NR_readdir            204 /* Linux Specific                              */
+#define __NR_readahead          205 /* Linux Specific                              */
+#define __NR_socketcall         206 /* Linux Specific                              */
+#define __NR_syslog             207 /* Linux Specific                              */
+#define __NR_lookup_dcookie     208 /* Linux Specific                              */
+#define __NR_fadvise64          209 /* Linux Specific                              */
+#define __NR_fadvise64_64       210 /* Linux Specific                              */
+#define __NR_tgkill             211 /* Linux Specific                              */
+#define __NR_waitpid            212 /* Linux Specific                              */
+#define __NR_swapoff            213 /* Linux Specific                              */
+#define __NR_sysinfo            214 /* Linux Specific                              */
+#define __NR_ipc                215 /* Linux Specific                              */
+#define __NR_sigreturn          216 /* Linux Specific                              */
+#define __NR_clone              217 /* Linux Specific                              */
+#define __NR_ioprio_get         218 /* Linux Specific                              */
+#define __NR_adjtimex           219 /* Linux Specific                              */
+#define __NR_sigprocmask        220 /* Linux Specific                              */
+#define __NR_create_module      221 /* Linux Specific                              */
+#define __NR_delete_module      222 /* Linux Specific                              */
+#define __NR_get_kernel_syms    223 /* Linux Specific                              */
+#define __NR_getpgid            224 /* Linux Specific                              */
+#define __NR_bdflush            225 /* Linux Specific                              */
+#define __NR_sysfs              226 /* Linux Specific                              */
+#define __NR_afs_syscall        227 /* Linux Specific                              */
+#define __NR_setfsuid           228 /* Linux Specific                              */
+#define __NR_setfsgid           229 /* Linux Specific                              */
+#define __NR__newselect         230 /* Linux Specific                              */
+#define __NR_time               231 /* Linux Specific                              */
+#define __NR_splice             232 /* Linux Specific                              */
+#define __NR_stime              233 /* Linux Specific                              */
+#define __NR_statfs64           234 /* Linux Specific                              */
+#define __NR_fstatfs64          235 /* Linux Specific                              */
+#define __NR__llseek            236 /* Linux Specific                              */
+#define __NR_mlock              237
+#define __NR_munlock            238
+#define __NR_mlockall           239
+#define __NR_munlockall         240
+#define __NR_sched_setparam     241
+#define __NR_sched_getparam     242
+#define __NR_sched_setscheduler 243
+#define __NR_sched_getscheduler 244
+#define __NR_sched_yield        245
+#define __NR_sched_get_priority_max 246
+#define __NR_sched_get_priority_min 247
+#define __NR_sched_rr_get_interval  248
+#define __NR_nanosleep          249
+#define __NR_mremap             250
+#define __NR__sysctl            251
+#define __NR_getsid             252
+#define __NR_fdatasync          253
+#define __NR_nfsservctl         254
+#define __NR_sync_file_range	255
+#define __NR_clock_settime	256
+#define __NR_clock_gettime	257
+#define __NR_clock_getres	258
+#define __NR_clock_nanosleep	259
+#define __NR_sched_getaffinity	260
+#define __NR_sched_setaffinity	261
+#define __NR_timer_settime	262
+#define __NR_timer_gettime	263
+#define __NR_timer_getoverrun	264
+#define __NR_timer_delete	265
+#define __NR_timer_create	266
+/* #define __NR_vserver		267 Reserved for VSERVER */
+#define __NR_io_setup		268
+#define __NR_io_destroy		269
+#define __NR_io_submit		270
+#define __NR_io_cancel		271
+#define __NR_io_getevents	272
+#define __NR_mq_open		273
+#define __NR_mq_unlink		274
+#define __NR_mq_timedsend	275
+#define __NR_mq_timedreceive	276
+#define __NR_mq_notify		277
+#define __NR_mq_getsetattr	278
+#define __NR_waitid		279
+#define __NR_tee		280
+#define __NR_add_key		281
+#define __NR_request_key	282
+#define __NR_keyctl		283
+#define __NR_openat		284
+#define __NR_mkdirat		285
+#define __NR_mknodat		286
+#define __NR_fchownat		287
+#define __NR_futimesat		288
+#define __NR_fstatat64		289
+#define __NR_unlinkat		290
+#define __NR_renameat		291
+#define __NR_linkat		292
+#define __NR_symlinkat		293
+#define __NR_readlinkat		294
+#define __NR_fchmodat		295
+#define __NR_faccessat		296
+#define __NR_pselect6		297
+#define __NR_ppoll		298
+#define __NR_unshare		299
+#define __NR_set_robust_list	300
+#define __NR_get_robust_list	301
+#define __NR_migrate_pages	302
+#define __NR_mbind		303
+#define __NR_get_mempolicy	304
+#define __NR_set_mempolicy	305
+#define __NR_kexec_load		306
+#define __NR_move_pages		307
+#define __NR_getcpu		308
+#define __NR_epoll_pwait	309
+#define __NR_utimensat		310
+#define __NR_signalfd		311
+#define __NR_timerfd_create	312
+#define __NR_eventfd		313
+#define __NR_fallocate		314
+#define __NR_timerfd_settime	315
+#define __NR_timerfd_gettime	316
+
+#define NR_SYSCALLS		317
+
+/* Sparc 32-bit only has the "setresuid32", "getresuid32" variants,
+ * it never had the plain ones and there is no value to adding those
+ * old versions into the syscall table.
+ */
+#define __IGNORE_setresuid
+#define __IGNORE_getresuid
+#define __IGNORE_setresgid
+#define __IGNORE_getresgid
+
+#ifdef __KERNEL__
+#define __ARCH_WANT_IPC_PARSE_VERSION
+#define __ARCH_WANT_OLD_READDIR
+#define __ARCH_WANT_STAT64
+#define __ARCH_WANT_SYS_ALARM
+#define __ARCH_WANT_SYS_GETHOSTNAME
+#define __ARCH_WANT_SYS_PAUSE
+#define __ARCH_WANT_SYS_SGETMASK
+#define __ARCH_WANT_SYS_SIGNAL
+#define __ARCH_WANT_SYS_TIME
+#define __ARCH_WANT_SYS_UTIME
+#define __ARCH_WANT_SYS_WAITPID
+#define __ARCH_WANT_SYS_SOCKETCALL
+#define __ARCH_WANT_SYS_FADVISE64
+#define __ARCH_WANT_SYS_GETPGRP
+#define __ARCH_WANT_SYS_LLSEEK
+#define __ARCH_WANT_SYS_NICE
+#define __ARCH_WANT_SYS_OLDUMOUNT
+#define __ARCH_WANT_SYS_SIGPENDING
+#define __ARCH_WANT_SYS_SIGPROCMASK
+#define __ARCH_WANT_SYS_RT_SIGSUSPEND
+
+/*
+ * "Conditional" syscalls
+ *
+ * What we want is __attribute__((weak,alias("sys_ni_syscall"))),
+ * but it doesn't work on all toolchains, so we just do it by hand
+ */
+#define cond_syscall(x) asm(".weak\t" #x "\n\t.set\t" #x ",sys_ni_syscall")
+
+#endif /* __KERNEL__ */
+#endif /* _SPARC_UNISTD_H */
diff --git a/include/asm-sparc/unistd_64.h b/include/asm-sparc/unistd_64.h
new file mode 100644
index 00000000000..13be4453a1f
--- /dev/null
+++ b/include/asm-sparc/unistd_64.h
@@ -0,0 +1,373 @@
+#ifndef _SPARC64_UNISTD_H
+#define _SPARC64_UNISTD_H
+
+/*
+ * System calls under the Sparc.
+ *
+ * Don't be scared by the ugly clobbers, it is the only way I can
+ * think of right now to force the arguments into fixed registers
+ * before the trap into the system call with gcc 'asm' statements.
+ *
+ * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
+ *
+ * SunOS compatibility based upon preliminary work which is:
+ *
+ * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu)
+ */
+
+#define __NR_restart_syscall      0 /* Linux Specific				   */
+#define __NR_exit                 1 /* Common                                      */
+#define __NR_fork                 2 /* Common                                      */
+#define __NR_read                 3 /* Common                                      */
+#define __NR_write                4 /* Common                                      */
+#define __NR_open                 5 /* Common                                      */
+#define __NR_close                6 /* Common                                      */
+#define __NR_wait4                7 /* Common                                      */
+#define __NR_creat                8 /* Common                                      */
+#define __NR_link                 9 /* Common                                      */
+#define __NR_unlink              10 /* Common                                      */
+#define __NR_execv               11 /* SunOS Specific                              */
+#define __NR_chdir               12 /* Common                                      */
+#define __NR_chown		 13 /* Common					   */
+#define __NR_mknod               14 /* Common                                      */
+#define __NR_chmod               15 /* Common                                      */
+#define __NR_lchown              16 /* Common                                      */
+#define __NR_brk                 17 /* Common                                      */
+#define __NR_perfctr             18 /* Performance counter operations              */
+#define __NR_lseek               19 /* Common                                      */
+#define __NR_getpid              20 /* Common                                      */
+#define __NR_capget		 21 /* Linux Specific				   */
+#define __NR_capset		 22 /* Linux Specific				   */
+#define __NR_setuid              23 /* Implemented via setreuid in SunOS           */
+#define __NR_getuid              24 /* Common                                      */
+#define __NR_vmsplice	         25 /* ENOSYS under SunOS			   */
+#define __NR_ptrace              26 /* Common                                      */
+#define __NR_alarm               27 /* Implemented via setitimer in SunOS          */
+#define __NR_sigaltstack	 28 /* Common					   */
+#define __NR_pause               29 /* Is sigblock(0)->sigpause() in SunOS         */
+#define __NR_utime               30 /* Implemented via utimes() under SunOS        */
+/* #define __NR_lchown32         31    Linux sparc32 specific                      */
+/* #define __NR_fchown32         32    Linux sparc32 specific                      */
+#define __NR_access              33 /* Common                                      */
+#define __NR_nice                34 /* Implemented via get/setpriority() in SunOS  */
+/* #define __NR_chown32          35    Linux sparc32 specific                      */
+#define __NR_sync                36 /* Common                                      */
+#define __NR_kill                37 /* Common                                      */
+#define __NR_stat                38 /* Common                                      */
+#define __NR_sendfile		 39 /* Linux Specific				   */
+#define __NR_lstat               40 /* Common                                      */
+#define __NR_dup                 41 /* Common                                      */
+#define __NR_pipe                42 /* Common                                      */
+#define __NR_times               43 /* Implemented via getrusage() in SunOS        */
+/* #define __NR_getuid32         44    Linux sparc32 specific                      */
+#define __NR_umount2             45 /* Linux Specific                              */
+#define __NR_setgid              46 /* Implemented via setregid() in SunOS         */
+#define __NR_getgid              47 /* Common                                      */
+#define __NR_signal              48 /* Implemented via sigvec() in SunOS           */
+#define __NR_geteuid             49 /* SunOS calls getuid()                        */
+#define __NR_getegid             50 /* SunOS calls getgid()                        */
+#define __NR_acct                51 /* Common                                      */
+#define __NR_memory_ordering	 52 /* Linux Specific				   */
+/* #define __NR_getgid32         53    Linux sparc32 specific                      */
+#define __NR_ioctl               54 /* Common                                      */
+#define __NR_reboot              55 /* Common                                      */
+/* #define __NR_mmap2		 56    Linux sparc32 Specific                      */
+#define __NR_symlink             57 /* Common                                      */
+#define __NR_readlink            58 /* Common                                      */
+#define __NR_execve              59 /* Common                                      */
+#define __NR_umask               60 /* Common                                      */
+#define __NR_chroot              61 /* Common                                      */
+#define __NR_fstat               62 /* Common                                      */
+#define __NR_fstat64             63 /* Linux Specific                              */
+#define __NR_getpagesize         64 /* Common                                      */
+#define __NR_msync               65 /* Common in newer 1.3.x revs...               */
+#define __NR_vfork               66 /* Common                                      */
+#define __NR_pread64             67 /* Linux Specific                              */
+#define __NR_pwrite64            68 /* Linux Specific                              */
+/* #define __NR_geteuid32        69    Linux sparc32, sbrk under SunOS             */
+/* #define __NR_getegid32        70    Linux sparc32, sstk under SunOS             */
+#define __NR_mmap                71 /* Common                                      */
+/* #define __NR_setreuid32       72    Linux sparc32, vadvise under SunOS          */
+#define __NR_munmap              73 /* Common                                      */
+#define __NR_mprotect            74 /* Common                                      */
+#define __NR_madvise             75 /* Common                                      */
+#define __NR_vhangup             76 /* Common                                      */
+/* #define __NR_truncate64       77    Linux sparc32 Specific			   */
+#define __NR_mincore             78 /* Common                                      */
+#define __NR_getgroups           79 /* Common                                      */
+#define __NR_setgroups           80 /* Common                                      */
+#define __NR_getpgrp             81 /* Common                                      */
+/* #define __NR_setgroups32      82    Linux sparc32, setpgrp under SunOS          */
+#define __NR_setitimer           83 /* Common                                      */
+/* #define __NR_ftruncate64      84    Linux sparc32 Specific			   */
+#define __NR_swapon              85 /* Common                                      */
+#define __NR_getitimer           86 /* Common                                      */
+/* #define __NR_setuid32         87    Linux sparc32, gethostname under SunOS      */
+#define __NR_sethostname         88 /* Common                                      */
+/* #define __NR_setgid32         89    Linux sparc32, getdtablesize under SunOS    */
+#define __NR_dup2                90 /* Common                                      */
+/* #define __NR_setfsuid32       91    Linux sparc32, getdopt under SunOS          */
+#define __NR_fcntl               92 /* Common                                      */
+#define __NR_select              93 /* Common                                      */
+/* #define __NR_setfsgid32       94    Linux sparc32, setdopt under SunOS          */
+#define __NR_fsync               95 /* Common                                      */
+#define __NR_setpriority         96 /* Common                                      */
+#define __NR_socket              97 /* Common                                      */
+#define __NR_connect             98 /* Common                                      */
+#define __NR_accept              99 /* Common                                      */
+#define __NR_getpriority        100 /* Common                                      */
+#define __NR_rt_sigreturn       101 /* Linux Specific                              */
+#define __NR_rt_sigaction       102 /* Linux Specific                              */
+#define __NR_rt_sigprocmask     103 /* Linux Specific                              */
+#define __NR_rt_sigpending      104 /* Linux Specific                              */
+#define __NR_rt_sigtimedwait    105 /* Linux Specific                              */
+#define __NR_rt_sigqueueinfo    106 /* Linux Specific                              */
+#define __NR_rt_sigsuspend      107 /* Linux Specific                              */
+#define __NR_setresuid          108 /* Linux Specific, sigvec under SunOS	   */
+#define __NR_getresuid          109 /* Linux Specific, sigblock under SunOS	   */
+#define __NR_setresgid          110 /* Linux Specific, sigsetmask under SunOS	   */
+#define __NR_getresgid          111 /* Linux Specific, sigpause under SunOS	   */
+/* #define __NR_setregid32       75    Linux sparc32, sigstack under SunOS         */
+#define __NR_recvmsg            113 /* Common                                      */
+#define __NR_sendmsg            114 /* Common                                      */
+/* #define __NR_getgroups32     115    Linux sparc32, vtrace under SunOS           */
+#define __NR_gettimeofday       116 /* Common                                      */
+#define __NR_getrusage          117 /* Common                                      */
+#define __NR_getsockopt         118 /* Common                                      */
+#define __NR_getcwd		119 /* Linux Specific				   */
+#define __NR_readv              120 /* Common                                      */
+#define __NR_writev             121 /* Common                                      */
+#define __NR_settimeofday       122 /* Common                                      */
+#define __NR_fchown             123 /* Common                                      */
+#define __NR_fchmod             124 /* Common                                      */
+#define __NR_recvfrom           125 /* Common                                      */
+#define __NR_setreuid           126 /* Common                                      */
+#define __NR_setregid           127 /* Common                                      */
+#define __NR_rename             128 /* Common                                      */
+#define __NR_truncate           129 /* Common                                      */
+#define __NR_ftruncate          130 /* Common                                      */
+#define __NR_flock              131 /* Common                                      */
+#define __NR_lstat64		132 /* Linux Specific                              */
+#define __NR_sendto             133 /* Common                                      */
+#define __NR_shutdown           134 /* Common                                      */
+#define __NR_socketpair         135 /* Common                                      */
+#define __NR_mkdir              136 /* Common                                      */
+#define __NR_rmdir              137 /* Common                                      */
+#define __NR_utimes             138 /* SunOS Specific                              */
+#define __NR_stat64		139 /* Linux Specific			           */
+#define __NR_sendfile64         140 /* adjtime under SunOS                         */
+#define __NR_getpeername        141 /* Common                                      */
+#define __NR_futex              142 /* gethostid under SunOS                       */
+#define __NR_gettid             143 /* ENOSYS under SunOS                          */
+#define __NR_getrlimit		144 /* Common                                      */
+#define __NR_setrlimit          145 /* Common                                      */
+#define __NR_pivot_root		146 /* Linux Specific, killpg under SunOS          */
+#define __NR_prctl		147 /* ENOSYS under SunOS                          */
+#define __NR_pciconfig_read	148 /* ENOSYS under SunOS                          */
+#define __NR_pciconfig_write	149 /* ENOSYS under SunOS                          */
+#define __NR_getsockname        150 /* Common                                      */
+#define __NR_inotify_init       151 /* Linux specific                              */
+#define __NR_inotify_add_watch  152 /* Linux specific                              */
+#define __NR_poll               153 /* Common                                      */
+#define __NR_getdents64		154 /* Linux specific				   */
+/* #define __NR_fcntl64         155    Linux sparc32 Specific                      */
+#define __NR_inotify_rm_watch   156 /* Linux specific				   */
+#define __NR_statfs             157 /* Common                                      */
+#define __NR_fstatfs            158 /* Common                                      */
+#define __NR_umount             159 /* Common                                      */
+#define __NR_sched_set_affinity 160 /* Linux specific, async_daemon under SunOS    */
+#define __NR_sched_get_affinity 161 /* Linux specific, getfh under SunOS           */
+#define __NR_getdomainname      162 /* SunOS Specific                              */
+#define __NR_setdomainname      163 /* Common                                      */
+#define __NR_utrap_install	164 /* SYSV ABI/v9 required			   */
+#define __NR_quotactl           165 /* Common                                      */
+#define __NR_set_tid_address    166 /* Linux specific, exportfs under SunOS        */
+#define __NR_mount              167 /* Common                                      */
+#define __NR_ustat              168 /* Common                                      */
+#define __NR_setxattr           169 /* SunOS: semsys                               */
+#define __NR_lsetxattr          170 /* SunOS: msgsys                               */
+#define __NR_fsetxattr          171 /* SunOS: shmsys                               */
+#define __NR_getxattr           172 /* SunOS: auditsys                             */
+#define __NR_lgetxattr          173 /* SunOS: rfssys                               */
+#define __NR_getdents           174 /* Common                                      */
+#define __NR_setsid             175 /* Common                                      */
+#define __NR_fchdir             176 /* Common                                      */
+#define __NR_fgetxattr          177 /* SunOS: fchroot                              */
+#define __NR_listxattr          178 /* SunOS: vpixsys                              */
+#define __NR_llistxattr         179 /* SunOS: aioread                              */
+#define __NR_flistxattr         180 /* SunOS: aiowrite                             */
+#define __NR_removexattr        181 /* SunOS: aiowait                              */
+#define __NR_lremovexattr       182 /* SunOS: aiocancel                            */
+#define __NR_sigpending         183 /* Common                                      */
+#define __NR_query_module	184 /* Linux Specific				   */
+#define __NR_setpgid            185 /* Common                                      */
+#define __NR_fremovexattr       186 /* SunOS: pathconf                             */
+#define __NR_tkill              187 /* SunOS: fpathconf                            */
+#define __NR_exit_group		188 /* Linux specific, sysconf undef SunOS         */
+#define __NR_uname              189 /* Linux Specific                              */
+#define __NR_init_module        190 /* Linux Specific                              */
+#define __NR_personality        191 /* Linux Specific                              */
+#define __NR_remap_file_pages   192 /* Linux Specific                              */
+#define __NR_epoll_create       193 /* Linux Specific                              */
+#define __NR_epoll_ctl          194 /* Linux Specific                              */
+#define __NR_epoll_wait         195 /* Linux Specific                              */
+#define __NR_ioprio_set         196 /* Linux Specific                              */
+#define __NR_getppid            197 /* Linux Specific                              */
+#define __NR_sigaction          198 /* Linux Specific                              */
+#define __NR_sgetmask           199 /* Linux Specific                              */
+#define __NR_ssetmask           200 /* Linux Specific                              */
+#define __NR_sigsuspend         201 /* Linux Specific                              */
+#define __NR_oldlstat           202 /* Linux Specific                              */
+#define __NR_uselib             203 /* Linux Specific                              */
+#define __NR_readdir            204 /* Linux Specific                              */
+#define __NR_readahead          205 /* Linux Specific                              */
+#define __NR_socketcall         206 /* Linux Specific                              */
+#define __NR_syslog             207 /* Linux Specific                              */
+#define __NR_lookup_dcookie     208 /* Linux Specific                              */
+#define __NR_fadvise64          209 /* Linux Specific                              */
+#define __NR_fadvise64_64       210 /* Linux Specific                              */
+#define __NR_tgkill             211 /* Linux Specific                              */
+#define __NR_waitpid            212 /* Linux Specific                              */
+#define __NR_swapoff            213 /* Linux Specific                              */
+#define __NR_sysinfo            214 /* Linux Specific                              */
+#define __NR_ipc                215 /* Linux Specific                              */
+#define __NR_sigreturn          216 /* Linux Specific                              */
+#define __NR_clone              217 /* Linux Specific                              */
+#define __NR_ioprio_get         218 /* Linux Specific                              */
+#define __NR_adjtimex           219 /* Linux Specific                              */
+#define __NR_sigprocmask        220 /* Linux Specific                              */
+#define __NR_create_module      221 /* Linux Specific                              */
+#define __NR_delete_module      222 /* Linux Specific                              */
+#define __NR_get_kernel_syms    223 /* Linux Specific                              */
+#define __NR_getpgid            224 /* Linux Specific                              */
+#define __NR_bdflush            225 /* Linux Specific                              */
+#define __NR_sysfs              226 /* Linux Specific                              */
+#define __NR_afs_syscall        227 /* Linux Specific                              */
+#define __NR_setfsuid           228 /* Linux Specific                              */
+#define __NR_setfsgid           229 /* Linux Specific                              */
+#define __NR__newselect         230 /* Linux Specific                              */
+#ifdef __KERNEL__
+#define __NR_time		231 /* Linux sparc32                               */
+#endif
+#define __NR_splice             232 /* Linux Specific                              */
+#define __NR_stime              233 /* Linux Specific                              */
+#define __NR_statfs64           234 /* Linux Specific                              */
+#define __NR_fstatfs64          235 /* Linux Specific                              */
+#define __NR__llseek            236 /* Linux Specific                              */
+#define __NR_mlock              237
+#define __NR_munlock            238
+#define __NR_mlockall           239
+#define __NR_munlockall         240
+#define __NR_sched_setparam     241
+#define __NR_sched_getparam     242
+#define __NR_sched_setscheduler 243
+#define __NR_sched_getscheduler 244
+#define __NR_sched_yield        245
+#define __NR_sched_get_priority_max 246
+#define __NR_sched_get_priority_min 247
+#define __NR_sched_rr_get_interval  248
+#define __NR_nanosleep          249
+#define __NR_mremap             250
+#define __NR__sysctl            251
+#define __NR_getsid             252
+#define __NR_fdatasync          253
+#define __NR_nfsservctl         254
+#define __NR_sync_file_range	255
+#define __NR_clock_settime	256
+#define __NR_clock_gettime	257
+#define __NR_clock_getres	258
+#define __NR_clock_nanosleep	259
+#define __NR_sched_getaffinity	260
+#define __NR_sched_setaffinity	261
+#define __NR_timer_settime	262
+#define __NR_timer_gettime	263
+#define __NR_timer_getoverrun	264
+#define __NR_timer_delete	265
+#define __NR_timer_create	266
+/* #define __NR_vserver		267 Reserved for VSERVER */
+#define __NR_io_setup		268
+#define __NR_io_destroy		269
+#define __NR_io_submit		270
+#define __NR_io_cancel		271
+#define __NR_io_getevents	272
+#define __NR_mq_open		273
+#define __NR_mq_unlink		274
+#define __NR_mq_timedsend	275
+#define __NR_mq_timedreceive	276
+#define __NR_mq_notify		277
+#define __NR_mq_getsetattr	278
+#define __NR_waitid		279
+#define __NR_tee		280
+#define __NR_add_key		281
+#define __NR_request_key	282
+#define __NR_keyctl		283
+#define __NR_openat		284
+#define __NR_mkdirat		285
+#define __NR_mknodat		286
+#define __NR_fchownat		287
+#define __NR_futimesat		288
+#define __NR_fstatat64		289
+#define __NR_unlinkat		290
+#define __NR_renameat		291
+#define __NR_linkat		292
+#define __NR_symlinkat		293
+#define __NR_readlinkat		294
+#define __NR_fchmodat		295
+#define __NR_faccessat		296
+#define __NR_pselect6		297
+#define __NR_ppoll		298
+#define __NR_unshare		299
+#define __NR_set_robust_list	300
+#define __NR_get_robust_list	301
+#define __NR_migrate_pages	302
+#define __NR_mbind		303
+#define __NR_get_mempolicy	304
+#define __NR_set_mempolicy	305
+#define __NR_kexec_load		306
+#define __NR_move_pages		307
+#define __NR_getcpu		308
+#define __NR_epoll_pwait	309
+#define __NR_utimensat		310
+#define __NR_signalfd		311
+#define __NR_timerfd_create	312
+#define __NR_eventfd		313
+#define __NR_fallocate		314
+#define __NR_timerfd_settime	315
+#define __NR_timerfd_gettime	316
+
+#define NR_SYSCALLS		317
+
+#ifdef __KERNEL__
+#define __ARCH_WANT_IPC_PARSE_VERSION
+#define __ARCH_WANT_OLD_READDIR
+#define __ARCH_WANT_STAT64
+#define __ARCH_WANT_SYS_ALARM
+#define __ARCH_WANT_SYS_GETHOSTNAME
+#define __ARCH_WANT_SYS_PAUSE
+#define __ARCH_WANT_SYS_SGETMASK
+#define __ARCH_WANT_SYS_SIGNAL
+#define __ARCH_WANT_SYS_TIME
+#define __ARCH_WANT_COMPAT_SYS_TIME
+#define __ARCH_WANT_SYS_UTIME
+#define __ARCH_WANT_SYS_WAITPID
+#define __ARCH_WANT_SYS_SOCKETCALL
+#define __ARCH_WANT_SYS_FADVISE64
+#define __ARCH_WANT_SYS_GETPGRP
+#define __ARCH_WANT_SYS_LLSEEK
+#define __ARCH_WANT_SYS_NICE
+#define __ARCH_WANT_SYS_OLDUMOUNT
+#define __ARCH_WANT_SYS_SIGPENDING
+#define __ARCH_WANT_SYS_SIGPROCMASK
+#define __ARCH_WANT_SYS_RT_SIGSUSPEND
+#define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND
+
+/*
+ * "Conditional" syscalls
+ *
+ * What we want is __attribute__((weak,alias("sys_ni_syscall"))),
+ * but it doesn't work on all toolchains, so we just do it by hand
+ */
+#define cond_syscall(x) asm(".weak\t" #x "\n\t.set\t" #x ",sys_ni_syscall")
+
+#endif /* __KERNEL__ */
+#endif /* _SPARC64_UNISTD_H */
diff --git a/include/asm-sparc/upa.h b/include/asm-sparc/upa.h
new file mode 100644
index 00000000000..5b1633223f9
--- /dev/null
+++ b/include/asm-sparc/upa.h
@@ -0,0 +1,109 @@
+#ifndef _SPARC64_UPA_H
+#define _SPARC64_UPA_H
+
+#include <asm/asi.h>
+
+/* UPA level registers and defines. */
+
+/* UPA Config Register */
+#define UPA_CONFIG_RESV		0xffffffffc0000000 /* Reserved.                    */
+#define UPA_CONFIG_PCON		0x000000003fc00000 /* Depth of various sys queues. */
+#define UPA_CONFIG_MID		0x00000000003e0000 /* Module ID.                   */
+#define UPA_CONFIG_PCAP		0x000000000001ffff /* Port Capabilities.           */
+
+/* UPA Port ID Register */
+#define UPA_PORTID_FNP		0xff00000000000000 /* Hardcoded to 0xfc on ultra.  */
+#define UPA_PORTID_RESV		0x00fffff800000000 /* Reserved.                    */
+#define UPA_PORTID_ECCVALID     0x0000000400000000 /* Zero if mod can generate ECC */
+#define UPA_PORTID_ONEREAD      0x0000000200000000 /* Set if mod generates P_RASB  */
+#define UPA_PORTID_PINTRDQ      0x0000000180000000 /* # outstanding P_INT_REQ's    */
+#define UPA_PORTID_PREQDQ       0x000000007e000000 /* slave-wr's to mod supported  */
+#define UPA_PORTID_PREQRD       0x0000000001e00000 /* # incoming P_REQ's supported */
+#define UPA_PORTID_UPACAP       0x00000000001f0000 /* UPA capabilities of mod      */
+#define UPA_PORTID_ID           0x000000000000ffff /* Module Identification bits  */
+
+/* UPA I/O space accessors */
+#if defined(__KERNEL__) && !defined(__ASSEMBLY__)
+static inline unsigned char _upa_readb(unsigned long addr)
+{
+	unsigned char ret;
+
+	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* upa_readb */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline unsigned short _upa_readw(unsigned long addr)
+{
+	unsigned short ret;
+
+	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* upa_readw */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline unsigned int _upa_readl(unsigned long addr)
+{
+	unsigned int ret;
+
+	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* upa_readl */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline unsigned long _upa_readq(unsigned long addr)
+{
+	unsigned long ret;
+
+	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* upa_readq */"
+			     : "=r" (ret)
+			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+
+	return ret;
+}
+
+static inline void _upa_writeb(unsigned char b, unsigned long addr)
+{
+	__asm__ __volatile__("stba\t%0, [%1] %2\t/* upa_writeb */"
+			     : /* no outputs */
+			     : "r" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _upa_writew(unsigned short w, unsigned long addr)
+{
+	__asm__ __volatile__("stha\t%0, [%1] %2\t/* upa_writew */"
+			     : /* no outputs */
+			     : "r" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _upa_writel(unsigned int l, unsigned long addr)
+{
+	__asm__ __volatile__("stwa\t%0, [%1] %2\t/* upa_writel */"
+			     : /* no outputs */
+			     : "r" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+static inline void _upa_writeq(unsigned long q, unsigned long addr)
+{
+	__asm__ __volatile__("stxa\t%0, [%1] %2\t/* upa_writeq */"
+			     : /* no outputs */
+			     : "r" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
+}
+
+#define upa_readb(__addr)		(_upa_readb((unsigned long)(__addr)))
+#define upa_readw(__addr)		(_upa_readw((unsigned long)(__addr)))
+#define upa_readl(__addr)		(_upa_readl((unsigned long)(__addr)))
+#define upa_readq(__addr)		(_upa_readq((unsigned long)(__addr)))
+#define upa_writeb(__b, __addr)		(_upa_writeb((__b), (unsigned long)(__addr)))
+#define upa_writew(__w, __addr)		(_upa_writew((__w), (unsigned long)(__addr)))
+#define upa_writel(__l, __addr)		(_upa_writel((__l), (unsigned long)(__addr)))
+#define upa_writeq(__q, __addr)		(_upa_writeq((__q), (unsigned long)(__addr)))
+#endif /* __KERNEL__ && !__ASSEMBLY__ */
+
+#endif /* !(_SPARC64_UPA_H) */
diff --git a/include/asm-sparc/utrap.h b/include/asm-sparc/utrap.h
new file mode 100644
index 00000000000..9da37babbe5
--- /dev/null
+++ b/include/asm-sparc/utrap.h
@@ -0,0 +1,51 @@
+/*
+ * include/asm-sparc64/utrap.h
+ *
+ * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
+ */
+
+#ifndef __ASM_SPARC64_UTRAP_H
+#define __ASM_SPARC64_UTRAP_H
+
+#define UT_INSTRUCTION_EXCEPTION		1
+#define UT_INSTRUCTION_ERROR			2
+#define UT_INSTRUCTION_PROTECTION		3
+#define UT_ILLTRAP_INSTRUCTION			4
+#define UT_ILLEGAL_INSTRUCTION			5
+#define UT_PRIVILEGED_OPCODE			6
+#define UT_FP_DISABLED				7
+#define UT_FP_EXCEPTION_IEEE_754		8
+#define UT_FP_EXCEPTION_OTHER			9
+#define UT_TAG_OVERVIEW				10
+#define UT_DIVISION_BY_ZERO			11
+#define UT_DATA_EXCEPTION			12
+#define UT_DATA_ERROR				13
+#define UT_DATA_PROTECTION			14
+#define UT_MEM_ADDRESS_NOT_ALIGNED		15
+#define UT_PRIVILEGED_ACTION			16
+#define UT_ASYNC_DATA_ERROR			17
+#define UT_TRAP_INSTRUCTION_16			18
+#define UT_TRAP_INSTRUCTION_17			19
+#define UT_TRAP_INSTRUCTION_18			20
+#define UT_TRAP_INSTRUCTION_19			21
+#define UT_TRAP_INSTRUCTION_20			22
+#define UT_TRAP_INSTRUCTION_21			23
+#define UT_TRAP_INSTRUCTION_22			24
+#define UT_TRAP_INSTRUCTION_23			25
+#define UT_TRAP_INSTRUCTION_24			26
+#define UT_TRAP_INSTRUCTION_25			27
+#define UT_TRAP_INSTRUCTION_26			28
+#define UT_TRAP_INSTRUCTION_27			29
+#define UT_TRAP_INSTRUCTION_28			30
+#define UT_TRAP_INSTRUCTION_29			31
+#define UT_TRAP_INSTRUCTION_30			32
+#define UT_TRAP_INSTRUCTION_31			33
+
+#define	UTH_NOCHANGE				(-1)
+
+#ifndef __ASSEMBLY__
+typedef int utrap_entry_t;
+typedef void *utrap_handler_t;
+#endif /* __ASSEMBLY__ */
+
+#endif /* !(__ASM_SPARC64_PROCESSOR_H) */
diff --git a/include/asm-sparc/vaddrs.h b/include/asm-sparc/vaddrs.h
index f6ca4779056..a22fed5a3c6 100644
--- a/include/asm-sparc/vaddrs.h
+++ b/include/asm-sparc/vaddrs.h
@@ -34,11 +34,6 @@
 #define IOBASE_VADDR		0xfe000000
 #define IOBASE_END		0xfe600000
 
-#define VMALLOC_START		0xfe600000
-
-/* XXX Alter this when I get around to fixing sun4c - Anton */
-#define VMALLOC_END		0xffc00000
-
 /*
  * On the sun4/4c we need a place
  * to reliably map locked down kernel data.  This includes the
diff --git a/include/asm-sparc/vio.h b/include/asm-sparc/vio.h
new file mode 100644
index 00000000000..d4de32f0f8a
--- /dev/null
+++ b/include/asm-sparc/vio.h
@@ -0,0 +1,406 @@
+#ifndef _SPARC64_VIO_H
+#define _SPARC64_VIO_H
+
+#include <linux/kernel.h>
+#include <linux/device.h>
+#include <linux/mod_devicetable.h>
+#include <linux/timer.h>
+#include <linux/spinlock.h>
+#include <linux/completion.h>
+#include <linux/list.h>
+#include <linux/log2.h>
+
+#include <asm/ldc.h>
+#include <asm/mdesc.h>
+
+struct vio_msg_tag {
+	u8			type;
+#define VIO_TYPE_CTRL		0x01
+#define VIO_TYPE_DATA		0x02
+#define VIO_TYPE_ERR		0x04
+
+	u8			stype;
+#define VIO_SUBTYPE_INFO	0x01
+#define VIO_SUBTYPE_ACK		0x02
+#define VIO_SUBTYPE_NACK	0x04
+
+	u16			stype_env;
+#define VIO_VER_INFO		0x0001
+#define VIO_ATTR_INFO		0x0002
+#define VIO_DRING_REG		0x0003
+#define VIO_DRING_UNREG		0x0004
+#define VIO_RDX			0x0005
+#define VIO_PKT_DATA		0x0040
+#define VIO_DESC_DATA		0x0041
+#define VIO_DRING_DATA		0x0042
+#define VNET_MCAST_INFO		0x0101
+
+	u32		sid;
+};
+
+struct vio_rdx {
+	struct vio_msg_tag	tag;
+	u64			resv[6];
+};
+
+struct vio_ver_info {
+	struct vio_msg_tag	tag;
+	u16			major;
+	u16			minor;
+	u8			dev_class;
+#define VDEV_NETWORK		0x01
+#define VDEV_NETWORK_SWITCH	0x02
+#define VDEV_DISK		0x03
+#define VDEV_DISK_SERVER	0x04
+
+	u8			resv1[3];
+	u64			resv2[5];
+};
+
+struct vio_dring_register {
+	struct vio_msg_tag	tag;
+	u64			dring_ident;
+	u32			num_descr;
+	u32			descr_size;
+	u16			options;
+#define VIO_TX_DRING		0x0001
+#define VIO_RX_DRING		0x0002
+	u16			resv;
+	u32			num_cookies;
+	struct ldc_trans_cookie	cookies[0];
+};
+
+struct vio_dring_unregister {
+	struct vio_msg_tag	tag;
+	u64			dring_ident;
+	u64			resv[5];
+};
+
+/* Data transfer modes */
+#define VIO_PKT_MODE		0x01 /* Packet based transfer	*/
+#define VIO_DESC_MODE		0x02 /* In-band descriptors	*/
+#define VIO_DRING_MODE		0x03 /* Descriptor rings	*/
+
+struct vio_dring_data {
+	struct vio_msg_tag	tag;
+	u64			seq;
+	u64			dring_ident;
+	u32			start_idx;
+	u32			end_idx;
+	u8			state;
+#define VIO_DRING_ACTIVE	0x01
+#define VIO_DRING_STOPPED	0x02
+
+	u8			__pad1;
+	u16			__pad2;
+	u32			__pad3;
+	u64			__par4[2];
+};
+
+struct vio_dring_hdr {
+	u8			state;
+#define VIO_DESC_FREE		0x01
+#define VIO_DESC_READY		0x02
+#define VIO_DESC_ACCEPTED	0x03
+#define VIO_DESC_DONE		0x04
+	u8			ack;
+#define VIO_ACK_ENABLE		0x01
+#define VIO_ACK_DISABLE		0x00
+
+	u16			__pad1;
+	u32			__pad2;
+};
+
+/* VIO disk specific structures and defines */
+struct vio_disk_attr_info {
+	struct vio_msg_tag	tag;
+	u8			xfer_mode;
+	u8			vdisk_type;
+#define VD_DISK_TYPE_SLICE	0x01 /* Slice in block device	*/
+#define VD_DISK_TYPE_DISK	0x02 /* Entire block device	*/
+	u16			resv1;
+	u32			vdisk_block_size;
+	u64			operations;
+	u64			vdisk_size;
+	u64			max_xfer_size;
+	u64			resv2[2];
+};
+
+struct vio_disk_desc {
+	struct vio_dring_hdr	hdr;
+	u64			req_id;
+	u8			operation;
+#define VD_OP_BREAD		0x01 /* Block read			*/
+#define VD_OP_BWRITE		0x02 /* Block write			*/
+#define VD_OP_FLUSH		0x03 /* Flush disk contents		*/
+#define VD_OP_GET_WCE		0x04 /* Get write-cache status		*/
+#define VD_OP_SET_WCE		0x05 /* Enable/disable write-cache	*/
+#define VD_OP_GET_VTOC		0x06 /* Get VTOC			*/
+#define VD_OP_SET_VTOC		0x07 /* Set VTOC			*/
+#define VD_OP_GET_DISKGEOM	0x08 /* Get disk geometry		*/
+#define VD_OP_SET_DISKGEOM	0x09 /* Set disk geometry		*/
+#define VD_OP_SCSICMD		0x0a /* SCSI control command		*/
+#define VD_OP_GET_DEVID		0x0b /* Get device ID			*/
+#define VD_OP_GET_EFI		0x0c /* Get EFI				*/
+#define VD_OP_SET_EFI		0x0d /* Set EFI				*/
+	u8			slice;
+	u16			resv1;
+	u32			status;
+	u64			offset;
+	u64			size;
+	u32			ncookies;
+	u32			resv2;
+	struct ldc_trans_cookie	cookies[0];
+};
+
+#define VIO_DISK_VNAME_LEN	8
+#define VIO_DISK_ALABEL_LEN	128
+#define VIO_DISK_NUM_PART	8
+
+struct vio_disk_vtoc {
+	u8			volume_name[VIO_DISK_VNAME_LEN];
+	u16			sector_size;
+	u16			num_partitions;
+	u8			ascii_label[VIO_DISK_ALABEL_LEN];
+	struct {
+		u16		id;
+		u16		perm_flags;
+		u32		resv;
+		u64		start_block;
+		u64		num_blocks;
+	} partitions[VIO_DISK_NUM_PART];
+};
+
+struct vio_disk_geom {
+	u16			num_cyl; /* Num data cylinders		*/
+	u16			alt_cyl; /* Num alternate cylinders	*/
+	u16			beg_cyl; /* Cyl off of fixed head area	*/
+	u16			num_hd;  /* Num heads			*/
+	u16			num_sec; /* Num sectors			*/
+	u16			ifact;   /* Interleave factor		*/
+	u16			apc;     /* Alts per cylinder (SCSI)	*/
+	u16			rpm;	 /* Revolutions per minute	*/
+	u16			phy_cyl; /* Num physical cylinders	*/
+	u16			wr_skip; /* Num sects to skip, writes	*/
+	u16			rd_skip; /* Num sects to skip, writes	*/
+};
+
+struct vio_disk_devid {
+	u16			resv;
+	u16			type;
+	u32			len;
+	char			id[0];
+};
+
+struct vio_disk_efi {
+	u64			lba;
+	u64			len;
+	char			data[0];
+};
+
+/* VIO net specific structures and defines */
+struct vio_net_attr_info {
+	struct vio_msg_tag	tag;
+	u8			xfer_mode;
+	u8			addr_type;
+#define VNET_ADDR_ETHERMAC	0x01
+	u16			ack_freq;
+	u32			resv1;
+	u64			addr;
+	u64			mtu;
+	u64			resv2[3];
+};
+
+#define VNET_NUM_MCAST		7
+
+struct vio_net_mcast_info {
+	struct vio_msg_tag	tag;
+	u8			set;
+	u8			count;
+	u8			mcast_addr[VNET_NUM_MCAST * 6];
+	u32			resv;
+};
+
+struct vio_net_desc {
+	struct vio_dring_hdr	hdr;
+	u32			size;
+	u32			ncookies;
+	struct ldc_trans_cookie	cookies[0];
+};
+
+#define VIO_MAX_RING_COOKIES	24
+
+struct vio_dring_state {
+	u64			ident;
+	void			*base;
+	u64			snd_nxt;
+	u64			rcv_nxt;
+	u32			entry_size;
+	u32			num_entries;
+	u32			prod;
+	u32			cons;
+	u32			pending;
+	int			ncookies;
+	struct ldc_trans_cookie	cookies[VIO_MAX_RING_COOKIES];
+};
+
+static inline void *vio_dring_cur(struct vio_dring_state *dr)
+{
+	return dr->base + (dr->entry_size * dr->prod);
+}
+
+static inline void *vio_dring_entry(struct vio_dring_state *dr,
+				    unsigned int index)
+{
+	return dr->base + (dr->entry_size * index);
+}
+
+static inline u32 vio_dring_avail(struct vio_dring_state *dr,
+				  unsigned int ring_size)
+{
+	BUILD_BUG_ON(!is_power_of_2(ring_size));
+
+	return (dr->pending -
+		((dr->prod - dr->cons) & (ring_size - 1)));
+}
+
+#define VIO_MAX_TYPE_LEN	32
+#define VIO_MAX_COMPAT_LEN	64
+
+struct vio_dev {
+	u64			mp;
+	struct device_node	*dp;
+
+	char			type[VIO_MAX_TYPE_LEN];
+	char			compat[VIO_MAX_COMPAT_LEN];
+	int			compat_len;
+
+	u64			dev_no;
+
+	unsigned long		channel_id;
+
+	unsigned int		tx_irq;
+	unsigned int		rx_irq;
+
+	struct device		dev;
+};
+
+struct vio_driver {
+	struct list_head		node;
+	const struct vio_device_id	*id_table;
+	int (*probe)(struct vio_dev *dev, const struct vio_device_id *id);
+	int (*remove)(struct vio_dev *dev);
+	void (*shutdown)(struct vio_dev *dev);
+	unsigned long			driver_data;
+	struct device_driver		driver;
+};
+
+struct vio_version {
+	u16		major;
+	u16		minor;
+};
+
+struct vio_driver_state;
+struct vio_driver_ops {
+	int	(*send_attr)(struct vio_driver_state *vio);
+	int	(*handle_attr)(struct vio_driver_state *vio, void *pkt);
+	void	(*handshake_complete)(struct vio_driver_state *vio);
+};
+
+struct vio_completion {
+	struct completion	com;
+	int			err;
+	int			waiting_for;
+};
+
+struct vio_driver_state {
+	/* Protects VIO handshake and, optionally, driver private state.  */
+	spinlock_t		lock;
+
+	struct ldc_channel	*lp;
+
+	u32			_peer_sid;
+	u32			_local_sid;
+	struct vio_dring_state	drings[2];
+#define VIO_DRIVER_TX_RING	0
+#define VIO_DRIVER_RX_RING	1
+
+	u8			hs_state;
+#define VIO_HS_INVALID		0x00
+#define VIO_HS_GOTVERS		0x01
+#define VIO_HS_GOT_ATTR		0x04
+#define VIO_HS_SENT_DREG	0x08
+#define VIO_HS_SENT_RDX		0x10
+#define VIO_HS_GOT_RDX_ACK	0x20
+#define VIO_HS_GOT_RDX		0x40
+#define VIO_HS_SENT_RDX_ACK	0x80
+#define VIO_HS_COMPLETE		(VIO_HS_GOT_RDX_ACK | VIO_HS_SENT_RDX_ACK)
+
+	u8			dev_class;
+
+	u8			dr_state;
+#define VIO_DR_STATE_TXREG	0x01
+#define VIO_DR_STATE_RXREG	0x02
+#define VIO_DR_STATE_TXREQ	0x10
+#define VIO_DR_STATE_RXREQ	0x20
+
+	u8			debug;
+#define VIO_DEBUG_HS		0x01
+#define VIO_DEBUG_DATA		0x02
+
+	void			*desc_buf;
+	unsigned int		desc_buf_len;
+
+	struct vio_completion	*cmp;
+
+	struct vio_dev		*vdev;
+
+	struct timer_list	timer;
+
+	struct vio_version	ver;
+
+	struct vio_version	*ver_table;
+	int			ver_table_entries;
+
+	char			*name;
+
+	struct vio_driver_ops	*ops;
+};
+
+#define viodbg(TYPE, f, a...) \
+do {	if (vio->debug & VIO_DEBUG_##TYPE) \
+		printk(KERN_INFO "vio: ID[%lu] " f, \
+		       vio->vdev->channel_id, ## a); \
+} while (0)
+
+extern int vio_register_driver(struct vio_driver *drv);
+extern void vio_unregister_driver(struct vio_driver *drv);
+
+static inline struct vio_driver *to_vio_driver(struct device_driver *drv)
+{
+	return container_of(drv, struct vio_driver, driver);
+}
+
+static inline struct vio_dev *to_vio_dev(struct device *dev)
+{
+	return container_of(dev, struct vio_dev, dev);
+}
+
+extern int vio_ldc_send(struct vio_driver_state *vio, void *data, int len);
+extern void vio_link_state_change(struct vio_driver_state *vio, int event);
+extern void vio_conn_reset(struct vio_driver_state *vio);
+extern int vio_control_pkt_engine(struct vio_driver_state *vio, void *pkt);
+extern int vio_validate_sid(struct vio_driver_state *vio,
+			    struct vio_msg_tag *tp);
+extern u32 vio_send_sid(struct vio_driver_state *vio);
+extern int vio_ldc_alloc(struct vio_driver_state *vio,
+			 struct ldc_channel_config *base_cfg, void *event_arg);
+extern void vio_ldc_free(struct vio_driver_state *vio);
+extern int vio_driver_init(struct vio_driver_state *vio, struct vio_dev *vdev,
+			   u8 dev_class, struct vio_version *ver_table,
+			   int ver_table_size, struct vio_driver_ops *ops,
+			   char *name);
+
+extern void vio_port_up(struct vio_driver_state *vio);
+
+#endif /* _SPARC64_VIO_H */
diff --git a/include/asm-sparc/visasm.h b/include/asm-sparc/visasm.h
new file mode 100644
index 00000000000..de797b9bf55
--- /dev/null
+++ b/include/asm-sparc/visasm.h
@@ -0,0 +1,62 @@
+#ifndef _SPARC64_VISASM_H
+#define _SPARC64_VISASM_H
+
+/* visasm.h:  FPU saving macros for VIS routines
+ *
+ * Copyright (C) 1998 Jakub Jelinek (jj@ultra.linux.cz)
+ */
+
+#include <asm/pstate.h>
+#include <asm/ptrace.h>
+
+/* Clobbers %o5, %g1, %g2, %g3, %g7, %icc, %xcc */
+
+#define VISEntry					\
+	rd		%fprs, %o5;			\
+	andcc		%o5, (FPRS_FEF|FPRS_DU), %g0;	\
+	be,pt		%icc, 297f;			\
+	 sethi		%hi(297f), %g7;			\
+	sethi		%hi(VISenter), %g1;		\
+	jmpl		%g1 + %lo(VISenter), %g0;	\
+	 or		%g7, %lo(297f), %g7;		\
+297:	wr		%g0, FPRS_FEF, %fprs;		\
+
+#define VISExit						\
+	wr		%g0, 0, %fprs;
+
+/* Clobbers %o5, %g1, %g2, %g3, %g7, %icc, %xcc.
+ * Must preserve %o5 between VISEntryHalf and VISExitHalf */
+
+#define VISEntryHalf					\
+	rd		%fprs, %o5;			\
+	andcc		%o5, FPRS_FEF, %g0;		\
+	be,pt		%icc, 297f;			\
+	 sethi		%hi(298f), %g7;			\
+	sethi		%hi(VISenterhalf), %g1;		\
+	jmpl		%g1 + %lo(VISenterhalf), %g0;	\
+	 or		%g7, %lo(298f), %g7;		\
+	clr		%o5;				\
+297:	wr		%o5, FPRS_FEF, %fprs;		\
+298:
+
+#define VISExitHalf					\
+	wr		%o5, 0, %fprs;
+
+#ifndef __ASSEMBLY__
+static inline void save_and_clear_fpu(void) {
+	__asm__ __volatile__ (
+"		rd %%fprs, %%o5\n"
+"		andcc %%o5, %0, %%g0\n"
+"		be,pt %%icc, 299f\n"
+"		 sethi %%hi(298f), %%g7\n"
+"		sethi %%hi(VISenter), %%g1\n"
+"		jmpl %%g1 + %%lo(VISenter), %%g0\n"
+"		 or %%g7, %%lo(298f), %%g7\n"
+"	298:	wr %%g0, 0, %%fprs\n"
+"	299:\n"
+"		" : : "i" (FPRS_FEF|FPRS_DU) :
+		"o5", "g1", "g2", "g3", "g7", "cc");
+}
+#endif
+
+#endif /* _SPARC64_ASI_H */
diff --git a/include/asm-sparc/watchdog.h b/include/asm-sparc/watchdog.h
new file mode 100644
index 00000000000..5baf2d3919c
--- /dev/null
+++ b/include/asm-sparc/watchdog.h
@@ -0,0 +1,31 @@
+/*
+ *
+ * watchdog - Driver interface for the hardware watchdog timers
+ * present on Sun Microsystems boardsets
+ *
+ * Copyright (c) 2000 Eric Brower <ebrower@usa.net>
+ *
+ */
+
+#ifndef _SPARC64_WATCHDOG_H
+#define _SPARC64_WATCHDOG_H
+
+#include <linux/watchdog.h>
+
+/* Solaris compatibility ioctls--
+ * Ref. <linux/watchdog.h> for standard linux watchdog ioctls
+ */
+#define WIOCSTART _IO (WATCHDOG_IOCTL_BASE, 10)		/* Start Timer		*/
+#define WIOCSTOP  _IO (WATCHDOG_IOCTL_BASE, 11)		/* Stop Timer		*/
+#define WIOCGSTAT _IOR(WATCHDOG_IOCTL_BASE, 12, int)/* Get Timer Status	*/
+
+/* Status flags from WIOCGSTAT ioctl
+ */
+#define WD_FREERUN	0x01	/* timer is running, interrupts disabled	*/
+#define WD_EXPIRED	0x02	/* timer has expired						*/
+#define WD_RUNNING	0x04	/* timer is running, interrupts enabled		*/
+#define WD_STOPPED	0x08	/* timer has not been started				*/
+#define WD_SERVICED 0x10	/* timer interrupt was serviced				*/
+
+#endif /* ifndef _SPARC64_WATCHDOG_H */
+
diff --git a/include/asm-sparc/xor.h b/include/asm-sparc/xor.h
index f34b2cfa820..35089a838c3 100644
--- a/include/asm-sparc/xor.h
+++ b/include/asm-sparc/xor.h
@@ -1,269 +1,8 @@
-/*
- * include/asm-sparc/xor.h
- *
- * Optimized RAID-5 checksumming functions for 32-bit Sparc.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2, or (at your option)
- * any later version.
- *
- * You should have received a copy of the GNU General Public License
- * (for example /usr/src/linux/COPYING); if not, write to the Free
- * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- */
-
-/*
- * High speed xor_block operation for RAID4/5 utilizing the
- * ldd/std SPARC instructions.
- *
- * Copyright (C) 1999 Jakub Jelinek (jj@ultra.linux.cz)
- */
-
-static void
-sparc_2(unsigned long bytes, unsigned long *p1, unsigned long *p2)
-{
-	int lines = bytes / (sizeof (long)) / 8;
-
-	do {
-		__asm__ __volatile__(
-		  "ldd [%0 + 0x00], %%g2\n\t"
-		  "ldd [%0 + 0x08], %%g4\n\t"
-		  "ldd [%0 + 0x10], %%o0\n\t"
-		  "ldd [%0 + 0x18], %%o2\n\t"
-		  "ldd [%1 + 0x00], %%o4\n\t"
-		  "ldd [%1 + 0x08], %%l0\n\t"
-		  "ldd [%1 + 0x10], %%l2\n\t"
-		  "ldd [%1 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "std %%g2, [%0 + 0x00]\n\t"
-		  "std %%g4, [%0 + 0x08]\n\t"
-		  "std %%o0, [%0 + 0x10]\n\t"
-		  "std %%o2, [%0 + 0x18]\n"
-		:
-		: "r" (p1), "r" (p2)
-		: "g2", "g3", "g4", "g5",
-		  "o0", "o1", "o2", "o3", "o4", "o5",
-		  "l0", "l1", "l2", "l3", "l4", "l5");
-		p1 += 8;
-		p2 += 8;
-	} while (--lines > 0);
-}
-
-static void
-sparc_3(unsigned long bytes, unsigned long *p1, unsigned long *p2,
-	unsigned long *p3)
-{
-	int lines = bytes / (sizeof (long)) / 8;
-
-	do {
-		__asm__ __volatile__(
-		  "ldd [%0 + 0x00], %%g2\n\t"
-		  "ldd [%0 + 0x08], %%g4\n\t"
-		  "ldd [%0 + 0x10], %%o0\n\t"
-		  "ldd [%0 + 0x18], %%o2\n\t"
-		  "ldd [%1 + 0x00], %%o4\n\t"
-		  "ldd [%1 + 0x08], %%l0\n\t"
-		  "ldd [%1 + 0x10], %%l2\n\t"
-		  "ldd [%1 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%2 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%2 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%2 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%2 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "std %%g2, [%0 + 0x00]\n\t"
-		  "std %%g4, [%0 + 0x08]\n\t"
-		  "std %%o0, [%0 + 0x10]\n\t"
-		  "std %%o2, [%0 + 0x18]\n"
-		:
-		: "r" (p1), "r" (p2), "r" (p3)
-		: "g2", "g3", "g4", "g5",
-		  "o0", "o1", "o2", "o3", "o4", "o5",
-		  "l0", "l1", "l2", "l3", "l4", "l5");
-		p1 += 8;
-		p2 += 8;
-		p3 += 8;
-	} while (--lines > 0);
-}
-
-static void
-sparc_4(unsigned long bytes, unsigned long *p1, unsigned long *p2,
-	unsigned long *p3, unsigned long *p4)
-{
-	int lines = bytes / (sizeof (long)) / 8;
-
-	do {
-		__asm__ __volatile__(
-		  "ldd [%0 + 0x00], %%g2\n\t"
-		  "ldd [%0 + 0x08], %%g4\n\t"
-		  "ldd [%0 + 0x10], %%o0\n\t"
-		  "ldd [%0 + 0x18], %%o2\n\t"
-		  "ldd [%1 + 0x00], %%o4\n\t"
-		  "ldd [%1 + 0x08], %%l0\n\t"
-		  "ldd [%1 + 0x10], %%l2\n\t"
-		  "ldd [%1 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%2 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%2 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%2 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%2 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%3 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%3 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%3 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%3 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "std %%g2, [%0 + 0x00]\n\t"
-		  "std %%g4, [%0 + 0x08]\n\t"
-		  "std %%o0, [%0 + 0x10]\n\t"
-		  "std %%o2, [%0 + 0x18]\n"
-		:
-		: "r" (p1), "r" (p2), "r" (p3), "r" (p4)
-		: "g2", "g3", "g4", "g5",
-		  "o0", "o1", "o2", "o3", "o4", "o5",
-		  "l0", "l1", "l2", "l3", "l4", "l5");
-		p1 += 8;
-		p2 += 8;
-		p3 += 8;
-		p4 += 8;
-	} while (--lines > 0);
-}
-
-static void
-sparc_5(unsigned long bytes, unsigned long *p1, unsigned long *p2,
-	unsigned long *p3, unsigned long *p4, unsigned long *p5)
-{
-	int lines = bytes / (sizeof (long)) / 8;
-
-	do {
-		__asm__ __volatile__(
-		  "ldd [%0 + 0x00], %%g2\n\t"
-		  "ldd [%0 + 0x08], %%g4\n\t"
-		  "ldd [%0 + 0x10], %%o0\n\t"
-		  "ldd [%0 + 0x18], %%o2\n\t"
-		  "ldd [%1 + 0x00], %%o4\n\t"
-		  "ldd [%1 + 0x08], %%l0\n\t"
-		  "ldd [%1 + 0x10], %%l2\n\t"
-		  "ldd [%1 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%2 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%2 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%2 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%2 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%3 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%3 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%3 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%3 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "ldd [%4 + 0x00], %%o4\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "ldd [%4 + 0x08], %%l0\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "ldd [%4 + 0x10], %%l2\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "ldd [%4 + 0x18], %%l4\n\t"
-		  "xor %%g2, %%o4, %%g2\n\t"
-		  "xor %%g3, %%o5, %%g3\n\t"
-		  "xor %%g4, %%l0, %%g4\n\t"
-		  "xor %%g5, %%l1, %%g5\n\t"
-		  "xor %%o0, %%l2, %%o0\n\t"
-		  "xor %%o1, %%l3, %%o1\n\t"
-		  "xor %%o2, %%l4, %%o2\n\t"
-		  "xor %%o3, %%l5, %%o3\n\t"
-		  "std %%g2, [%0 + 0x00]\n\t"
-		  "std %%g4, [%0 + 0x08]\n\t"
-		  "std %%o0, [%0 + 0x10]\n\t"
-		  "std %%o2, [%0 + 0x18]\n"
-		:
-		: "r" (p1), "r" (p2), "r" (p3), "r" (p4), "r" (p5)
-		: "g2", "g3", "g4", "g5",
-		  "o0", "o1", "o2", "o3", "o4", "o5",
-		  "l0", "l1", "l2", "l3", "l4", "l5");
-		p1 += 8;
-		p2 += 8;
-		p3 += 8;
-		p4 += 8;
-		p5 += 8;
-	} while (--lines > 0);
-}
-
-static struct xor_block_template xor_block_SPARC = {
-	.name	= "SPARC",
-	.do_2	= sparc_2,
-	.do_3	= sparc_3,
-	.do_4	= sparc_4,
-	.do_5	= sparc_5,
-};
-
-/* For grins, also test the generic routines.  */
-#include <asm-generic/xor.h>
-
-#undef XOR_TRY_TEMPLATES
-#define XOR_TRY_TEMPLATES				\
-	do {						\
-		xor_speed(&xor_block_8regs);		\
-		xor_speed(&xor_block_32regs);		\
-		xor_speed(&xor_block_SPARC);		\
-	} while (0)
+#ifndef ___ASM_SPARC_XOR_H
+#define ___ASM_SPARC_XOR_H
+#if defined(__sparc__) && defined(__arch64__)
+#include <asm-sparc/xor_64.h>
+#else
+#include <asm-sparc/xor_32.h>
+#endif
+#endif
diff --git a/include/asm-sparc/xor_32.h b/include/asm-sparc/xor_32.h
new file mode 100644
index 00000000000..f34b2cfa820
--- /dev/null
+++ b/include/asm-sparc/xor_32.h
@@ -0,0 +1,269 @@
+/*
+ * include/asm-sparc/xor.h
+ *
+ * Optimized RAID-5 checksumming functions for 32-bit Sparc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * You should have received a copy of the GNU General Public License
+ * (for example /usr/src/linux/COPYING); if not, write to the Free
+ * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+/*
+ * High speed xor_block operation for RAID4/5 utilizing the
+ * ldd/std SPARC instructions.
+ *
+ * Copyright (C) 1999 Jakub Jelinek (jj@ultra.linux.cz)
+ */
+
+static void
+sparc_2(unsigned long bytes, unsigned long *p1, unsigned long *p2)
+{
+	int lines = bytes / (sizeof (long)) / 8;
+
+	do {
+		__asm__ __volatile__(
+		  "ldd [%0 + 0x00], %%g2\n\t"
+		  "ldd [%0 + 0x08], %%g4\n\t"
+		  "ldd [%0 + 0x10], %%o0\n\t"
+		  "ldd [%0 + 0x18], %%o2\n\t"
+		  "ldd [%1 + 0x00], %%o4\n\t"
+		  "ldd [%1 + 0x08], %%l0\n\t"
+		  "ldd [%1 + 0x10], %%l2\n\t"
+		  "ldd [%1 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "std %%g2, [%0 + 0x00]\n\t"
+		  "std %%g4, [%0 + 0x08]\n\t"
+		  "std %%o0, [%0 + 0x10]\n\t"
+		  "std %%o2, [%0 + 0x18]\n"
+		:
+		: "r" (p1), "r" (p2)
+		: "g2", "g3", "g4", "g5",
+		  "o0", "o1", "o2", "o3", "o4", "o5",
+		  "l0", "l1", "l2", "l3", "l4", "l5");
+		p1 += 8;
+		p2 += 8;
+	} while (--lines > 0);
+}
+
+static void
+sparc_3(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	unsigned long *p3)
+{
+	int lines = bytes / (sizeof (long)) / 8;
+
+	do {
+		__asm__ __volatile__(
+		  "ldd [%0 + 0x00], %%g2\n\t"
+		  "ldd [%0 + 0x08], %%g4\n\t"
+		  "ldd [%0 + 0x10], %%o0\n\t"
+		  "ldd [%0 + 0x18], %%o2\n\t"
+		  "ldd [%1 + 0x00], %%o4\n\t"
+		  "ldd [%1 + 0x08], %%l0\n\t"
+		  "ldd [%1 + 0x10], %%l2\n\t"
+		  "ldd [%1 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%2 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%2 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%2 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%2 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "std %%g2, [%0 + 0x00]\n\t"
+		  "std %%g4, [%0 + 0x08]\n\t"
+		  "std %%o0, [%0 + 0x10]\n\t"
+		  "std %%o2, [%0 + 0x18]\n"
+		:
+		: "r" (p1), "r" (p2), "r" (p3)
+		: "g2", "g3", "g4", "g5",
+		  "o0", "o1", "o2", "o3", "o4", "o5",
+		  "l0", "l1", "l2", "l3", "l4", "l5");
+		p1 += 8;
+		p2 += 8;
+		p3 += 8;
+	} while (--lines > 0);
+}
+
+static void
+sparc_4(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	unsigned long *p3, unsigned long *p4)
+{
+	int lines = bytes / (sizeof (long)) / 8;
+
+	do {
+		__asm__ __volatile__(
+		  "ldd [%0 + 0x00], %%g2\n\t"
+		  "ldd [%0 + 0x08], %%g4\n\t"
+		  "ldd [%0 + 0x10], %%o0\n\t"
+		  "ldd [%0 + 0x18], %%o2\n\t"
+		  "ldd [%1 + 0x00], %%o4\n\t"
+		  "ldd [%1 + 0x08], %%l0\n\t"
+		  "ldd [%1 + 0x10], %%l2\n\t"
+		  "ldd [%1 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%2 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%2 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%2 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%2 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%3 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%3 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%3 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%3 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "std %%g2, [%0 + 0x00]\n\t"
+		  "std %%g4, [%0 + 0x08]\n\t"
+		  "std %%o0, [%0 + 0x10]\n\t"
+		  "std %%o2, [%0 + 0x18]\n"
+		:
+		: "r" (p1), "r" (p2), "r" (p3), "r" (p4)
+		: "g2", "g3", "g4", "g5",
+		  "o0", "o1", "o2", "o3", "o4", "o5",
+		  "l0", "l1", "l2", "l3", "l4", "l5");
+		p1 += 8;
+		p2 += 8;
+		p3 += 8;
+		p4 += 8;
+	} while (--lines > 0);
+}
+
+static void
+sparc_5(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	unsigned long *p3, unsigned long *p4, unsigned long *p5)
+{
+	int lines = bytes / (sizeof (long)) / 8;
+
+	do {
+		__asm__ __volatile__(
+		  "ldd [%0 + 0x00], %%g2\n\t"
+		  "ldd [%0 + 0x08], %%g4\n\t"
+		  "ldd [%0 + 0x10], %%o0\n\t"
+		  "ldd [%0 + 0x18], %%o2\n\t"
+		  "ldd [%1 + 0x00], %%o4\n\t"
+		  "ldd [%1 + 0x08], %%l0\n\t"
+		  "ldd [%1 + 0x10], %%l2\n\t"
+		  "ldd [%1 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%2 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%2 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%2 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%2 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%3 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%3 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%3 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%3 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "ldd [%4 + 0x00], %%o4\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "ldd [%4 + 0x08], %%l0\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "ldd [%4 + 0x10], %%l2\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "ldd [%4 + 0x18], %%l4\n\t"
+		  "xor %%g2, %%o4, %%g2\n\t"
+		  "xor %%g3, %%o5, %%g3\n\t"
+		  "xor %%g4, %%l0, %%g4\n\t"
+		  "xor %%g5, %%l1, %%g5\n\t"
+		  "xor %%o0, %%l2, %%o0\n\t"
+		  "xor %%o1, %%l3, %%o1\n\t"
+		  "xor %%o2, %%l4, %%o2\n\t"
+		  "xor %%o3, %%l5, %%o3\n\t"
+		  "std %%g2, [%0 + 0x00]\n\t"
+		  "std %%g4, [%0 + 0x08]\n\t"
+		  "std %%o0, [%0 + 0x10]\n\t"
+		  "std %%o2, [%0 + 0x18]\n"
+		:
+		: "r" (p1), "r" (p2), "r" (p3), "r" (p4), "r" (p5)
+		: "g2", "g3", "g4", "g5",
+		  "o0", "o1", "o2", "o3", "o4", "o5",
+		  "l0", "l1", "l2", "l3", "l4", "l5");
+		p1 += 8;
+		p2 += 8;
+		p3 += 8;
+		p4 += 8;
+		p5 += 8;
+	} while (--lines > 0);
+}
+
+static struct xor_block_template xor_block_SPARC = {
+	.name	= "SPARC",
+	.do_2	= sparc_2,
+	.do_3	= sparc_3,
+	.do_4	= sparc_4,
+	.do_5	= sparc_5,
+};
+
+/* For grins, also test the generic routines.  */
+#include <asm-generic/xor.h>
+
+#undef XOR_TRY_TEMPLATES
+#define XOR_TRY_TEMPLATES				\
+	do {						\
+		xor_speed(&xor_block_8regs);		\
+		xor_speed(&xor_block_32regs);		\
+		xor_speed(&xor_block_SPARC);		\
+	} while (0)
diff --git a/include/asm-sparc/xor_64.h b/include/asm-sparc/xor_64.h
new file mode 100644
index 00000000000..a0233884fc9
--- /dev/null
+++ b/include/asm-sparc/xor_64.h
@@ -0,0 +1,70 @@
+/*
+ * include/asm-sparc64/xor.h
+ *
+ * High speed xor_block operation for RAID4/5 utilizing the
+ * UltraSparc Visual Instruction Set and Niagara block-init
+ * twin-load instructions.
+ *
+ * Copyright (C) 1997, 1999 Jakub Jelinek (jj@ultra.linux.cz)
+ * Copyright (C) 2006 David S. Miller <davem@davemloft.net>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * You should have received a copy of the GNU General Public License
+ * (for example /usr/src/linux/COPYING); if not, write to the Free
+ * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include <asm/spitfire.h>
+
+extern void xor_vis_2(unsigned long, unsigned long *, unsigned long *);
+extern void xor_vis_3(unsigned long, unsigned long *, unsigned long *,
+		      unsigned long *);
+extern void xor_vis_4(unsigned long, unsigned long *, unsigned long *,
+		      unsigned long *, unsigned long *);
+extern void xor_vis_5(unsigned long, unsigned long *, unsigned long *,
+		      unsigned long *, unsigned long *, unsigned long *);
+
+/* XXX Ugh, write cheetah versions... -DaveM */
+
+static struct xor_block_template xor_block_VIS = {
+        .name	= "VIS",
+        .do_2	= xor_vis_2,
+        .do_3	= xor_vis_3,
+        .do_4	= xor_vis_4,
+        .do_5	= xor_vis_5,
+};
+
+extern void xor_niagara_2(unsigned long, unsigned long *, unsigned long *);
+extern void xor_niagara_3(unsigned long, unsigned long *, unsigned long *,
+			  unsigned long *);
+extern void xor_niagara_4(unsigned long, unsigned long *, unsigned long *,
+			  unsigned long *, unsigned long *);
+extern void xor_niagara_5(unsigned long, unsigned long *, unsigned long *,
+			  unsigned long *, unsigned long *, unsigned long *);
+
+static struct xor_block_template xor_block_niagara = {
+        .name	= "Niagara",
+        .do_2	= xor_niagara_2,
+        .do_3	= xor_niagara_3,
+        .do_4	= xor_niagara_4,
+        .do_5	= xor_niagara_5,
+};
+
+#undef XOR_TRY_TEMPLATES
+#define XOR_TRY_TEMPLATES				\
+	do {						\
+		xor_speed(&xor_block_VIS);		\
+		xor_speed(&xor_block_niagara);		\
+	} while (0)
+
+/* For VIS for everything except Niagara.  */
+#define XOR_SELECT_TEMPLATE(FASTEST) \
+	((tlb_type == hypervisor && \
+	  (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 || \
+	   sun4v_chip_type == SUN4V_CHIP_NIAGARA2)) ? \
+	 &xor_block_niagara : \
+	 &xor_block_VIS)
diff --git a/include/asm-sparc64/Kbuild b/include/asm-sparc64/Kbuild
index dce1cf9a931..6cdaf9d33b3 100644
--- a/include/asm-sparc64/Kbuild
+++ b/include/asm-sparc64/Kbuild
@@ -1,23 +1 @@
-include include/asm-generic/Kbuild.asm
-
-ALTARCH := sparc
-ARCHDEF := defined __sparc__ && defined __arch64__
-ALTARCHDEF := defined __sparc__ && !defined __arch64__
-
-header-y += apb.h
-header-y += asi.h
-header-y += bbc.h
-header-y += bpp.h
-header-y += display7seg.h
-header-y += envctrl.h
-header-y += openprom.h
-header-y += openpromio.h
-header-y += psrcompat.h
-header-y += pstate.h
-header-y += reg.h
-header-y += uctx.h
-header-y += utrap.h
-header-y += watchdog.h
-
-unifdef-y += fbio.h
-unifdef-y += perfctr.h
+# dummy file to avoid breaking make headers_install
diff --git a/include/asm-sparc64/agp.h b/include/asm-sparc64/agp.h
index e9fcf0e781e..eb8d4b3f516 100644
--- a/include/asm-sparc64/agp.h
+++ b/include/asm-sparc64/agp.h
@@ -1,20 +1 @@
-#ifndef AGP_H
-#define AGP_H 1
-
-/* dummy for now */
-
-#define map_page_into_agp(page) 
-#define unmap_page_from_agp(page) 
-#define flush_agp_cache() mb()
-
-/* Convert a physical address to an address suitable for the GART. */
-#define phys_to_gart(x) (x)
-#define gart_to_phys(x) (x)
-
-/* GATT allocation. Returns/accepts GATT kernel virtual address. */
-#define alloc_gatt_pages(order)		\
-	((char *)__get_free_pages(GFP_KERNEL, (order)))
-#define free_gatt_pages(table, order)	\
-	free_pages((unsigned long)(table), (order))
-
-#endif
+#include <asm-sparc/agp.h>
diff --git a/include/asm-sparc64/apb.h b/include/asm-sparc64/apb.h
index 8f3b57db810..5e236ca6e49 100644
--- a/include/asm-sparc64/apb.h
+++ b/include/asm-sparc64/apb.h
@@ -1,36 +1 @@
-/*
- * apb.h: Advanced PCI Bridge Configuration Registers and Bits
- *
- * Copyright (C) 1998  Eddie C. Dost  (ecd@skynet.be)
- */
-
-#ifndef _SPARC64_APB_H
-#define _SPARC64_APB_H
-
-#define APB_TICK_REGISTER			0xb0
-#define APB_INT_ACK				0xb8
-#define APB_PRIMARY_MASTER_RETRY_LIMIT		0xc0
-#define APB_DMA_ASFR				0xc8
-#define APB_DMA_AFAR				0xd0
-#define APB_PIO_TARGET_RETRY_LIMIT		0xd8
-#define APB_PIO_TARGET_LATENCY_TIMER		0xd9
-#define APB_DMA_TARGET_RETRY_LIMIT		0xda
-#define APB_DMA_TARGET_LATENCY_TIMER		0xdb
-#define APB_SECONDARY_MASTER_RETRY_LIMIT	0xdc
-#define APB_SECONDARY_CONTROL			0xdd
-#define APB_IO_ADDRESS_MAP			0xde
-#define APB_MEM_ADDRESS_MAP			0xdf
-
-#define APB_PCI_CONTROL_LOW			0xe0
-#  define APB_PCI_CTL_LOW_ARB_PARK			(1 << 21)
-#  define APB_PCI_CTL_LOW_ERRINT_EN			(1 << 8)
-
-#define APB_PCI_CONTROL_HIGH			0xe4
-#  define APB_PCI_CTL_HIGH_SERR				(1 << 2)
-#  define APB_PCI_CTL_HIGH_ARBITER_EN			(1 << 0)
-
-#define APB_PIO_ASFR				0xe8
-#define APB_PIO_AFAR				0xf0
-#define APB_DIAG_REGISTER			0xf8
-
-#endif /* !(_SPARC64_APB_H) */
+#include <asm-sparc/apb.h>
diff --git a/include/asm-sparc64/asi.h b/include/asm-sparc64/asi.h
index bc57c405e7d..9b7110c516e 100644
--- a/include/asm-sparc64/asi.h
+++ b/include/asm-sparc64/asi.h
@@ -1,160 +1 @@
-#ifndef _SPARC64_ASI_H
-#define _SPARC64_ASI_H
-
-/* asi.h:  Address Space Identifier values for the V9.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-/* V9 Architecture mandary ASIs. */
-#define ASI_N			0x04 /* Nucleus				*/
-#define ASI_NL			0x0c /* Nucleus, little endian		*/
-#define ASI_AIUP		0x10 /* Primary, user			*/
-#define ASI_AIUS		0x11 /* Secondary, user			*/
-#define ASI_AIUPL		0x18 /* Primary, user, little endian	*/
-#define ASI_AIUSL		0x19 /* Secondary, user, little endian	*/
-#define ASI_P			0x80 /* Primary, implicit		*/
-#define ASI_S			0x81 /* Secondary, implicit		*/
-#define ASI_PNF			0x82 /* Primary, no fault		*/
-#define ASI_SNF			0x83 /* Secondary, no fault		*/
-#define ASI_PL			0x88 /* Primary, implicit, l-endian	*/
-#define ASI_SL			0x89 /* Secondary, implicit, l-endian	*/
-#define ASI_PNFL		0x8a /* Primary, no fault, l-endian	*/
-#define ASI_SNFL		0x8b /* Secondary, no fault, l-endian	*/
-
-/* SpitFire and later extended ASIs.  The "(III)" marker designates
- * UltraSparc-III and later specific ASIs.  The "(CMT)" marker designates
- * Chip Multi Threading specific ASIs.  "(NG)" designates Niagara specific
- * ASIs, "(4V)" designates SUN4V specific ASIs.
- */
-#define ASI_PHYS_USE_EC		0x14 /* PADDR, E-cachable		*/
-#define ASI_PHYS_BYPASS_EC_E	0x15 /* PADDR, E-bit			*/
-#define ASI_BLK_AIUP_4V		0x16 /* (4V) Prim, user, block ld/st	*/
-#define ASI_BLK_AIUS_4V		0x17 /* (4V) Sec, user, block ld/st	*/
-#define ASI_PHYS_USE_EC_L	0x1c /* PADDR, E-cachable, little endian*/
-#define ASI_PHYS_BYPASS_EC_E_L	0x1d /* PADDR, E-bit, little endian	*/
-#define ASI_BLK_AIUP_L_4V	0x1e /* (4V) Prim, user, block, l-endian*/
-#define ASI_BLK_AIUS_L_4V	0x1f /* (4V) Sec, user, block, l-endian	*/
-#define ASI_SCRATCHPAD		0x20 /* (4V) Scratch Pad Registers	*/
-#define ASI_MMU			0x21 /* (4V) MMU Context Registers	*/
-#define ASI_BLK_INIT_QUAD_LDD_AIUS 0x23 /* (NG) init-store, twin load,
-					 * secondary, user
-					 */
-#define ASI_NUCLEUS_QUAD_LDD	0x24 /* Cachable, qword load		*/
-#define ASI_QUEUE		0x25 /* (4V) Interrupt Queue Registers	*/
-#define ASI_QUAD_LDD_PHYS_4V	0x26 /* (4V) Physical, qword load	*/
-#define ASI_NUCLEUS_QUAD_LDD_L	0x2c /* Cachable, qword load, l-endian 	*/
-#define ASI_QUAD_LDD_PHYS_L_4V	0x2e /* (4V) Phys, qword load, l-endian	*/
-#define ASI_PCACHE_DATA_STATUS	0x30 /* (III) PCache data stat RAM diag	*/
-#define ASI_PCACHE_DATA		0x31 /* (III) PCache data RAM diag	*/
-#define ASI_PCACHE_TAG		0x32 /* (III) PCache tag RAM diag	*/
-#define ASI_PCACHE_SNOOP_TAG	0x33 /* (III) PCache snoop tag RAM diag	*/
-#define ASI_QUAD_LDD_PHYS	0x34 /* (III+) PADDR, qword load	*/
-#define ASI_WCACHE_VALID_BITS	0x38 /* (III) WCache Valid Bits diag	*/
-#define ASI_WCACHE_DATA		0x39 /* (III) WCache data RAM diag	*/
-#define ASI_WCACHE_TAG		0x3a /* (III) WCache tag RAM diag	*/
-#define ASI_WCACHE_SNOOP_TAG	0x3b /* (III) WCache snoop tag RAM diag	*/
-#define ASI_QUAD_LDD_PHYS_L	0x3c /* (III+) PADDR, qw-load, l-endian	*/
-#define ASI_SRAM_FAST_INIT	0x40 /* (III+) Fast SRAM init		*/
-#define ASI_CORE_AVAILABLE	0x41 /* (CMT) LP Available		*/
-#define ASI_CORE_ENABLE_STAT	0x41 /* (CMT) LP Enable Status		*/
-#define ASI_CORE_ENABLE		0x41 /* (CMT) LP Enable RW		*/
-#define ASI_XIR_STEERING	0x41 /* (CMT) XIR Steering RW		*/
-#define ASI_CORE_RUNNING_RW	0x41 /* (CMT) LP Running RW		*/
-#define ASI_CORE_RUNNING_W1S	0x41 /* (CMT) LP Running Write-One Set	*/
-#define ASI_CORE_RUNNING_W1C	0x41 /* (CMT) LP Running Write-One Clr	*/
-#define ASI_CORE_RUNNING_STAT	0x41 /* (CMT) LP Running Status		*/
-#define ASI_CMT_ERROR_STEERING	0x41 /* (CMT) Error Steering RW		*/
-#define ASI_DCACHE_INVALIDATE	0x42 /* (III) DCache Invalidate diag	*/
-#define ASI_DCACHE_UTAG		0x43 /* (III) DCache uTag diag		*/
-#define ASI_DCACHE_SNOOP_TAG	0x44 /* (III) DCache snoop tag RAM diag	*/
-#define ASI_LSU_CONTROL		0x45 /* Load-store control unit		*/
-#define ASI_DCU_CONTROL_REG	0x45 /* (III) DCache Unit Control reg	*/
-#define ASI_DCACHE_DATA		0x46 /* DCache data-ram diag access	*/
-#define ASI_DCACHE_TAG		0x47 /* Dcache tag/valid ram diag access*/
-#define ASI_INTR_DISPATCH_STAT	0x48 /* IRQ vector dispatch status	*/
-#define ASI_INTR_RECEIVE	0x49 /* IRQ vector receive status	*/
-#define ASI_UPA_CONFIG		0x4a /* UPA config space		*/
-#define ASI_JBUS_CONFIG		0x4a /* (IIIi) JBUS Config Register	*/
-#define ASI_SAFARI_CONFIG	0x4a /* (III) Safari Config Register	*/
-#define ASI_SAFARI_ADDRESS	0x4a /* (III) Safari Address Register	*/
-#define ASI_ESTATE_ERROR_EN	0x4b /* E-cache error enable space	*/
-#define ASI_AFSR		0x4c /* Async fault status register	*/
-#define ASI_AFAR		0x4d /* Async fault address register	*/
-#define ASI_EC_TAG_DATA		0x4e /* E-cache tag/valid ram diag acc	*/
-#define ASI_IMMU		0x50 /* Insn-MMU main register space	*/
-#define ASI_IMMU_TSB_8KB_PTR	0x51 /* Insn-MMU 8KB TSB pointer reg	*/
-#define ASI_IMMU_TSB_64KB_PTR	0x52 /* Insn-MMU 64KB TSB pointer reg	*/
-#define ASI_ITLB_DATA_IN	0x54 /* Insn-MMU TLB data in reg	*/
-#define ASI_ITLB_DATA_ACCESS	0x55 /* Insn-MMU TLB data access reg	*/
-#define ASI_ITLB_TAG_READ	0x56 /* Insn-MMU TLB tag read reg	*/
-#define ASI_IMMU_DEMAP		0x57 /* Insn-MMU TLB demap		*/
-#define ASI_DMMU		0x58 /* Data-MMU main register space	*/
-#define ASI_DMMU_TSB_8KB_PTR	0x59 /* Data-MMU 8KB TSB pointer reg	*/
-#define ASI_DMMU_TSB_64KB_PTR	0x5a /* Data-MMU 16KB TSB pointer reg	*/
-#define ASI_DMMU_TSB_DIRECT_PTR	0x5b /* Data-MMU TSB direct pointer reg	*/
-#define ASI_DTLB_DATA_IN	0x5c /* Data-MMU TLB data in reg	*/
-#define ASI_DTLB_DATA_ACCESS	0x5d /* Data-MMU TLB data access reg	*/
-#define ASI_DTLB_TAG_READ	0x5e /* Data-MMU TLB tag read reg	*/
-#define ASI_DMMU_DEMAP		0x5f /* Data-MMU TLB demap		*/
-#define ASI_IIU_INST_TRAP	0x60 /* (III) Instruction Breakpoint	*/
-#define ASI_INTR_ID		0x63 /* (CMT) Interrupt ID register	*/
-#define ASI_CORE_ID		0x63 /* (CMT) LP ID register		*/
-#define ASI_CESR_ID		0x63 /* (CMT) CESR ID register		*/
-#define ASI_IC_INSTR		0x66 /* Insn cache instrucion ram diag	*/
-#define ASI_IC_TAG		0x67 /* Insn cache tag/valid ram diag 	*/
-#define ASI_IC_STAG		0x68 /* (III) Insn cache snoop tag ram	*/
-#define ASI_IC_PRE_DECODE	0x6e /* Insn cache pre-decode ram diag	*/
-#define ASI_IC_NEXT_FIELD	0x6f /* Insn cache next-field ram diag	*/
-#define ASI_BRPRED_ARRAY	0x6f /* (III) Branch Prediction RAM diag*/
-#define ASI_BLK_AIUP		0x70 /* Primary, user, block load/store	*/
-#define ASI_BLK_AIUS		0x71 /* Secondary, user, block ld/st	*/
-#define ASI_MCU_CTRL_REG	0x72 /* (III) Memory controller regs	*/
-#define ASI_EC_DATA		0x74 /* (III) E-cache data staging reg	*/
-#define ASI_EC_CTRL		0x75 /* (III) E-cache control reg	*/
-#define ASI_EC_W		0x76 /* E-cache diag write access	*/
-#define ASI_UDB_ERROR_W		0x77 /* External UDB error regs W	*/
-#define ASI_UDB_CONTROL_W	0x77 /* External UDB control regs W	*/
-#define ASI_INTR_W		0x77 /* IRQ vector dispatch write	*/
-#define ASI_INTR_DATAN_W	0x77 /* (III) Out irq vector data reg N	*/
-#define ASI_INTR_DISPATCH_W	0x77 /* (III) Interrupt vector dispatch	*/
-#define ASI_BLK_AIUPL		0x78 /* Primary, user, little, blk ld/st*/
-#define ASI_BLK_AIUSL		0x79 /* Secondary, user, little, blk ld/st*/
-#define ASI_EC_R		0x7e /* E-cache diag read access	*/
-#define ASI_UDBH_ERROR_R	0x7f /* External UDB error regs rd hi	*/
-#define ASI_UDBL_ERROR_R	0x7f /* External UDB error regs rd low	*/
-#define ASI_UDBH_CONTROL_R	0x7f /* External UDB control regs rd hi	*/
-#define ASI_UDBL_CONTROL_R	0x7f /* External UDB control regs rd low*/
-#define ASI_INTR_R		0x7f /* IRQ vector dispatch read	*/
-#define ASI_INTR_DATAN_R	0x7f /* (III) In irq vector data reg N	*/
-#define ASI_PST8_P		0xc0 /* Primary, 8 8-bit, partial	*/
-#define ASI_PST8_S		0xc1 /* Secondary, 8 8-bit, partial	*/
-#define ASI_PST16_P		0xc2 /* Primary, 4 16-bit, partial	*/
-#define ASI_PST16_S		0xc3 /* Secondary, 4 16-bit, partial	*/
-#define ASI_PST32_P		0xc4 /* Primary, 2 32-bit, partial	*/
-#define ASI_PST32_S		0xc5 /* Secondary, 2 32-bit, partial	*/
-#define ASI_PST8_PL		0xc8 /* Primary, 8 8-bit, partial, L	*/
-#define ASI_PST8_SL		0xc9 /* Secondary, 8 8-bit, partial, L	*/
-#define ASI_PST16_PL		0xca /* Primary, 4 16-bit, partial, L	*/
-#define ASI_PST16_SL		0xcb /* Secondary, 4 16-bit, partial, L	*/
-#define ASI_PST32_PL		0xcc /* Primary, 2 32-bit, partial, L	*/
-#define ASI_PST32_SL		0xcd /* Secondary, 2 32-bit, partial, L	*/
-#define ASI_FL8_P		0xd0 /* Primary, 1 8-bit, fpu ld/st	*/
-#define ASI_FL8_S		0xd1 /* Secondary, 1 8-bit, fpu ld/st	*/
-#define ASI_FL16_P		0xd2 /* Primary, 1 16-bit, fpu ld/st	*/
-#define ASI_FL16_S		0xd3 /* Secondary, 1 16-bit, fpu ld/st	*/
-#define ASI_FL8_PL		0xd8 /* Primary, 1 8-bit, fpu ld/st, L	*/
-#define ASI_FL8_SL		0xd9 /* Secondary, 1 8-bit, fpu ld/st, L*/
-#define ASI_FL16_PL		0xda /* Primary, 1 16-bit, fpu ld/st, L	*/
-#define ASI_FL16_SL		0xdb /* Secondary, 1 16-bit, fpu ld/st,L*/
-#define ASI_BLK_COMMIT_P	0xe0 /* Primary, blk store commit	*/
-#define ASI_BLK_COMMIT_S	0xe1 /* Secondary, blk store commit	*/
-#define ASI_BLK_INIT_QUAD_LDD_P	0xe2 /* (NG) init-store, twin load,
-				      * primary, implicit
-				      */
-#define ASI_BLK_P		0xf0 /* Primary, blk ld/st		*/
-#define ASI_BLK_S		0xf1 /* Secondary, blk ld/st		*/
-#define ASI_BLK_PL		0xf8 /* Primary, blk ld/st, little	*/
-#define ASI_BLK_SL		0xf9 /* Secondary, blk ld/st, little	*/
-
-#endif /* _SPARC64_ASI_H */
+#include <asm-sparc/asi.h>
diff --git a/include/asm-sparc64/atomic.h b/include/asm-sparc64/atomic.h
index 2c71ec4a3b1..f5126826ba3 100644
--- a/include/asm-sparc64/atomic.h
+++ b/include/asm-sparc64/atomic.h
@@ -1,128 +1 @@
-/* atomic.h: Thankfully the V9 is at least reasonable for this
- *           stuff.
- *
- * Copyright (C) 1996, 1997, 2000 David S. Miller (davem@redhat.com)
- */
-
-#ifndef __ARCH_SPARC64_ATOMIC__
-#define __ARCH_SPARC64_ATOMIC__
-
-#include <linux/types.h>
-#include <asm/system.h>
-
-typedef struct { volatile int counter; } atomic_t;
-typedef struct { volatile __s64 counter; } atomic64_t;
-
-#define ATOMIC_INIT(i)		{ (i) }
-#define ATOMIC64_INIT(i)	{ (i) }
-
-#define atomic_read(v)		((v)->counter)
-#define atomic64_read(v)	((v)->counter)
-
-#define atomic_set(v, i)	(((v)->counter) = i)
-#define atomic64_set(v, i)	(((v)->counter) = i)
-
-extern void atomic_add(int, atomic_t *);
-extern void atomic64_add(int, atomic64_t *);
-extern void atomic_sub(int, atomic_t *);
-extern void atomic64_sub(int, atomic64_t *);
-
-extern int atomic_add_ret(int, atomic_t *);
-extern int atomic64_add_ret(int, atomic64_t *);
-extern int atomic_sub_ret(int, atomic_t *);
-extern int atomic64_sub_ret(int, atomic64_t *);
-
-#define atomic_dec_return(v) atomic_sub_ret(1, v)
-#define atomic64_dec_return(v) atomic64_sub_ret(1, v)
-
-#define atomic_inc_return(v) atomic_add_ret(1, v)
-#define atomic64_inc_return(v) atomic64_add_ret(1, v)
-
-#define atomic_sub_return(i, v) atomic_sub_ret(i, v)
-#define atomic64_sub_return(i, v) atomic64_sub_ret(i, v)
-
-#define atomic_add_return(i, v) atomic_add_ret(i, v)
-#define atomic64_add_return(i, v) atomic64_add_ret(i, v)
-
-/*
- * atomic_inc_and_test - increment and test
- * @v: pointer of type atomic_t
- *
- * Atomically increments @v by 1
- * and returns true if the result is zero, or false for all
- * other cases.
- */
-#define atomic_inc_and_test(v) (atomic_inc_return(v) == 0)
-#define atomic64_inc_and_test(v) (atomic64_inc_return(v) == 0)
-
-#define atomic_sub_and_test(i, v) (atomic_sub_ret(i, v) == 0)
-#define atomic64_sub_and_test(i, v) (atomic64_sub_ret(i, v) == 0)
-
-#define atomic_dec_and_test(v) (atomic_sub_ret(1, v) == 0)
-#define atomic64_dec_and_test(v) (atomic64_sub_ret(1, v) == 0)
-
-#define atomic_inc(v) atomic_add(1, v)
-#define atomic64_inc(v) atomic64_add(1, v)
-
-#define atomic_dec(v) atomic_sub(1, v)
-#define atomic64_dec(v) atomic64_sub(1, v)
-
-#define atomic_add_negative(i, v) (atomic_add_ret(i, v) < 0)
-#define atomic64_add_negative(i, v) (atomic64_add_ret(i, v) < 0)
-
-#define atomic_cmpxchg(v, o, n) (cmpxchg(&((v)->counter), (o), (n)))
-#define atomic_xchg(v, new) (xchg(&((v)->counter), new))
-
-static inline int atomic_add_unless(atomic_t *v, int a, int u)
-{
-	int c, old;
-	c = atomic_read(v);
-	for (;;) {
-		if (unlikely(c == (u)))
-			break;
-		old = atomic_cmpxchg((v), c, c + (a));
-		if (likely(old == c))
-			break;
-		c = old;
-	}
-	return c != (u);
-}
-
-#define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
-
-#define atomic64_cmpxchg(v, o, n) \
-	((__typeof__((v)->counter))cmpxchg(&((v)->counter), (o), (n)))
-#define atomic64_xchg(v, new) (xchg(&((v)->counter), new))
-
-static inline int atomic64_add_unless(atomic64_t *v, long a, long u)
-{
-	long c, old;
-	c = atomic64_read(v);
-	for (;;) {
-		if (unlikely(c == (u)))
-			break;
-		old = atomic64_cmpxchg((v), c, c + (a));
-		if (likely(old == c))
-			break;
-		c = old;
-	}
-	return c != (u);
-}
-
-#define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
-
-/* Atomic operations are already serializing */
-#ifdef CONFIG_SMP
-#define smp_mb__before_atomic_dec()	membar_storeload_loadload();
-#define smp_mb__after_atomic_dec()	membar_storeload_storestore();
-#define smp_mb__before_atomic_inc()	membar_storeload_loadload();
-#define smp_mb__after_atomic_inc()	membar_storeload_storestore();
-#else
-#define smp_mb__before_atomic_dec()	barrier()
-#define smp_mb__after_atomic_dec()	barrier()
-#define smp_mb__before_atomic_inc()	barrier()
-#define smp_mb__after_atomic_inc()	barrier()
-#endif
-
-#include <asm-generic/atomic.h>
-#endif /* !(__ARCH_SPARC64_ATOMIC__) */
+#include <asm-sparc/atomic.h>
diff --git a/include/asm-sparc64/auxio.h b/include/asm-sparc64/auxio.h
index c4100494c7a..46c9042f30b 100644
--- a/include/asm-sparc64/auxio.h
+++ b/include/asm-sparc64/auxio.h
@@ -1,100 +1 @@
-/*
- * auxio.h:  Definitions and code for the Auxiliary I/O registers.
- *
- * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
- *
- * Refactoring for unified NCR/PCIO support 2002 Eric Brower (ebrower@usa.net)
- */
-#ifndef _SPARC64_AUXIO_H
-#define _SPARC64_AUXIO_H
-
-/* AUXIO implementations:
- * sbus-based NCR89C105 "Slavio"
- *	LED/Floppy (AUX1) register
- *	Power (AUX2) register
- *
- * ebus-based auxio on PCIO 
- *	LED Auxio Register
- *	Power Auxio Register
- *
- * Register definitions from NCR _NCR89C105 Chip Specification_
- * 
- * SLAVIO AUX1 @ 0x1900000
- * -------------------------------------------------
- * | (R) | (R) |  D  | (R) |  E  |  M  |  T  |  L  |
- * -------------------------------------------------
- * (R) - bit 7:6,4 are reserved and should be masked in s/w
- *  D  - Floppy Density Sense (1=high density) R/O
- *  E  - Link Test Enable, directly reflected on AT&T 7213 LTE pin
- *  M  - Monitor/Mouse Mux, directly reflected on MON_MSE_MUX pin
- *  T  - Terminal Count: sends TC pulse to 82077 floppy controller
- *  L  - System LED on front panel (0=off, 1=on) 
- */
-#define AUXIO_AUX1_MASK		0xc0 /* Mask bits 		*/
-#define AUXIO_AUX1_FDENS	0x20 /* Floppy Density Sense	*/
-#define AUXIO_AUX1_LTE 		0x08 /* Link Test Enable 	*/
-#define AUXIO_AUX1_MMUX		0x04 /* Monitor/Mouse Mux	*/
-#define AUXIO_AUX1_FTCNT	0x02 /* Terminal Count, 	*/
-#define AUXIO_AUX1_LED		0x01 /* System LED		*/
-
-/* SLAVIO AUX2 @ 0x1910000
- * -------------------------------------------------
- * | (R) | (R) |  D  | (R) | (R) | (R) |  C  |  F  |
- * -------------------------------------------------
- * (R) - bits 7:6,4:2 are reserved and should be masked in s/w
- *  D  - Power Failure Detect (1=power fail)
- *  C  - Clear Power Failure Detect Int (1=clear)
- *  F  - Power Off (1=power off)
- */
-#define AUXIO_AUX2_MASK		0xdc /* Mask Bits		*/
-#define AUXIO_AUX2_PFAILDET	0x20 /* Power Fail Detect	*/
-#define AUXIO_AUX2_PFAILCLR 	0x02 /* Clear Pwr Fail Det Intr	*/
-#define AUXIO_AUX2_PWR_OFF	0x01 /* Power Off		*/
-
-/* Register definitions from Sun Microsystems _PCIO_ p/n 802-7837
- *
- * PCIO LED Auxio @ 0x726000
- * -------------------------------------------------
- * |             31:1 Unused                 | LED |
- * -------------------------------------------------
- * Bits 31:1 unused
- * LED - System LED on front panel (0=off, 1=on)
- */
-#define AUXIO_PCIO_LED		0x01 /* System LED 		*/ 
-
-/* PCIO Power Auxio @ 0x724000
- * -------------------------------------------------
- * |             31:2 Unused           | CPO | SPO |
- * -------------------------------------------------
- * Bits 31:2 unused
- * CPO - Courtesy Power Off (1=off)
- * SPO - System Power Off   (1=off)
- */
-#define AUXIO_PCIO_CPWR_OFF	0x02 /* Courtesy Power Off	*/
-#define AUXIO_PCIO_SPWR_OFF	0x01 /* System Power Off	*/
-
-#ifndef __ASSEMBLY__
-
-extern void __iomem *auxio_register;
-
-#define AUXIO_LTE_ON	1
-#define AUXIO_LTE_OFF	0
-
-/* auxio_set_lte - Set Link Test Enable (TPE Link Detect)
- *
- * on - AUXIO_LTE_ON or AUXIO_LTE_OFF
- */
-extern void auxio_set_lte(int on);
-
-#define AUXIO_LED_ON	1
-#define AUXIO_LED_OFF	0
-
-/* auxio_set_led - Set system front panel LED 
- *
- * on - AUXIO_LED_ON or AUXIO_LED_OFF
- */
-extern void auxio_set_led(int on);
-
-#endif /* ifndef __ASSEMBLY__ */ 
-
-#endif /* !(_SPARC64_AUXIO_H) */
+#include <asm-sparc/auxio.h>
diff --git a/include/asm-sparc64/backoff.h b/include/asm-sparc64/backoff.h
index fa1fdf67e35..8ee26d947e0 100644
--- a/include/asm-sparc64/backoff.h
+++ b/include/asm-sparc64/backoff.h
@@ -1,31 +1 @@
-#ifndef _SPARC64_BACKOFF_H
-#define _SPARC64_BACKOFF_H
-
-#define BACKOFF_LIMIT	(4 * 1024)
-
-#ifdef CONFIG_SMP
-
-#define BACKOFF_SETUP(reg)	\
-	mov	1, reg
-
-#define BACKOFF_SPIN(reg, tmp, label)	\
-	mov	reg, tmp; \
-88:	brnz,pt	tmp, 88b; \
-	 sub	tmp, 1, tmp; \
-	set	BACKOFF_LIMIT, tmp; \
-	cmp	reg, tmp; \
-	bg,pn	%xcc, label; \
-	 nop; \
-	ba,pt	%xcc, label; \
-	 sllx	reg, 1, reg;
-
-#else
-
-#define BACKOFF_SETUP(reg)
-#define BACKOFF_SPIN(reg, tmp, label) \
-	ba,pt	%xcc, label; \
-	 nop;
-
-#endif
-
-#endif /* _SPARC64_BACKOFF_H */
+#include <asm-sparc/backoff.h>
diff --git a/include/asm-sparc64/bbc.h b/include/asm-sparc64/bbc.h
index 423a85800aa..06e8b630651 100644
--- a/include/asm-sparc64/bbc.h
+++ b/include/asm-sparc64/bbc.h
@@ -1,225 +1 @@
-/*
- * bbc.h: Defines for BootBus Controller found on UltraSPARC-III
- *        systems.
- *
- * Copyright (C) 2000 David S. Miller (davem@redhat.com)
- */
-
-#ifndef _SPARC64_BBC_H
-#define _SPARC64_BBC_H
-
-/* Register sizes are indicated by "B" (Byte, 1-byte),
- * "H" (Half-word, 2 bytes), "W" (Word, 4 bytes) or
- * "Q" (Quad, 8 bytes) inside brackets.
- */
-
-#define BBC_AID		0x00	/* [B] Agent ID			*/
-#define BBC_DEVP	0x01	/* [B] Device Present		*/
-#define BBC_ARB		0x02	/* [B] Arbitration		*/
-#define BBC_QUIESCE	0x03	/* [B] Quiesce			*/
-#define BBC_WDACTION	0x04	/* [B] Watchdog Action		*/
-#define BBC_SPG		0x06	/* [B] Soft POR Gen		*/
-#define BBC_SXG		0x07	/* [B] Soft XIR Gen		*/
-#define BBC_PSRC	0x08	/* [W] POR Source		*/
-#define BBC_XSRC	0x0c	/* [B] XIR Source		*/
-#define BBC_CSC		0x0d	/* [B] Clock Synthesizers Control*/
-#define BBC_ES_CTRL	0x0e	/* [H] Energy Star Control	*/
-#define BBC_ES_ACT	0x10	/* [W] E* Assert Change Time	*/
-#define BBC_ES_DACT	0x14	/* [B] E* De-Assert Change Time	*/
-#define BBC_ES_DABT	0x15	/* [B] E* De-Assert Bypass Time	*/
-#define BBC_ES_ABT	0x16	/* [H] E* Assert Bypass Time	*/
-#define BBC_ES_PST	0x18	/* [W] E* PLL Settle Time	*/
-#define BBC_ES_FSL	0x1c	/* [W] E* Frequency Switch Latency*/
-#define BBC_EBUST	0x20	/* [Q] EBUS Timing		*/
-#define BBC_JTAG_CMD	0x28	/* [W] JTAG+ Command		*/
-#define BBC_JTAG_CTRL	0x2c	/* [B] JTAG+ Control		*/
-#define BBC_I2C_SEL	0x2d	/* [B] I2C Selection		*/
-#define BBC_I2C_0_S1	0x2e	/* [B] I2C ctrlr-0 reg S1	*/
-#define BBC_I2C_0_S0	0x2f	/* [B] I2C ctrlr-0 regs S0,S0',S2,S3*/
-#define BBC_I2C_1_S1	0x30	/* [B] I2C ctrlr-1 reg S1	*/
-#define BBC_I2C_1_S0	0x31	/* [B] I2C ctrlr-1 regs S0,S0',S2,S3*/
-#define BBC_KBD_BEEP	0x32	/* [B] Keyboard Beep		*/
-#define BBC_KBD_BCNT	0x34	/* [W] Keyboard Beep Counter	*/
-
-#define BBC_REGS_SIZE	0x40
-
-/* There is a 2K scratch ram area at offset 0x80000 but I doubt
- * we will use it for anything.
- */
-
-/* Agent ID register.  This register shows the Safari Agent ID
- * for the processors.  The value returned depends upon which
- * cpu is reading the register.
- */
-#define BBC_AID_ID	0x07	/* Safari ID		*/
-#define BBC_AID_RESV	0xf8	/* Reserved		*/
-
-/* Device Present register.  One can determine which cpus are actually
- * present in the machine by interrogating this register.
- */
-#define BBC_DEVP_CPU0	0x01	/* Processor 0 present	*/
-#define BBC_DEVP_CPU1	0x02	/* Processor 1 present	*/
-#define BBC_DEVP_CPU2	0x04	/* Processor 2 present	*/
-#define BBC_DEVP_CPU3	0x08	/* Processor 3 present	*/
-#define BBC_DEVP_RESV	0xf0	/* Reserved		*/
-
-/* Arbitration register.  This register is used to block access to
- * the BBC from a particular cpu.
- */
-#define BBC_ARB_CPU0	0x01	/* Enable cpu 0 BBC arbitratrion */
-#define BBC_ARB_CPU1	0x02	/* Enable cpu 1 BBC arbitratrion */
-#define BBC_ARB_CPU2	0x04	/* Enable cpu 2 BBC arbitratrion */
-#define BBC_ARB_CPU3	0x08	/* Enable cpu 3 BBC arbitratrion */
-#define BBC_ARB_RESV	0xf0	/* Reserved			 */
-
-/* Quiesce register.  Bus and BBC segments for cpus can be disabled
- * with this register, ie. for hot plugging.
- */
-#define BBC_QUIESCE_S02	0x01	/* Quiesce Safari segment for cpu 0 and 2 */
-#define BBC_QUIESCE_S13	0x02	/* Quiesce Safari segment for cpu 1 and 3 */
-#define BBC_QUIESCE_B02	0x04	/* Quiesce BBC segment for cpu 0 and 2    */
-#define BBC_QUIESCE_B13	0x08	/* Quiesce BBC segment for cpu 1 and 3    */
-#define BBC_QUIESCE_FD0 0x10	/* Disable Fatal_Error[0] reporting	  */
-#define BBC_QUIESCE_FD1 0x20	/* Disable Fatal_Error[1] reporting	  */
-#define BBC_QUIESCE_FD2 0x40	/* Disable Fatal_Error[2] reporting	  */
-#define BBC_QUIESCE_FD3 0x80	/* Disable Fatal_Error[3] reporting	  */
-
-/* Watchdog Action register.  When the watchdog device timer expires
- * a line is enabled to the BBC.  The action BBC takes when this line
- * is asserted can be controlled by this regiser.
- */
-#define BBC_WDACTION_RST  0x01	/* When set, watchdog causes system reset.
-				 * When clear, BBC ignores watchdog signal.
-				 */
-#define BBC_WDACTION_RESV 0xfe	/* Reserved */
-
-/* Soft_POR_GEN register.  The POR (Power On Reset) signal may be asserted
- * for specific processors or all processors via this register.
- */
-#define BBC_SPG_CPU0	0x01 /* Assert POR for processor 0	*/
-#define BBC_SPG_CPU1	0x02 /* Assert POR for processor 1	*/
-#define BBC_SPG_CPU2	0x04 /* Assert POR for processor 2	*/
-#define BBC_SPG_CPU3	0x08 /* Assert POR for processor 3	*/
-#define BBC_SPG_CPUALL	0x10 /* Reset all processors and reset
-			      * the entire system.
-			      */
-#define BBC_SPG_RESV	0xe0 /* Reserved			*/
-
-/* Soft_XIR_GEN register.  The XIR (eXternally Initiated Reset) signal
- * may be asserted to specific processors via this register.
- */
-#define BBC_SXG_CPU0	0x01 /* Assert XIR for processor 0	*/
-#define BBC_SXG_CPU1	0x02 /* Assert XIR for processor 1	*/
-#define BBC_SXG_CPU2	0x04 /* Assert XIR for processor 2	*/
-#define BBC_SXG_CPU3	0x08 /* Assert XIR for processor 3	*/
-#define BBC_SXG_RESV	0xf0 /* Reserved			*/
-
-/* POR Source register.  One may identify the cause of the most recent
- * reset by reading this register.
- */
-#define BBC_PSRC_SPG0	0x0001 /* CPU 0 reset via BBC_SPG register	*/
-#define BBC_PSRC_SPG1	0x0002 /* CPU 1 reset via BBC_SPG register	*/
-#define BBC_PSRC_SPG2	0x0004 /* CPU 2 reset via BBC_SPG register	*/
-#define BBC_PSRC_SPG3	0x0008 /* CPU 3 reset via BBC_SPG register	*/
-#define BBC_PSRC_SPGSYS	0x0010 /* System reset via BBC_SPG register	*/
-#define BBC_PSRC_JTAG	0x0020 /* System reset via JTAG+		*/
-#define BBC_PSRC_BUTTON	0x0040 /* System reset via push-button dongle	*/
-#define BBC_PSRC_PWRUP	0x0080 /* System reset via power-up		*/
-#define BBC_PSRC_FE0	0x0100 /* CPU 0 reported Fatal_Error		*/
-#define BBC_PSRC_FE1	0x0200 /* CPU 1 reported Fatal_Error		*/
-#define BBC_PSRC_FE2	0x0400 /* CPU 2 reported Fatal_Error		*/
-#define BBC_PSRC_FE3	0x0800 /* CPU 3 reported Fatal_Error		*/
-#define BBC_PSRC_FE4	0x1000 /* Schizo reported Fatal_Error		*/
-#define BBC_PSRC_FE5	0x2000 /* Safari device 5 reported Fatal_Error	*/
-#define BBC_PSRC_FE6	0x4000 /* CPMS reported Fatal_Error		*/
-#define BBC_PSRC_SYNTH	0x8000 /* System reset when on-board clock synthesizers
-				* were updated.
-				*/
-#define BBC_PSRC_WDT   0x10000 /* System reset via Super I/O watchdog	*/
-#define BBC_PSRC_RSC   0x20000 /* System reset via RSC remote monitoring
-				* device
-				*/
-
-/* XIR Source register.  The source of an XIR event sent to a processor may
- * be determined via this register.
- */
-#define BBC_XSRC_SXG0	0x01	/* CPU 0 received XIR via Soft_XIR_GEN reg */
-#define BBC_XSRC_SXG1	0x02	/* CPU 1 received XIR via Soft_XIR_GEN reg */
-#define BBC_XSRC_SXG2	0x04	/* CPU 2 received XIR via Soft_XIR_GEN reg */
-#define BBC_XSRC_SXG3	0x08	/* CPU 3 received XIR via Soft_XIR_GEN reg */
-#define BBC_XSRC_JTAG	0x10	/* All CPUs received XIR via JTAG+         */
-#define BBC_XSRC_W_OR_B	0x20	/* All CPUs received XIR either because:
-				 * a) Super I/O watchdog fired, or
-				 * b) XIR push button was activated
-				 */
-#define BBC_XSRC_RESV	0xc0	/* Reserved				   */
-
-/* Clock Synthesizers Control register.  This register provides the big-bang
- * programming interface to the two clock synthesizers of the machine.
- */
-#define BBC_CSC_SLOAD	0x01	/* Directly connected to S_LOAD pins	*/
-#define BBC_CSC_SDATA	0x02	/* Directly connected to S_DATA pins	*/
-#define BBC_CSC_SCLOCK	0x04	/* Directly connected to S_CLOCK pins	*/
-#define BBC_CSC_RESV	0x78	/* Reserved				*/
-#define BBC_CSC_RST	0x80	/* Generate system reset when S_LOAD==1	*/
-
-/* Energy Star Control register.  This register is used to generate the
- * clock frequency change trigger to the main system devices (Schizo and
- * the processors).  The transition occurs when bits in this register
- * go from 0 to 1, only one bit must be set at once else no action
- * occurs.  Basically the sequence of events is:
- * a) Choose new frequency: full, 1/2 or 1/32
- * b) Program this desired frequency into the cpus and Schizo.
- * c) Set the same value in this register.
- * d) 16 system clocks later, clear this register.
- */
-#define BBC_ES_CTRL_1_1		0x01	/* Full frequency	*/
-#define BBC_ES_CTRL_1_2		0x02	/* 1/2 frequency	*/
-#define BBC_ES_CTRL_1_32	0x20	/* 1/32 frequency	*/
-#define BBC_ES_RESV		0xdc	/* Reserved		*/
-
-/* Energy Star Assert Change Time register.  This determines the number
- * of BBC clock cycles (which is half the system frequency) between
- * the detection of FREEZE_ACK being asserted and the assertion of
- * the CLK_CHANGE_L[2:0] signals.
- */
-#define BBC_ES_ACT_VAL	0xff
-
-/* Energy Star Assert Bypass Time register.  This determines the number
- * of BBC clock cycles (which is half the system frequency) between
- * the assertion of the CLK_CHANGE_L[2:0] signals and the assertion of
- * the ESTAR_PLL_BYPASS signal.
- */
-#define BBC_ES_ABT_VAL	0xffff
-
-/* Energy Star PLL Settle Time register.  This determines the number of
- * BBC clock cycles (which is half the system frequency) between the
- * de-assertion of CLK_CHANGE_L[2:0] and the de-assertion of the FREEZE_L
- * signal.
- */
-#define BBC_ES_PST_VAL	0xffffffff
-
-/* Energy Star Frequency Switch Latency register.  This is the number of
- * BBC clocks between the de-assertion of CLK_CHANGE_L[2:0] and the first
- * edge of the Safari clock at the new frequency.
- */
-#define BBC_ES_FSL_VAL	0xffffffff
-
-/* Keyboard Beep control register.  This is a simple enabler for the audio
- * beep sound.
- */
-#define BBC_KBD_BEEP_ENABLE	0x01 /* Enable beep	*/
-#define BBC_KBD_BEEP_RESV	0xfe /* Reserved	*/
-
-/* Keyboard Beep Counter register.  There is a free-running counter inside
- * the BBC which runs at half the system clock.  The bit set in this register
- * determines when the audio sound is generated.  So for example if bit
- * 10 is set, the audio beep will oscillate at 1/(2**12).  The keyboard beep
- * generator automatically selects a different bit to use if the system clock
- * is changed via Energy Star.
- */
-#define BBC_KBD_BCNT_BITS	0x0007fc00
-#define BBC_KBC_BCNT_RESV	0xfff803ff
-
-#endif /* _SPARC64_BBC_H */
-
+#include <asm-sparc/bbc.h>
diff --git a/include/asm-sparc64/bitops.h b/include/asm-sparc64/bitops.h
index bb87b808022..204404355bd 100644
--- a/include/asm-sparc64/bitops.h
+++ b/include/asm-sparc64/bitops.h
@@ -1,107 +1 @@
-/*
- * bitops.h: Bit string operations on the V9.
- *
- * Copyright 1996, 1997 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC64_BITOPS_H
-#define _SPARC64_BITOPS_H
-
-#ifndef _LINUX_BITOPS_H
-#error only <linux/bitops.h> can be included directly
-#endif
-
-#include <linux/compiler.h>
-#include <asm/byteorder.h>
-
-extern int test_and_set_bit(unsigned long nr, volatile unsigned long *addr);
-extern int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr);
-extern int test_and_change_bit(unsigned long nr, volatile unsigned long *addr);
-extern void set_bit(unsigned long nr, volatile unsigned long *addr);
-extern void clear_bit(unsigned long nr, volatile unsigned long *addr);
-extern void change_bit(unsigned long nr, volatile unsigned long *addr);
-
-#include <asm-generic/bitops/non-atomic.h>
-
-#ifdef CONFIG_SMP
-#define smp_mb__before_clear_bit()	membar_storeload_loadload()
-#define smp_mb__after_clear_bit()	membar_storeload_storestore()
-#else
-#define smp_mb__before_clear_bit()	barrier()
-#define smp_mb__after_clear_bit()	barrier()
-#endif
-
-#include <asm-generic/bitops/ffz.h>
-#include <asm-generic/bitops/__ffs.h>
-#include <asm-generic/bitops/fls.h>
-#include <asm-generic/bitops/__fls.h>
-#include <asm-generic/bitops/fls64.h>
-
-#ifdef __KERNEL__
-
-#include <asm-generic/bitops/sched.h>
-#include <asm-generic/bitops/ffs.h>
-
-/*
- * hweightN: returns the hamming weight (i.e. the number
- * of bits set) of a N-bit word
- */
-
-#ifdef ULTRA_HAS_POPULATION_COUNT
-
-static inline unsigned int hweight64(unsigned long w)
-{
-	unsigned int res;
-
-	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w));
-	return res;
-}
-
-static inline unsigned int hweight32(unsigned int w)
-{
-	unsigned int res;
-
-	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xffffffff));
-	return res;
-}
-
-static inline unsigned int hweight16(unsigned int w)
-{
-	unsigned int res;
-
-	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xffff));
-	return res;
-}
-
-static inline unsigned int hweight8(unsigned int w)
-{
-	unsigned int res;
-
-	__asm__ ("popc %1,%0" : "=r" (res) : "r" (w & 0xff));
-	return res;
-}
-
-#else
-
-#include <asm-generic/bitops/hweight.h>
-
-#endif
-#include <asm-generic/bitops/lock.h>
-#endif /* __KERNEL__ */
-
-#include <asm-generic/bitops/find.h>
-
-#ifdef __KERNEL__
-
-#include <asm-generic/bitops/ext2-non-atomic.h>
-
-#define ext2_set_bit_atomic(lock,nr,addr) \
-	test_and_set_bit((nr) ^ 0x38,(unsigned long *)(addr))
-#define ext2_clear_bit_atomic(lock,nr,addr) \
-	test_and_clear_bit((nr) ^ 0x38,(unsigned long *)(addr))
-
-#include <asm-generic/bitops/minix.h>
-
-#endif /* __KERNEL__ */
-
-#endif /* defined(_SPARC64_BITOPS_H) */
+#include <asm-sparc/bitops.h>
diff --git a/include/asm-sparc64/cacheflush.h b/include/asm-sparc64/cacheflush.h
index 122e4058dd9..cf5b6b3e8a5 100644
--- a/include/asm-sparc64/cacheflush.h
+++ b/include/asm-sparc64/cacheflush.h
@@ -1,76 +1 @@
-#ifndef _SPARC64_CACHEFLUSH_H
-#define _SPARC64_CACHEFLUSH_H
-
-#include <asm/page.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/mm.h>
-
-/* Cache flush operations. */
-
-/* These are the same regardless of whether this is an SMP kernel or not. */
-#define flush_cache_mm(__mm) \
-	do { if ((__mm) == current->mm) flushw_user(); } while(0)
-#define flush_cache_dup_mm(mm) flush_cache_mm(mm)
-#define flush_cache_range(vma, start, end) \
-	flush_cache_mm((vma)->vm_mm)
-#define flush_cache_page(vma, page, pfn) \
-	flush_cache_mm((vma)->vm_mm)
-
-/* 
- * On spitfire, the icache doesn't snoop local stores and we don't
- * use block commit stores (which invalidate icache lines) during
- * module load, so we need this.
- */
-extern void flush_icache_range(unsigned long start, unsigned long end);
-extern void __flush_icache_page(unsigned long);
-
-extern void __flush_dcache_page(void *addr, int flush_icache);
-extern void flush_dcache_page_impl(struct page *page);
-#ifdef CONFIG_SMP
-extern void smp_flush_dcache_page_impl(struct page *page, int cpu);
-extern void flush_dcache_page_all(struct mm_struct *mm, struct page *page);
-#else
-#define smp_flush_dcache_page_impl(page,cpu) flush_dcache_page_impl(page)
-#define flush_dcache_page_all(mm,page) flush_dcache_page_impl(page)
-#endif
-
-extern void __flush_dcache_range(unsigned long start, unsigned long end);
-extern void flush_dcache_page(struct page *page);
-
-#define flush_icache_page(vma, pg)	do { } while(0)
-#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
-
-extern void flush_ptrace_access(struct vm_area_struct *, struct page *,
-				unsigned long uaddr, void *kaddr,
-				unsigned long len, int write);
-
-#define copy_to_user_page(vma, page, vaddr, dst, src, len)		\
-	do {								\
-		flush_cache_page(vma, vaddr, page_to_pfn(page));	\
-		memcpy(dst, src, len);					\
-		flush_ptrace_access(vma, page, vaddr, src, len, 0);	\
-	} while (0)
-
-#define copy_from_user_page(vma, page, vaddr, dst, src, len) 		\
-	do {								\
-		flush_cache_page(vma, vaddr, page_to_pfn(page));	\
-		memcpy(dst, src, len);					\
-		flush_ptrace_access(vma, page, vaddr, dst, len, 1);	\
-	} while (0)
-
-#define flush_dcache_mmap_lock(mapping)		do { } while (0)
-#define flush_dcache_mmap_unlock(mapping)	do { } while (0)
-
-#define flush_cache_vmap(start, end)		do { } while (0)
-#define flush_cache_vunmap(start, end)		do { } while (0)
-
-#ifdef CONFIG_DEBUG_PAGEALLOC
-/* internal debugging function */
-void kernel_map_pages(struct page *page, int numpages, int enable);
-#endif
-
-#endif /* !__ASSEMBLY__ */
-
-#endif /* _SPARC64_CACHEFLUSH_H */
+#include <asm-sparc/cacheflush.h>
diff --git a/include/asm-sparc64/chafsr.h b/include/asm-sparc64/chafsr.h
index 85c69b38220..aaab97562a3 100644
--- a/include/asm-sparc64/chafsr.h
+++ b/include/asm-sparc64/chafsr.h
@@ -1,241 +1 @@
-#ifndef _SPARC64_CHAFSR_H
-#define _SPARC64_CHAFSR_H
-
-/* Cheetah Asynchronous Fault Status register, ASI=0x4C VA<63:0>=0x0 */
-
-/* Comments indicate which processor variants on which the bit definition
- * is valid.  Codes are:
- * ch	-->	cheetah
- * ch+	-->	cheetah plus
- * jp	-->	jalapeno
- */
-
-/* All bits of this register except M_SYNDROME and E_SYNDROME are
- * read, write 1 to clear.  M_SYNDROME and E_SYNDROME are read-only.
- */
-
-/* Software bit set by linux trap handlers to indicate that the trap was
- * signalled at %tl >= 1.
- */
-#define CHAFSR_TL1		(1UL << 63UL) /* n/a */
-
-/* Unmapped error from system bus for prefetch queue or
- * store queue read operation
- */
-#define CHPAFSR_DTO		(1UL << 59UL) /* ch+ */
-
-/* Bus error from system bus for prefetch queue or store queue
- * read operation
- */
-#define CHPAFSR_DBERR		(1UL << 58UL) /* ch+ */
-
-/* Hardware corrected E-cache Tag ECC error */
-#define CHPAFSR_THCE		(1UL << 57UL) /* ch+ */
-/* System interface protocol error, hw timeout caused */
-#define JPAFSR_JETO		(1UL << 57UL) /* jp */
-
-/* SW handled correctable E-cache Tag ECC error */
-#define CHPAFSR_TSCE		(1UL << 56UL) /* ch+ */
-/* Parity error on system snoop results */
-#define JPAFSR_SCE		(1UL << 56UL) /* jp */
-
-/* Uncorrectable E-cache Tag ECC error */
-#define CHPAFSR_TUE		(1UL << 55UL) /* ch+ */
-/* System interface protocol error, illegal command detected */
-#define JPAFSR_JEIC		(1UL << 55UL) /* jp */
-
-/* Uncorrectable system bus data ECC error due to prefetch
- * or store fill request
- */
-#define CHPAFSR_DUE		(1UL << 54UL) /* ch+ */
-/* System interface protocol error, illegal ADTYPE detected */
-#define JPAFSR_JEIT		(1UL << 54UL) /* jp */
-
-/* Multiple errors of the same type have occurred.  This bit is set when
- * an uncorrectable error or a SW correctable error occurs and the status
- * bit to report that error is already set.  When multiple errors of
- * different types are indicated by setting multiple status bits.
- *
- * This bit is not set if multiple HW corrected errors with the same
- * status bit occur, only uncorrectable and SW correctable ones have
- * this behavior.
- *
- * This bit is not set when multiple ECC errors happen within a single
- * 64-byte system bus transaction.  Only the first ECC error in a 16-byte
- * subunit will be logged.  All errors in subsequent 16-byte subunits
- * from the same 64-byte transaction are ignored.
- */
-#define CHAFSR_ME		(1UL << 53UL) /* ch,ch+,jp */
-
-/* Privileged state error has occurred.  This is a capture of PSTATE.PRIV
- * at the time the error is detected.
- */
-#define CHAFSR_PRIV		(1UL << 52UL) /* ch,ch+,jp */
-
-/* The following bits 51 (CHAFSR_PERR) to 33 (CHAFSR_CE) are sticky error
- * bits and record the most recently detected errors.  Bits accumulate
- * errors that have been detected since the last write to clear the bit.
- */
-
-/* System interface protocol error.  The processor asserts its' ERROR
- * pin when this event occurs and it also logs a specific cause code
- * into a JTAG scannable flop.
- */
-#define CHAFSR_PERR		(1UL << 51UL) /* ch,ch+,jp */
-
-/* Internal processor error.  The processor asserts its' ERROR
- * pin when this event occurs and it also logs a specific cause code
- * into a JTAG scannable flop.
- */
-#define CHAFSR_IERR		(1UL << 50UL) /* ch,ch+,jp */
-
-/* System request parity error on incoming address */
-#define CHAFSR_ISAP		(1UL << 49UL) /* ch,ch+,jp */
-
-/* HW Corrected system bus MTAG ECC error */
-#define CHAFSR_EMC		(1UL << 48UL) /* ch,ch+ */
-/* Parity error on L2 cache tag SRAM */
-#define JPAFSR_ETP		(1UL << 48UL) /* jp */
-
-/* Uncorrectable system bus MTAG ECC error */
-#define CHAFSR_EMU		(1UL << 47UL) /* ch,ch+ */
-/* Out of range memory error has occurred */
-#define JPAFSR_OM		(1UL << 47UL) /* jp */
-
-/* HW Corrected system bus data ECC error for read of interrupt vector */
-#define CHAFSR_IVC		(1UL << 46UL) /* ch,ch+ */
-/* Error due to unsupported store */
-#define JPAFSR_UMS		(1UL << 46UL) /* jp */
-
-/* Uncorrectable system bus data ECC error for read of interrupt vector */
-#define CHAFSR_IVU		(1UL << 45UL) /* ch,ch+,jp */
-
-/* Unmapped error from system bus */
-#define CHAFSR_TO		(1UL << 44UL) /* ch,ch+,jp */
-
-/* Bus error response from system bus */
-#define CHAFSR_BERR		(1UL << 43UL) /* ch,ch+,jp */
-
-/* SW Correctable E-cache ECC error for instruction fetch or data access
- * other than block load.
- */
-#define CHAFSR_UCC		(1UL << 42UL) /* ch,ch+,jp */
-
-/* Uncorrectable E-cache ECC error for instruction fetch or data access
- * other than block load.
- */
-#define CHAFSR_UCU		(1UL << 41UL) /* ch,ch+,jp */
-
-/* Copyout HW Corrected ECC error */
-#define CHAFSR_CPC		(1UL << 40UL) /* ch,ch+,jp */
-
-/* Copyout Uncorrectable ECC error */
-#define CHAFSR_CPU		(1UL << 39UL) /* ch,ch+,jp */
-
-/* HW Corrected ECC error from E-cache for writeback */
-#define CHAFSR_WDC		(1UL << 38UL) /* ch,ch+,jp */
-
-/* Uncorrectable ECC error from E-cache for writeback */
-#define CHAFSR_WDU		(1UL << 37UL) /* ch,ch+,jp */
-
-/* HW Corrected ECC error from E-cache for store merge or block load */
-#define CHAFSR_EDC		(1UL << 36UL) /* ch,ch+,jp */
-
-/* Uncorrectable ECC error from E-cache for store merge or block load */
-#define CHAFSR_EDU		(1UL << 35UL) /* ch,ch+,jp */
-
-/* Uncorrectable system bus data ECC error for read of memory or I/O */
-#define CHAFSR_UE		(1UL << 34UL) /* ch,ch+,jp */
-
-/* HW Corrected system bus data ECC error for read of memory or I/O */
-#define CHAFSR_CE		(1UL << 33UL) /* ch,ch+,jp */
-
-/* Uncorrectable ECC error from remote cache/memory */
-#define JPAFSR_RUE		(1UL << 32UL) /* jp */
-
-/* Correctable ECC error from remote cache/memory */
-#define JPAFSR_RCE		(1UL << 31UL) /* jp */
-
-/* JBUS parity error on returned read data */
-#define JPAFSR_BP		(1UL << 30UL) /* jp */
-
-/* JBUS parity error on data for writeback or block store */
-#define JPAFSR_WBP		(1UL << 29UL) /* jp */
-
-/* Foreign read to DRAM incurring correctable ECC error */
-#define JPAFSR_FRC		(1UL << 28UL) /* jp */
-
-/* Foreign read to DRAM incurring uncorrectable ECC error */
-#define JPAFSR_FRU		(1UL << 27UL) /* jp */
-
-#define CHAFSR_ERRORS		(CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP | CHAFSR_EMC | \
-				 CHAFSR_EMU | CHAFSR_IVC | CHAFSR_IVU | CHAFSR_TO | \
-				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | CHAFSR_CPC | \
-				 CHAFSR_CPU | CHAFSR_WDC | CHAFSR_WDU | CHAFSR_EDC | \
-				 CHAFSR_EDU | CHAFSR_UE | CHAFSR_CE)
-#define CHPAFSR_ERRORS		(CHPAFSR_DTO | CHPAFSR_DBERR | CHPAFSR_THCE | \
-				 CHPAFSR_TSCE | CHPAFSR_TUE | CHPAFSR_DUE | \
-				 CHAFSR_PERR | CHAFSR_IERR | CHAFSR_ISAP | CHAFSR_EMC | \
-				 CHAFSR_EMU | CHAFSR_IVC | CHAFSR_IVU | CHAFSR_TO | \
-				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | CHAFSR_CPC | \
-				 CHAFSR_CPU | CHAFSR_WDC | CHAFSR_WDU | CHAFSR_EDC | \
-				 CHAFSR_EDU | CHAFSR_UE | CHAFSR_CE)
-#define JPAFSR_ERRORS		(JPAFSR_JETO | JPAFSR_SCE | JPAFSR_JEIC | \
-				 JPAFSR_JEIT | CHAFSR_PERR | CHAFSR_IERR | \
-				 CHAFSR_ISAP | JPAFSR_ETP | JPAFSR_OM | \
-				 JPAFSR_UMS | CHAFSR_IVU | CHAFSR_TO | \
-				 CHAFSR_BERR | CHAFSR_UCC | CHAFSR_UCU | \
-				 CHAFSR_CPC | CHAFSR_CPU | CHAFSR_WDC | \
-				 CHAFSR_WDU | CHAFSR_EDC | CHAFSR_EDU | \
-				 CHAFSR_UE | CHAFSR_CE | JPAFSR_RUE | \
-				 JPAFSR_RCE | JPAFSR_BP | JPAFSR_WBP | \
-				 JPAFSR_FRC | JPAFSR_FRU)
-
-/* Active JBUS request signal when error occurred */
-#define JPAFSR_JBREQ		(0x7UL << 24UL) /* jp */
-#define JPAFSR_JBREQ_SHIFT	24UL
-
-/* L2 cache way information */
-#define JPAFSR_ETW		(0x3UL << 22UL) /* jp */
-#define JPAFSR_ETW_SHIFT	22UL
-
-/* System bus MTAG ECC syndrome.  This field captures the status of the
- * first occurrence of the highest-priority error according to the M_SYND
- * overwrite policy.  After the AFSR sticky bit, corresponding to the error
- * for which the M_SYND is reported, is cleared, the contents of the M_SYND
- * field will be unchanged by will be unfrozen for further error capture.
- */
-#define CHAFSR_M_SYNDROME	(0xfUL << 16UL) /* ch,ch+,jp */
-#define CHAFSR_M_SYNDROME_SHIFT	16UL
-
-/* Agenid Id of the foreign device causing the UE/CE errors */
-#define JPAFSR_AID		(0x1fUL << 9UL) /* jp */
-#define JPAFSR_AID_SHIFT	9UL
-
-/* System bus or E-cache data ECC syndrome.  This field captures the status
- * of the first occurrence of the highest-priority error according to the
- * E_SYND overwrite policy.  After the AFSR sticky bit, corresponding to the
- * error for which the E_SYND is reported, is cleare, the contents of the E_SYND
- * field will be unchanged but will be unfrozen for further error capture.
- */
-#define CHAFSR_E_SYNDROME	(0x1ffUL << 0UL) /* ch,ch+,jp */
-#define CHAFSR_E_SYNDROME_SHIFT	0UL
-
-/* The AFSR must be explicitly cleared by software, it is not cleared automatically
- * by a read.  Writes to bits <51:33> with bits set will clear the corresponding
- * bits in the AFSR.  Bits associated with disrupting traps must be cleared before
- * interrupts are re-enabled to prevent multiple traps for the same error.  I.e.
- * PSTATE.IE and AFSR bits control delivery of disrupting traps.
- *
- * Since there is only one AFAR, when multiple events have been logged by the
- * bits in the AFSR, at most one of these events will have its status captured
- * in the AFAR.  The highest priority of those event bits will get AFAR logging.
- * The AFAR will be unlocked and available to capture the address of another event
- * as soon as the one bit in AFSR that corresponds to the event logged in AFAR is
- * cleared.  For example, if AFSR.CE is detected, then AFSR.UE (which overwrites
- * the AFAR), and AFSR.UE is cleared by not AFSR.CE, then the AFAR will be unlocked
- * and ready for another event, even though AFSR.CE is still set.  The same rules
- * also apply to the M_SYNDROME and E_SYNDROME fields of the AFSR.
- */
-
-#endif /* _SPARC64_CHAFSR_H */
+#include <asm-sparc/chafsr.h>
diff --git a/include/asm-sparc64/checksum.h b/include/asm-sparc64/checksum.h
index b290564c8ce..c3966c5e29d 100644
--- a/include/asm-sparc64/checksum.h
+++ b/include/asm-sparc64/checksum.h
@@ -1,167 +1 @@
-#ifndef __SPARC64_CHECKSUM_H
-#define __SPARC64_CHECKSUM_H
-
-/*  checksum.h:  IP/UDP/TCP checksum routines on the V9.
- *
- *  Copyright(C) 1995 Linus Torvalds
- *  Copyright(C) 1995 Miguel de Icaza
- *  Copyright(C) 1996 David S. Miller
- *  Copyright(C) 1996 Eddie C. Dost
- *  Copyright(C) 1997 Jakub Jelinek
- *
- * derived from:
- *	Alpha checksum c-code
- *      ix86 inline assembly
- *      RFC1071 Computing the Internet Checksum
- */
-
-#include <linux/in6.h>
-#include <asm/uaccess.h>
-
-/* computes the checksum of a memory block at buff, length len,
- * and adds in "sum" (32-bit)
- *
- * returns a 32-bit number suitable for feeding into itself
- * or csum_tcpudp_magic
- *
- * this function must be called with even lengths, except
- * for the last fragment, which may be odd
- *
- * it's best to have buff aligned on a 32-bit boundary
- */
-extern __wsum csum_partial(const void * buff, int len, __wsum sum);
-
-/* the same as csum_partial, but copies from user space while it
- * checksums
- *
- * here even more important to align src and dst on a 32-bit (or even
- * better 64-bit) boundary
- */
-extern __wsum csum_partial_copy_nocheck(const void *src, void *dst,
-					      int len, __wsum sum);
-
-extern long __csum_partial_copy_from_user(const void __user *src,
-					  void *dst, int len,
-					  __wsum sum);
-
-static inline __wsum
-csum_partial_copy_from_user(const void __user *src,
-			    void *dst, int len,
-			    __wsum sum, int *err)
-{
-	long ret = __csum_partial_copy_from_user(src, dst, len, sum);
-	if (ret < 0)
-		*err = -EFAULT;
-	return (__force __wsum) ret;
-}
-
-/* 
- *	Copy and checksum to user
- */
-#define HAVE_CSUM_COPY_USER
-extern long __csum_partial_copy_to_user(const void *src,
-					void __user *dst, int len,
-					  __wsum sum);
-
-static inline __wsum
-csum_and_copy_to_user(const void *src,
-		      void __user *dst, int len,
-		      __wsum sum, int *err)
-{
-	long ret = __csum_partial_copy_to_user(src, dst, len, sum);
-	if (ret < 0)
-		*err = -EFAULT;
-	return (__force __wsum) ret;
-}
-
-/* ihl is always 5 or greater, almost always is 5, and iph is word aligned
- * the majority of the time.
- */
-extern __sum16 ip_fast_csum(const void *iph, unsigned int ihl);
-
-/* Fold a partial checksum without adding pseudo headers. */
-static inline __sum16 csum_fold(__wsum sum)
-{
-	unsigned int tmp;
-
-	__asm__ __volatile__(
-"	addcc		%0, %1, %1\n"
-"	srl		%1, 16, %1\n"
-"	addc		%1, %%g0, %1\n"
-"	xnor		%%g0, %1, %0\n"
-	: "=&r" (sum), "=r" (tmp)
-	: "0" (sum), "1" ((__force u32)sum<<16)
-	: "cc");
-	return (__force __sum16)sum;
-}
-
-static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
-					       unsigned int len,
-					       unsigned short proto,
-					       __wsum sum)
-{
-	__asm__ __volatile__(
-"	addcc		%1, %0, %0\n"
-"	addccc		%2, %0, %0\n"
-"	addccc		%3, %0, %0\n"
-"	addc		%0, %%g0, %0\n"
-	: "=r" (sum), "=r" (saddr)
-	: "r" (daddr), "r" (proto + len), "0" (sum), "1" (saddr)
-	: "cc");
-	return sum;
-}
-
-/*
- * computes the checksum of the TCP/UDP pseudo-header
- * returns a 16-bit checksum, already complemented
- */
-static inline __sum16 csum_tcpudp_magic(__be32 saddr, __be32 daddr,
-						   unsigned short len,
-						   unsigned short proto,
-						   __wsum sum)
-{
-	return csum_fold(csum_tcpudp_nofold(saddr,daddr,len,proto,sum));
-}
-
-#define _HAVE_ARCH_IPV6_CSUM
-
-static inline __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
-				      const struct in6_addr *daddr,
-				      __u32 len, unsigned short proto,
-				      __wsum sum)
-{
-	__asm__ __volatile__ (
-"	addcc		%3, %4, %%g7\n"
-"	addccc		%5, %%g7, %%g7\n"
-"	lduw		[%2 + 0x0c], %%g2\n"
-"	lduw		[%2 + 0x08], %%g3\n"
-"	addccc		%%g2, %%g7, %%g7\n"
-"	lduw		[%2 + 0x04], %%g2\n"
-"	addccc		%%g3, %%g7, %%g7\n"
-"	lduw		[%2 + 0x00], %%g3\n"
-"	addccc		%%g2, %%g7, %%g7\n"
-"	lduw		[%1 + 0x0c], %%g2\n"
-"	addccc		%%g3, %%g7, %%g7\n"
-"	lduw		[%1 + 0x08], %%g3\n"
-"	addccc		%%g2, %%g7, %%g7\n"
-"	lduw		[%1 + 0x04], %%g2\n"
-"	addccc		%%g3, %%g7, %%g7\n"
-"	lduw		[%1 + 0x00], %%g3\n"
-"	addccc		%%g2, %%g7, %%g7\n"
-"	addccc		%%g3, %%g7, %0\n"
-"	addc		0, %0, %0\n"
-	: "=&r" (sum)
-	: "r" (saddr), "r" (daddr), "r"(htonl(len)),
-	  "r"(htonl(proto)), "r"(sum)
-	: "g2", "g3", "g7", "cc");
-
-	return csum_fold(sum);
-}
-
-/* this routine is used for miscellaneous IP-like checksums, mainly in icmp.c */
-static inline __sum16 ip_compute_csum(const void *buff, int len)
-{
-	return csum_fold(csum_partial(buff, len, 0));
-}
-
-#endif /* !(__SPARC64_CHECKSUM_H) */
+#include <asm-sparc/checksum.h>
diff --git a/include/asm-sparc64/chmctrl.h b/include/asm-sparc64/chmctrl.h
index 859b4a4b0d3..eb757b483b3 100644
--- a/include/asm-sparc64/chmctrl.h
+++ b/include/asm-sparc64/chmctrl.h
@@ -1,183 +1 @@
-#ifndef _SPARC64_CHMCTRL_H
-#define _SPARC64_CHMCTRL_H
-
-/* Cheetah memory controller programmable registers. */
-#define CHMCTRL_TCTRL1		0x00 /* Memory Timing Control I		*/
-#define CHMCTRL_TCTRL2		0x08 /* Memory Timing Control II	*/
-#define CHMCTRL_TCTRL3		0x38 /* Memory Timing Control III	*/
-#define CHMCTRL_TCTRL4		0x40 /* Memory Timing Control IV	*/
-#define CHMCTRL_DECODE1		0x10 /* Memory Address Decode I		*/
-#define CHMCTRL_DECODE2		0x18 /* Memory Address Decode II	*/
-#define CHMCTRL_DECODE3		0x20 /* Memory Address Decode III	*/
-#define CHMCTRL_DECODE4		0x28 /* Memory Address Decode IV	*/
-#define CHMCTRL_MACTRL		0x30 /* Memory Address Control		*/
-
-/* Memory Timing Control I */
-#define TCTRL1_SDRAMCTL_DLY	0xf000000000000000UL
-#define TCTRL1_SDRAMCTL_DLY_SHIFT     60
-#define TCTRL1_SDRAMCLK_DLY	0x0e00000000000000UL
-#define TCTRL1_SDRAMCLK_DLY_SHIFT     57
-#define TCTRL1_R		0x0100000000000000UL
-#define TCTRL1_R_SHIFT 		      56
-#define TCTRL1_AUTORFR_CYCLE	0x00fe000000000000UL
-#define TCTRL1_AUTORFR_CYCLE_SHIFT    49
-#define TCTRL1_RD_WAIT		0x0001f00000000000UL
-#define TCTRL1_RD_WAIT_SHIFT	      44
-#define TCTRL1_PC_CYCLE		0x00000fc000000000UL
-#define TCTRL1_PC_CYCLE_SHIFT	      38
-#define TCTRL1_WR_MORE_RAS_PW	0x0000003f00000000UL
-#define TCTRL1_WR_MORE_RAS_PW_SHIFT   32
-#define TCTRL1_RD_MORE_RAW_PW	0x00000000fc000000UL
-#define TCTRL1_RD_MORE_RAS_PW_SHIFT   26
-#define TCTRL1_ACT_WR_DLY	0x0000000003f00000UL
-#define TCTRL1_ACT_WR_DLY_SHIFT	      20
-#define TCTRL1_ACT_RD_DLY	0x00000000000fc000UL
-#define TCTRL1_ACT_RD_DLY_SHIFT	      14
-#define TCTRL1_BANK_PRESENT	0x0000000000003000UL
-#define TCTRL1_BANK_PRESENT_SHIFT     12
-#define TCTRL1_RFR_INT		0x0000000000000ff8UL
-#define TCTRL1_RFR_INT_SHIFT	      3
-#define TCTRL1_SET_MODE_REG	0x0000000000000004UL
-#define TCTRL1_SET_MODE_REG_SHIFT     2
-#define TCTRL1_RFR_ENABLE	0x0000000000000002UL
-#define TCTRL1_RFR_ENABLE_SHIFT	      1
-#define TCTRL1_PRECHG_ALL	0x0000000000000001UL
-#define TCTRL1_PRECHG_ALL_SHIFT	      0
-
-/* Memory Timing Control II */
-#define TCTRL2_WR_MSEL_DLY	0xfc00000000000000UL
-#define TCTRL2_WR_MSEL_DLY_SHIFT      58
-#define TCTRL2_RD_MSEL_DLY	0x03f0000000000000UL
-#define TCTRL2_RD_MSEL_DLY_SHIFT      52
-#define TCTRL2_WRDATA_THLD	0x000c000000000000UL
-#define TCTRL2_WRDATA_THLD_SHIFT      50
-#define TCTRL2_RDWR_RD_TI_DLY	0x0003f00000000000UL
-#define TCTRL2_RDWR_RD_TI_DLY_SHIFT   44
-#define TCTRL2_AUTOPRECHG_ENBL	0x0000080000000000UL
-#define TCTRL2_AUTOPRECHG_ENBL_SHIFT  43
-#define TCTRL2_RDWR_PI_MORE_DLY	0x000007c000000000UL
-#define TCTRL2_RDWR_PI_MORE_DLY_SHIFT 38
-#define TCTRL2_RDWR_1_DLY	0x0000003f00000000UL
-#define TCTRL2_RDWR_1_DLY_SHIFT       32
-#define TCTRL2_WRWR_PI_MORE_DLY	0x00000000f8000000UL
-#define TCTRL2_WRWR_PI_MORE_DLY_SHIFT 27
-#define TCTRL2_WRWR_1_DLY	0x0000000007e00000UL
-#define TCTRL2_WRWR_1_DLY_SHIFT       21
-#define TCTRL2_RDWR_RD_PI_MORE_DLY 0x00000000001f0000UL
-#define TCTRL2_RDWR_RD_PI_MORE_DLY_SHIFT 16
-#define TCTRL2_R		0x0000000000008000UL
-#define TCTRL2_R_SHIFT		      15
-#define TCTRL2_SDRAM_MODE_REG_DATA 0x0000000000007fffUL
-#define TCTRL2_SDRAM_MODE_REG_DATA_SHIFT 0
-
-/* Memory Timing Control III */
-#define TCTRL3_SDRAM_CTL_DLY	0xf000000000000000UL
-#define TCTRL3_SDRAM_CTL_DLY_SHIFT    60
-#define TCTRL3_SDRAM_CLK_DLY	0x0e00000000000000UL
-#define TCTRL3_SDRAM_CLK_DLY_SHIFT    57
-#define TCTRL3_R		0x0100000000000000UL
-#define TCTRL3_R_SHIFT		      56
-#define TCTRL3_AUTO_RFR_CYCLE	0x00fe000000000000UL
-#define TCTRL3_AUTO_RFR_CYCLE_SHIFT   49
-#define TCTRL3_RD_WAIT		0x0001f00000000000UL
-#define TCTRL3_RD_WAIT_SHIFT	      44
-#define TCTRL3_PC_CYCLE		0x00000fc000000000UL
-#define TCTRL3_PC_CYCLE_SHIFT	      38
-#define TCTRL3_WR_MORE_RAW_PW	0x0000003f00000000UL
-#define TCTRL3_WR_MORE_RAW_PW_SHIFT   32
-#define TCTRL3_RD_MORE_RAW_PW	0x00000000fc000000UL
-#define TCTRL3_RD_MORE_RAW_PW_SHIFT   26
-#define TCTRL3_ACT_WR_DLY	0x0000000003f00000UL
-#define TCTRL3_ACT_WR_DLY_SHIFT       20
-#define TCTRL3_ACT_RD_DLY	0x00000000000fc000UL
-#define TCTRL3_ACT_RD_DLY_SHIFT       14
-#define TCTRL3_BANK_PRESENT	0x0000000000003000UL
-#define TCTRL3_BANK_PRESENT_SHIFT     12
-#define TCTRL3_RFR_INT		0x0000000000000ff8UL
-#define TCTRL3_RFR_INT_SHIFT	      3
-#define TCTRL3_SET_MODE_REG	0x0000000000000004UL
-#define TCTRL3_SET_MODE_REG_SHIFT     2
-#define TCTRL3_RFR_ENABLE	0x0000000000000002UL
-#define TCTRL3_RFR_ENABLE_SHIFT       1
-#define TCTRL3_PRECHG_ALL	0x0000000000000001UL
-#define TCTRL3_PRECHG_ALL_SHIFT	      0
-
-/* Memory Timing Control IV */
-#define TCTRL4_WR_MSEL_DLY	0xfc00000000000000UL
-#define TCTRL4_WR_MSEL_DLY_SHIFT      58
-#define TCTRL4_RD_MSEL_DLY	0x03f0000000000000UL
-#define TCTRL4_RD_MSEL_DLY_SHIFT      52
-#define TCTRL4_WRDATA_THLD	0x000c000000000000UL
-#define TCTRL4_WRDATA_THLD_SHIFT      50
-#define TCTRL4_RDWR_RD_RI_DLY	0x0003f00000000000UL
-#define TCTRL4_RDWR_RD_RI_DLY_SHIFT   44
-#define TCTRL4_AUTO_PRECHG_ENBL	0x0000080000000000UL
-#define TCTRL4_AUTO_PRECHG_ENBL_SHIFT 43
-#define TCTRL4_RD_WR_PI_MORE_DLY 0x000007c000000000UL
-#define TCTRL4_RD_WR_PI_MORE_DLY_SHIFT 38
-#define TCTRL4_RD_WR_TI_DLY	0x0000003f00000000UL
-#define TCTRL4_RD_WR_TI_DLY_SHIFT     32
-#define TCTRL4_WR_WR_PI_MORE_DLY 0x00000000f8000000UL
-#define TCTRL4_WR_WR_PI_MORE_DLY_SHIFT 27
-#define TCTRL4_WR_WR_TI_DLY	0x0000000007e00000UL
-#define TCTRL4_WR_WR_TI_DLY_SHIFT     21
-#define TCTRL4_RDWR_RD_PI_MORE_DLY 0x00000000001f000UL0
-#define TCTRL4_RDWR_RD_PI_MORE_DLY_SHIFT 16
-#define TCTRL4_R		0x0000000000008000UL
-#define TCTRL4_R_SHIFT		      15
-#define TCTRL4_SDRAM_MODE_REG_DATA 0x0000000000007fffUL
-#define TCTRL4_SDRAM_MODE_REG_DATA_SHIFT 0
-
-/* All 4 memory address decoding registers have the
- * same layout.
- */
-#define MEM_DECODE_VALID	0x8000000000000000UL /* Valid */
-#define MEM_DECODE_VALID_SHIFT	      63
-#define MEM_DECODE_UK		0x001ffe0000000000UL /* Upper mask */
-#define MEM_DECODE_UK_SHIFT	      41
-#define MEM_DECODE_UM		0x0000001ffff00000UL /* Upper match */
-#define MEM_DECODE_UM_SHIFT	      20
-#define MEM_DECODE_LK		0x000000000003c000UL /* Lower mask */
-#define MEM_DECODE_LK_SHIFT	      14
-#define MEM_DECODE_LM		0x0000000000000f00UL /* Lower match */
-#define MEM_DECODE_LM_SHIFT           8
-
-#define PA_UPPER_BITS		0x000007fffc000000UL
-#define PA_UPPER_BITS_SHIFT	26
-#define PA_LOWER_BITS		0x00000000000003c0UL
-#define PA_LOWER_BITS_SHIFT	6
-
-#define MACTRL_R0		         0x8000000000000000UL
-#define MACTRL_R0_SHIFT		         63
-#define MACTRL_ADDR_LE_PW                0x7000000000000000UL
-#define MACTRL_ADDR_LE_PW_SHIFT		 60
-#define MACTRL_CMD_PW                    0x0f00000000000000UL
-#define MACTRL_CMD_PW_SHIFT		 56
-#define MACTRL_HALF_MODE_WR_MSEL_DLY     0x00fc000000000000UL
-#define MACTRL_HALF_MODE_WR_MSEL_DLY_SHIFT 50
-#define MACTRL_HALF_MODE_RD_MSEL_DLY     0x0003f00000000000UL
-#define MACTRL_HALF_MODE_RD_MSEL_DLY_SHIFT 44
-#define MACTRL_HALF_MODE_SDRAM_CTL_DLY   0x00000f0000000000UL
-#define MACTRL_HALF_MODE_SDRAM_CTL_DLY_SHIFT 40
-#define MACTRL_HALF_MODE_SDRAM_CLK_DLY   0x000000e000000000UL
-#define MACTRL_HALF_MODE_SDRAM_CLK_DLY_SHIFT 37
-#define MACTRL_R1                        0x0000001000000000UL
-#define MACTRL_R1_SHIFT                      36
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B3 0x0000000f00000000UL
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B3_SHIFT 32
-#define MACTRL_ENC_INTLV_B3              0x00000000f8000000UL
-#define MACTRL_ENC_INTLV_B3_SHIFT              27
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B2 0x0000000007800000UL
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B2_SHIFT 23
-#define MACTRL_ENC_INTLV_B2              0x00000000007c0000UL
-#define MACTRL_ENC_INTLV_B2_SHIFT              18
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B1 0x000000000003c000UL
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B1_SHIFT 14
-#define MACTRL_ENC_INTLV_B1              0x0000000000003e00UL
-#define MACTRL_ENC_INTLV_B1_SHIFT               9
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B0 0x00000000000001e0UL
-#define MACTRL_BANKSEL_N_ROWADDR_SIZE_B0_SHIFT  5
-#define MACTRL_ENC_INTLV_B0              0x000000000000001fUL
-#define MACTRL_ENC_INTLV_B0_SHIFT               0
-
-#endif /* _SPARC64_CHMCTRL_H */
+#include <asm-sparc/chmctrl.h>
diff --git a/include/asm-sparc64/cmt.h b/include/asm-sparc64/cmt.h
index 870db592857..b19b445cb81 100644
--- a/include/asm-sparc64/cmt.h
+++ b/include/asm-sparc64/cmt.h
@@ -1,59 +1 @@
-#ifndef _SPARC64_CMT_H
-#define _SPARC64_CMT_H
-
-/* cmt.h: Chip Multi-Threading register definitions
- *
- * Copyright (C) 2004 David S. Miller (davem@redhat.com)
- */
-
-/* ASI_CORE_ID - private */
-#define LP_ID		0x0000000000000010UL
-#define  LP_ID_MAX	0x00000000003f0000UL
-#define  LP_ID_ID	0x000000000000003fUL
-
-/* ASI_INTR_ID - private */
-#define LP_INTR_ID	0x0000000000000000UL
-#define  LP_INTR_ID_ID	0x00000000000003ffUL
-
-/* ASI_CESR_ID - private */
-#define CESR_ID		0x0000000000000040UL
-#define  CESR_ID_ID	0x00000000000000ffUL
-
-/* ASI_CORE_AVAILABLE - shared */
-#define LP_AVAIL	0x0000000000000000UL
-#define  LP_AVAIL_1	0x0000000000000002UL
-#define  LP_AVAIL_0	0x0000000000000001UL
-
-/* ASI_CORE_ENABLE_STATUS - shared */
-#define LP_ENAB_STAT	0x0000000000000010UL
-#define  LP_ENAB_STAT_1	0x0000000000000002UL
-#define  LP_ENAB_STAT_0	0x0000000000000001UL
-
-/* ASI_CORE_ENABLE - shared */
-#define LP_ENAB		0x0000000000000020UL
-#define  LP_ENAB_1	0x0000000000000002UL
-#define  LP_ENAB_0	0x0000000000000001UL
-
-/* ASI_CORE_RUNNING - shared */
-#define LP_RUNNING_RW	0x0000000000000050UL
-#define LP_RUNNING_W1S	0x0000000000000060UL
-#define LP_RUNNING_W1C	0x0000000000000068UL
-#define  LP_RUNNING_1	0x0000000000000002UL
-#define  LP_RUNNING_0	0x0000000000000001UL
-
-/* ASI_CORE_RUNNING_STAT - shared */
-#define LP_RUN_STAT	0x0000000000000058UL
-#define  LP_RUN_STAT_1	0x0000000000000002UL
-#define  LP_RUN_STAT_0	0x0000000000000001UL
-
-/* ASI_XIR_STEERING - shared */
-#define LP_XIR_STEER	0x0000000000000030UL
-#define  LP_XIR_STEER_1	0x0000000000000002UL
-#define  LP_XIR_STEER_0	0x0000000000000001UL
-
-/* ASI_CMT_ERROR_STEERING - shared */
-#define CMT_ER_STEER	0x0000000000000040UL
-#define  CMT_ER_STEER_1	0x0000000000000002UL
-#define  CMT_ER_STEER_0	0x0000000000000001UL
-
-#endif /* _SPARC64_CMT_H */
+#include <asm-sparc/cmt.h>
diff --git a/include/asm-sparc64/compat.h b/include/asm-sparc64/compat.h
index f260b58f5ce..8c155d22195 100644
--- a/include/asm-sparc64/compat.h
+++ b/include/asm-sparc64/compat.h
@@ -1,243 +1 @@
-#ifndef _ASM_SPARC64_COMPAT_H
-#define _ASM_SPARC64_COMPAT_H
-/*
- * Architecture specific compatibility types
- */
-#include <linux/types.h>
-
-#define COMPAT_USER_HZ	100
-
-typedef u32		compat_size_t;
-typedef s32		compat_ssize_t;
-typedef s32		compat_time_t;
-typedef s32		compat_clock_t;
-typedef s32		compat_pid_t;
-typedef u16		__compat_uid_t;
-typedef u16		__compat_gid_t;
-typedef u32		__compat_uid32_t;
-typedef u32		__compat_gid32_t;
-typedef u16		compat_mode_t;
-typedef u32		compat_ino_t;
-typedef u16		compat_dev_t;
-typedef s32		compat_off_t;
-typedef s64		compat_loff_t;
-typedef s16		compat_nlink_t;
-typedef u16		compat_ipc_pid_t;
-typedef s32		compat_daddr_t;
-typedef u32		compat_caddr_t;
-typedef __kernel_fsid_t	compat_fsid_t;
-typedef s32		compat_key_t;
-typedef s32		compat_timer_t;
-
-typedef s32		compat_int_t;
-typedef s32		compat_long_t;
-typedef s64		compat_s64;
-typedef u32		compat_uint_t;
-typedef u32		compat_ulong_t;
-typedef u64		compat_u64;
-
-struct compat_timespec {
-	compat_time_t	tv_sec;
-	s32		tv_nsec;
-};
-
-struct compat_timeval {
-	compat_time_t	tv_sec;
-	s32		tv_usec;
-};
-
-struct compat_stat {
-	compat_dev_t	st_dev;
-	compat_ino_t	st_ino;
-	compat_mode_t	st_mode;
-	compat_nlink_t	st_nlink;
-	__compat_uid_t	st_uid;
-	__compat_gid_t	st_gid;
-	compat_dev_t	st_rdev;
-	compat_off_t	st_size;
-	compat_time_t	st_atime;
-	compat_ulong_t	st_atime_nsec;
-	compat_time_t	st_mtime;
-	compat_ulong_t	st_mtime_nsec;
-	compat_time_t	st_ctime;
-	compat_ulong_t	st_ctime_nsec;
-	compat_off_t	st_blksize;
-	compat_off_t	st_blocks;
-	u32		__unused4[2];
-};
-
-struct compat_stat64 {
-	unsigned long long	st_dev;
-
-	unsigned long long	st_ino;
-
-	unsigned int	st_mode;
-	unsigned int	st_nlink;
-
-	unsigned int	st_uid;
-	unsigned int	st_gid;
-
-	unsigned long long	st_rdev;
-
-	unsigned char	__pad3[8];
-
-	long long	st_size;
-	unsigned int	st_blksize;
-
-	unsigned char	__pad4[8];
-	unsigned int	st_blocks;
-
-	unsigned int	st_atime;
-	unsigned int	st_atime_nsec;
-
-	unsigned int	st_mtime;
-	unsigned int	st_mtime_nsec;
-
-	unsigned int	st_ctime;
-	unsigned int	st_ctime_nsec;
-
-	unsigned int	__unused4;
-	unsigned int	__unused5;
-};
-
-struct compat_flock {
-	short		l_type;
-	short		l_whence;
-	compat_off_t	l_start;
-	compat_off_t	l_len;
-	compat_pid_t	l_pid;
-	short		__unused;
-};
-
-#define F_GETLK64	12
-#define F_SETLK64	13
-#define F_SETLKW64	14
-
-struct compat_flock64 {
-	short		l_type;
-	short		l_whence;
-	compat_loff_t	l_start;
-	compat_loff_t	l_len;
-	compat_pid_t	l_pid;
-	short		__unused;
-};
-
-struct compat_statfs {
-	int		f_type;
-	int		f_bsize;
-	int		f_blocks;
-	int		f_bfree;
-	int		f_bavail;
-	int		f_files;
-	int		f_ffree;
-	compat_fsid_t	f_fsid;
-	int		f_namelen;	/* SunOS ignores this field. */
-	int		f_frsize;
-	int		f_spare[5];
-};
-
-#define COMPAT_RLIM_INFINITY 0x7fffffff
-
-typedef u32		compat_old_sigset_t;
-
-#define _COMPAT_NSIG		64
-#define _COMPAT_NSIG_BPW	32
-
-typedef u32		compat_sigset_word;
-
-#define COMPAT_OFF_T_MAX	0x7fffffff
-#define COMPAT_LOFF_T_MAX	0x7fffffffffffffffL
-
-/*
- * A pointer passed in from user mode. This should not
- * be used for syscall parameters, just declare them
- * as pointers because the syscall entry code will have
- * appropriately converted them already.
- */
-typedef	u32		compat_uptr_t;
-
-static inline void __user *compat_ptr(compat_uptr_t uptr)
-{
-	return (void __user *)(unsigned long)uptr;
-}
-
-static inline compat_uptr_t ptr_to_compat(void __user *uptr)
-{
-	return (u32)(unsigned long)uptr;
-}
-
-static inline void __user *compat_alloc_user_space(long len)
-{
-	struct pt_regs *regs = current_thread_info()->kregs;
-	unsigned long usp = regs->u_regs[UREG_I6];
-
-	if (!(test_thread_flag(TIF_32BIT)))
-		usp += STACK_BIAS;
-	else
-		usp &= 0xffffffffUL;
-
-	usp -= len;
-	usp &= ~0x7UL;
-
-	return (void __user *) usp;
-}
-
-struct compat_ipc64_perm {
-	compat_key_t key;
-	__compat_uid32_t uid;
-	__compat_gid32_t gid;
-	__compat_uid32_t cuid;
-	__compat_gid32_t cgid;
-	unsigned short __pad1;
-	compat_mode_t mode;
-	unsigned short __pad2;
-	unsigned short seq;
-	unsigned long __unused1;	/* yes they really are 64bit pads */
-	unsigned long __unused2;
-};
-
-struct compat_semid64_ds {
-	struct compat_ipc64_perm sem_perm;
-	unsigned int	__pad1;
-	compat_time_t	sem_otime;
-	unsigned int	__pad2;
-	compat_time_t	sem_ctime;
-	u32		sem_nsems;
-	u32		__unused1;
-	u32		__unused2;
-};
-
-struct compat_msqid64_ds {
-	struct compat_ipc64_perm msg_perm;
-	unsigned int	__pad1;
-	compat_time_t	msg_stime;
-	unsigned int	__pad2;
-	compat_time_t	msg_rtime;
-	unsigned int	__pad3;
-	compat_time_t	msg_ctime;
-	unsigned int	msg_cbytes;
-	unsigned int	msg_qnum;
-	unsigned int	msg_qbytes;
-	compat_pid_t	msg_lspid;
-	compat_pid_t	msg_lrpid;
-	unsigned int	__unused1;
-	unsigned int	__unused2;
-};
-
-struct compat_shmid64_ds {
-	struct compat_ipc64_perm shm_perm;
-	unsigned int	__pad1;
-	compat_time_t	shm_atime;
-	unsigned int	__pad2;
-	compat_time_t	shm_dtime;
-	unsigned int	__pad3;
-	compat_time_t	shm_ctime;
-	compat_size_t	shm_segsz;
-	compat_pid_t	shm_cpid;
-	compat_pid_t	shm_lpid;
-	unsigned int	shm_nattch;
-	unsigned int	__unused1;
-	unsigned int	__unused2;
-};
-
-#endif /* _ASM_SPARC64_COMPAT_H */
+#include <asm-sparc/compat.h>
diff --git a/include/asm-sparc64/compat_signal.h b/include/asm-sparc64/compat_signal.h
index b759eab9b51..7187dcc8cac 100644
--- a/include/asm-sparc64/compat_signal.h
+++ b/include/asm-sparc64/compat_signal.h
@@ -1,29 +1 @@
-#ifndef _COMPAT_SIGNAL_H
-#define _COMPAT_SIGNAL_H
-
-#include <linux/compat.h>
-#include <asm/signal.h>
-
-#ifdef CONFIG_COMPAT
-struct __new_sigaction32 {
-	unsigned		sa_handler;
-	unsigned int    	sa_flags;
-	unsigned		sa_restorer;     /* not used by Linux/SPARC yet */
-	compat_sigset_t 	sa_mask;
-};
-
-struct __old_sigaction32 {
-	unsigned		sa_handler;
-	compat_old_sigset_t  	sa_mask;
-	unsigned int    	sa_flags;
-	unsigned		sa_restorer;     /* not used by Linux/SPARC yet */
-};
-
-typedef struct sigaltstack32 {
-	u32			ss_sp;
-	int			ss_flags;
-	compat_size_t		ss_size;
-} stack_t32;
-#endif
-
-#endif /* !(_COMPAT_SIGNAL_H) */
+#include <asm-sparc/compat_signal.h>
diff --git a/include/asm-sparc64/cpudata.h b/include/asm-sparc64/cpudata.h
index 532975ecfe1..3220e134a57 100644
--- a/include/asm-sparc64/cpudata.h
+++ b/include/asm-sparc64/cpudata.h
@@ -1,240 +1 @@
-/* cpudata.h: Per-cpu parameters.
- *
- * Copyright (C) 2003, 2005, 2006 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC64_CPUDATA_H
-#define _SPARC64_CPUDATA_H
-
-#include <asm/hypervisor.h>
-#include <asm/asi.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/percpu.h>
-#include <linux/threads.h>
-
-typedef struct {
-	/* Dcache line 1 */
-	unsigned int	__softirq_pending; /* must be 1st, see rtrap.S */
-	unsigned int	__pad0;
-	unsigned long	clock_tick;	/* %tick's per second */
-	unsigned long	__pad;
-	unsigned int	__pad1;
-	unsigned int	__pad2;
-
-	/* Dcache line 2, rarely used */
-	unsigned int	dcache_size;
-	unsigned int	dcache_line_size;
-	unsigned int	icache_size;
-	unsigned int	icache_line_size;
-	unsigned int	ecache_size;
-	unsigned int	ecache_line_size;
-	int		core_id;
-	int		proc_id;
-} cpuinfo_sparc;
-
-DECLARE_PER_CPU(cpuinfo_sparc, __cpu_data);
-#define cpu_data(__cpu)		per_cpu(__cpu_data, (__cpu))
-#define local_cpu_data()	__get_cpu_var(__cpu_data)
-
-/* Trap handling code needs to get at a few critical values upon
- * trap entry and to process TSB misses.  These cannot be in the
- * per_cpu() area as we really need to lock them into the TLB and
- * thus make them part of the main kernel image.  As a result we
- * try to make this as small as possible.
- *
- * This is padded out and aligned to 64-bytes to avoid false sharing
- * on SMP.
- */
-
-/* If you modify the size of this structure, please update
- * TRAP_BLOCK_SZ_SHIFT below.
- */
-struct thread_info;
-struct trap_per_cpu {
-/* D-cache line 1: Basic thread information, cpu and device mondo queues */
-	struct thread_info	*thread;
-	unsigned long		pgd_paddr;
-	unsigned long		cpu_mondo_pa;
-	unsigned long		dev_mondo_pa;
-
-/* D-cache line 2: Error Mondo Queue and kernel buffer pointers */
-	unsigned long		resum_mondo_pa;
-	unsigned long		resum_kernel_buf_pa;
-	unsigned long		nonresum_mondo_pa;
-	unsigned long		nonresum_kernel_buf_pa;
-
-/* Dcache lines 3, 4, 5, and 6: Hypervisor Fault Status */
-	struct hv_fault_status	fault_info;
-
-/* Dcache line 7: Physical addresses of CPU send mondo block and CPU list.  */
-	unsigned long		cpu_mondo_block_pa;
-	unsigned long		cpu_list_pa;
-	unsigned long		tsb_huge;
-	unsigned long		tsb_huge_temp;
-
-/* Dcache line 8: IRQ work list, and keep trap_block a power-of-2 in size.  */
-	unsigned long		irq_worklist_pa;
-	unsigned int		cpu_mondo_qmask;
-	unsigned int		dev_mondo_qmask;
-	unsigned int		resum_qmask;
-	unsigned int		nonresum_qmask;
-	void			*hdesc;
-} __attribute__((aligned(64)));
-extern struct trap_per_cpu trap_block[NR_CPUS];
-extern void init_cur_cpu_trap(struct thread_info *);
-extern void setup_tba(void);
-extern int ncpus_probed;
-extern void __init cpu_probe(void);
-extern const struct seq_operations cpuinfo_op;
-
-extern unsigned long real_hard_smp_processor_id(void);
-
-struct cpuid_patch_entry {
-	unsigned int	addr;
-	unsigned int	cheetah_safari[4];
-	unsigned int	cheetah_jbus[4];
-	unsigned int	starfire[4];
-	unsigned int	sun4v[4];
-};
-extern struct cpuid_patch_entry __cpuid_patch, __cpuid_patch_end;
-
-struct sun4v_1insn_patch_entry {
-	unsigned int	addr;
-	unsigned int	insn;
-};
-extern struct sun4v_1insn_patch_entry __sun4v_1insn_patch,
-	__sun4v_1insn_patch_end;
-
-struct sun4v_2insn_patch_entry {
-	unsigned int	addr;
-	unsigned int	insns[2];
-};
-extern struct sun4v_2insn_patch_entry __sun4v_2insn_patch,
-	__sun4v_2insn_patch_end;
-
-#endif /* !(__ASSEMBLY__) */
-
-#define TRAP_PER_CPU_THREAD		0x00
-#define TRAP_PER_CPU_PGD_PADDR		0x08
-#define TRAP_PER_CPU_CPU_MONDO_PA	0x10
-#define TRAP_PER_CPU_DEV_MONDO_PA	0x18
-#define TRAP_PER_CPU_RESUM_MONDO_PA	0x20
-#define TRAP_PER_CPU_RESUM_KBUF_PA	0x28
-#define TRAP_PER_CPU_NONRESUM_MONDO_PA	0x30
-#define TRAP_PER_CPU_NONRESUM_KBUF_PA	0x38
-#define TRAP_PER_CPU_FAULT_INFO		0x40
-#define TRAP_PER_CPU_CPU_MONDO_BLOCK_PA	0xc0
-#define TRAP_PER_CPU_CPU_LIST_PA	0xc8
-#define TRAP_PER_CPU_TSB_HUGE		0xd0
-#define TRAP_PER_CPU_TSB_HUGE_TEMP	0xd8
-#define TRAP_PER_CPU_IRQ_WORKLIST_PA	0xe0
-#define TRAP_PER_CPU_CPU_MONDO_QMASK	0xe8
-#define TRAP_PER_CPU_DEV_MONDO_QMASK	0xec
-#define TRAP_PER_CPU_RESUM_QMASK	0xf0
-#define TRAP_PER_CPU_NONRESUM_QMASK	0xf4
-
-#define TRAP_BLOCK_SZ_SHIFT		8
-
-#include <asm/scratchpad.h>
-
-#define __GET_CPUID(REG)				\
-	/* Spitfire implementation (default). */	\
-661:	ldxa		[%g0] ASI_UPA_CONFIG, REG;	\
-	srlx		REG, 17, REG;			\
-	 and		REG, 0x1f, REG;			\
-	nop;						\
-	.section	.cpuid_patch, "ax";		\
-	/* Instruction location. */			\
-	.word		661b;				\
-	/* Cheetah Safari implementation. */		\
-	ldxa		[%g0] ASI_SAFARI_CONFIG, REG;	\
-	srlx		REG, 17, REG;			\
-	and		REG, 0x3ff, REG;		\
-	nop;						\
-	/* Cheetah JBUS implementation. */		\
-	ldxa		[%g0] ASI_JBUS_CONFIG, REG;	\
-	srlx		REG, 17, REG;			\
-	and		REG, 0x1f, REG;			\
-	nop;						\
-	/* Starfire implementation. */			\
-	sethi		%hi(0x1fff40000d0 >> 9), REG;	\
-	sllx		REG, 9, REG;			\
-	or		REG, 0xd0, REG;			\
-	lduwa		[REG] ASI_PHYS_BYPASS_EC_E, REG;\
-	/* sun4v implementation. */			\
-	mov		SCRATCHPAD_CPUID, REG;		\
-	ldxa		[REG] ASI_SCRATCHPAD, REG;	\
-	nop;						\
-	nop;						\
-	.previous;
-
-#ifdef CONFIG_SMP
-
-#define TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	__GET_CPUID(TMP)			\
-	sethi	%hi(trap_block), DEST;		\
-	sllx	TMP, TRAP_BLOCK_SZ_SHIFT, TMP;	\
-	or	DEST, %lo(trap_block), DEST;	\
-	add	DEST, TMP, DEST;		\
-
-/* Clobbers TMP, current address space PGD phys address into DEST.  */
-#define TRAP_LOAD_PGD_PHYS(DEST, TMP)		\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	ldx	[DEST + TRAP_PER_CPU_PGD_PADDR], DEST;
-
-/* Clobbers TMP, loads local processor's IRQ work area into DEST.  */
-#define TRAP_LOAD_IRQ_WORK_PA(DEST, TMP)	\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	add	DEST, TRAP_PER_CPU_IRQ_WORKLIST_PA, DEST;
-
-/* Clobbers TMP, loads DEST with current thread info pointer.  */
-#define TRAP_LOAD_THREAD_REG(DEST, TMP)		\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	ldx	[DEST + TRAP_PER_CPU_THREAD], DEST;
-
-/* Given the current thread info pointer in THR, load the per-cpu
- * area base of the current processor into DEST.  REG1, REG2, and REG3 are
- * clobbered.
- *
- * You absolutely cannot use DEST as a temporary in this code.  The
- * reason is that traps can happen during execution, and return from
- * trap will load the fully resolved DEST per-cpu base.  This can corrupt
- * the calculations done by the macro mid-stream.
- */
-#define LOAD_PER_CPU_BASE(DEST, THR, REG1, REG2, REG3)	\
-	lduh	[THR + TI_CPU], REG1;			\
-	sethi	%hi(__per_cpu_shift), REG3;		\
-	sethi	%hi(__per_cpu_base), REG2;		\
-	ldx	[REG3 + %lo(__per_cpu_shift)], REG3;	\
-	ldx	[REG2 + %lo(__per_cpu_base)], REG2;	\
-	sllx	REG1, REG3, REG3;			\
-	add	REG3, REG2, DEST;
-
-#else
-
-#define TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	sethi	%hi(trap_block), DEST;		\
-	or	DEST, %lo(trap_block), DEST;	\
-
-/* Uniprocessor versions, we know the cpuid is zero.  */
-#define TRAP_LOAD_PGD_PHYS(DEST, TMP)		\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	ldx	[DEST + TRAP_PER_CPU_PGD_PADDR], DEST;
-
-/* Clobbers TMP, loads local processor's IRQ work area into DEST.  */
-#define TRAP_LOAD_IRQ_WORK_PA(DEST, TMP)	\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	add	DEST, TRAP_PER_CPU_IRQ_WORKLIST_PA, DEST;
-
-#define TRAP_LOAD_THREAD_REG(DEST, TMP)		\
-	TRAP_LOAD_TRAP_BLOCK(DEST, TMP)		\
-	ldx	[DEST + TRAP_PER_CPU_THREAD], DEST;
-
-/* No per-cpu areas on uniprocessor, so no need to load DEST.  */
-#define LOAD_PER_CPU_BASE(DEST, THR, REG1, REG2, REG3)
-
-#endif /* !(CONFIG_SMP) */
-
-#endif /* _SPARC64_CPUDATA_H */
+#include <asm-sparc/cpudata.h>
diff --git a/include/asm-sparc64/dcr.h b/include/asm-sparc64/dcr.h
index 620c9ba642e..d67613b1f5f 100644
--- a/include/asm-sparc64/dcr.h
+++ b/include/asm-sparc64/dcr.h
@@ -1,14 +1 @@
-#ifndef _SPARC64_DCR_H
-#define _SPARC64_DCR_H
-
-/* UltraSparc-III/III+ Dispatch Control Register, ASR 0x12 */
-#define DCR_DPE		0x0000000000001000 /* III+: D$ Parity Error Enable	*/
-#define DCR_OBS		0x0000000000000fc0 /* Observability Bus Controls	*/
-#define DCR_BPE		0x0000000000000020 /* Branch Predict Enable		*/
-#define DCR_RPE		0x0000000000000010 /* Return Address Prediction Enable	*/
-#define DCR_SI		0x0000000000000008 /* Single Instruction Disable	*/
-#define DCR_IPE		0x0000000000000004 /* III+: I$ Parity Error Enable	*/
-#define DCR_IFPOE	0x0000000000000002 /* IRQ FP Operation Enable		*/
-#define DCR_MS		0x0000000000000001 /* Multi-Scalar dispatch		*/
-
-#endif /* _SPARC64_DCR_H */
+#include <asm-sparc/dcr.h>
diff --git a/include/asm-sparc64/dcu.h b/include/asm-sparc64/dcu.h
index 0f704e106a1..28853f4968d 100644
--- a/include/asm-sparc64/dcu.h
+++ b/include/asm-sparc64/dcu.h
@@ -1,27 +1 @@
-#ifndef _SPARC64_DCU_H
-#define _SPARC64_DCU_H
-
-#include <linux/const.h>
-
-/* UltraSparc-III Data Cache Unit Control Register */
-#define DCU_CP	_AC(0x0002000000000000,UL) /* Phys Cache Enable w/o mmu	*/
-#define DCU_CV	_AC(0x0001000000000000,UL) /* Virt Cache Enable w/o mmu	*/
-#define DCU_ME	_AC(0x0000800000000000,UL) /* NC-store Merging Enable	*/
-#define DCU_RE	_AC(0x0000400000000000,UL) /* RAW bypass Enable		*/
-#define DCU_PE	_AC(0x0000200000000000,UL) /* PCache Enable		*/
-#define DCU_HPE	_AC(0x0000100000000000,UL) /* HW prefetch Enable	*/
-#define DCU_SPE	_AC(0x0000080000000000,UL) /* SW prefetch Enable	*/
-#define DCU_SL	_AC(0x0000040000000000,UL) /* Secondary ld-steering Enab*/
-#define DCU_WE	_AC(0x0000020000000000,UL) /* WCache enable		*/
-#define DCU_PM	_AC(0x000001fe00000000,UL) /* PA Watchpoint Byte Mask	*/
-#define DCU_VM	_AC(0x00000001fe000000,UL) /* VA Watchpoint Byte Mask	*/
-#define DCU_PR	_AC(0x0000000001000000,UL) /* PA Watchpoint Read Enable	*/
-#define DCU_PW	_AC(0x0000000000800000,UL) /* PA Watchpoint Write Enable*/
-#define DCU_VR	_AC(0x0000000000400000,UL) /* VA Watchpoint Read Enable	*/
-#define DCU_VW	_AC(0x0000000000200000,UL) /* VA Watchpoint Write Enable*/
-#define DCU_DM	_AC(0x0000000000000008,UL) /* DMMU Enable		*/
-#define DCU_IM	_AC(0x0000000000000004,UL) /* IMMU Enable		*/
-#define DCU_DC	_AC(0x0000000000000002,UL) /* Data Cache Enable		*/
-#define DCU_IC	_AC(0x0000000000000001,UL) /* Instruction Cache Enable	*/
-
-#endif /* _SPARC64_DCU_H */
+#include <asm-sparc/dcu.h>
diff --git a/include/asm-sparc64/delay.h b/include/asm-sparc64/delay.h
index a77aa622d76..33dc5589d84 100644
--- a/include/asm-sparc64/delay.h
+++ b/include/asm-sparc64/delay.h
@@ -1,17 +1 @@
-/* delay.h: Linux delay routines on sparc64.
- *
- * Copyright (C) 1996, 2004, 2007 David S. Miller (davem@davemloft.net).
- */
-
-#ifndef _SPARC64_DELAY_H
-#define _SPARC64_DELAY_H
-
-#ifndef __ASSEMBLY__
-
-extern void __delay(unsigned long loops);
-extern void udelay(unsigned long usecs);
-#define mdelay(n)	udelay((n) * 1000)
-
-#endif /* !__ASSEMBLY__ */
-
-#endif /* _SPARC64_DELAY_H */
+#include <asm-sparc/delay.h>
diff --git a/include/asm-sparc64/display7seg.h b/include/asm-sparc64/display7seg.h
index c066a8964ea..e74f046b41d 100644
--- a/include/asm-sparc64/display7seg.h
+++ b/include/asm-sparc64/display7seg.h
@@ -1,79 +1 @@
-/*
- *
- * display7seg - Driver interface for the 7-segment display
- * present on Sun Microsystems CP1400 and CP1500
- *
- * Copyright (c) 2000 Eric Brower <ebrower@usa.net>
- *
- */
-
-#ifndef __display7seg_h__
-#define __display7seg_h__
-
-#define D7S_IOC	'p'
-
-#define D7SIOCRD _IOR(D7S_IOC, 0x45, int)	/* Read device state	*/
-#define D7SIOCWR _IOW(D7S_IOC, 0x46, int)	/* Write device state	*/
-#define D7SIOCTM _IO (D7S_IOC, 0x47)		/* Translate mode (FLIP)*/
-
-/*
- * ioctl flag definitions
- *
- * POINT	- Toggle decimal point	(0=absent 1=present)
- * ALARM	- Toggle alarm LED 		(0=green  1=red)
- * FLIP		- Toggle inverted mode 	(0=normal 1=flipped) 
- * bits 0-4	- Character displayed	(see definitions below)
- *
- * Display segments are defined as follows, 
- * subject to D7S_FLIP register state:
- *
- *    a
- *   ---
- * f|   |b
- *   -g-
- * e|   |c
- *   ---
- *    d
- */
-
-#define D7S_POINT	(1 << 7)	/* Decimal point*/
-#define D7S_ALARM	(1 << 6)	/* Alarm LED 	*/
-#define D7S_FLIP	(1 << 5)	/* Flip display */
-
-#define D7S_0		0x00		/* Numerals 0-9 */
-#define D7S_1		0x01
-#define D7S_2		0x02
-#define D7S_3		0x03
-#define D7S_4		0x04
-#define D7S_5		0x05
-#define D7S_6		0x06
-#define D7S_7		0x07
-#define D7S_8		0x08
-#define D7S_9		0x09
-#define D7S_A		0x0A		/* Letters A-F, H, L, P */
-#define D7S_B		0x0B
-#define D7S_C		0x0C
-#define D7S_D		0x0D
-#define D7S_E		0x0E
-#define D7S_F		0x0F
-#define D7S_H		0x10
-#define D7S_E2		0x11
-#define D7S_L		0x12
-#define D7S_P		0x13
-#define D7S_SEGA	0x14		/* Individual segments */
-#define D7S_SEGB	0x15
-#define D7S_SEGC	0x16
-#define D7S_SEGD	0x17
-#define D7S_SEGE	0x18
-#define D7S_SEGF	0x19
-#define D7S_SEGG	0x1A
-#define D7S_SEGABFG 0x1B		/* Segment groupings */
-#define D7S_SEGCDEG	0x1C
-#define D7S_SEGBCEF 0x1D
-#define D7S_SEGADG	0x1E
-#define D7S_BLANK	0x1F		/* Clear all segments */
-
-#define D7S_MIN_VAL	0x0
-#define D7S_MAX_VAL	0x1F
-
-#endif /* ifndef __display7seg_h__ */
+#include <asm-sparc/display7seg.h>
diff --git a/include/asm-sparc64/dma-mapping.h b/include/asm-sparc64/dma-mapping.h
index 38cbec76a33..380b7b63147 100644
--- a/include/asm-sparc64/dma-mapping.h
+++ b/include/asm-sparc64/dma-mapping.h
@@ -1,154 +1 @@
-#ifndef _ASM_SPARC64_DMA_MAPPING_H
-#define _ASM_SPARC64_DMA_MAPPING_H
-
-#include <linux/scatterlist.h>
-#include <linux/mm.h>
-
-#define DMA_ERROR_CODE	(~(dma_addr_t)0x0)
-
-struct dma_ops {
-	void *(*alloc_coherent)(struct device *dev, size_t size,
-				dma_addr_t *dma_handle, gfp_t flag);
-	void (*free_coherent)(struct device *dev, size_t size,
-			      void *cpu_addr, dma_addr_t dma_handle);
-	dma_addr_t (*map_single)(struct device *dev, void *cpu_addr,
-				 size_t size,
-				 enum dma_data_direction direction);
-	void (*unmap_single)(struct device *dev, dma_addr_t dma_addr,
-			     size_t size,
-			     enum dma_data_direction direction);
-	int (*map_sg)(struct device *dev, struct scatterlist *sg, int nents,
-		      enum dma_data_direction direction);
-	void (*unmap_sg)(struct device *dev, struct scatterlist *sg,
-			 int nhwentries,
-			 enum dma_data_direction direction);
-	void (*sync_single_for_cpu)(struct device *dev,
-				    dma_addr_t dma_handle, size_t size,
-				    enum dma_data_direction direction);
-	void (*sync_sg_for_cpu)(struct device *dev, struct scatterlist *sg,
-				int nelems,
-				enum dma_data_direction direction);
-};
-extern const struct dma_ops *dma_ops;
-
-extern int dma_supported(struct device *dev, u64 mask);
-extern int dma_set_mask(struct device *dev, u64 dma_mask);
-
-static inline void *dma_alloc_coherent(struct device *dev, size_t size,
-				       dma_addr_t *dma_handle, gfp_t flag)
-{
-	return dma_ops->alloc_coherent(dev, size, dma_handle, flag);
-}
-
-static inline void dma_free_coherent(struct device *dev, size_t size,
-				     void *cpu_addr, dma_addr_t dma_handle)
-{
-	dma_ops->free_coherent(dev, size, cpu_addr, dma_handle);
-}
-
-static inline dma_addr_t dma_map_single(struct device *dev, void *cpu_addr,
-					size_t size,
-					enum dma_data_direction direction)
-{
-	return dma_ops->map_single(dev, cpu_addr, size, direction);
-}
-
-static inline void dma_unmap_single(struct device *dev, dma_addr_t dma_addr,
-				    size_t size,
-				    enum dma_data_direction direction)
-{
-	dma_ops->unmap_single(dev, dma_addr, size, direction);
-}
-
-static inline dma_addr_t dma_map_page(struct device *dev, struct page *page,
-				      unsigned long offset, size_t size,
-				      enum dma_data_direction direction)
-{
-	return dma_ops->map_single(dev, page_address(page) + offset,
-				   size, direction);
-}
-
-static inline void dma_unmap_page(struct device *dev, dma_addr_t dma_address,
-				  size_t size,
-				  enum dma_data_direction direction)
-{
-	dma_ops->unmap_single(dev, dma_address, size, direction);
-}
-
-static inline int dma_map_sg(struct device *dev, struct scatterlist *sg,
-			     int nents, enum dma_data_direction direction)
-{
-	return dma_ops->map_sg(dev, sg, nents, direction);
-}
-
-static inline void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
-				int nents, enum dma_data_direction direction)
-{
-	dma_ops->unmap_sg(dev, sg, nents, direction);
-}
-
-static inline void dma_sync_single_for_cpu(struct device *dev,
-					   dma_addr_t dma_handle, size_t size,
-					   enum dma_data_direction direction)
-{
-	dma_ops->sync_single_for_cpu(dev, dma_handle, size, direction);
-}
-
-static inline void dma_sync_single_for_device(struct device *dev,
-					      dma_addr_t dma_handle,
-					      size_t size,
-					      enum dma_data_direction direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-static inline void dma_sync_single_range_for_cpu(struct device *dev,
-						 dma_addr_t dma_handle,
-						 unsigned long offset,
-						 size_t size,
-						 enum dma_data_direction direction)
-{
-	dma_sync_single_for_cpu(dev, dma_handle+offset, size, direction);
-}
-
-static inline void dma_sync_single_range_for_device(struct device *dev,
-						    dma_addr_t dma_handle,
-						    unsigned long offset,
-						    size_t size,
-						    enum dma_data_direction direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-
-static inline void dma_sync_sg_for_cpu(struct device *dev,
-				       struct scatterlist *sg, int nelems,
-				       enum dma_data_direction direction)
-{
-	dma_ops->sync_sg_for_cpu(dev, sg, nelems, direction);
-}
-
-static inline void dma_sync_sg_for_device(struct device *dev,
-					  struct scatterlist *sg, int nelems,
-					  enum dma_data_direction direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-static inline int dma_mapping_error(dma_addr_t dma_addr)
-{
-	return (dma_addr == DMA_ERROR_CODE);
-}
-
-static inline int dma_get_cache_alignment(void)
-{
-	/* no easy way to get cache size on all processors, so return
-	 * the maximum possible, to be safe */
-	return (1 << INTERNODE_CACHE_SHIFT);
-}
-
-#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
-#define dma_free_noncoherent(d, s, v, h) dma_free_coherent(d, s, v, h)
-#define dma_is_consistent(d, h)	(1)
-
-#endif /* _ASM_SPARC64_DMA_MAPPING_H */
+#include <asm-sparc/dma-mapping.h>
diff --git a/include/asm-sparc64/dma.h b/include/asm-sparc64/dma.h
index 9d4c024bd3b..2e36248e6b5 100644
--- a/include/asm-sparc64/dma.h
+++ b/include/asm-sparc64/dma.h
@@ -1,205 +1 @@
-/*
- * include/asm-sparc64/dma.h
- *
- * Copyright 1996 (C) David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _ASM_SPARC64_DMA_H
-#define _ASM_SPARC64_DMA_H
-
-#include <linux/kernel.h>
-#include <linux/types.h>
-#include <linux/spinlock.h>
-
-#include <asm/sbus.h>
-#include <asm/delay.h>
-#include <asm/oplib.h>
-
-/* These are irrelevant for Sparc DMA, but we leave it in so that
- * things can compile.
- */
-#define MAX_DMA_CHANNELS 8
-#define DMA_MODE_READ    1
-#define DMA_MODE_WRITE   2
-#define MAX_DMA_ADDRESS  (~0UL)
-
-/* Useful constants */
-#define SIZE_16MB      (16*1024*1024)
-#define SIZE_64K       (64*1024)
-
-/* SBUS DMA controller reg offsets */
-#define DMA_CSR		0x00UL		/* rw  DMA control/status register    0x00   */
-#define DMA_ADDR	0x04UL		/* rw  DMA transfer address register  0x04   */
-#define DMA_COUNT	0x08UL		/* rw  DMA transfer count register    0x08   */
-#define DMA_TEST	0x0cUL		/* rw  DMA test/debug register        0x0c   */
-
-/* DVMA chip revisions */
-enum dvma_rev {
-	dvmarev0,
-	dvmaesc1,
-	dvmarev1,
-	dvmarev2,
-	dvmarev3,
-	dvmarevplus,
-	dvmahme
-};
-
-#define DMA_HASCOUNT(rev)  ((rev)==dvmaesc1)
-
-/* Linux DMA information structure, filled during probe. */
-struct sbus_dma {
-	struct sbus_dma *next;
-	struct sbus_dev *sdev;
-	void __iomem *regs;
-
-	/* Status, misc info */
-	int node;                /* Prom node for this DMA device */
-	int running;             /* Are we doing DMA now? */
-	int allocated;           /* Are we "owned" by anyone yet? */
-
-	/* Transfer information. */
-	u32 addr;                /* Start address of current transfer */
-	int nbytes;              /* Size of current transfer */
-	int realbytes;           /* For splitting up large transfers, etc. */
-
-	/* DMA revision */
-	enum dvma_rev revision;
-};
-
-extern struct sbus_dma *dma_chain;
-
-/* Broken hardware... */
-#define DMA_ISBROKEN(dma)    ((dma)->revision == dvmarev1)
-#define DMA_ISESC1(dma)      ((dma)->revision == dvmaesc1)
-
-/* Main routines in dma.c */
-extern void dvma_init(struct sbus_bus *);
-
-/* Fields in the cond_reg register */
-/* First, the version identification bits */
-#define DMA_DEVICE_ID    0xf0000000        /* Device identification bits */
-#define DMA_VERS0        0x00000000        /* Sunray DMA version */
-#define DMA_ESCV1        0x40000000        /* DMA ESC Version 1 */
-#define DMA_VERS1        0x80000000        /* DMA rev 1 */
-#define DMA_VERS2        0xa0000000        /* DMA rev 2 */
-#define DMA_VERHME       0xb0000000        /* DMA hme gate array */
-#define DMA_VERSPLUS     0x90000000        /* DMA rev 1 PLUS */
-
-#define DMA_HNDL_INTR    0x00000001        /* An IRQ needs to be handled */
-#define DMA_HNDL_ERROR   0x00000002        /* We need to take an error */
-#define DMA_FIFO_ISDRAIN 0x0000000c        /* The DMA FIFO is draining */
-#define DMA_INT_ENAB     0x00000010        /* Turn on interrupts */
-#define DMA_FIFO_INV     0x00000020        /* Invalidate the FIFO */
-#define DMA_ACC_SZ_ERR   0x00000040        /* The access size was bad */
-#define DMA_FIFO_STDRAIN 0x00000040        /* DMA_VERS1 Drain the FIFO */
-#define DMA_RST_SCSI     0x00000080        /* Reset the SCSI controller */
-#define DMA_RST_ENET     DMA_RST_SCSI      /* Reset the ENET controller */
-#define DMA_ST_WRITE     0x00000100        /* write from device to memory */
-#define DMA_ENABLE       0x00000200        /* Fire up DMA, handle requests */
-#define DMA_PEND_READ    0x00000400        /* DMA_VERS1/0/PLUS Pending Read */
-#define DMA_ESC_BURST    0x00000800        /* 1=16byte 0=32byte */
-#define DMA_READ_AHEAD   0x00001800        /* DMA read ahead partial longword */
-#define DMA_DSBL_RD_DRN  0x00001000        /* No EC drain on slave reads */
-#define DMA_BCNT_ENAB    0x00002000        /* If on, use the byte counter */
-#define DMA_TERM_CNTR    0x00004000        /* Terminal counter */
-#define DMA_SCSI_SBUS64  0x00008000        /* HME: Enable 64-bit SBUS mode. */
-#define DMA_CSR_DISAB    0x00010000        /* No FIFO drains during csr */
-#define DMA_SCSI_DISAB   0x00020000        /* No FIFO drains during reg */
-#define DMA_DSBL_WR_INV  0x00020000        /* No EC inval. on slave writes */
-#define DMA_ADD_ENABLE   0x00040000        /* Special ESC DVMA optimization */
-#define DMA_E_BURSTS	 0x000c0000	   /* ENET: SBUS r/w burst mask */
-#define DMA_E_BURST32	 0x00040000	   /* ENET: SBUS 32 byte r/w burst */
-#define DMA_E_BURST16	 0x00000000	   /* ENET: SBUS 16 byte r/w burst */
-#define DMA_BRST_SZ      0x000c0000        /* SCSI: SBUS r/w burst size */
-#define DMA_BRST64       0x000c0000        /* SCSI: 64byte bursts (HME on UltraSparc only) */
-#define DMA_BRST32       0x00040000        /* SCSI: 32byte bursts */
-#define DMA_BRST16       0x00000000        /* SCSI: 16byte bursts */
-#define DMA_BRST0        0x00080000        /* SCSI: no bursts (non-HME gate arrays) */
-#define DMA_ADDR_DISAB   0x00100000        /* No FIFO drains during addr */
-#define DMA_2CLKS        0x00200000        /* Each transfer = 2 clock ticks */
-#define DMA_3CLKS        0x00400000        /* Each transfer = 3 clock ticks */
-#define DMA_EN_ENETAUI   DMA_3CLKS         /* Put lance into AUI-cable mode */
-#define DMA_CNTR_DISAB   0x00800000        /* No IRQ when DMA_TERM_CNTR set */
-#define DMA_AUTO_NADDR   0x01000000        /* Use "auto nxt addr" feature */
-#define DMA_SCSI_ON      0x02000000        /* Enable SCSI dma */
-#define DMA_PARITY_OFF   0x02000000        /* HME: disable parity checking */
-#define DMA_LOADED_ADDR  0x04000000        /* Address has been loaded */
-#define DMA_LOADED_NADDR 0x08000000        /* Next address has been loaded */
-#define DMA_RESET_FAS366 0x08000000        /* HME: Assert RESET to FAS366 */
-
-/* Values describing the burst-size property from the PROM */
-#define DMA_BURST1       0x01
-#define DMA_BURST2       0x02
-#define DMA_BURST4       0x04
-#define DMA_BURST8       0x08
-#define DMA_BURST16      0x10
-#define DMA_BURST32      0x20
-#define DMA_BURST64      0x40
-#define DMA_BURSTBITS    0x7f
-
-/* Determine highest possible final transfer address given a base */
-#define DMA_MAXEND(addr) (0x01000000UL-(((unsigned long)(addr))&0x00ffffffUL))
-
-/* Yes, I hack a lot of elisp in my spare time... */
-#define DMA_ERROR_P(regs)  ((sbus_readl((regs) + DMA_CSR) & DMA_HNDL_ERROR))
-#define DMA_IRQ_P(regs)    ((sbus_readl((regs) + DMA_CSR)) & (DMA_HNDL_INTR | DMA_HNDL_ERROR))
-#define DMA_WRITE_P(regs)  ((sbus_readl((regs) + DMA_CSR) & DMA_ST_WRITE))
-#define DMA_OFF(__regs)		\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp &= ~DMA_ENABLE; \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-#define DMA_INTSOFF(__regs)	\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp &= ~DMA_INT_ENAB; \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-#define DMA_INTSON(__regs)	\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp |= DMA_INT_ENAB; \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-#define DMA_PUNTFIFO(__regs)	\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp |= DMA_FIFO_INV; \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-#define DMA_SETSTART(__regs, __addr)	\
-	sbus_writel((u32)(__addr), (__regs) + DMA_ADDR);
-#define DMA_BEGINDMA_W(__regs)	\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp |= (DMA_ST_WRITE|DMA_ENABLE|DMA_INT_ENAB); \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-#define DMA_BEGINDMA_R(__regs)	\
-do {	u32 tmp = sbus_readl((__regs) + DMA_CSR); \
-	tmp |= (DMA_ENABLE|DMA_INT_ENAB); \
-	tmp &= ~DMA_ST_WRITE; \
-	sbus_writel(tmp, (__regs) + DMA_CSR); \
-} while(0)
-
-/* For certain DMA chips, we need to disable ints upon irq entry
- * and turn them back on when we are done.  So in any ESP interrupt
- * handler you *must* call DMA_IRQ_ENTRY upon entry and DMA_IRQ_EXIT
- * when leaving the handler.  You have been warned...
- */
-#define DMA_IRQ_ENTRY(dma, dregs) do { \
-        if(DMA_ISBROKEN(dma)) DMA_INTSOFF(dregs); \
-   } while (0)
-
-#define DMA_IRQ_EXIT(dma, dregs) do { \
-	if(DMA_ISBROKEN(dma)) DMA_INTSON(dregs); \
-   } while(0)
-
-#define for_each_dvma(dma) \
-        for((dma) = dma_chain; (dma); (dma) = (dma)->next)
-
-/* From PCI */
-
-#ifdef CONFIG_PCI
-extern int isa_dma_bridge_buggy;
-#else
-#define isa_dma_bridge_buggy 	(0)
-#endif
-
-#endif /* !(_ASM_SPARC64_DMA_H) */
+#include <asm-sparc/dma.h>
diff --git a/include/asm-sparc64/ebus.h b/include/asm-sparc64/ebus.h
index fcc62b97ced..d7d476158bd 100644
--- a/include/asm-sparc64/ebus.h
+++ b/include/asm-sparc64/ebus.h
@@ -1,94 +1 @@
-/*
- * ebus.h: PCI to Ebus pseudo driver software state.
- *
- * Copyright (C) 1997 Eddie C. Dost (ecd@skynet.be)
- * Copyright (C) 1999 David S. Miller (davem@redhat.com)
- */
-
-#ifndef __SPARC64_EBUS_H
-#define __SPARC64_EBUS_H
-
-#include <asm/oplib.h>
-#include <asm/prom.h>
-#include <asm/of_device.h>
-
-struct linux_ebus_child {
-	struct linux_ebus_child		*next;
-	struct linux_ebus_device	*parent;
-	struct linux_ebus		*bus;
-	struct device_node		*prom_node;
-	struct resource			 resource[PROMREG_MAX];
-	int				 num_addrs;
-	unsigned int			 irqs[PROMINTR_MAX];
-	int				 num_irqs;
-};
-
-struct linux_ebus_device {
-	struct of_device		ofdev;
-	struct linux_ebus_device	*next;
-	struct linux_ebus_child		*children;
-	struct linux_ebus		*bus;
-	struct device_node		*prom_node;
-	struct resource			 resource[PROMREG_MAX];
-	int				 num_addrs;
-	unsigned int			 irqs[PROMINTR_MAX];
-	int				 num_irqs;
-};
-#define to_ebus_device(d) container_of(d, struct linux_ebus_device, ofdev.dev)
-
-struct linux_ebus {
-	struct of_device		ofdev;
-	struct linux_ebus		*next;
-	struct linux_ebus_device	*devices;
-	struct pci_dev			*self;
-	int				 index;
-	int				 is_rio;
-	struct device_node		*prom_node;
-};
-#define to_ebus(d) container_of(d, struct linux_ebus, ofdev.dev)
-
-struct ebus_dma_info {
-	spinlock_t	lock;
-	void __iomem	*regs;
-
-	unsigned int	flags;
-#define EBUS_DMA_FLAG_USE_EBDMA_HANDLER		0x00000001
-#define EBUS_DMA_FLAG_TCI_DISABLE		0x00000002
-
-	/* These are only valid is EBUS_DMA_FLAG_USE_EBDMA_HANDLER is
-	 * set.
-	 */
-	void (*callback)(struct ebus_dma_info *p, int event, void *cookie);
-	void *client_cookie;
-	unsigned int	irq;
-#define EBUS_DMA_EVENT_ERROR	1
-#define EBUS_DMA_EVENT_DMA	2
-#define EBUS_DMA_EVENT_DEVICE	4
-
-	unsigned char	name[64];
-};
-
-extern int ebus_dma_register(struct ebus_dma_info *p);
-extern int ebus_dma_irq_enable(struct ebus_dma_info *p, int on);
-extern void ebus_dma_unregister(struct ebus_dma_info *p);
-extern int ebus_dma_request(struct ebus_dma_info *p, dma_addr_t bus_addr,
-			    size_t len);
-extern void ebus_dma_prepare(struct ebus_dma_info *p, int write);
-extern unsigned int ebus_dma_residue(struct ebus_dma_info *p);
-extern unsigned int ebus_dma_addr(struct ebus_dma_info *p);
-extern void ebus_dma_enable(struct ebus_dma_info *p, int on);
-
-extern struct linux_ebus		*ebus_chain;
-
-extern void ebus_init(void);
-
-#define for_each_ebus(bus)						\
-        for((bus) = ebus_chain; (bus); (bus) = (bus)->next)
-
-#define for_each_ebusdev(dev, bus)					\
-        for((dev) = (bus)->devices; (dev); (dev) = (dev)->next)
-
-#define for_each_edevchild(dev, child)					\
-        for((child) = (dev)->children; (child); (child) = (child)->next)
-
-#endif /* !(__SPARC64_EBUS_H) */
+#include <asm-sparc/ebus.h>
diff --git a/include/asm-sparc64/elf.h b/include/asm-sparc64/elf.h
index 0818a1308f4..f256d9472c8 100644
--- a/include/asm-sparc64/elf.h
+++ b/include/asm-sparc64/elf.h
@@ -1,217 +1 @@
-#ifndef __ASM_SPARC64_ELF_H
-#define __ASM_SPARC64_ELF_H
-
-/*
- * ELF register definitions..
- */
-
-#include <asm/ptrace.h>
-#include <asm/processor.h>
-#include <asm/uaccess.h>
-#include <asm/spitfire.h>
-
-/*
- * Sparc section types
- */
-#define STT_REGISTER		13
-
-/*
- * Sparc ELF relocation types
- */
-#define	R_SPARC_NONE		0
-#define	R_SPARC_8		1
-#define	R_SPARC_16		2
-#define	R_SPARC_32		3
-#define	R_SPARC_DISP8		4
-#define	R_SPARC_DISP16		5
-#define	R_SPARC_DISP32		6
-#define	R_SPARC_WDISP30		7
-#define	R_SPARC_WDISP22		8
-#define	R_SPARC_HI22		9
-#define	R_SPARC_22		10
-#define	R_SPARC_13		11
-#define	R_SPARC_LO10		12
-#define	R_SPARC_GOT10		13
-#define	R_SPARC_GOT13		14
-#define	R_SPARC_GOT22		15
-#define	R_SPARC_PC10		16
-#define	R_SPARC_PC22		17
-#define	R_SPARC_WPLT30		18
-#define	R_SPARC_COPY		19
-#define	R_SPARC_GLOB_DAT	20
-#define	R_SPARC_JMP_SLOT	21
-#define	R_SPARC_RELATIVE	22
-#define	R_SPARC_UA32		23
-#define R_SPARC_PLT32		24
-#define R_SPARC_HIPLT22		25
-#define R_SPARC_LOPLT10		26
-#define R_SPARC_PCPLT32		27
-#define R_SPARC_PCPLT22		28
-#define R_SPARC_PCPLT10		29
-#define R_SPARC_10		30
-#define R_SPARC_11		31
-#define R_SPARC_64		32
-#define R_SPARC_OLO10		33
-#define R_SPARC_WDISP16		40
-#define R_SPARC_WDISP19		41
-#define R_SPARC_7		43
-#define R_SPARC_5		44
-#define R_SPARC_6		45
-
-/* Bits present in AT_HWCAP, primarily for Sparc32.  */
-
-#define HWCAP_SPARC_FLUSH       1    /* CPU supports flush instruction. */
-#define HWCAP_SPARC_STBAR       2
-#define HWCAP_SPARC_SWAP        4
-#define HWCAP_SPARC_MULDIV      8
-#define HWCAP_SPARC_V9		16
-#define HWCAP_SPARC_ULTRA3	32
-#define HWCAP_SPARC_BLKINIT	64
-#define HWCAP_SPARC_N2		128
-
-#define CORE_DUMP_USE_REGSET
-
-/*
- * These are used to set parameters in the core dumps.
- */
-#define ELF_ARCH		EM_SPARCV9
-#define ELF_CLASS		ELFCLASS64
-#define ELF_DATA		ELFDATA2MSB
-
-/* Format of 64-bit elf_gregset_t is:
- * 	G0 --> G7
- * 	O0 --> O7
- * 	L0 --> L7
- * 	I0 --> I7
- *	TSTATE
- *	TPC
- *	TNPC
- *	Y
- */
-typedef unsigned long elf_greg_t;
-#define ELF_NGREG 36
-typedef elf_greg_t elf_gregset_t[ELF_NGREG];
-
-typedef struct {
-	unsigned long	pr_regs[32];
-	unsigned long	pr_fsr;
-	unsigned long	pr_gsr;
-	unsigned long	pr_fprs;
-} elf_fpregset_t;
-
-/* Format of 32-bit elf_gregset_t is:
- * 	G0 --> G7
- *	O0 --> O7
- *	L0 --> L7
- *	I0 --> I7
- *	PSR, PC, nPC, Y, WIM, TBR
- */
-typedef unsigned int compat_elf_greg_t;
-#define COMPAT_ELF_NGREG 38
-typedef compat_elf_greg_t compat_elf_gregset_t[COMPAT_ELF_NGREG];
-
-typedef struct {
-	union {
-		unsigned int	pr_regs[32];
-		unsigned long	pr_dregs[16];
-	} pr_fr;
-	unsigned int __unused;
-	unsigned int	pr_fsr;
-	unsigned char	pr_qcnt;
-	unsigned char	pr_q_entrysize;
-	unsigned char	pr_en;
-	unsigned int	pr_q[64];
-} compat_elf_fpregset_t;
-
-/* UltraSparc extensions.  Still unused, but will be eventually.  */
-typedef struct {
-	unsigned int pr_type;
-	unsigned int pr_align;
-	union {
-		struct {
-			union {
-				unsigned int	pr_regs[32];
-				unsigned long	pr_dregs[16];
-				long double	pr_qregs[8];
-			} pr_xfr;
-		} pr_v8p;
-		unsigned int	pr_xfsr;
-		unsigned int	pr_fprs;
-		unsigned int	pr_xg[8];
-		unsigned int	pr_xo[8];
-		unsigned long	pr_tstate;
-		unsigned int	pr_filler[8];
-	} pr_un;
-} elf_xregset_t;
-
-/*
- * This is used to ensure we don't load something for the wrong architecture.
- */
-#define elf_check_arch(x)		((x)->e_machine == ELF_ARCH)
-#define compat_elf_check_arch(x)	((x)->e_machine == EM_SPARC || \
-					 (x)->e_machine == EM_SPARC32PLUS)
-#define compat_start_thread		start_thread32
-
-#define USE_ELF_CORE_DUMP
-#define ELF_EXEC_PAGESIZE	PAGE_SIZE
-
-/* This is the location that an ET_DYN program is loaded if exec'ed.  Typical
-   use of this is to invoke "./ld.so someprog" to test out a new version of
-   the loader.  We need to make sure that it is out of the way of the program
-   that it will "exec", and that there is sufficient room for the brk.  */
-
-#define ELF_ET_DYN_BASE		0x0000010000000000UL
-#define COMPAT_ELF_ET_DYN_BASE	0x0000000070000000UL
-
-
-/* This yields a mask that user programs can use to figure out what
-   instruction set this cpu supports.  */
-
-/* On Ultra, we support all of the v8 capabilities. */
-static inline unsigned int sparc64_elf_hwcap(void)
-{
-	unsigned int cap = (HWCAP_SPARC_FLUSH | HWCAP_SPARC_STBAR |
-			    HWCAP_SPARC_SWAP | HWCAP_SPARC_MULDIV |
-			    HWCAP_SPARC_V9);
-
-	if (tlb_type == cheetah || tlb_type == cheetah_plus)
-		cap |= HWCAP_SPARC_ULTRA3;
-	else if (tlb_type == hypervisor) {
-		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 ||
-		    sun4v_chip_type == SUN4V_CHIP_NIAGARA2)
-			cap |= HWCAP_SPARC_BLKINIT;
-		if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2)
-			cap |= HWCAP_SPARC_N2;
-	}
-
-	return cap;
-}
-
-#define ELF_HWCAP	sparc64_elf_hwcap();
-
-/* This yields a string that ld.so will use to load implementation
-   specific libraries for optimization.  This is more specific in
-   intent than poking at uname or /proc/cpuinfo.  */
-
-#define ELF_PLATFORM	(NULL)
-
-#define SET_PERSONALITY(ex, ibcs2)			\
-do {	unsigned long new_flags = current_thread_info()->flags; \
-	new_flags &= _TIF_32BIT;			\
-	if ((ex).e_ident[EI_CLASS] == ELFCLASS32)	\
-		new_flags |= _TIF_32BIT;		\
-	else						\
-		new_flags &= ~_TIF_32BIT;		\
-	if ((current_thread_info()->flags & _TIF_32BIT) \
-	    != new_flags)				\
-		set_thread_flag(TIF_ABI_PENDING);	\
-	else						\
-		clear_thread_flag(TIF_ABI_PENDING);	\
-	/* flush_thread will update pgd cache */	\
-	if (ibcs2)					\
-		set_personality(PER_SVR4);		\
-	else if (current->personality != PER_LINUX32)	\
-		set_personality(PER_LINUX);		\
-} while (0)
-
-#endif /* !(__ASM_SPARC64_ELF_H) */
+#include <asm-sparc/elf.h>
diff --git a/include/asm-sparc64/envctrl.h b/include/asm-sparc64/envctrl.h
index a5668a082b1..a2cc0ca334b 100644
--- a/include/asm-sparc64/envctrl.h
+++ b/include/asm-sparc64/envctrl.h
@@ -1,103 +1 @@
-/*
- *
- * envctrl.h: Definitions for access to the i2c environment
- *            monitoring on Ultrasparc systems.
- *
- * Copyright (C) 1998  Eddie C. Dost  (ecd@skynet.be)
- * Copyright (C) 2000  Vinh Truong  (vinh.truong@eng.sun.com)
- * VT - Add all ioctl commands and environment status definitions 
- * VT - Add application note 
- */
-#ifndef _SPARC64_ENVCTRL_H
-#define _SPARC64_ENVCTRL_H 1
-
-#include <linux/ioctl.h>
-
-/* Application note:
- *
- * The driver supports 4 operations: open(), close(), ioctl(), read()
- * The device name is /dev/envctrl.
- * Below is sample usage:
- *
- *	fd = open("/dev/envtrl", O_RDONLY);
- *	if (ioctl(fd, ENVCTRL_READ_SHUTDOWN_TEMPERATURE, 0) < 0)
- *		printf("error\n");
- *	ret = read(fd, buf, 10);
- *	close(fd);
- *
- * Notice in the case of cpu voltage and temperature, the default is
- * cpu0.  If we need to know the info of cpu1, cpu2, cpu3, we need to
- * pass in cpu number in ioctl() last parameter.  For example, to
- * get the voltage of cpu2:
- *
- *	ioctlbuf[0] = 2;
- *	if (ioctl(fd, ENVCTRL_READ_CPU_VOLTAGE, ioctlbuf) < 0)
- *		printf("error\n");
- *	ret = read(fd, buf, 10);
- *
- * All the return values are in ascii.  So check read return value
- * and do appropriate conversions in your application.
- */
-
-/* IOCTL commands */
-
-/* Note: these commands reflect possible monitor features.
- * Some boards choose to support some of the features only.
- */
-#define ENVCTRL_RD_CPU_TEMPERATURE	_IOR('p', 0x40, int)
-#define ENVCTRL_RD_CPU_VOLTAGE		_IOR('p', 0x41, int)
-#define ENVCTRL_RD_FAN_STATUS		_IOR('p', 0x42, int)
-#define ENVCTRL_RD_WARNING_TEMPERATURE	_IOR('p', 0x43, int)
-#define ENVCTRL_RD_SHUTDOWN_TEMPERATURE	_IOR('p', 0x44, int)
-#define ENVCTRL_RD_VOLTAGE_STATUS	_IOR('p', 0x45, int)
-#define ENVCTRL_RD_SCSI_TEMPERATURE	_IOR('p', 0x46, int)
-#define ENVCTRL_RD_ETHERNET_TEMPERATURE	_IOR('p', 0x47, int)
-#define ENVCTRL_RD_MTHRBD_TEMPERATURE	_IOR('p', 0x48, int)
-
-#define ENVCTRL_RD_GLOBALADDRESS	_IOR('p', 0x49, int)
-
-/* Read return values for a voltage status request. */
-#define ENVCTRL_VOLTAGE_POWERSUPPLY_GOOD	0x01
-#define ENVCTRL_VOLTAGE_BAD			0x02
-#define ENVCTRL_POWERSUPPLY_BAD			0x03
-#define ENVCTRL_VOLTAGE_POWERSUPPLY_BAD		0x04
-
-/* Read return values for a fan status request.
- * A failure match means either the fan fails or
- * the fan is not connected.  Some boards have optional
- * connectors to connect extra fans.
- *
- * There are maximum 8 monitor fans.  Some are cpu fans
- * some are system fans.  The mask below only indicates
- * fan by order number.
- * Below is a sample application:
- *
- *	if (ioctl(fd, ENVCTRL_READ_FAN_STATUS, 0) < 0) {
- *		printf("ioctl fan failed\n");
- *	}
- *	if (read(fd, rslt, 1) <= 0) {
- *		printf("error or fan not monitored\n");
- *	} else {
- *		if (rslt[0] == ENVCTRL_ALL_FANS_GOOD) {
- *			printf("all fans good\n");
- *	} else if (rslt[0] == ENVCTRL_ALL_FANS_BAD) {
- *		printf("all fans bad\n");
- *	} else {
- *		if (rslt[0] & ENVCTRL_FAN0_FAILURE_MASK) {
- *			printf("fan 0 failed or not connected\n");
- *	}
- *	......
- */  
-
-#define ENVCTRL_ALL_FANS_GOOD			0x00
-#define ENVCTRL_FAN0_FAILURE_MASK		0x01
-#define ENVCTRL_FAN1_FAILURE_MASK		0x02
-#define ENVCTRL_FAN2_FAILURE_MASK		0x04
-#define ENVCTRL_FAN3_FAILURE_MASK		0x08
-#define ENVCTRL_FAN4_FAILURE_MASK		0x10
-#define ENVCTRL_FAN5_FAILURE_MASK		0x20
-#define ENVCTRL_FAN6_FAILURE_MASK		0x40
-#define ENVCTRL_FAN7_FAILURE_MASK		0x80
-#define ENVCTRL_ALL_FANS_BAD 			0xFF
-
-#endif /* !(_SPARC64_ENVCTRL_H) */
+#include <asm-sparc/envctrl.h>
diff --git a/include/asm-sparc64/estate.h b/include/asm-sparc64/estate.h
index 520c08560d1..bedd0ef5f19 100644
--- a/include/asm-sparc64/estate.h
+++ b/include/asm-sparc64/estate.h
@@ -1,49 +1 @@
-#ifndef _SPARC64_ESTATE_H
-#define _SPARC64_ESTATE_H
-
-/* UltraSPARC-III E-cache Error Enable */
-#define ESTATE_ERROR_FMT	0x0000000000040000 /* Force MTAG ECC		*/
-#define ESTATE_ERROR_FMESS	0x000000000003c000 /* Forced MTAG ECC val	*/
-#define ESTATE_ERROR_FMD	0x0000000000002000 /* Force DATA ECC		*/
-#define ESTATE_ERROR_FDECC	0x0000000000001ff0 /* Forced DATA ECC val	*/
-#define ESTATE_ERROR_UCEEN	0x0000000000000008 /* See below			*/
-#define ESTATE_ERROR_NCEEN	0x0000000000000002 /* See below			*/
-#define ESTATE_ERROR_CEEN	0x0000000000000001 /* See below			*/
-
-/* UCEEN enables the fast_ECC_error trap for: 1) software correctable E-cache
- * errors 2) uncorrectable E-cache errors.  Such events only occur on reads
- * of the E-cache by the local processor for: 1) data loads 2) instruction
- * fetches 3) atomic operations.  Such events _cannot_ occur for: 1) merge
- * 2) writeback 2) copyout.  The AFSR bits associated with these traps are
- * UCC and UCU.
- */
-
-/* NCEEN enables instruction_access_error, data_access_error, and ECC_error traps
- * for uncorrectable ECC errors and system errors.
- *
- * Uncorrectable system bus data error or MTAG ECC error, system bus TimeOUT,
- * or system bus BusERR:
- * 1) As the result of an instruction fetch, will generate instruction_access_error
- * 2) As the result of a load etc. will generate data_access_error.
- * 3) As the result of store merge completion, writeback, or copyout will
- *    generate a disrupting ECC_error trap.
- * 4) As the result of such errors on instruction vector fetch can generate any
- *    of the 3 trap types.
- *
- * The AFSR bits associated with these traps are EMU, EDU, WDU, CPU, IVU, UE,
- * BERR, and TO.
- */
-
-/* CEEN enables the ECC_error trap for hardware corrected ECC errors.  System bus
- * reads resulting in a hardware corrected data or MTAG ECC error will generate an
- * ECC_error disrupting trap with this bit enabled.
- *
- * This same trap will also be generated when a hardware corrected ECC error results
- * during store merge, writeback, and copyout operations.
- */
-
-/* In general, if the trap enable bits above are disabled the AFSR bits will still
- * log the events even though the trap will not be generated by the processor.
- */
-
-#endif /* _SPARC64_ESTATE_H */
+#include <asm-sparc/estate.h>
diff --git a/include/asm-sparc64/fbio.h b/include/asm-sparc64/fbio.h
index b9215a0907d..c17edf8c7bc 100644
--- a/include/asm-sparc64/fbio.h
+++ b/include/asm-sparc64/fbio.h
@@ -1,330 +1 @@
-#ifndef __LINUX_FBIO_H
-#define __LINUX_FBIO_H
-
-#include <linux/compiler.h>
-#include <linux/types.h>
-
-/* Constants used for fbio SunOS compatibility */
-/* (C) 1996 Miguel de Icaza */
-
-/* Frame buffer types */
-#define FBTYPE_NOTYPE           -1
-#define FBTYPE_SUN1BW           0   /* mono */
-#define FBTYPE_SUN1COLOR        1 
-#define FBTYPE_SUN2BW           2 
-#define FBTYPE_SUN2COLOR        3 
-#define FBTYPE_SUN2GP           4 
-#define FBTYPE_SUN5COLOR        5 
-#define FBTYPE_SUN3COLOR        6 
-#define FBTYPE_MEMCOLOR         7 
-#define FBTYPE_SUN4COLOR        8 
- 
-#define FBTYPE_NOTSUN1          9 
-#define FBTYPE_NOTSUN2          10
-#define FBTYPE_NOTSUN3          11
- 
-#define FBTYPE_SUNFAST_COLOR    12  /* cg6 */
-#define FBTYPE_SUNROP_COLOR     13
-#define FBTYPE_SUNFB_VIDEO      14
-#define FBTYPE_SUNGIFB          15
-#define FBTYPE_SUNGPLAS         16
-#define FBTYPE_SUNGP3           17
-#define FBTYPE_SUNGT            18
-#define FBTYPE_SUNLEO           19      /* zx Leo card */
-#define FBTYPE_MDICOLOR         20      /* cg14 */
-#define FBTYPE_TCXCOLOR		21	/* SUNW,tcx card */
-
-#define FBTYPE_LASTPLUSONE      21	/* This is not last + 1 in fact... */
-
-/* Does not seem to be listed in the Sun file either */
-#define FBTYPE_CREATOR          22
-#define FBTYPE_PCI_IGA1682	23
-#define FBTYPE_P9100COLOR	24
-
-#define FBTYPE_PCI_GENERIC	1000
-#define FBTYPE_PCI_MACH64	1001
-
-/* fbio ioctls */
-/* Returned by FBIOGTYPE */
-struct  fbtype {
-        int     fb_type;        /* fb type, see above */
-        int     fb_height;      /* pixels */
-        int     fb_width;       /* pixels */
-        int     fb_depth;
-        int     fb_cmsize;      /* color map entries */
-        int     fb_size;        /* fb size in bytes */
-};
-#define FBIOGTYPE _IOR('F', 0, struct fbtype)
-
-struct  fbcmap {
-        int             index;          /* first element (0 origin) */
-        int             count;
-        unsigned char   __user *red;
-        unsigned char   __user *green;
-        unsigned char   __user *blue;
-};
-
-#ifdef __KERNEL__
-#define FBIOPUTCMAP_SPARC _IOW('F', 3, struct fbcmap)
-#define FBIOGETCMAP_SPARC _IOW('F', 4, struct fbcmap)
-#else
-#define FBIOPUTCMAP _IOW('F', 3, struct fbcmap)
-#define FBIOGETCMAP _IOW('F', 4, struct fbcmap)
-#endif
-
-/* # of device specific values */
-#define FB_ATTR_NDEVSPECIFIC    8
-/* # of possible emulations */
-#define FB_ATTR_NEMUTYPES       4
- 
-struct fbsattr {
-        int     flags;
-        int     emu_type;	/* -1 if none */
-        int     dev_specific[FB_ATTR_NDEVSPECIFIC];
-};
- 
-struct fbgattr {
-        int     real_type;	/* real frame buffer type */
-        int     owner;		/* unknown */
-        struct fbtype fbtype;	/* real frame buffer fbtype */
-        struct fbsattr sattr;   
-        int     emu_types[FB_ATTR_NEMUTYPES]; /* supported emulations */
-};
-#define FBIOSATTR  _IOW('F', 5, struct fbgattr) /* Unsupported: */
-#define FBIOGATTR  _IOR('F', 6, struct fbgattr)	/* supported */
-
-#define FBIOSVIDEO _IOW('F', 7, int)
-#define FBIOGVIDEO _IOR('F', 8, int)
-
-struct fbcursor {
-        short set;              /* what to set, choose from the list above */
-        short enable;           /* cursor on/off */
-        struct fbcurpos pos;    /* cursor position */
-        struct fbcurpos hot;    /* cursor hot spot */
-        struct fbcmap cmap;     /* color map info */
-        struct fbcurpos size;   /* cursor bit map size */
-        char __user *image;     /* cursor image bits */
-        char __user *mask;      /* cursor mask bits */
-};
-
-/* set/get cursor attributes/shape */
-#define FBIOSCURSOR     _IOW('F', 24, struct fbcursor)
-#define FBIOGCURSOR     _IOWR('F', 25, struct fbcursor)
- 
-/* set/get cursor position */
-#define FBIOSCURPOS     _IOW('F', 26, struct fbcurpos)
-#define FBIOGCURPOS     _IOW('F', 27, struct fbcurpos)
- 
-/* get max cursor size */
-#define FBIOGCURMAX     _IOR('F', 28, struct fbcurpos)
-
-/* wid manipulation */
-struct fb_wid_alloc {
-#define FB_WID_SHARED_8		0
-#define FB_WID_SHARED_24	1
-#define FB_WID_DBL_8		2
-#define FB_WID_DBL_24		3
-	__u32	wa_type;
-	__s32	wa_index;	/* Set on return */
-	__u32	wa_count;	
-};
-struct fb_wid_item {
-	__u32	wi_type;
-	__s32	wi_index;
-	__u32	wi_attrs;
-	__u32	wi_values[32];
-};
-struct fb_wid_list {
-	__u32	wl_flags;
-	__u32	wl_count;
-	struct fb_wid_item	*wl_list;
-};
-
-#define FBIO_WID_ALLOC	_IOWR('F', 30, struct fb_wid_alloc)
-#define FBIO_WID_FREE	_IOW('F', 31, struct fb_wid_alloc)
-#define FBIO_WID_PUT	_IOW('F', 32, struct fb_wid_list)
-#define FBIO_WID_GET	_IOWR('F', 33, struct fb_wid_list)
-
-/* Creator ioctls */
-#define FFB_IOCTL	('F'<<8)
-#define FFB_SYS_INFO		(FFB_IOCTL|80)
-#define FFB_CLUTREAD		(FFB_IOCTL|81)
-#define FFB_CLUTPOST		(FFB_IOCTL|82)
-#define FFB_SETDIAGMODE		(FFB_IOCTL|83)
-#define FFB_GETMONITORID	(FFB_IOCTL|84)
-#define FFB_GETVIDEOMODE	(FFB_IOCTL|85)
-#define FFB_SETVIDEOMODE	(FFB_IOCTL|86)
-#define FFB_SETSERVER		(FFB_IOCTL|87)
-#define FFB_SETOVCTL		(FFB_IOCTL|88)
-#define FFB_GETOVCTL		(FFB_IOCTL|89)
-#define FFB_GETSAXNUM		(FFB_IOCTL|90)
-#define FFB_FBDEBUG		(FFB_IOCTL|91)
-
-/* Cg14 ioctls */
-#define MDI_IOCTL          ('M'<<8)
-#define MDI_RESET          (MDI_IOCTL|1)
-#define MDI_GET_CFGINFO    (MDI_IOCTL|2)
-#define MDI_SET_PIXELMODE  (MDI_IOCTL|3)
-#    define MDI_32_PIX     32
-#    define MDI_16_PIX     16
-#    define MDI_8_PIX      8
-
-struct mdi_cfginfo {
-	int     mdi_ncluts;     /* Number of implemented CLUTs in this MDI */
-        int     mdi_type;       /* FBTYPE name */
-        int     mdi_height;     /* height */
-        int     mdi_width;      /* widht */
-        int     mdi_size;       /* available ram */
-        int     mdi_mode;       /* 8bpp, 16bpp or 32bpp */
-        int     mdi_pixfreq;    /* pixel clock (from PROM) */
-};
-
-/* SparcLinux specific ioctl for the MDI, should be replaced for
- * the SET_XLUT/SET_CLUTn ioctls instead
- */
-#define MDI_CLEAR_XLUT       (MDI_IOCTL|9)
-
-/* leo & ffb ioctls */
-struct fb_clut_alloc {
-	__u32	clutid;	/* Set on return */
- 	__u32	flag;
- 	__u32	index;
-};
-
-struct fb_clut {
-#define FB_CLUT_WAIT	0x00000001	/* Not yet implemented */
- 	__u32	flag;
- 	__u32	clutid;
- 	__u32	offset;
- 	__u32	count;
- 	char *	red;
- 	char *	green;
- 	char *	blue;
-};
-
-struct fb_clut32 {
- 	__u32	flag;
- 	__u32	clutid;
- 	__u32	offset;
- 	__u32	count;
- 	__u32	red;
- 	__u32	green;
- 	__u32	blue;
-};
-
-#define LEO_CLUTALLOC	_IOWR('L', 53, struct fb_clut_alloc)
-#define LEO_CLUTFREE	_IOW('L', 54, struct fb_clut_alloc)
-#define LEO_CLUTREAD	_IOW('L', 55, struct fb_clut)
-#define LEO_CLUTPOST	_IOW('L', 56, struct fb_clut)
-#define LEO_SETGAMMA	_IOW('L', 68, int) /* Not yet implemented */
-#define LEO_GETGAMMA	_IOR('L', 69, int) /* Not yet implemented */
-
-#ifdef __KERNEL__
-/* Addresses on the fd of a cgsix that are mappable */
-#define CG6_FBC    0x70000000
-#define CG6_TEC    0x70001000
-#define CG6_BTREGS 0x70002000
-#define CG6_FHC    0x70004000
-#define CG6_THC    0x70005000
-#define CG6_ROM    0x70006000
-#define CG6_RAM    0x70016000
-#define CG6_DHC    0x80000000
-
-#define CG3_MMAP_OFFSET 0x4000000
-
-/* Addresses on the fd of a tcx that are mappable */
-#define TCX_RAM8BIT   		0x00000000
-#define TCX_RAM24BIT   		0x01000000
-#define TCX_UNK3   		0x10000000
-#define TCX_UNK4   		0x20000000
-#define TCX_CONTROLPLANE   	0x28000000
-#define TCX_UNK6   		0x30000000
-#define TCX_UNK7   		0x38000000
-#define TCX_TEC    		0x70000000
-#define TCX_BTREGS 		0x70002000
-#define TCX_THC    		0x70004000
-#define TCX_DHC    		0x70008000
-#define TCX_ALT	   		0x7000a000
-#define TCX_SYNC   		0x7000e000
-#define TCX_UNK2    		0x70010000
-
-/* CG14 definitions */
-
-/* Offsets into the OBIO space: */
-#define CG14_REGS        0       /* registers */
-#define CG14_CURSORREGS  0x1000  /* cursor registers */
-#define CG14_DACREGS     0x2000  /* DAC registers */
-#define CG14_XLUT        0x3000  /* X Look Up Table -- ??? */
-#define CG14_CLUT1       0x4000  /* Color Look Up Table */
-#define CG14_CLUT2       0x5000  /* Color Look Up Table */
-#define CG14_CLUT3       0x6000  /* Color Look Up Table */
-#define CG14_AUTO	 0xf000
-
-#endif /* KERNEL */
-
-/* These are exported to userland for applications to use */
-/* Mappable offsets for the cg14: control registers */
-#define MDI_DIRECT_MAP 0x10000000
-#define MDI_CTLREG_MAP 0x20000000
-#define MDI_CURSOR_MAP 0x30000000
-#define MDI_SHDW_VRT_MAP 0x40000000
-
-/* Mappable offsets for the cg14: frame buffer resolutions */
-/* 32 bits */
-#define MDI_CHUNKY_XBGR_MAP 0x50000000
-#define MDI_CHUNKY_BGR_MAP 0x60000000
-
-/* 16 bits */
-#define MDI_PLANAR_X16_MAP 0x70000000
-#define MDI_PLANAR_C16_MAP 0x80000000
-
-/* 8 bit is done as CG3 MMAP offset */
-/* 32 bits, planar */
-#define MDI_PLANAR_X32_MAP 0x90000000
-#define MDI_PLANAR_B32_MAP 0xa0000000
-#define MDI_PLANAR_G32_MAP 0xb0000000
-#define MDI_PLANAR_R32_MAP 0xc0000000
-
-/* Mappable offsets on leo */
-#define LEO_SS0_MAP            0x00000000
-#define LEO_LC_SS0_USR_MAP     0x00800000
-#define LEO_LD_SS0_MAP         0x00801000
-#define LEO_LX_CURSOR_MAP      0x00802000
-#define LEO_SS1_MAP            0x00803000
-#define LEO_LC_SS1_USR_MAP     0x01003000
-#define LEO_LD_SS1_MAP         0x01004000
-#define LEO_UNK_MAP            0x01005000
-#define LEO_LX_KRN_MAP         0x01006000
-#define LEO_LC_SS0_KRN_MAP     0x01007000
-#define LEO_LC_SS1_KRN_MAP     0x01008000
-#define LEO_LD_GBL_MAP         0x01009000
-#define LEO_UNK2_MAP           0x0100a000
-
-#ifdef __KERNEL__
-struct  fbcmap32 {
-	int             index;          /* first element (0 origin) */
-	int             count;
-	u32		red;
-	u32		green;
-	u32		blue;
-};
-
-#define FBIOPUTCMAP32	_IOW('F', 3, struct fbcmap32)
-#define FBIOGETCMAP32	_IOW('F', 4, struct fbcmap32)
-
-struct fbcursor32 {
-	short set;		/* what to set, choose from the list above */
-	short enable;		/* cursor on/off */
-	struct fbcurpos pos;	/* cursor position */
-	struct fbcurpos hot;	/* cursor hot spot */
-	struct fbcmap32 cmap;	/* color map info */
-	struct fbcurpos size;	/* cursor bit map size */
-	u32	image;		/* cursor image bits */
-	u32	mask;		/* cursor mask bits */
-};
-
-#define FBIOSCURSOR32	_IOW('F', 24, struct fbcursor32)
-#define FBIOGCURSOR32	_IOW('F', 25, struct fbcursor32)
-#endif
-
-#endif /* __LINUX_FBIO_H */
+#include <asm-sparc/fbio.h>
diff --git a/include/asm-sparc64/fcntl.h b/include/asm-sparc64/fcntl.h
index 8a09ca7aa2f..8b1beae48cd 100644
--- a/include/asm-sparc64/fcntl.h
+++ b/include/asm-sparc64/fcntl.h
@@ -1,35 +1 @@
-#ifndef _SPARC64_FCNTL_H
-#define _SPARC64_FCNTL_H
-
-/* open/fcntl - O_SYNC is only implemented on blocks devices and on files
-   located on an ext2 file system */
-#define O_NDELAY	0x0004
-#define O_APPEND	0x0008
-#define FASYNC		0x0040	/* fcntl, for BSD compatibility */
-#define O_CREAT		0x0200	/* not fcntl */
-#define O_TRUNC		0x0400	/* not fcntl */
-#define O_EXCL		0x0800	/* not fcntl */
-#define O_SYNC		0x2000
-#define O_NONBLOCK	0x4000
-#define O_NOCTTY	0x8000	/* not fcntl */
-#define O_LARGEFILE	0x40000
-#define O_DIRECT        0x100000 /* direct disk access hint */
-#define O_NOATIME	0x200000
-#define O_CLOEXEC	0x400000
-
-#define F_GETOWN	5	/*  for sockets. */
-#define F_SETOWN	6	/*  for sockets. */
-#define F_GETLK		7
-#define F_SETLK		8
-#define F_SETLKW	9
-
-/* for posix fcntl() and lockf() */
-#define F_RDLCK		1
-#define F_WRLCK		2
-#define F_UNLCK		3
-
-#define __ARCH_FLOCK_PAD	short __unused;
-
-#include <asm-generic/fcntl.h>
-
-#endif /* !(_SPARC64_FCNTL_H) */
+#include <asm-sparc/fcntl.h>
diff --git a/include/asm-sparc64/fhc.h b/include/asm-sparc64/fhc.h
index ddffcdfbc98..73eb04c19c4 100644
--- a/include/asm-sparc64/fhc.h
+++ b/include/asm-sparc64/fhc.h
@@ -1,131 +1 @@
-/*
- * fhc.h: Structures for central/fhc pseudo driver on Sunfire/Starfire/Wildfire.
- *
- * Copyright (C) 1997, 1999 David S. Miller (davem@redhat.com)
- */
-
-#ifndef _SPARC64_FHC_H
-#define _SPARC64_FHC_H
-
-#include <linux/timer.h>
-
-#include <asm/oplib.h>
-#include <asm/prom.h>
-#include <asm/upa.h>
-
-struct linux_fhc;
-
-/* Clock board register offsets. */
-#define CLOCK_CTRL	0x00UL	/* Main control */
-#define CLOCK_STAT1	0x10UL	/* Status one */
-#define CLOCK_STAT2	0x20UL	/* Status two */
-#define CLOCK_PWRSTAT	0x30UL	/* Power status */
-#define CLOCK_PWRPRES	0x40UL	/* Power presence */
-#define CLOCK_TEMP	0x50UL	/* Temperature */
-#define CLOCK_IRQDIAG	0x60UL	/* IRQ diagnostics */
-#define CLOCK_PWRSTAT2	0x70UL	/* Power status two */
-
-#define CLOCK_CTRL_LLED		0x04	/* Left LED, 0 == on */
-#define CLOCK_CTRL_MLED		0x02	/* Mid LED, 1 == on */
-#define CLOCK_CTRL_RLED		0x01	/* RIght LED, 1 == on */
-
-struct linux_central {
-	struct linux_fhc		*child;
-	unsigned long			cfreg;
-	unsigned long			clkregs;
-	unsigned long			clkver;
-	int				slots;
-	struct device_node		*prom_node;
-
-	struct linux_prom_ranges	central_ranges[PROMREG_MAX];
-	int				num_central_ranges;
-};
-
-/* Firehose controller register offsets */
-struct fhc_regs {
-	unsigned long			pregs;	/* FHC internal regs */
-#define FHC_PREGS_ID	0x00UL	/* FHC ID */
-#define  FHC_ID_VERS		0xf0000000 /* Version of this FHC		*/
-#define  FHC_ID_PARTID		0x0ffff000 /* Part ID code (0x0f9f == FHC)	*/
-#define  FHC_ID_MANUF		0x0000007e /* Manufacturer (0x3e == SUN's JEDEC)*/
-#define  FHC_ID_RESV		0x00000001 /* Read as one			*/
-#define FHC_PREGS_RCS	0x10UL	/* FHC Reset Control/Status Register */
-#define  FHC_RCS_POR		0x80000000 /* Last reset was a power cycle	*/
-#define  FHC_RCS_SPOR		0x40000000 /* Last reset was sw power on reset	*/
-#define  FHC_RCS_SXIR		0x20000000 /* Last reset was sw XIR reset	*/
-#define  FHC_RCS_BPOR		0x10000000 /* Last reset was due to POR button	*/
-#define  FHC_RCS_BXIR		0x08000000 /* Last reset was due to XIR button	*/
-#define  FHC_RCS_WEVENT		0x04000000 /* CPU reset was due to wakeup event	*/
-#define  FHC_RCS_CFATAL		0x02000000 /* Centerplane Fatal Error signalled	*/
-#define  FHC_RCS_FENAB		0x01000000 /* Fatal errors elicit system reset	*/
-#define FHC_PREGS_CTRL	0x20UL	/* FHC Control Register */
-#define  FHC_CONTROL_ICS	0x00100000 /* Ignore Centerplane Signals	*/
-#define  FHC_CONTROL_FRST	0x00080000 /* Fatal Error Reset Enable		*/
-#define  FHC_CONTROL_LFAT	0x00040000 /* AC/DC signalled a local error	*/
-#define  FHC_CONTROL_SLINE	0x00010000 /* Firmware Synchronization Line	*/
-#define  FHC_CONTROL_DCD	0x00008000 /* DC-->DC Converter Disable		*/
-#define  FHC_CONTROL_POFF	0x00004000 /* AC/DC Controller PLL Disable	*/
-#define  FHC_CONTROL_FOFF	0x00002000 /* FHC Controller PLL Disable	*/
-#define  FHC_CONTROL_AOFF	0x00001000 /* CPU A SRAM/SBD Low Power Mode	*/
-#define  FHC_CONTROL_BOFF	0x00000800 /* CPU B SRAM/SBD Low Power Mode	*/
-#define  FHC_CONTROL_PSOFF	0x00000400 /* Turns off this FHC's power supply	*/
-#define  FHC_CONTROL_IXIST	0x00000200 /* 0=FHC tells clock board it exists	*/
-#define  FHC_CONTROL_XMSTR	0x00000100 /* 1=Causes this FHC to be XIR master*/
-#define  FHC_CONTROL_LLED	0x00000040 /* 0=Left LED ON			*/
-#define  FHC_CONTROL_MLED	0x00000020 /* 1=Middle LED ON			*/
-#define  FHC_CONTROL_RLED	0x00000010 /* 1=Right LED			*/
-#define  FHC_CONTROL_BPINS	0x00000003 /* Spare Bidirectional Pins		*/
-#define FHC_PREGS_BSR	0x30UL	/* FHC Board Status Register */
-#define  FHC_BSR_DA64		0x00040000 /* Port A: 0=128bit 1=64bit data path */
-#define  FHC_BSR_DB64		0x00020000 /* Port B: 0=128bit 1=64bit data path */
-#define  FHC_BSR_BID		0x0001e000 /* Board ID                           */
-#define  FHC_BSR_SA		0x00001c00 /* Port A UPA Speed (from the pins)   */
-#define  FHC_BSR_SB		0x00000380 /* Port B UPA Speed (from the pins)   */
-#define  FHC_BSR_NDIAG		0x00000040 /* Not in Diag Mode                   */
-#define  FHC_BSR_NTBED		0x00000020 /* Not in TestBED Mode                */
-#define  FHC_BSR_NIA		0x0000001c /* Jumper, bit 18 in PROM space       */
-#define  FHC_BSR_SI		0x00000001 /* Spare input pin value              */
-#define FHC_PREGS_ECC	0x40UL	/* FHC ECC Control Register (16 bits) */
-#define FHC_PREGS_JCTRL	0xf0UL	/* FHC JTAG Control Register */
-#define  FHC_JTAG_CTRL_MENAB	0x80000000 /* Indicates this is JTAG Master	 */
-#define  FHC_JTAG_CTRL_MNONE	0x40000000 /* Indicates no JTAG Master present	 */
-#define FHC_PREGS_JCMD	0x100UL	/* FHC JTAG Command Register */
-	unsigned long			ireg;	/* FHC IGN reg */
-#define FHC_IREG_IGN	0x00UL	/* This FHC's IGN */
-	unsigned long			ffregs;	/* FHC fanfail regs */
-#define FHC_FFREGS_IMAP	0x00UL	/* FHC Fanfail IMAP */
-#define FHC_FFREGS_ICLR	0x10UL	/* FHC Fanfail ICLR */
-	unsigned long			sregs;	/* FHC system regs */
-#define FHC_SREGS_IMAP	0x00UL	/* FHC System IMAP */
-#define FHC_SREGS_ICLR	0x10UL	/* FHC System ICLR */
-	unsigned long			uregs;	/* FHC uart regs */
-#define FHC_UREGS_IMAP	0x00UL	/* FHC Uart IMAP */
-#define FHC_UREGS_ICLR	0x10UL	/* FHC Uart ICLR */
-	unsigned long			tregs;	/* FHC TOD regs */
-#define FHC_TREGS_IMAP	0x00UL	/* FHC TOD IMAP */
-#define FHC_TREGS_ICLR	0x10UL	/* FHC TOD ICLR */
-};
-
-struct linux_fhc {
-	struct linux_fhc		*next;
-	struct linux_central		*parent;	/* NULL if not central FHC */
-	struct fhc_regs			fhc_regs;
-	int				board;
-	int				jtag_master;
-	struct device_node		*prom_node;
-
-	struct linux_prom_ranges	fhc_ranges[PROMREG_MAX];
-	int				num_fhc_ranges;
-};
-
-extern struct linux_central *central_bus;
-
-extern void apply_central_ranges(struct linux_central *central, 
-				 struct linux_prom_registers *regs,
-				 int nregs);
-
-extern void apply_fhc_ranges(struct linux_fhc *fhc, 
-			     struct linux_prom_registers *regs,
-			     int nregs);
-
-#endif /* !(_SPARC64_FHC_H) */
+#include <asm-sparc/fhc.h>
diff --git a/include/asm-sparc64/floppy.h b/include/asm-sparc64/floppy.h
index ca19f80a9b7..21487811443 100644
--- a/include/asm-sparc64/floppy.h
+++ b/include/asm-sparc64/floppy.h
@@ -1,782 +1 @@
-/* floppy.h: Sparc specific parts of the Floppy driver.
- *
- * Copyright (C) 1996, 2007 David S. Miller (davem@davemloft.net)
- * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- *
- * Ultra/PCI support added: Sep 1997  Eddie C. Dost  (ecd@skynet.be)
- */
-
-#ifndef __ASM_SPARC64_FLOPPY_H
-#define __ASM_SPARC64_FLOPPY_H
-
-#include <linux/init.h>
-#include <linux/pci.h>
-
-#include <asm/page.h>
-#include <asm/pgtable.h>
-#include <asm/system.h>
-#include <asm/idprom.h>
-#include <asm/oplib.h>
-#include <asm/auxio.h>
-#include <asm/sbus.h>
-#include <asm/irq.h>
-
-
-/*
- * Define this to enable exchanging drive 0 and 1 if only drive 1 is
- * probed on PCI machines.
- */
-#undef PCI_FDC_SWAP_DRIVES
-
-
-/* References:
- * 1) Netbsd Sun floppy driver.
- * 2) NCR 82077 controller manual
- * 3) Intel 82077 controller manual
- */
-struct sun_flpy_controller {
-	volatile unsigned char status1_82077; /* Auxiliary Status reg. 1 */
-	volatile unsigned char status2_82077; /* Auxiliary Status reg. 2 */
-	volatile unsigned char dor_82077;     /* Digital Output reg. */
-	volatile unsigned char tapectl_82077; /* Tape Control reg */
-	volatile unsigned char status_82077;  /* Main Status Register. */
-#define drs_82077              status_82077   /* Digital Rate Select reg. */
-	volatile unsigned char data_82077;    /* Data fifo. */
-	volatile unsigned char ___unused;
-	volatile unsigned char dir_82077;     /* Digital Input reg. */
-#define dcr_82077              dir_82077      /* Config Control reg. */
-};
-
-/* You'll only ever find one controller on an Ultra anyways. */
-static struct sun_flpy_controller *sun_fdc = (struct sun_flpy_controller *)-1;
-unsigned long fdc_status;
-static struct sbus_dev *floppy_sdev = NULL;
-
-struct sun_floppy_ops {
-	unsigned char	(*fd_inb) (unsigned long port);
-	void		(*fd_outb) (unsigned char value, unsigned long port);
-	void		(*fd_enable_dma) (void);
-	void		(*fd_disable_dma) (void);
-	void		(*fd_set_dma_mode) (int);
-	void		(*fd_set_dma_addr) (char *);
-	void		(*fd_set_dma_count) (int);
-	unsigned int	(*get_dma_residue) (void);
-	int		(*fd_request_irq) (void);
-	void		(*fd_free_irq) (void);
-	int		(*fd_eject) (int);
-};
-
-static struct sun_floppy_ops sun_fdops;
-
-#define fd_inb(port)              sun_fdops.fd_inb(port)
-#define fd_outb(value,port)       sun_fdops.fd_outb(value,port)
-#define fd_enable_dma()           sun_fdops.fd_enable_dma()
-#define fd_disable_dma()          sun_fdops.fd_disable_dma()
-#define fd_request_dma()          (0) /* nothing... */
-#define fd_free_dma()             /* nothing... */
-#define fd_clear_dma_ff()         /* nothing... */
-#define fd_set_dma_mode(mode)     sun_fdops.fd_set_dma_mode(mode)
-#define fd_set_dma_addr(addr)     sun_fdops.fd_set_dma_addr(addr)
-#define fd_set_dma_count(count)   sun_fdops.fd_set_dma_count(count)
-#define get_dma_residue(x)        sun_fdops.get_dma_residue()
-#define fd_cacheflush(addr, size) /* nothing... */
-#define fd_request_irq()          sun_fdops.fd_request_irq()
-#define fd_free_irq()             sun_fdops.fd_free_irq()
-#define fd_eject(drive)           sun_fdops.fd_eject(drive)
-
-/* Super paranoid... */
-#undef HAVE_DISABLE_HLT
-
-static int sun_floppy_types[2] = { 0, 0 };
-
-/* Here is where we catch the floppy driver trying to initialize,
- * therefore this is where we call the PROM device tree probing
- * routine etc. on the Sparc.
- */
-#define FLOPPY0_TYPE		sun_floppy_init()
-#define FLOPPY1_TYPE		sun_floppy_types[1]
-
-#define FDC1			((unsigned long)sun_fdc)
-
-#define N_FDC    1
-#define N_DRIVE  8
-
-/* No 64k boundary crossing problems on the Sparc. */
-#define CROSS_64KB(a,s) (0)
-
-static unsigned char sun_82077_fd_inb(unsigned long port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to read unknown port %lx\n", port);
-		panic("floppy: Port bolixed.");
-	case 4: /* FD_STATUS */
-		return sbus_readb(&sun_fdc->status_82077) & ~STATUS_DMA;
-	case 5: /* FD_DATA */
-		return sbus_readb(&sun_fdc->data_82077);
-	case 7: /* FD_DIR */
-		/* XXX: Is DCL on 0x80 in sun4m? */
-		return sbus_readb(&sun_fdc->dir_82077);
-	};
-	panic("sun_82072_fd_inb: How did I get here?");
-}
-
-static void sun_82077_fd_outb(unsigned char value, unsigned long port)
-{
-	udelay(5);
-	switch(port & 7) {
-	default:
-		printk("floppy: Asked to write to unknown port %lx\n", port);
-		panic("floppy: Port bolixed.");
-	case 2: /* FD_DOR */
-		/* Happily, the 82077 has a real DOR register. */
-		sbus_writeb(value, &sun_fdc->dor_82077);
-		break;
-	case 5: /* FD_DATA */
-		sbus_writeb(value, &sun_fdc->data_82077);
-		break;
-	case 7: /* FD_DCR */
-		sbus_writeb(value, &sun_fdc->dcr_82077);
-		break;
-	case 4: /* FD_STATUS */
-		sbus_writeb(value, &sun_fdc->status_82077);
-		break;
-	};
-	return;
-}
-
-/* For pseudo-dma (Sun floppy drives have no real DMA available to
- * them so we must eat the data fifo bytes directly ourselves) we have
- * three state variables.  doing_pdma tells our inline low-level
- * assembly floppy interrupt entry point whether it should sit and eat
- * bytes from the fifo or just transfer control up to the higher level
- * floppy interrupt c-code.  I tried very hard but I could not get the
- * pseudo-dma to work in c-code without getting many overruns and
- * underruns.  If non-zero, doing_pdma encodes the direction of
- * the transfer for debugging.  1=read 2=write
- */
-unsigned char *pdma_vaddr;
-unsigned long pdma_size;
-volatile int doing_pdma = 0;
-
-/* This is software state */
-char *pdma_base = NULL;
-unsigned long pdma_areasize;
-
-/* Common routines to all controller types on the Sparc. */
-static void sun_fd_disable_dma(void)
-{
-	doing_pdma = 0;
-	if (pdma_base) {
-		mmu_unlockarea(pdma_base, pdma_areasize);
-		pdma_base = NULL;
-	}
-}
-
-static void sun_fd_set_dma_mode(int mode)
-{
-	switch(mode) {
-	case DMA_MODE_READ:
-		doing_pdma = 1;
-		break;
-	case DMA_MODE_WRITE:
-		doing_pdma = 2;
-		break;
-	default:
-		printk("Unknown dma mode %d\n", mode);
-		panic("floppy: Giving up...");
-	}
-}
-
-static void sun_fd_set_dma_addr(char *buffer)
-{
-	pdma_vaddr = buffer;
-}
-
-static void sun_fd_set_dma_count(int length)
-{
-	pdma_size = length;
-}
-
-static void sun_fd_enable_dma(void)
-{
-	pdma_vaddr = mmu_lockarea(pdma_vaddr, pdma_size);
-	pdma_base = pdma_vaddr;
-	pdma_areasize = pdma_size;
-}
-
-irqreturn_t sparc_floppy_irq(int irq, void *dev_cookie)
-{
-	if (likely(doing_pdma)) {
-		void __iomem *stat = (void __iomem *) fdc_status;
-		unsigned char *vaddr = pdma_vaddr;
-		unsigned long size = pdma_size;
-		u8 val;
-
-		while (size) {
-			val = readb(stat);
-			if (unlikely(!(val & 0x80))) {
-				pdma_vaddr = vaddr;
-				pdma_size = size;
-				return IRQ_HANDLED;
-			}
-			if (unlikely(!(val & 0x20))) {
-				pdma_vaddr = vaddr;
-				pdma_size = size;
-				doing_pdma = 0;
-				goto main_interrupt;
-			}
-			if (val & 0x40) {
-				/* read */
-				*vaddr++ = readb(stat + 1);
-			} else {
-				unsigned char data = *vaddr++;
-
-				/* write */
-				writeb(data, stat + 1);
-			}
-			size--;
-		}
-
-		pdma_vaddr = vaddr;
-		pdma_size = size;
-
-		/* Send Terminal Count pulse to floppy controller. */
-		val = readb(auxio_register);
-		val |= AUXIO_AUX1_FTCNT;
-		writeb(val, auxio_register);
-		val &= ~AUXIO_AUX1_FTCNT;
-		writeb(val, auxio_register);
-
-		doing_pdma = 0;
-	}
-
-main_interrupt:
-	return floppy_interrupt(irq, dev_cookie);
-}
-
-static int sun_fd_request_irq(void)
-{
-	static int once = 0;
-	int error;
-
-	if(!once) {
-		once = 1;
-
-		error = request_irq(FLOPPY_IRQ, sparc_floppy_irq, 
-				    IRQF_DISABLED, "floppy", NULL);
-
-		return ((error == 0) ? 0 : -1);
-	}
-	return 0;
-}
-
-static void sun_fd_free_irq(void)
-{
-}
-
-static unsigned int sun_get_dma_residue(void)
-{
-	/* XXX This isn't really correct. XXX */
-	return 0;
-}
-
-static int sun_fd_eject(int drive)
-{
-	set_dor(0x00, 0xff, 0x90);
-	udelay(500);
-	set_dor(0x00, 0x6f, 0x00);
-	udelay(500);
-	return 0;
-}
-
-#ifdef CONFIG_PCI
-#include <asm/ebus.h>
-#include <asm/ns87303.h>
-
-static struct ebus_dma_info sun_pci_fd_ebus_dma;
-static struct pci_dev *sun_pci_ebus_dev;
-static int sun_pci_broken_drive = -1;
-
-struct sun_pci_dma_op {
-	unsigned int 	addr;
-	int		len;
-	int		direction;
-	char		*buf;
-};
-static struct sun_pci_dma_op sun_pci_dma_current = { -1U, 0, 0, NULL};
-static struct sun_pci_dma_op sun_pci_dma_pending = { -1U, 0, 0, NULL};
-
-extern irqreturn_t floppy_interrupt(int irq, void *dev_id);
-
-static unsigned char sun_pci_fd_inb(unsigned long port)
-{
-	udelay(5);
-	return inb(port);
-}
-
-static void sun_pci_fd_outb(unsigned char val, unsigned long port)
-{
-	udelay(5);
-	outb(val, port);
-}
-
-static void sun_pci_fd_broken_outb(unsigned char val, unsigned long port)
-{
-	udelay(5);
-	/*
-	 * XXX: Due to SUN's broken floppy connector on AX and AXi
-	 *      we need to turn on MOTOR_0 also, if the floppy is
-	 *      jumpered to DS1 (like most PC floppies are). I hope
-	 *      this does not hurt correct hardware like the AXmp.
-	 *      (Eddie, Sep 12 1998).
-	 */
-	if (port == ((unsigned long)sun_fdc) + 2) {
-		if (((val & 0x03) == sun_pci_broken_drive) && (val & 0x20)) {
-			val |= 0x10;
-		}
-	}
-	outb(val, port);
-}
-
-#ifdef PCI_FDC_SWAP_DRIVES
-static void sun_pci_fd_lde_broken_outb(unsigned char val, unsigned long port)
-{
-	udelay(5);
-	/*
-	 * XXX: Due to SUN's broken floppy connector on AX and AXi
-	 *      we need to turn on MOTOR_0 also, if the floppy is
-	 *      jumpered to DS1 (like most PC floppies are). I hope
-	 *      this does not hurt correct hardware like the AXmp.
-	 *      (Eddie, Sep 12 1998).
-	 */
-	if (port == ((unsigned long)sun_fdc) + 2) {
-		if (((val & 0x03) == sun_pci_broken_drive) && (val & 0x10)) {
-			val &= ~(0x03);
-			val |= 0x21;
-		}
-	}
-	outb(val, port);
-}
-#endif /* PCI_FDC_SWAP_DRIVES */
-
-static void sun_pci_fd_enable_dma(void)
-{
-	BUG_ON((NULL == sun_pci_dma_pending.buf) 	||
-	    (0	  == sun_pci_dma_pending.len) 	||
-	    (0	  == sun_pci_dma_pending.direction));
-
-	sun_pci_dma_current.buf = sun_pci_dma_pending.buf;
-	sun_pci_dma_current.len = sun_pci_dma_pending.len;
-	sun_pci_dma_current.direction = sun_pci_dma_pending.direction;
-
-	sun_pci_dma_pending.buf  = NULL;
-	sun_pci_dma_pending.len  = 0;
-	sun_pci_dma_pending.direction = 0;
-	sun_pci_dma_pending.addr = -1U;
-
-	sun_pci_dma_current.addr = 
-		pci_map_single(sun_pci_ebus_dev,
-			       sun_pci_dma_current.buf,
-			       sun_pci_dma_current.len,
-			       sun_pci_dma_current.direction);
-
-	ebus_dma_enable(&sun_pci_fd_ebus_dma, 1);
-
-	if (ebus_dma_request(&sun_pci_fd_ebus_dma,
-			     sun_pci_dma_current.addr,
-			     sun_pci_dma_current.len))
-		BUG();
-}
-
-static void sun_pci_fd_disable_dma(void)
-{
-	ebus_dma_enable(&sun_pci_fd_ebus_dma, 0);
-	if (sun_pci_dma_current.addr != -1U)
-		pci_unmap_single(sun_pci_ebus_dev,
-				 sun_pci_dma_current.addr,
-				 sun_pci_dma_current.len,
-				 sun_pci_dma_current.direction);
-	sun_pci_dma_current.addr = -1U;
-}
-
-static void sun_pci_fd_set_dma_mode(int mode)
-{
-	if (mode == DMA_MODE_WRITE)
-		sun_pci_dma_pending.direction = PCI_DMA_TODEVICE;
-	else
-		sun_pci_dma_pending.direction = PCI_DMA_FROMDEVICE;
-
-	ebus_dma_prepare(&sun_pci_fd_ebus_dma, mode != DMA_MODE_WRITE);
-}
-
-static void sun_pci_fd_set_dma_count(int length)
-{
-	sun_pci_dma_pending.len = length;
-}
-
-static void sun_pci_fd_set_dma_addr(char *buffer)
-{
-	sun_pci_dma_pending.buf = buffer;
-}
-
-static unsigned int sun_pci_get_dma_residue(void)
-{
-	return ebus_dma_residue(&sun_pci_fd_ebus_dma);
-}
-
-static int sun_pci_fd_request_irq(void)
-{
-	return ebus_dma_irq_enable(&sun_pci_fd_ebus_dma, 1);
-}
-
-static void sun_pci_fd_free_irq(void)
-{
-	ebus_dma_irq_enable(&sun_pci_fd_ebus_dma, 0);
-}
-
-static int sun_pci_fd_eject(int drive)
-{
-	return -EINVAL;
-}
-
-void sun_pci_fd_dma_callback(struct ebus_dma_info *p, int event, void *cookie)
-{
-	floppy_interrupt(0, NULL);
-}
-
-/*
- * Floppy probing, we'd like to use /dev/fd0 for a single Floppy on PCI,
- * even if this is configured using DS1, thus looks like /dev/fd1 with
- * the cabling used in Ultras.
- */
-#define DOR	(port + 2)
-#define MSR	(port + 4)
-#define FIFO	(port + 5)
-
-static void sun_pci_fd_out_byte(unsigned long port, unsigned char val,
-			        unsigned long reg)
-{
-	unsigned char status;
-	int timeout = 1000;
-
-	while (!((status = inb(MSR)) & 0x80) && --timeout)
-		udelay(100);
-	outb(val, reg);
-}
-
-static unsigned char sun_pci_fd_sensei(unsigned long port)
-{
-	unsigned char result[2] = { 0x70, 0x00 };
-	unsigned char status;
-	int i = 0;
-
-	sun_pci_fd_out_byte(port, 0x08, FIFO);
-	do {
-		int timeout = 1000;
-
-		while (!((status = inb(MSR)) & 0x80) && --timeout)
-			udelay(100);
-
-		if (!timeout)
-			break;
-
-		if ((status & 0xf0) == 0xd0)
-			result[i++] = inb(FIFO);
-		else
-			break;
-	} while (i < 2);
-
-	return result[0];
-}
-
-static void sun_pci_fd_reset(unsigned long port)
-{
-	unsigned char mask = 0x00;
-	unsigned char status;
-	int timeout = 10000;
-
-	outb(0x80, MSR);
-	do {
-		status = sun_pci_fd_sensei(port);
-		if ((status & 0xc0) == 0xc0)
-			mask |= 1 << (status & 0x03);
-		else
-			udelay(100);
-	} while ((mask != 0x0f) && --timeout);
-}
-
-static int sun_pci_fd_test_drive(unsigned long port, int drive)
-{
-	unsigned char status, data;
-	int timeout = 1000;
-	int ready;
-
-	sun_pci_fd_reset(port);
-
-	data = (0x10 << drive) | 0x0c | drive;
-	sun_pci_fd_out_byte(port, data, DOR);
-
-	sun_pci_fd_out_byte(port, 0x07, FIFO);
-	sun_pci_fd_out_byte(port, drive & 0x03, FIFO);
-
-	do {
-		udelay(100);
-		status = sun_pci_fd_sensei(port);
-	} while (((status & 0xc0) == 0x80) && --timeout);
-
-	if (!timeout)
-		ready = 0;
-	else
-		ready = (status & 0x10) ? 0 : 1;
-
-	sun_pci_fd_reset(port);
-	return ready;
-}
-#undef FIFO
-#undef MSR
-#undef DOR
-
-#endif /* CONFIG_PCI */
-
-#ifdef CONFIG_PCI
-static int __init ebus_fdthree_p(struct linux_ebus_device *edev)
-{
-	if (!strcmp(edev->prom_node->name, "fdthree"))
-		return 1;
-	if (!strcmp(edev->prom_node->name, "floppy")) {
-		const char *compat;
-
-		compat = of_get_property(edev->prom_node,
-					 "compatible", NULL);
-		if (compat && !strcmp(compat, "fdthree"))
-			return 1;
-	}
-	return 0;
-}
-#endif
-
-static unsigned long __init sun_floppy_init(void)
-{
-	char state[128];
-	struct sbus_bus *bus;
-	struct sbus_dev *sdev = NULL;
-	static int initialized = 0;
-
-	if (initialized)
-		return sun_floppy_types[0];
-	initialized = 1;
-
-	for_all_sbusdev (sdev, bus) {
-		if (!strcmp(sdev->prom_name, "SUNW,fdtwo")) 
-			break;
-	}
-	if(sdev) {
-		floppy_sdev = sdev;
-		FLOPPY_IRQ = sdev->irqs[0];
-	} else {
-#ifdef CONFIG_PCI
-		struct linux_ebus *ebus;
-		struct linux_ebus_device *edev = NULL;
-		unsigned long config = 0;
-		void __iomem *auxio_reg;
-		const char *state_prop;
-
-		for_each_ebus(ebus) {
-			for_each_ebusdev(edev, ebus) {
-				if (ebus_fdthree_p(edev))
-					goto ebus_done;
-			}
-		}
-	ebus_done:
-		if (!edev)
-			return 0;
-
-		state_prop = of_get_property(edev->prom_node, "status", NULL);
-		if (state_prop && !strncmp(state_prop, "disabled", 8))
-			return 0;
-			
-		FLOPPY_IRQ = edev->irqs[0];
-
-		/* Make sure the high density bit is set, some systems
-		 * (most notably Ultra5/Ultra10) come up with it clear.
-		 */
-		auxio_reg = (void __iomem *) edev->resource[2].start;
-		writel(readl(auxio_reg)|0x2, auxio_reg);
-
-		sun_pci_ebus_dev = ebus->self;
-
-		spin_lock_init(&sun_pci_fd_ebus_dma.lock);
-
-		/* XXX ioremap */
-		sun_pci_fd_ebus_dma.regs = (void __iomem *)
-			edev->resource[1].start;
-		if (!sun_pci_fd_ebus_dma.regs)
-			return 0;
-
-		sun_pci_fd_ebus_dma.flags = (EBUS_DMA_FLAG_USE_EBDMA_HANDLER |
-					     EBUS_DMA_FLAG_TCI_DISABLE);
-		sun_pci_fd_ebus_dma.callback = sun_pci_fd_dma_callback;
-		sun_pci_fd_ebus_dma.client_cookie = NULL;
-		sun_pci_fd_ebus_dma.irq = FLOPPY_IRQ;
-		strcpy(sun_pci_fd_ebus_dma.name, "floppy");
-		if (ebus_dma_register(&sun_pci_fd_ebus_dma))
-			return 0;
-
-		/* XXX ioremap */
-		sun_fdc = (struct sun_flpy_controller *)edev->resource[0].start;
-
-		sun_fdops.fd_inb = sun_pci_fd_inb;
-		sun_fdops.fd_outb = sun_pci_fd_outb;
-
-		can_use_virtual_dma = use_virtual_dma = 0;
-		sun_fdops.fd_enable_dma = sun_pci_fd_enable_dma;
-		sun_fdops.fd_disable_dma = sun_pci_fd_disable_dma;
-		sun_fdops.fd_set_dma_mode = sun_pci_fd_set_dma_mode;
-		sun_fdops.fd_set_dma_addr = sun_pci_fd_set_dma_addr;
-		sun_fdops.fd_set_dma_count = sun_pci_fd_set_dma_count;
-		sun_fdops.get_dma_residue = sun_pci_get_dma_residue;
-
-		sun_fdops.fd_request_irq = sun_pci_fd_request_irq;
-		sun_fdops.fd_free_irq = sun_pci_fd_free_irq;
-
-		sun_fdops.fd_eject = sun_pci_fd_eject;
-
-        	fdc_status = (unsigned long) &sun_fdc->status_82077;
-
-		/*
-		 * XXX: Find out on which machines this is really needed.
-		 */
-		if (1) {
-			sun_pci_broken_drive = 1;
-			sun_fdops.fd_outb = sun_pci_fd_broken_outb;
-		}
-
-		allowed_drive_mask = 0;
-		if (sun_pci_fd_test_drive((unsigned long)sun_fdc, 0))
-			sun_floppy_types[0] = 4;
-		if (sun_pci_fd_test_drive((unsigned long)sun_fdc, 1))
-			sun_floppy_types[1] = 4;
-
-		/*
-		 * Find NS87303 SuperIO config registers (through ecpp).
-		 */
-		for_each_ebus(ebus) {
-			for_each_ebusdev(edev, ebus) {
-				if (!strcmp(edev->prom_node->name, "ecpp")) {
-					config = edev->resource[1].start;
-					goto config_done;
-				}
-			}
-		}
-	config_done:
-
-		/*
-		 * Sanity check, is this really the NS87303?
-		 */
-		switch (config & 0x3ff) {
-		case 0x02e:
-		case 0x15c:
-		case 0x26e:
-		case 0x398:
-			break;
-		default:
-			config = 0;
-		}
-
-		if (!config)
-			return sun_floppy_types[0];
-
-		/* Enable PC-AT mode. */
-		ns87303_modify(config, ASC, 0, 0xc0);
-
-#ifdef PCI_FDC_SWAP_DRIVES
-		/*
-		 * If only Floppy 1 is present, swap drives.
-		 */
-		if (!sun_floppy_types[0] && sun_floppy_types[1]) {
-			/*
-			 * Set the drive exchange bit in FCR on NS87303,
-			 * make sure other bits are sane before doing so.
-			 */
-			ns87303_modify(config, FER, FER_EDM, 0);
-			ns87303_modify(config, ASC, ASC_DRV2_SEL, 0);
-			ns87303_modify(config, FCR, 0, FCR_LDE);
-
-			config = sun_floppy_types[0];
-			sun_floppy_types[0] = sun_floppy_types[1];
-			sun_floppy_types[1] = config;
-
-			if (sun_pci_broken_drive != -1) {
-				sun_pci_broken_drive = 1 - sun_pci_broken_drive;
-				sun_fdops.fd_outb = sun_pci_fd_lde_broken_outb;
-			}
-		}
-#endif /* PCI_FDC_SWAP_DRIVES */
-
-		return sun_floppy_types[0];
-#else
-		return 0;
-#endif
-	}
-	prom_getproperty(sdev->prom_node, "status", state, sizeof(state));
-	if(!strncmp(state, "disabled", 8))
-		return 0;
-
-	/*
-	 * We cannot do sbus_ioremap here: it does request_region,
-	 * which the generic floppy driver tries to do once again.
-	 * But we must use the sdev resource values as they have
-	 * had parent ranges applied.
-	 */
-	sun_fdc = (struct sun_flpy_controller *)
-		(sdev->resource[0].start +
-		 ((sdev->resource[0].flags & 0x1ffUL) << 32UL));
-
-	/* Last minute sanity check... */
-	if(sbus_readb(&sun_fdc->status1_82077) == 0xff) {
-		sun_fdc = (struct sun_flpy_controller *)-1;
-		return 0;
-	}
-
-        sun_fdops.fd_inb = sun_82077_fd_inb;
-        sun_fdops.fd_outb = sun_82077_fd_outb;
-
-	can_use_virtual_dma = use_virtual_dma = 1;
-	sun_fdops.fd_enable_dma = sun_fd_enable_dma;
-	sun_fdops.fd_disable_dma = sun_fd_disable_dma;
-	sun_fdops.fd_set_dma_mode = sun_fd_set_dma_mode;
-	sun_fdops.fd_set_dma_addr = sun_fd_set_dma_addr;
-	sun_fdops.fd_set_dma_count = sun_fd_set_dma_count;
-	sun_fdops.get_dma_residue = sun_get_dma_residue;
-
-	sun_fdops.fd_request_irq = sun_fd_request_irq;
-	sun_fdops.fd_free_irq = sun_fd_free_irq;
-
-	sun_fdops.fd_eject = sun_fd_eject;
-
-        fdc_status = (unsigned long) &sun_fdc->status_82077;
-
-	/* Success... */
-	allowed_drive_mask = 0x01;
-	sun_floppy_types[0] = 4;
-	sun_floppy_types[1] = 0;
-
-	return sun_floppy_types[0];
-}
-
-#define EXTRA_FLOPPY_PARAMS
-
-static DEFINE_SPINLOCK(dma_spin_lock);
-
-#define claim_dma_lock() \
-({	unsigned long flags; \
-	spin_lock_irqsave(&dma_spin_lock, flags); \
-	flags; \
-})
-
-#define release_dma_lock(__flags) \
-	spin_unlock_irqrestore(&dma_spin_lock, __flags);
-
-#endif /* !(__ASM_SPARC64_FLOPPY_H) */
+#include <asm-sparc/floppy.h>
diff --git a/include/asm-sparc64/fpumacro.h b/include/asm-sparc64/fpumacro.h
index cc463fec806..30d6d0f68bc 100644
--- a/include/asm-sparc64/fpumacro.h
+++ b/include/asm-sparc64/fpumacro.h
@@ -1,33 +1 @@
-/* fpumacro.h: FPU related macros.
- *
- * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC64_FPUMACRO_H
-#define _SPARC64_FPUMACRO_H
-
-#include <asm/asi.h>
-#include <asm/visasm.h>
-
-struct fpustate {
-	u32	regs[64];
-};
-
-#define FPUSTATE (struct fpustate *)(current_thread_info()->fpregs)
-
-static inline unsigned long fprs_read(void)
-{
-	unsigned long retval;
-
-	__asm__ __volatile__("rd %%fprs, %0" : "=r" (retval));
-
-	return retval;
-}
-
-static inline void fprs_write(unsigned long val)
-{
-	__asm__ __volatile__("wr %0, 0x0, %%fprs" : : "r" (val));
-}
-
-#endif /* !(_SPARC64_FPUMACRO_H) */
+#include <asm-sparc/fpumacro.h>
diff --git a/include/asm-sparc64/futex.h b/include/asm-sparc64/futex.h
index d8378935ae9..1ceb0bb2fe5 100644
--- a/include/asm-sparc64/futex.h
+++ b/include/asm-sparc64/futex.h
@@ -1,110 +1 @@
-#ifndef _SPARC64_FUTEX_H
-#define _SPARC64_FUTEX_H
-
-#include <linux/futex.h>
-#include <linux/uaccess.h>
-#include <asm/errno.h>
-#include <asm/system.h>
-
-#define __futex_cas_op(insn, ret, oldval, uaddr, oparg)	\
-	__asm__ __volatile__(				\
-	"\n1:	lduwa	[%3] %%asi, %2\n"		\
-	"	" insn "\n"				\
-	"2:	casa	[%3] %%asi, %2, %1\n"		\
-	"	cmp	%2, %1\n"			\
-	"	bne,pn	%%icc, 1b\n"			\
-	"	 mov	0, %0\n"			\
-	"3:\n"						\
-	"	.section .fixup,#alloc,#execinstr\n"	\
-	"	.align	4\n"				\
-	"4:	sethi	%%hi(3b), %0\n"			\
-	"	jmpl	%0 + %%lo(3b), %%g0\n"		\
-	"	 mov	%5, %0\n"			\
-	"	.previous\n"				\
-	"	.section __ex_table,\"a\"\n"		\
-	"	.align	4\n"				\
-	"	.word	1b, 4b\n"			\
-	"	.word	2b, 4b\n"			\
-	"	.previous\n"				\
-	: "=&r" (ret), "=&r" (oldval), "=&r" (tem)	\
-	: "r" (uaddr), "r" (oparg), "i" (-EFAULT)	\
-	: "memory")
-
-static inline int futex_atomic_op_inuser(int encoded_op, int __user *uaddr)
-{
-	int op = (encoded_op >> 28) & 7;
-	int cmp = (encoded_op >> 24) & 15;
-	int oparg = (encoded_op << 8) >> 20;
-	int cmparg = (encoded_op << 20) >> 20;
-	int oldval = 0, ret, tem;
-
-	if (unlikely(!access_ok(VERIFY_WRITE, uaddr, sizeof(int))))
-		return -EFAULT;
-	if (unlikely((((unsigned long) uaddr) & 0x3UL)))
-		return -EINVAL;
-
-	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
-		oparg = 1 << oparg;
-
-	pagefault_disable();
-
-	switch (op) {
-	case FUTEX_OP_SET:
-		__futex_cas_op("mov\t%4, %1", ret, oldval, uaddr, oparg);
-		break;
-	case FUTEX_OP_ADD:
-		__futex_cas_op("add\t%2, %4, %1", ret, oldval, uaddr, oparg);
-		break;
-	case FUTEX_OP_OR:
-		__futex_cas_op("or\t%2, %4, %1", ret, oldval, uaddr, oparg);
-		break;
-	case FUTEX_OP_ANDN:
-		__futex_cas_op("and\t%2, %4, %1", ret, oldval, uaddr, oparg);
-		break;
-	case FUTEX_OP_XOR:
-		__futex_cas_op("xor\t%2, %4, %1", ret, oldval, uaddr, oparg);
-		break;
-	default:
-		ret = -ENOSYS;
-	}
-
-	pagefault_enable();
-
-	if (!ret) {
-		switch (cmp) {
-		case FUTEX_OP_CMP_EQ: ret = (oldval == cmparg); break;
-		case FUTEX_OP_CMP_NE: ret = (oldval != cmparg); break;
-		case FUTEX_OP_CMP_LT: ret = (oldval < cmparg); break;
-		case FUTEX_OP_CMP_GE: ret = (oldval >= cmparg); break;
-		case FUTEX_OP_CMP_LE: ret = (oldval <= cmparg); break;
-		case FUTEX_OP_CMP_GT: ret = (oldval > cmparg); break;
-		default: ret = -ENOSYS;
-		}
-	}
-	return ret;
-}
-
-static inline int
-futex_atomic_cmpxchg_inatomic(int __user *uaddr, int oldval, int newval)
-{
-	__asm__ __volatile__(
-	"\n1:	casa	[%3] %%asi, %2, %0\n"
-	"2:\n"
-	"	.section .fixup,#alloc,#execinstr\n"
-	"	.align	4\n"
-	"3:	sethi	%%hi(2b), %0\n"
-	"	jmpl	%0 + %%lo(2b), %%g0\n"
-	"	 mov	%4, %0\n"
-	"	.previous\n"
-	"	.section __ex_table,\"a\"\n"
-	"	.align	4\n"
-	"	.word	1b, 3b\n"
-	"	.previous\n"
-	: "=r" (newval)
-	: "0" (newval), "r" (oldval), "r" (uaddr), "i" (-EFAULT)
-	: "memory");
-
-	return newval;
-}
-
-#endif /* !(_SPARC64_FUTEX_H) */
+#include <asm-sparc/futex.h>
diff --git a/include/asm-sparc64/hardirq.h b/include/asm-sparc64/hardirq.h
index 7c29fd1a87a..63dca3db11f 100644
--- a/include/asm-sparc64/hardirq.h
+++ b/include/asm-sparc64/hardirq.h
@@ -1,19 +1 @@
-/* hardirq.h: 64-bit Sparc hard IRQ support.
- *
- * Copyright (C) 1997, 1998, 2005 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef __SPARC64_HARDIRQ_H
-#define __SPARC64_HARDIRQ_H
-
-#include <asm/cpudata.h>
-
-#define __ARCH_IRQ_STAT
-#define local_softirq_pending() \
-	(local_cpu_data().__softirq_pending)
-
-void ack_bad_irq(unsigned int irq);
-
-#define HARDIRQ_BITS	8
-
-#endif /* !(__SPARC64_HARDIRQ_H) */
+#include <asm-sparc/hardirq.h>
diff --git a/include/asm-sparc64/head.h b/include/asm-sparc64/head.h
index 10e9dabc4c4..2254c09e53f 100644
--- a/include/asm-sparc64/head.h
+++ b/include/asm-sparc64/head.h
@@ -1,76 +1 @@
-#ifndef _SPARC64_HEAD_H
-#define _SPARC64_HEAD_H
-
-#include <asm/pstate.h>
-
-	/* wrpr	%g0, val, %gl */
-#define SET_GL(val)	\
-	.word	0xa1902000 | val
-
-	/* rdpr %gl, %gN */
-#define GET_GL_GLOBAL(N)	\
-	.word	0x81540000 | (N << 25)
-
-#define KERNBASE	0x400000
-
-#define	PTREGS_OFF	(STACK_BIAS + STACKFRAME_SZ)
-
-#define __CHEETAH_ID	0x003e0014
-#define __JALAPENO_ID	0x003e0016
-#define __SERRANO_ID	0x003e0022
-
-#define CHEETAH_MANUF		0x003e
-#define CHEETAH_IMPL		0x0014 /* Ultra-III   */
-#define CHEETAH_PLUS_IMPL	0x0015 /* Ultra-III+  */
-#define JALAPENO_IMPL		0x0016 /* Ultra-IIIi  */
-#define JAGUAR_IMPL		0x0018 /* Ultra-IV    */
-#define PANTHER_IMPL		0x0019 /* Ultra-IV+   */
-#define SERRANO_IMPL		0x0022 /* Ultra-IIIi+ */
-
-#define BRANCH_IF_SUN4V(tmp1,label)		\
-	sethi	%hi(is_sun4v), %tmp1;		\
-	lduw	[%tmp1 + %lo(is_sun4v)], %tmp1; \
-	brnz,pn	%tmp1, label;			\
-	 nop
-
-#define BRANCH_IF_CHEETAH_BASE(tmp1,tmp2,label)	\
-	rdpr	%ver, %tmp1;			\
-	sethi	%hi(__CHEETAH_ID), %tmp2;	\
-	srlx	%tmp1, 32, %tmp1;		\
-	or	%tmp2, %lo(__CHEETAH_ID), %tmp2;\
-	cmp	%tmp1, %tmp2;			\
-	be,pn	%icc, label;			\
-	 nop;
-
-#define BRANCH_IF_JALAPENO(tmp1,tmp2,label)	\
-	rdpr	%ver, %tmp1;			\
-	sethi	%hi(__JALAPENO_ID), %tmp2;	\
-	srlx	%tmp1, 32, %tmp1;		\
-	or	%tmp2, %lo(__JALAPENO_ID), %tmp2;\
-	cmp	%tmp1, %tmp2;			\
-	be,pn	%icc, label;			\
-	 nop;
-
-#define BRANCH_IF_CHEETAH_PLUS_OR_FOLLOWON(tmp1,tmp2,label)	\
-	rdpr	%ver, %tmp1;			\
-	srlx	%tmp1, (32 + 16), %tmp2;	\
-	cmp	%tmp2, CHEETAH_MANUF;		\
-	bne,pt	%xcc, 99f;			\
-	 sllx	%tmp1, 16, %tmp1;		\
-	srlx	%tmp1, (32 + 16), %tmp2;	\
-	cmp	%tmp2, CHEETAH_PLUS_IMPL;	\
-	bgeu,pt	%xcc, label;			\
-99:	 nop;
-
-#define BRANCH_IF_ANY_CHEETAH(tmp1,tmp2,label)	\
-	rdpr	%ver, %tmp1;			\
-	srlx	%tmp1, (32 + 16), %tmp2;	\
-	cmp	%tmp2, CHEETAH_MANUF;		\
-	bne,pt	%xcc, 99f;			\
-	 sllx	%tmp1, 16, %tmp1;		\
-	srlx	%tmp1, (32 + 16), %tmp2;	\
-	cmp	%tmp2, CHEETAH_IMPL;		\
-	bgeu,pt	%xcc, label;			\
-99:	 nop;
-
-#endif /* !(_SPARC64_HEAD_H) */
+#include <asm-sparc/head.h>
diff --git a/include/asm-sparc64/hugetlb.h b/include/asm-sparc64/hugetlb.h
index 412af58926a..21d8f0a9c24 100644
--- a/include/asm-sparc64/hugetlb.h
+++ b/include/asm-sparc64/hugetlb.h
@@ -1,84 +1 @@
-#ifndef _ASM_SPARC64_HUGETLB_H
-#define _ASM_SPARC64_HUGETLB_H
-
-#include <asm/page.h>
-
-
-void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
-		     pte_t *ptep, pte_t pte);
-
-pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr,
-			      pte_t *ptep);
-
-void hugetlb_prefault_arch_hook(struct mm_struct *mm);
-
-static inline int is_hugepage_only_range(struct mm_struct *mm,
-					 unsigned long addr,
-					 unsigned long len) {
-	return 0;
-}
-
-/*
- * If the arch doesn't supply something else, assume that hugepage
- * size aligned regions are ok without further preparation.
- */
-static inline int prepare_hugepage_range(unsigned long addr, unsigned long len)
-{
-	if (len & ~HPAGE_MASK)
-		return -EINVAL;
-	if (addr & ~HPAGE_MASK)
-		return -EINVAL;
-	return 0;
-}
-
-static inline void hugetlb_free_pgd_range(struct mmu_gather **tlb,
-					  unsigned long addr, unsigned long end,
-					  unsigned long floor,
-					  unsigned long ceiling)
-{
-	free_pgd_range(tlb, addr, end, floor, ceiling);
-}
-
-static inline void huge_ptep_clear_flush(struct vm_area_struct *vma,
-					 unsigned long addr, pte_t *ptep)
-{
-}
-
-static inline int huge_pte_none(pte_t pte)
-{
-	return pte_none(pte);
-}
-
-static inline pte_t huge_pte_wrprotect(pte_t pte)
-{
-	return pte_wrprotect(pte);
-}
-
-static inline void huge_ptep_set_wrprotect(struct mm_struct *mm,
-					   unsigned long addr, pte_t *ptep)
-{
-	ptep_set_wrprotect(mm, addr, ptep);
-}
-
-static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma,
-					     unsigned long addr, pte_t *ptep,
-					     pte_t pte, int dirty)
-{
-	return ptep_set_access_flags(vma, addr, ptep, pte, dirty);
-}
-
-static inline pte_t huge_ptep_get(pte_t *ptep)
-{
-	return *ptep;
-}
-
-static inline int arch_prepare_hugepage(struct page *page)
-{
-	return 0;
-}
-
-static inline void arch_release_hugepage(struct page *page)
-{
-}
-
-#endif /* _ASM_SPARC64_HUGETLB_H */
+#include <asm-sparc/hugetlb.h>
diff --git a/include/asm-sparc64/hvtramp.h b/include/asm-sparc64/hvtramp.h
index b2b9b947b3a..fb46bfe934a 100644
--- a/include/asm-sparc64/hvtramp.h
+++ b/include/asm-sparc64/hvtramp.h
@@ -1,37 +1 @@
-#ifndef _SPARC64_HVTRAP_H
-#define _SPARC64_HVTRAP_H
-
-#ifndef __ASSEMBLY__
-
-#include <linux/types.h>
-
-struct hvtramp_mapping {
-	__u64		vaddr;
-	__u64		tte;
-};
-
-struct hvtramp_descr {
-	__u32			cpu;
-	__u32			num_mappings;
-	__u64			fault_info_va;
-	__u64			fault_info_pa;
-	__u64			thread_reg;
-	struct hvtramp_mapping	maps[1];
-};
-
-extern void hv_cpu_startup(unsigned long hvdescr_pa);
-
-#endif
-
-#define HVTRAMP_DESCR_CPU		0x00
-#define HVTRAMP_DESCR_NUM_MAPPINGS	0x04
-#define HVTRAMP_DESCR_FAULT_INFO_VA	0x08
-#define HVTRAMP_DESCR_FAULT_INFO_PA	0x10
-#define HVTRAMP_DESCR_THREAD_REG	0x18
-#define HVTRAMP_DESCR_MAPS		0x20
-
-#define HVTRAMP_MAPPING_VADDR		0x00
-#define HVTRAMP_MAPPING_TTE		0x08
-#define HVTRAMP_MAPPING_SIZE		0x10
-
-#endif /* _SPARC64_HVTRAP_H */
+#include <asm-sparc/hvtramp.h>
diff --git a/include/asm-sparc64/hw_irq.h b/include/asm-sparc64/hw_irq.h
index 8e44a836082..16920a291f5 100644
--- a/include/asm-sparc64/hw_irq.h
+++ b/include/asm-sparc64/hw_irq.h
@@ -1,4 +1 @@
-#ifndef __ASM_SPARC64_HW_IRQ_H
-#define __ASM_SPARC64_HW_IRQ_H
-
-#endif
+#include <asm-sparc/hw_irq.h>
diff --git a/include/asm-sparc64/hypervisor.h b/include/asm-sparc64/hypervisor.h
index 3ad45dff52f..fe7e51a9e42 100644
--- a/include/asm-sparc64/hypervisor.h
+++ b/include/asm-sparc64/hypervisor.h
@@ -1,2945 +1 @@
-#ifndef _SPARC64_HYPERVISOR_H
-#define _SPARC64_HYPERVISOR_H
-
-/* Sun4v hypervisor interfaces and defines.
- *
- * Hypervisor calls are made via traps to software traps number 0x80
- * and above.  Registers %o0 to %o5 serve as argument, status, and
- * return value registers.
- *
- * There are two kinds of these traps.  First there are the normal
- * "fast traps" which use software trap 0x80 and encode the function
- * to invoke by number in register %o5.  Argument and return value
- * handling is as follows:
- *
- * -----------------------------------------------
- * |  %o5  | function number |     undefined     |
- * |  %o0  |   argument 0    |   return status   |
- * |  %o1  |   argument 1    |   return value 1  |
- * |  %o2  |   argument 2    |   return value 2  |
- * |  %o3  |   argument 3    |   return value 3  |
- * |  %o4  |   argument 4    |   return value 4  |
- * -----------------------------------------------
- *
- * The second type are "hyper-fast traps" which encode the function
- * number in the software trap number itself.  So these use trap
- * numbers > 0x80.  The register usage for hyper-fast traps is as
- * follows:
- *
- * -----------------------------------------------
- * |  %o0  |   argument 0    |   return status   |
- * |  %o1  |   argument 1    |   return value 1  |
- * |  %o2  |   argument 2    |   return value 2  |
- * |  %o3  |   argument 3    |   return value 3  |
- * |  %o4  |   argument 4    |   return value 4  |
- * -----------------------------------------------
- *
- * Registers providing explicit arguments to the hypervisor calls
- * are volatile across the call.  Upon return their values are
- * undefined unless explicitly specified as containing a particular
- * return value by the specific call.  The return status is always
- * returned in register %o0, zero indicates a successful execution of
- * the hypervisor call and other values indicate an error status as
- * defined below.  So, for example, if a hyper-fast trap takes
- * arguments 0, 1, and 2, then %o0, %o1, and %o2 are volatile across
- * the call and %o3, %o4, and %o5 would be preserved.
- *
- * If the hypervisor trap is invalid, or the fast trap function number
- * is invalid, HV_EBADTRAP will be returned in %o0.  Also, all 64-bits
- * of the argument and return values are significant.
- */
-
-/* Trap numbers.  */
-#define HV_FAST_TRAP		0x80
-#define HV_MMU_MAP_ADDR_TRAP	0x83
-#define HV_MMU_UNMAP_ADDR_TRAP	0x84
-#define HV_TTRACE_ADDENTRY_TRAP	0x85
-#define HV_CORE_TRAP		0xff
-
-/* Error codes.  */
-#define HV_EOK				0  /* Successful return            */
-#define HV_ENOCPU			1  /* Invalid CPU id               */
-#define HV_ENORADDR			2  /* Invalid real address         */
-#define HV_ENOINTR			3  /* Invalid interrupt id         */
-#define HV_EBADPGSZ			4  /* Invalid pagesize encoding    */
-#define HV_EBADTSB			5  /* Invalid TSB description      */
-#define HV_EINVAL			6  /* Invalid argument             */
-#define HV_EBADTRAP			7  /* Invalid function number      */
-#define HV_EBADALIGN			8  /* Invalid address alignment    */
-#define HV_EWOULDBLOCK			9  /* Cannot complete w/o blocking */
-#define HV_ENOACCESS			10 /* No access to resource        */
-#define HV_EIO				11 /* I/O error                    */
-#define HV_ECPUERROR			12 /* CPU in error state           */
-#define HV_ENOTSUPPORTED		13 /* Function not supported       */
-#define HV_ENOMAP			14 /* No mapping found             */
-#define HV_ETOOMANY			15 /* Too many items specified     */
-#define HV_ECHANNEL			16 /* Invalid LDC channel          */
-#define HV_EBUSY			17 /* Resource busy                */
-
-/* mach_exit()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_EXIT
- * ARG0:	exit code
- * ERRORS:	This service does not return.
- *
- * Stop all CPUs in the virtual domain and place them into the stopped
- * state.  The 64-bit exit code may be passed to a service entity as
- * the domain's exit status.  On systems without a service entity, the
- * domain will undergo a reset, and the boot firmware will be
- * reloaded.
- *
- * This function will never return to the guest that invokes it.
- *
- * Note: By convention an exit code of zero denotes a successful exit by
- *       the guest code.  A non-zero exit code denotes a guest specific
- *       error indication.
- *
- */
-#define HV_FAST_MACH_EXIT		0x00
-
-#ifndef __ASSEMBLY__
-extern void sun4v_mach_exit(unsigned long exit_code);
-#endif
-
-/* Domain services.  */
-
-/* mach_desc()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_DESC
- * ARG0:	buffer
- * ARG1:	length
- * RET0:	status
- * RET1:	length
- * ERRORS:	HV_EBADALIGN	Buffer is badly aligned
- *		HV_ENORADDR	Buffer is to an illegal real address.
- *		HV_EINVAL	Buffer length is too small for complete
- *				machine description.
- *
- * Copy the most current machine description into the buffer indicated
- * by the real address in ARG0.  The buffer provided must be 16 byte
- * aligned.  Upon success or HV_EINVAL, this service returns the
- * actual size of the machine description in the RET1 return value.
- *
- * Note: A method of determining the appropriate buffer size for the
- *       machine description is to first call this service with a buffer
- *       length of 0 bytes.
- */
-#define HV_FAST_MACH_DESC		0x01
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mach_desc(unsigned long buffer_pa,
-				     unsigned long buf_len,
-				     unsigned long *real_buf_len);
-#endif
-
-/* mach_sir()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_SIR
- * ERRORS:	This service does not return.
- *
- * Perform a software initiated reset of the virtual machine domain.
- * All CPUs are captured as soon as possible, all hardware devices are
- * returned to the entry default state, and the domain is restarted at
- * the SIR (trap type 0x04) real trap table (RTBA) entry point on one
- * of the CPUs.  The single CPU restarted is selected as determined by
- * platform specific policy.  Memory is preserved across this
- * operation.
- */
-#define HV_FAST_MACH_SIR		0x02
-
-#ifndef __ASSEMBLY__
-extern void sun4v_mach_sir(void);
-#endif
-
-/* mach_set_watchdog()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_SET_WATCHDOG
- * ARG0:	timeout in milliseconds
- * RET0:	status
- * RET1:	time remaining in milliseconds
- *
- * A guest uses this API to set a watchdog timer.  Once the gues has set
- * the timer, it must call the timer service again either to disable or
- * postpone the expiration.  If the timer expires before being reset or
- * disabled, then the hypervisor take a platform specific action leading
- * to guest termination within a bounded time period.  The platform action
- * may include recovery actions such as reporting the expiration to a
- * Service Processor, and/or automatically restarting the gues.
- *
- * The 'timeout' parameter is specified in milliseconds, however the
- * implementated granularity is given by the 'watchdog-resolution'
- * property in the 'platform' node of the guest's machine description.
- * The largest allowed timeout value is specified by the
- * 'watchdog-max-timeout' property of the 'platform' node.
- *
- * If the 'timeout' argument is not zero, the watchdog timer is set to
- * expire after a minimum of 'timeout' milliseconds.
- *
- * If the 'timeout' argument is zero, the watchdog timer is disabled.
- *
- * If the 'timeout' value exceeds the value of the 'max-watchdog-timeout'
- * property, the hypervisor leaves the watchdog timer state unchanged,
- * and returns a status of EINVAL.
- *
- * The 'time remaining' return value is valid regardless of whether the
- * return status is EOK or EINVAL.  A non-zero return value indicates the
- * number of milliseconds that were remaining until the timer was to expire.
- * If less than one millisecond remains, the return value is '1'.  If the
- * watchdog timer was disabled at the time of the call, the return value is
- * zero.
- *
- * If the hypervisor cannot support the exact timeout value requested, but
- * can support a larger timeout value, the hypervisor may round the actual
- * timeout to a value larger than the requested timeout, consequently the
- * 'time remaining' return value may be larger than the previously requested
- * timeout value.
- *
- * Any guest OS debugger should be aware that the watchdog service may be in
- * use.  Consequently, it is recommended that the watchdog service is
- * disabled upon debugger entry (e.g. reaching a breakpoint), and then
- * re-enabled upon returning to normal execution.  The API has been designed
- * with this in mind, and the 'time remaining' result of the disable call may
- * be used directly as the timeout argument of the re-enable call.
- */
-#define HV_FAST_MACH_SET_WATCHDOG	0x05
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mach_set_watchdog(unsigned long timeout,
-					     unsigned long *orig_timeout);
-#endif
-
-/* CPU services.
- *
- * CPUs represent devices that can execute software threads.  A single
- * chip that contains multiple cores or strands is represented as
- * multiple CPUs with unique CPU identifiers.  CPUs are exported to
- * OBP via the machine description (and to the OS via the OBP device
- * tree).  CPUs are always in one of three states: stopped, running,
- * or error.
- *
- * A CPU ID is a pre-assigned 16-bit value that uniquely identifies a
- * CPU within a logical domain.  Operations that are to be performed
- * on multiple CPUs specify them via a CPU list.  A CPU list is an
- * array in real memory, of which each 16-bit word is a CPU ID.  CPU
- * lists are passed through the API as two arguments.  The first is
- * the number of entries (16-bit words) in the CPU list, and the
- * second is the (real address) pointer to the CPU ID list.
- */
-
-/* cpu_start()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_START
- * ARG0:	CPU ID
- * ARG1:	PC
- * ARG2:	RTBA
- * ARG3:	target ARG0
- * RET0:	status
- * ERRORS:	ENOCPU		Invalid CPU ID
- *		EINVAL		Target CPU ID is not in the stopped state
- *		ENORADDR	Invalid PC or RTBA real address
- *		EBADALIGN	Unaligned PC or unaligned RTBA
- *		EWOULDBLOCK	Starting resources are not available
- *
- * Start CPU with given CPU ID with PC in %pc and with a real trap
- * base address value of RTBA.  The indicated CPU must be in the
- * stopped state.  The supplied RTBA must be aligned on a 256 byte
- * boundary.  On successful completion, the specified CPU will be in
- * the running state and will be supplied with "target ARG0" in %o0
- * and RTBA in %tba.
- */
-#define HV_FAST_CPU_START		0x10
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_cpu_start(unsigned long cpuid,
-				     unsigned long pc,
-				     unsigned long rtba,
-				     unsigned long arg0);
-#endif
-
-/* cpu_stop()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_STOP
- * ARG0:	CPU ID
- * RET0:	status
- * ERRORS:	ENOCPU		Invalid CPU ID
- *		EINVAL		Target CPU ID is the current cpu
- *		EINVAL		Target CPU ID is not in the running state
- *		EWOULDBLOCK	Stopping resources are not available
- *		ENOTSUPPORTED	Not supported on this platform
- *
- * The specified CPU is stopped.  The indicated CPU must be in the
- * running state.  On completion, it will be in the stopped state.  It
- * is not legal to stop the current CPU.
- *
- * Note: As this service cannot be used to stop the current cpu, this service
- *       may not be used to stop the last running CPU in a domain.  To stop
- *       and exit a running domain, a guest must use the mach_exit() service.
- */
-#define HV_FAST_CPU_STOP		0x11
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_cpu_stop(unsigned long cpuid);
-#endif
-
-/* cpu_yield()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_YIELD
- * RET0:	status
- * ERRORS:	No possible error.
- *
- * Suspend execution on the current CPU.  Execution will resume when
- * an interrupt (device, %stick_compare, or cross-call) is targeted to
- * the CPU.  On some CPUs, this API may be used by the hypervisor to
- * save power by disabling hardware strands.
- */
-#define HV_FAST_CPU_YIELD		0x12
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_cpu_yield(void);
-#endif
-
-/* cpu_qconf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_QCONF
- * ARG0:	queue
- * ARG1:	base real address
- * ARG2:	number of entries
- * RET0:	status
- * ERRORS:	ENORADDR	Invalid base real address
- *		EINVAL		Invalid queue or number of entries is less
- *				than 2 or too large.
- *		EBADALIGN	Base real address is not correctly aligned
- *				for size.
- *
- * Configure the given queue to be placed at the given base real
- * address, with the given number of entries.  The number of entries
- * must be a power of 2.  The base real address must be aligned
- * exactly to match the queue size.  Each queue entry is 64 bytes
- * long, so for example a 32 entry queue must be aligned on a 2048
- * byte real address boundary.
- *
- * The specified queue is unconfigured if the number of entries is given
- * as zero.
- *
- * For the current version of this API service, the argument queue is defined
- * as follows:
- *
- *	queue		description
- *	-----		-------------------------
- *	0x3c		cpu mondo queue
- *	0x3d		device mondo queue
- *	0x3e		resumable error queue
- *	0x3f		non-resumable error queue
- *
- * Note: The maximum number of entries for each queue for a specific cpu may
- *       be determined from the machine description.
- */
-#define HV_FAST_CPU_QCONF		0x14
-#define  HV_CPU_QUEUE_CPU_MONDO		 0x3c
-#define  HV_CPU_QUEUE_DEVICE_MONDO	 0x3d
-#define  HV_CPU_QUEUE_RES_ERROR		 0x3e
-#define  HV_CPU_QUEUE_NONRES_ERROR	 0x3f
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_cpu_qconf(unsigned long type,
-				     unsigned long queue_paddr,
-				     unsigned long num_queue_entries);
-#endif
-
-/* cpu_qinfo()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_QINFO
- * ARG0:	queue
- * RET0:	status
- * RET1:	base real address
- * RET1:	number of entries
- * ERRORS:	EINVAL		Invalid queue
- *
- * Return the configuration info for the given queue.  The base real
- * address and number of entries of the defined queue are returned.
- * The queue argument values are the same as for cpu_qconf() above.
- *
- * If the specified queue is a valid queue number, but no queue has
- * been defined, the number of entries will be set to zero and the
- * base real address returned is undefined.
- */
-#define HV_FAST_CPU_QINFO		0x15
-
-/* cpu_mondo_send()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_MONDO_SEND
- * ARG0-1:	CPU list
- * ARG2:	data real address
- * RET0:	status
- * ERRORS:	EBADALIGN	Mondo data is not 64-byte aligned or CPU list
- *				is not 2-byte aligned.
- *		ENORADDR	Invalid data mondo address, or invalid cpu list
- *				address.
- *		ENOCPU		Invalid cpu in CPU list
- *		EWOULDBLOCK	Some or all of the listed CPUs did not receive
- *				the mondo
- *		ECPUERROR	One or more of the listed CPUs are in error
- *				state, use HV_FAST_CPU_STATE to see which ones
- *		EINVAL		CPU list includes caller's CPU ID
- *
- * Send a mondo interrupt to the CPUs in the given CPU list with the
- * 64-bytes at the given data real address.  The data must be 64-byte
- * aligned.  The mondo data will be delivered to the cpu_mondo queues
- * of the recipient CPUs.
- *
- * In all cases, error or not, the CPUs in the CPU list to which the
- * mondo has been successfully delivered will be indicated by having
- * their entry in CPU list updated with the value 0xffff.
- */
-#define HV_FAST_CPU_MONDO_SEND		0x42
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_cpu_mondo_send(unsigned long cpu_count, unsigned long cpu_list_pa, unsigned long mondo_block_pa);
-#endif
-
-/* cpu_myid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_MYID
- * RET0:	status
- * RET1:	CPU ID
- * ERRORS:	No errors defined.
- *
- * Return the hypervisor ID handle for the current CPU.  Use by a
- * virtual CPU to discover it's own identity.
- */
-#define HV_FAST_CPU_MYID		0x16
-
-/* cpu_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_STATE
- * ARG0:	CPU ID
- * RET0:	status
- * RET1:	state
- * ERRORS:	ENOCPU		Invalid CPU ID
- *
- * Retrieve the current state of the CPU with the given CPU ID.
- */
-#define HV_FAST_CPU_STATE		0x17
-#define  HV_CPU_STATE_STOPPED		 0x01
-#define  HV_CPU_STATE_RUNNING		 0x02
-#define  HV_CPU_STATE_ERROR		 0x03
-
-#ifndef __ASSEMBLY__
-extern long sun4v_cpu_state(unsigned long cpuid);
-#endif
-
-/* cpu_set_rtba()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_SET_RTBA
- * ARG0:	RTBA
- * RET0:	status
- * RET1:	previous RTBA
- * ERRORS:	ENORADDR	Invalid RTBA real address
- *		EBADALIGN	RTBA is incorrectly aligned for a trap table
- *
- * Set the real trap base address of the local cpu to the given RTBA.
- * The supplied RTBA must be aligned on a 256 byte boundary.  Upon
- * success the previous value of the RTBA is returned in RET1.
- *
- * Note: This service does not affect %tba
- */
-#define HV_FAST_CPU_SET_RTBA		0x18
-
-/* cpu_set_rtba()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CPU_GET_RTBA
- * RET0:	status
- * RET1:	previous RTBA
- * ERRORS:	No possible error.
- *
- * Returns the current value of RTBA in RET1.
- */
-#define HV_FAST_CPU_GET_RTBA		0x19
-
-/* MMU services.
- *
- * Layout of a TSB description for mmu_tsb_ctx{,non}0() calls.
- */
-#ifndef __ASSEMBLY__
-struct hv_tsb_descr {
-	unsigned short		pgsz_idx;
-	unsigned short		assoc;
-	unsigned int		num_ttes;	/* in TTEs */
-	unsigned int		ctx_idx;
-	unsigned int		pgsz_mask;
-	unsigned long		tsb_base;
-	unsigned long		resv;
-};
-#endif
-#define HV_TSB_DESCR_PGSZ_IDX_OFFSET	0x00
-#define HV_TSB_DESCR_ASSOC_OFFSET	0x02
-#define HV_TSB_DESCR_NUM_TTES_OFFSET	0x04
-#define HV_TSB_DESCR_CTX_IDX_OFFSET	0x08
-#define HV_TSB_DESCR_PGSZ_MASK_OFFSET	0x0c
-#define HV_TSB_DESCR_TSB_BASE_OFFSET	0x10
-#define HV_TSB_DESCR_RESV_OFFSET	0x18
-
-/* Page size bitmask.  */
-#define HV_PGSZ_MASK_8K			(1 << 0)
-#define HV_PGSZ_MASK_64K		(1 << 1)
-#define HV_PGSZ_MASK_512K		(1 << 2)
-#define HV_PGSZ_MASK_4MB		(1 << 3)
-#define HV_PGSZ_MASK_32MB		(1 << 4)
-#define HV_PGSZ_MASK_256MB		(1 << 5)
-#define HV_PGSZ_MASK_2GB		(1 << 6)
-#define HV_PGSZ_MASK_16GB		(1 << 7)
-
-/* Page size index.  The value given in the TSB descriptor must correspond
- * to the smallest page size specified in the pgsz_mask page size bitmask.
- */
-#define HV_PGSZ_IDX_8K			0
-#define HV_PGSZ_IDX_64K			1
-#define HV_PGSZ_IDX_512K		2
-#define HV_PGSZ_IDX_4MB			3
-#define HV_PGSZ_IDX_32MB		4
-#define HV_PGSZ_IDX_256MB		5
-#define HV_PGSZ_IDX_2GB			6
-#define HV_PGSZ_IDX_16GB		7
-
-/* MMU fault status area.
- *
- * MMU related faults have their status and fault address information
- * placed into a memory region made available by privileged code.  Each
- * virtual processor must make a mmu_fault_area_conf() call to tell the
- * hypervisor where that processor's fault status should be stored.
- *
- * The fault status block is a multiple of 64-bytes and must be aligned
- * on a 64-byte boundary.
- */
-#ifndef __ASSEMBLY__
-struct hv_fault_status {
-	unsigned long		i_fault_type;
-	unsigned long		i_fault_addr;
-	unsigned long		i_fault_ctx;
-	unsigned long		i_reserved[5];
-	unsigned long		d_fault_type;
-	unsigned long		d_fault_addr;
-	unsigned long		d_fault_ctx;
-	unsigned long		d_reserved[5];
-};
-#endif
-#define HV_FAULT_I_TYPE_OFFSET	0x00
-#define HV_FAULT_I_ADDR_OFFSET	0x08
-#define HV_FAULT_I_CTX_OFFSET	0x10
-#define HV_FAULT_D_TYPE_OFFSET	0x40
-#define HV_FAULT_D_ADDR_OFFSET	0x48
-#define HV_FAULT_D_CTX_OFFSET	0x50
-
-#define HV_FAULT_TYPE_FAST_MISS	1
-#define HV_FAULT_TYPE_FAST_PROT	2
-#define HV_FAULT_TYPE_MMU_MISS	3
-#define HV_FAULT_TYPE_INV_RA	4
-#define HV_FAULT_TYPE_PRIV_VIOL	5
-#define HV_FAULT_TYPE_PROT_VIOL	6
-#define HV_FAULT_TYPE_NFO	7
-#define HV_FAULT_TYPE_NFO_SEFF	8
-#define HV_FAULT_TYPE_INV_VA	9
-#define HV_FAULT_TYPE_INV_ASI	10
-#define HV_FAULT_TYPE_NC_ATOMIC	11
-#define HV_FAULT_TYPE_PRIV_ACT	12
-#define HV_FAULT_TYPE_RESV1	13
-#define HV_FAULT_TYPE_UNALIGNED	14
-#define HV_FAULT_TYPE_INV_PGSZ	15
-/* Values 16 --> -2 are reserved.  */
-#define HV_FAULT_TYPE_MULTIPLE	-1
-
-/* Flags argument for mmu_{map,unmap}_addr(), mmu_demap_{page,context,all}(),
- * and mmu_{map,unmap}_perm_addr().
- */
-#define HV_MMU_DMMU			0x01
-#define HV_MMU_IMMU			0x02
-#define HV_MMU_ALL			(HV_MMU_DMMU | HV_MMU_IMMU)
-
-/* mmu_map_addr()
- * TRAP:	HV_MMU_MAP_ADDR_TRAP
- * ARG0:	virtual address
- * ARG1:	mmu context
- * ARG2:	TTE
- * ARG3:	flags (HV_MMU_{IMMU,DMMU})
- * ERRORS:	EINVAL		Invalid virtual address, mmu context, or flags
- *		EBADPGSZ	Invalid page size value
- *		ENORADDR	Invalid real address in TTE
- *
- * Create a non-permanent mapping using the given TTE, virtual
- * address, and mmu context.  The flags argument determines which
- * (data, or instruction, or both) TLB the mapping gets loaded into.
- *
- * The behavior is undefined if the valid bit is clear in the TTE.
- *
- * Note: This API call is for privileged code to specify temporary translation
- *       mappings without the need to create and manage a TSB.
- */
-
-/* mmu_unmap_addr()
- * TRAP:	HV_MMU_UNMAP_ADDR_TRAP
- * ARG0:	virtual address
- * ARG1:	mmu context
- * ARG2:	flags (HV_MMU_{IMMU,DMMU})
- * ERRORS:	EINVAL		Invalid virtual address, mmu context, or flags
- *
- * Demaps the given virtual address in the given mmu context on this
- * CPU.  This function is intended to be used to demap pages mapped
- * with mmu_map_addr.  This service is equivalent to invoking
- * mmu_demap_page() with only the current CPU in the CPU list. The
- * flags argument determines which (data, or instruction, or both) TLB
- * the mapping gets unmapped from.
- *
- * Attempting to perform an unmap operation for a previously defined
- * permanent mapping will have undefined results.
- */
-
-/* mmu_tsb_ctx0()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_TSB_CTX0
- * ARG0:	number of TSB descriptions
- * ARG1:	TSB descriptions pointer
- * RET0:	status
- * ERRORS:	ENORADDR		Invalid TSB descriptions pointer or
- *					TSB base within a descriptor
- *		EBADALIGN		TSB descriptions pointer is not aligned
- *					to an 8-byte boundary, or TSB base
- *					within a descriptor is not aligned for
- *					the given TSB size
- *		EBADPGSZ		Invalid page size in a TSB descriptor
- *		EBADTSB			Invalid associativity or size in a TSB
- *					descriptor
- *		EINVAL			Invalid number of TSB descriptions, or
- *					invalid context index in a TSB
- *					descriptor, or index page size not
- *					equal to smallest page size in page
- *					size bitmask field.
- *
- * Configures the TSBs for the current CPU for virtual addresses with
- * context zero.  The TSB descriptions pointer is a pointer to an
- * array of the given number of TSB descriptions.
- *
- * Note: The maximum number of TSBs available to a virtual CPU is given by the
- *       mmu-max-#tsbs property of the cpu's corresponding "cpu" node in the
- *       machine description.
- */
-#define HV_FAST_MMU_TSB_CTX0		0x20
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mmu_tsb_ctx0(unsigned long num_descriptions,
-					unsigned long tsb_desc_ra);
-#endif
-
-/* mmu_tsb_ctxnon0()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_TSB_CTXNON0
- * ARG0:	number of TSB descriptions
- * ARG1:	TSB descriptions pointer
- * RET0:	status
- * ERRORS:	Same as for mmu_tsb_ctx0() above.
- *
- * Configures the TSBs for the current CPU for virtual addresses with
- * non-zero contexts.  The TSB descriptions pointer is a pointer to an
- * array of the given number of TSB descriptions.
- *
- * Note: A maximum of 16 TSBs may be specified in the TSB description list.
- */
-#define HV_FAST_MMU_TSB_CTXNON0		0x21
-
-/* mmu_demap_page()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_DEMAP_PAGE
- * ARG0:	reserved, must be zero
- * ARG1:	reserved, must be zero
- * ARG2:	virtual address
- * ARG3:	mmu context
- * ARG4:	flags (HV_MMU_{IMMU,DMMU})
- * RET0:	status
- * ERRORS:	EINVAL			Invalid virutal address, context, or
- *					flags value
- *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
- *
- * Demaps any page mapping of the given virtual address in the given
- * mmu context for the current virtual CPU.  Any virtually tagged
- * caches are guaranteed to be kept consistent.  The flags argument
- * determines which TLB (instruction, or data, or both) participate in
- * the operation.
- *
- * ARG0 and ARG1 are both reserved and must be set to zero.
- */
-#define HV_FAST_MMU_DEMAP_PAGE		0x22
-
-/* mmu_demap_ctx()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_DEMAP_CTX
- * ARG0:	reserved, must be zero
- * ARG1:	reserved, must be zero
- * ARG2:	mmu context
- * ARG3:	flags (HV_MMU_{IMMU,DMMU})
- * RET0:	status
- * ERRORS:	EINVAL			Invalid context or flags value
- *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
- *
- * Demaps all non-permanent virtual page mappings previously specified
- * for the given context for the current virtual CPU.  Any virtual
- * tagged caches are guaranteed to be kept consistent.  The flags
- * argument determines which TLB (instruction, or data, or both)
- * participate in the operation.
- *
- * ARG0 and ARG1 are both reserved and must be set to zero.
- */
-#define HV_FAST_MMU_DEMAP_CTX		0x23
-
-/* mmu_demap_all()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_DEMAP_ALL
- * ARG0:	reserved, must be zero
- * ARG1:	reserved, must be zero
- * ARG2:	flags (HV_MMU_{IMMU,DMMU})
- * RET0:	status
- * ERRORS:	EINVAL			Invalid flags value
- *		ENOTSUPPORTED		ARG0 or ARG1 is non-zero
- *
- * Demaps all non-permanent virtual page mappings previously specified
- * for the current virtual CPU.  Any virtual tagged caches are
- * guaranteed to be kept consistent.  The flags argument determines
- * which TLB (instruction, or data, or both) participate in the
- * operation.
- *
- * ARG0 and ARG1 are both reserved and must be set to zero.
- */
-#define HV_FAST_MMU_DEMAP_ALL		0x24
-
-#ifndef __ASSEMBLY__
-extern void sun4v_mmu_demap_all(void);
-#endif
-
-/* mmu_map_perm_addr()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_MAP_PERM_ADDR
- * ARG0:	virtual address
- * ARG1:	reserved, must be zero
- * ARG2:	TTE
- * ARG3:	flags (HV_MMU_{IMMU,DMMU})
- * RET0:	status
- * ERRORS:	EINVAL			Invalid virutal address or flags value
- *		EBADPGSZ		Invalid page size value
- *		ENORADDR		Invalid real address in TTE
- *		ETOOMANY		Too many mappings (max of 8 reached)
- *
- * Create a permanent mapping using the given TTE and virtual address
- * for context 0 on the calling virtual CPU.  A maximum of 8 such
- * permanent mappings may be specified by privileged code.  Mappings
- * may be removed with mmu_unmap_perm_addr().
- *
- * The behavior is undefined if a TTE with the valid bit clear is given.
- *
- * Note: This call is used to specify address space mappings for which
- *       privileged code does not expect to receive misses.  For example,
- *       this mechanism can be used to map kernel nucleus code and data.
- */
-#define HV_FAST_MMU_MAP_PERM_ADDR	0x25
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mmu_map_perm_addr(unsigned long vaddr,
-					     unsigned long set_to_zero,
-					     unsigned long tte,
-					     unsigned long flags);
-#endif
-
-/* mmu_fault_area_conf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_FAULT_AREA_CONF
- * ARG0:	real address
- * RET0:	status
- * RET1:	previous mmu fault area real address
- * ERRORS:	ENORADDR		Invalid real address
- *		EBADALIGN		Invalid alignment for fault area
- *
- * Configure the MMU fault status area for the calling CPU.  A 64-byte
- * aligned real address specifies where MMU fault status information
- * is placed.  The return value is the previously specified area, or 0
- * for the first invocation.  Specifying a fault area at real address
- * 0 is not allowed.
- */
-#define HV_FAST_MMU_FAULT_AREA_CONF	0x26
-
-/* mmu_enable()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_ENABLE
- * ARG0:	enable flag
- * ARG1:	return target address
- * RET0:	status
- * ERRORS:	ENORADDR		Invalid real address when disabling
- *					translation.
- *		EBADALIGN		The return target address is not
- *					aligned to an instruction.
- *		EINVAL			The enable flag request the current
- *					operating mode (e.g. disable if already
- *					disabled)
- *
- * Enable or disable virtual address translation for the calling CPU
- * within the virtual machine domain.  If the enable flag is zero,
- * translation is disabled, any non-zero value will enable
- * translation.
- *
- * When this function returns, the newly selected translation mode
- * will be active.  If the mmu is being enabled, then the return
- * target address is a virtual address else it is a real address.
- *
- * Upon successful completion, control will be returned to the given
- * return target address (ie. the cpu will jump to that address).  On
- * failure, the previous mmu mode remains and the trap simply returns
- * as normal with the appropriate error code in RET0.
- */
-#define HV_FAST_MMU_ENABLE		0x27
-
-/* mmu_unmap_perm_addr()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_UNMAP_PERM_ADDR
- * ARG0:	virtual address
- * ARG1:	reserved, must be zero
- * ARG2:	flags (HV_MMU_{IMMU,DMMU})
- * RET0:	status
- * ERRORS:	EINVAL			Invalid virutal address or flags value
- *		ENOMAP			Specified mapping was not found
- *
- * Demaps any permanent page mapping (established via
- * mmu_map_perm_addr()) at the given virtual address for context 0 on
- * the current virtual CPU.  Any virtual tagged caches are guaranteed
- * to be kept consistent.
- */
-#define HV_FAST_MMU_UNMAP_PERM_ADDR	0x28
-
-/* mmu_tsb_ctx0_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_TSB_CTX0_INFO
- * ARG0:	max TSBs
- * ARG1:	buffer pointer
- * RET0:	status
- * RET1:	number of TSBs
- * ERRORS:	EINVAL			Supplied buffer is too small
- *		EBADALIGN		The buffer pointer is badly aligned
- *		ENORADDR		Invalid real address for buffer pointer
- *
- * Return the TSB configuration as previous defined by mmu_tsb_ctx0()
- * into the provided buffer.  The size of the buffer is given in ARG1
- * in terms of the number of TSB description entries.
- *
- * Upon return, RET1 always contains the number of TSB descriptions
- * previously configured.  If zero TSBs were configured, EOK is
- * returned with RET1 containing 0.
- */
-#define HV_FAST_MMU_TSB_CTX0_INFO	0x29
-
-/* mmu_tsb_ctxnon0_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_TSB_CTXNON0_INFO
- * ARG0:	max TSBs
- * ARG1:	buffer pointer
- * RET0:	status
- * RET1:	number of TSBs
- * ERRORS:	EINVAL			Supplied buffer is too small
- *		EBADALIGN		The buffer pointer is badly aligned
- *		ENORADDR		Invalid real address for buffer pointer
- *
- * Return the TSB configuration as previous defined by
- * mmu_tsb_ctxnon0() into the provided buffer.  The size of the buffer
- * is given in ARG1 in terms of the number of TSB description entries.
- *
- * Upon return, RET1 always contains the number of TSB descriptions
- * previously configured.  If zero TSBs were configured, EOK is
- * returned with RET1 containing 0.
- */
-#define HV_FAST_MMU_TSB_CTXNON0_INFO	0x2a
-
-/* mmu_fault_area_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMU_FAULT_AREA_INFO
- * RET0:	status
- * RET1:	fault area real address
- * ERRORS:	No errors defined.
- *
- * Return the currently defined MMU fault status area for the current
- * CPU.  The real address of the fault status area is returned in
- * RET1, or 0 is returned in RET1 if no fault status area is defined.
- *
- * Note: mmu_fault_area_conf() may be called with the return value (RET1)
- *       from this service if there is a need to save and restore the fault
- *	 area for a cpu.
- */
-#define HV_FAST_MMU_FAULT_AREA_INFO	0x2b
-
-/* Cache and Memory services. */
-
-/* mem_scrub()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MEM_SCRUB
- * ARG0:	real address
- * ARG1:	length
- * RET0:	status
- * RET1:	length scrubbed
- * ERRORS:	ENORADDR	Invalid real address
- *		EBADALIGN	Start address or length are not correctly
- *				aligned
- *		EINVAL		Length is zero
- *
- * Zero the memory contents in the range real address to real address
- * plus length minus 1.  Also, valid ECC will be generated for that
- * memory address range.  Scrubbing is started at the given real
- * address, but may not scrub the entire given length.  The actual
- * length scrubbed will be returned in RET1.
- *
- * The real address and length must be aligned on an 8K boundary, or
- * contain the start address and length from a sun4v error report.
- *
- * Note: There are two uses for this function.  The first use is to block clear
- *       and initialize memory and the second is to scrub an u ncorrectable
- *       error reported via a resumable or non-resumable trap.  The second
- *       use requires the arguments to be equal to the real address and length
- *       provided in a sun4v memory error report.
- */
-#define HV_FAST_MEM_SCRUB		0x31
-
-/* mem_sync()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MEM_SYNC
- * ARG0:	real address
- * ARG1:	length
- * RET0:	status
- * RET1:	length synced
- * ERRORS:	ENORADDR	Invalid real address
- *		EBADALIGN	Start address or length are not correctly
- *				aligned
- *		EINVAL		Length is zero
- *
- * Force the next access within the real address to real address plus
- * length minus 1 to be fetches from main system memory.  Less than
- * the given length may be synced, the actual amount synced is
- * returned in RET1.  The real address and length must be aligned on
- * an 8K boundary.
- */
-#define HV_FAST_MEM_SYNC		0x32
-
-/* Time of day services.
- *
- * The hypervisor maintains the time of day on a per-domain basis.
- * Changing the time of day in one domain does not affect the time of
- * day on any other domain.
- *
- * Time is described by a single unsigned 64-bit word which is the
- * number of seconds since the UNIX Epoch (00:00:00 UTC, January 1,
- * 1970).
- */
-
-/* tod_get()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TOD_GET
- * RET0:	status
- * RET1:	TOD
- * ERRORS:	EWOULDBLOCK	TOD resource is temporarily unavailable
- *		ENOTSUPPORTED	If TOD not supported on this platform
- *
- * Return the current time of day.  May block if TOD access is
- * temporarily not possible.
- */
-#define HV_FAST_TOD_GET			0x50
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_tod_get(unsigned long *time);
-#endif
-
-/* tod_set()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TOD_SET
- * ARG0:	TOD
- * RET0:	status
- * ERRORS:	EWOULDBLOCK	TOD resource is temporarily unavailable
- *		ENOTSUPPORTED	If TOD not supported on this platform
- *
- * The current time of day is set to the value specified in ARG0.  May
- * block if TOD access is temporarily not possible.
- */
-#define HV_FAST_TOD_SET			0x51
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_tod_set(unsigned long time);
-#endif
-
-/* Console services */
-
-/* con_getchar()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CONS_GETCHAR
- * RET0:	status
- * RET1:	character
- * ERRORS:	EWOULDBLOCK	No character available.
- *
- * Returns a character from the console device.  If no character is
- * available then an EWOULDBLOCK error is returned.  If a character is
- * available, then the returned status is EOK and the character value
- * is in RET1.
- *
- * A virtual BREAK is represented by the 64-bit value -1.
- *
- * A virtual HUP signal is represented by the 64-bit value -2.
- */
-#define HV_FAST_CONS_GETCHAR		0x60
-
-/* con_putchar()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CONS_PUTCHAR
- * ARG0:	character
- * RET0:	status
- * ERRORS:	EINVAL		Illegal character
- *		EWOULDBLOCK	Output buffer currently full, would block
- *
- * Send a character to the console device.  Only character values
- * between 0 and 255 may be used.  Values outside this range are
- * invalid except for the 64-bit value -1 which is used to send a
- * virtual BREAK.
- */
-#define HV_FAST_CONS_PUTCHAR		0x61
-
-/* con_read()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CONS_READ
- * ARG0:	buffer real address
- * ARG1:	buffer size in bytes
- * RET0:	status
- * RET1:	bytes read or BREAK or HUP
- * ERRORS:	EWOULDBLOCK	No character available.
- *
- * Reads characters into a buffer from the console device.  If no
- * character is available then an EWOULDBLOCK error is returned.
- * If a character is available, then the returned status is EOK
- * and the number of bytes read into the given buffer is provided
- * in RET1.
- *
- * A virtual BREAK is represented by the 64-bit RET1 value -1.
- *
- * A virtual HUP signal is represented by the 64-bit RET1 value -2.
- *
- * If BREAK or HUP are indicated, no bytes were read into buffer.
- */
-#define HV_FAST_CONS_READ		0x62
-
-/* con_write()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_CONS_WRITE
- * ARG0:	buffer real address
- * ARG1:	buffer size in bytes
- * RET0:	status
- * RET1:	bytes written
- * ERRORS:	EWOULDBLOCK	Output buffer currently full, would block
- *
- * Send a characters in buffer to the console device.  Breaks must be
- * sent using con_putchar().
- */
-#define HV_FAST_CONS_WRITE		0x63
-
-#ifndef __ASSEMBLY__
-extern long sun4v_con_getchar(long *status);
-extern long sun4v_con_putchar(long c);
-extern long sun4v_con_read(unsigned long buffer,
-			   unsigned long size,
-			   unsigned long *bytes_read);
-extern unsigned long sun4v_con_write(unsigned long buffer,
-				     unsigned long size,
-				     unsigned long *bytes_written);
-#endif
-
-/* mach_set_soft_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_SET_SOFT_STATE
- * ARG0:	software state
- * ARG1:	software state description pointer
- * RET0:	status
- * ERRORS:	EINVAL		software state not valid or software state
- *				description is not NULL terminated
- *		ENORADDR	software state description pointer is not a
- *				valid real address
- *		EBADALIGNED	software state description is not correctly
- *				aligned
- *
- * This allows the guest to report it's soft state to the hypervisor.  There
- * are two primary components to this state.  The first part states whether
- * the guest software is running or not.  The second containts optional
- * details specific to the software.
- *
- * The software state argument is defined below in HV_SOFT_STATE_*, and
- * indicates whether the guest is operating normally or in a transitional
- * state.
- *
- * The software state description argument is a real address of a data buffer
- * of size 32-bytes aligned on a 32-byte boundary.  It is treated as a NULL
- * terminated 7-bit ASCII string of up to 31 characters not including the
- * NULL termination.
- */
-#define HV_FAST_MACH_SET_SOFT_STATE	0x70
-#define  HV_SOFT_STATE_NORMAL		 0x01
-#define  HV_SOFT_STATE_TRANSITION	 0x02
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mach_set_soft_state(unsigned long soft_state,
-					       unsigned long msg_string_ra);
-#endif
-
-/* mach_get_soft_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MACH_GET_SOFT_STATE
- * ARG0:	software state description pointer
- * RET0:	status
- * RET1:	software state
- * ERRORS:	ENORADDR	software state description pointer is not a
- *				valid real address
- *		EBADALIGNED	software state description is not correctly
- *				aligned
- *
- * Retrieve the current value of the guest's software state.  The rules
- * for the software state pointer are the same as for mach_set_soft_state()
- * above.
- */
-#define HV_FAST_MACH_GET_SOFT_STATE	0x71
-
-/* svc_send()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SVC_SEND
- * ARG0:	service ID
- * ARG1:	buffer real address
- * ARG2:	buffer size
- * RET0:	STATUS
- * RET1:	sent_bytes
- *
- * Be careful, all output registers are clobbered by this operation,
- * so for example it is not possible to save away a value in %o4
- * across the trap.
- */
-#define HV_FAST_SVC_SEND		0x80
-
-/* svc_recv()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SVC_RECV
- * ARG0:	service ID
- * ARG1:	buffer real address
- * ARG2:	buffer size
- * RET0:	STATUS
- * RET1:	recv_bytes
- *
- * Be careful, all output registers are clobbered by this operation,
- * so for example it is not possible to save away a value in %o4
- * across the trap.
- */
-#define HV_FAST_SVC_RECV		0x81
-
-/* svc_getstatus()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SVC_GETSTATUS
- * ARG0:	service ID
- * RET0:	STATUS
- * RET1:	status bits
- */
-#define HV_FAST_SVC_GETSTATUS		0x82
-
-/* svc_setstatus()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SVC_SETSTATUS
- * ARG0:	service ID
- * ARG1:	bits to set
- * RET0:	STATUS
- */
-#define HV_FAST_SVC_SETSTATUS		0x83
-
-/* svc_clrstatus()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SVC_CLRSTATUS
- * ARG0:	service ID
- * ARG1:	bits to clear
- * RET0:	STATUS
- */
-#define HV_FAST_SVC_CLRSTATUS		0x84
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_svc_send(unsigned long svc_id,
-				    unsigned long buffer,
-				    unsigned long buffer_size,
-				    unsigned long *sent_bytes);
-extern unsigned long sun4v_svc_recv(unsigned long svc_id,
-				    unsigned long buffer,
-				    unsigned long buffer_size,
-				    unsigned long *recv_bytes);
-extern unsigned long sun4v_svc_getstatus(unsigned long svc_id,
-					 unsigned long *status_bits);
-extern unsigned long sun4v_svc_setstatus(unsigned long svc_id,
-					 unsigned long status_bits);
-extern unsigned long sun4v_svc_clrstatus(unsigned long svc_id,
-					 unsigned long status_bits);
-#endif
-
-/* Trap trace services.
- *
- * The hypervisor provides a trap tracing capability for privileged
- * code running on each virtual CPU.  Privileged code provides a
- * round-robin trap trace queue within which the hypervisor writes
- * 64-byte entries detailing hyperprivileged traps taken n behalf of
- * privileged code.  This is provided as a debugging capability for
- * privileged code.
- *
- * The trap trace control structure is 64-bytes long and placed at the
- * start (offset 0) of the trap trace buffer, and is described as
- * follows:
- */
-#ifndef __ASSEMBLY__
-struct hv_trap_trace_control {
-	unsigned long		head_offset;
-	unsigned long		tail_offset;
-	unsigned long		__reserved[0x30 / sizeof(unsigned long)];
-};
-#endif
-#define HV_TRAP_TRACE_CTRL_HEAD_OFFSET	0x00
-#define HV_TRAP_TRACE_CTRL_TAIL_OFFSET	0x08
-
-/* The head offset is the offset of the most recently completed entry
- * in the trap-trace buffer.  The tail offset is the offset of the
- * next entry to be written.  The control structure is owned and
- * modified by the hypervisor.  A guest may not modify the control
- * structure contents.  Attempts to do so will result in undefined
- * behavior for the guest.
- *
- * Each trap trace buffer entry is layed out as follows:
- */
-#ifndef __ASSEMBLY__
-struct hv_trap_trace_entry {
-	unsigned char	type;		/* Hypervisor or guest entry?	*/
-	unsigned char	hpstate;	/* Hyper-privileged state	*/
-	unsigned char	tl;		/* Trap level			*/
-	unsigned char	gl;		/* Global register level	*/
-	unsigned short	tt;		/* Trap type			*/
-	unsigned short	tag;		/* Extended trap identifier	*/
-	unsigned long	tstate;		/* Trap state			*/
-	unsigned long	tick;		/* Tick				*/
-	unsigned long	tpc;		/* Trap PC			*/
-	unsigned long	f1;		/* Entry specific		*/
-	unsigned long	f2;		/* Entry specific		*/
-	unsigned long	f3;		/* Entry specific		*/
-	unsigned long	f4;		/* Entry specific		*/
-};
-#endif
-#define HV_TRAP_TRACE_ENTRY_TYPE	0x00
-#define HV_TRAP_TRACE_ENTRY_HPSTATE	0x01
-#define HV_TRAP_TRACE_ENTRY_TL		0x02
-#define HV_TRAP_TRACE_ENTRY_GL		0x03
-#define HV_TRAP_TRACE_ENTRY_TT		0x04
-#define HV_TRAP_TRACE_ENTRY_TAG		0x06
-#define HV_TRAP_TRACE_ENTRY_TSTATE	0x08
-#define HV_TRAP_TRACE_ENTRY_TICK	0x10
-#define HV_TRAP_TRACE_ENTRY_TPC		0x18
-#define HV_TRAP_TRACE_ENTRY_F1		0x20
-#define HV_TRAP_TRACE_ENTRY_F2		0x28
-#define HV_TRAP_TRACE_ENTRY_F3		0x30
-#define HV_TRAP_TRACE_ENTRY_F4		0x38
-
-/* The type field is encoded as follows.  */
-#define HV_TRAP_TYPE_UNDEF		0x00 /* Entry content undefined     */
-#define HV_TRAP_TYPE_HV			0x01 /* Hypervisor trap entry       */
-#define HV_TRAP_TYPE_GUEST		0xff /* Added via ttrace_addentry() */
-
-/* ttrace_buf_conf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TTRACE_BUF_CONF
- * ARG0:	real address
- * ARG1:	number of entries
- * RET0:	status
- * RET1:	number of entries
- * ERRORS:	ENORADDR	Invalid real address
- *		EINVAL		Size is too small
- *		EBADALIGN	Real address not aligned on 64-byte boundary
- *
- * Requests hypervisor trap tracing and declares a virtual CPU's trap
- * trace buffer to the hypervisor.  The real address supplies the real
- * base address of the trap trace queue and must be 64-byte aligned.
- * Specifying a value of 0 for the number of entries disables trap
- * tracing for the calling virtual CPU.  The buffer allocated must be
- * sized for a power of two number of 64-byte trap trace entries plus
- * an initial 64-byte control structure.
- * 
- * This may be invoked any number of times so that a virtual CPU may
- * relocate a trap trace buffer or create "snapshots" of information.
- *
- * If the real address is illegal or badly aligned, then trap tracing
- * is disabled and an error is returned.
- *
- * Upon failure with EINVAL, this service call returns in RET1 the
- * minimum number of buffer entries required.  Upon other failures
- * RET1 is undefined.
- */
-#define HV_FAST_TTRACE_BUF_CONF		0x90
-
-/* ttrace_buf_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TTRACE_BUF_INFO
- * RET0:	status
- * RET1:	real address
- * RET2:	size
- * ERRORS:	None defined.
- *
- * Returns the size and location of the previously declared trap-trace
- * buffer.  In the event that no buffer was previously defined, or the
- * buffer is disabled, this call will return a size of zero bytes.
- */
-#define HV_FAST_TTRACE_BUF_INFO		0x91
-
-/* ttrace_enable()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TTRACE_ENABLE
- * ARG0:	enable
- * RET0:	status
- * RET1:	previous enable state
- * ERRORS:	EINVAL		No trap trace buffer currently defined
- *
- * Enable or disable trap tracing, and return the previous enabled
- * state in RET1.  Future systems may define various flags for the
- * enable argument (ARG0), for the moment a guest should pass
- * "(uint64_t) -1" to enable, and "(uint64_t) 0" to disable all
- * tracing - which will ensure future compatability.
- */
-#define HV_FAST_TTRACE_ENABLE		0x92
-
-/* ttrace_freeze()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_TTRACE_FREEZE
- * ARG0:	freeze
- * RET0:	status
- * RET1:	previous freeze state
- * ERRORS:	EINVAL		No trap trace buffer currently defined
- *
- * Freeze or unfreeze trap tracing, returning the previous freeze
- * state in RET1.  A guest should pass a non-zero value to freeze and
- * a zero value to unfreeze all tracing.  The returned previous state
- * is 0 for not frozen and 1 for frozen.
- */
-#define HV_FAST_TTRACE_FREEZE		0x93
-
-/* ttrace_addentry()
- * TRAP:	HV_TTRACE_ADDENTRY_TRAP
- * ARG0:	tag (16-bits)
- * ARG1:	data word 0
- * ARG2:	data word 1
- * ARG3:	data word 2
- * ARG4:	data word 3
- * RET0:	status
- * ERRORS:	EINVAL		No trap trace buffer currently defined
- *
- * Add an entry to the trap trace buffer.  Upon return only ARG0/RET0
- * is modified - none of the other registers holding arguments are
- * volatile across this hypervisor service.
- */
-
-/* Core dump services.
- *
- * Since the hypervisor viraulizes and thus obscures a lot of the
- * physical machine layout and state, traditional OS crash dumps can
- * be difficult to diagnose especially when the problem is a
- * configuration error of some sort.
- *
- * The dump services provide an opaque buffer into which the
- * hypervisor can place it's internal state in order to assist in
- * debugging such situations.  The contents are opaque and extremely
- * platform and hypervisor implementation specific.  The guest, during
- * a core dump, requests that the hypervisor update any information in
- * the dump buffer in preparation to being dumped as part of the
- * domain's memory image.
- */
-
-/* dump_buf_update()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_DUMP_BUF_UPDATE
- * ARG0:	real address
- * ARG1:	size
- * RET0:	status
- * RET1:	required size of dump buffer
- * ERRORS:	ENORADDR	Invalid real address
- *		EBADALIGN	Real address is not aligned on a 64-byte
- *				boundary
- *		EINVAL		Size is non-zero but less than minimum size
- *				required
- *		ENOTSUPPORTED	Operation not supported on current logical
- *				domain
- *
- * Declare a domain dump buffer to the hypervisor.  The real address
- * provided for the domain dump buffer must be 64-byte aligned.  The
- * size specifies the size of the dump buffer and may be larger than
- * the minimum size specified in the machine description.  The
- * hypervisor will fill the dump buffer with opaque data.
- *
- * Note: A guest may elect to include dump buffer contents as part of a crash
- *       dump to assist with debugging.  This function may be called any number
- *       of times so that a guest may relocate a dump buffer, or create
- *       "snapshots" of any dump-buffer information.  Each call to
- *       dump_buf_update() atomically declares the new dump buffer to the
- *       hypervisor.
- *
- * A specified size of 0 unconfigures the dump buffer.  If the real
- * address is illegal or badly aligned, then any currently active dump
- * buffer is disabled and an error is returned.
- *
- * In the event that the call fails with EINVAL, RET1 contains the
- * minimum size requires by the hypervisor for a valid dump buffer.
- */
-#define HV_FAST_DUMP_BUF_UPDATE		0x94
-
-/* dump_buf_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_DUMP_BUF_INFO
- * RET0:	status
- * RET1:	real address of current dump buffer
- * RET2:	size of current dump buffer
- * ERRORS:	No errors defined.
- *
- * Return the currently configures dump buffer description.  A
- * returned size of 0 bytes indicates an undefined dump buffer.  In
- * this case the return address in RET1 is undefined.
- */
-#define HV_FAST_DUMP_BUF_INFO		0x95
-
-/* Device interrupt services.
- *
- * Device interrupts are allocated to system bus bridges by the hypervisor,
- * and described to OBP in the machine description.  OBP then describes
- * these interrupts to the OS via properties in the device tree.
- *
- * Terminology:
- *
- *	cpuid		Unique opaque value which represents a target cpu.
- *
- *	devhandle	Device handle.  It uniquely identifies a device, and
- *			consistes of the lower 28-bits of the hi-cell of the
- *			first entry of the device's "reg" property in the
- *			OBP device tree.
- *
- *	devino		Device interrupt number.  Specifies the relative
- *			interrupt number within the device.  The unique
- *			combination of devhandle and devino are used to
- *			identify a specific device interrupt.
- *
- *			Note: The devino value is the same as the values in the
- *			      "interrupts" property or "interrupt-map" property
- *			      in the OBP device tree for that device.
- *
- *	sysino		System interrupt number.  A 64-bit unsigned interger
- *			representing a unique interrupt within a virtual
- *			machine.
- *
- *	intr_state	A flag representing the interrupt state for a given
- *			sysino.  The state values are defined below.
- *
- *	intr_enabled	A flag representing the 'enabled' state for a given
- *			sysino.  The enable values are defined below.
- */
-
-#define HV_INTR_STATE_IDLE		0 /* Nothing pending */
-#define HV_INTR_STATE_RECEIVED		1 /* Interrupt received by hardware */
-#define HV_INTR_STATE_DELIVERED		2 /* Interrupt delivered to queue */
-
-#define HV_INTR_DISABLED		0 /* sysino not enabled */
-#define HV_INTR_ENABLED			1 /* sysino enabled */
-
-/* intr_devino_to_sysino()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_DEVINO2SYSINO
- * ARG0:	devhandle
- * ARG1:	devino
- * RET0:	status
- * RET1:	sysino
- * ERRORS:	EINVAL		Invalid devhandle/devino
- *
- * Converts a device specific interrupt number of the given
- * devhandle/devino into a system specific ino (sysino).
- */
-#define HV_FAST_INTR_DEVINO2SYSINO	0xa0
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_devino_to_sysino(unsigned long devhandle,
-					    unsigned long devino);
-#endif
-
-/* intr_getenabled()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_GETENABLED
- * ARG0:	sysino
- * RET0:	status
- * RET1:	intr_enabled (HV_INTR_{DISABLED,ENABLED})
- * ERRORS:	EINVAL		Invalid sysino
- *
- * Returns interrupt enabled state in RET1 for the interrupt defined
- * by the given sysino.
- */
-#define HV_FAST_INTR_GETENABLED		0xa1
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_getenabled(unsigned long sysino);
-#endif
-
-/* intr_setenabled()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_SETENABLED
- * ARG0:	sysino
- * ARG1:	intr_enabled (HV_INTR_{DISABLED,ENABLED})
- * RET0:	status
- * ERRORS:	EINVAL		Invalid sysino or intr_enabled value
- *
- * Set the 'enabled' state of the interrupt sysino.
- */
-#define HV_FAST_INTR_SETENABLED		0xa2
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_setenabled(unsigned long sysino, unsigned long intr_enabled);
-#endif
-
-/* intr_getstate()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_GETSTATE
- * ARG0:	sysino
- * RET0:	status
- * RET1:	intr_state (HV_INTR_STATE_*)
- * ERRORS:	EINVAL		Invalid sysino
- *
- * Returns current state of the interrupt defined by the given sysino.
- */
-#define HV_FAST_INTR_GETSTATE		0xa3
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_getstate(unsigned long sysino);
-#endif
-
-/* intr_setstate()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_SETSTATE
- * ARG0:	sysino
- * ARG1:	intr_state (HV_INTR_STATE_*)
- * RET0:	status
- * ERRORS:	EINVAL		Invalid sysino or intr_state value
- *
- * Sets the current state of the interrupt described by the given sysino
- * value.
- *
- * Note: Setting the state to HV_INTR_STATE_IDLE clears any pending
- *       interrupt for sysino.
- */
-#define HV_FAST_INTR_SETSTATE		0xa4
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_setstate(unsigned long sysino, unsigned long intr_state);
-#endif
-
-/* intr_gettarget()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_GETTARGET
- * ARG0:	sysino
- * RET0:	status
- * RET1:	cpuid
- * ERRORS:	EINVAL		Invalid sysino
- *
- * Returns CPU that is the current target of the interrupt defined by
- * the given sysino.  The CPU value returned is undefined if the target
- * has not been set via intr_settarget().
- */
-#define HV_FAST_INTR_GETTARGET		0xa5
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_gettarget(unsigned long sysino);
-#endif
-
-/* intr_settarget()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_INTR_SETTARGET
- * ARG0:	sysino
- * ARG1:	cpuid
- * RET0:	status
- * ERRORS:	EINVAL		Invalid sysino
- *		ENOCPU		Invalid cpuid
- *
- * Set the target CPU for the interrupt defined by the given sysino.
- */
-#define HV_FAST_INTR_SETTARGET		0xa6
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_intr_settarget(unsigned long sysino, unsigned long cpuid);
-#endif
-
-/* vintr_get_cookie()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_GET_COOKIE
- * ARG0:	device handle
- * ARG1:	device ino
- * RET0:	status
- * RET1:	cookie
- */
-#define HV_FAST_VINTR_GET_COOKIE	0xa7
-
-/* vintr_set_cookie()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_SET_COOKIE
- * ARG0:	device handle
- * ARG1:	device ino
- * ARG2:	cookie
- * RET0:	status
- */
-#define HV_FAST_VINTR_SET_COOKIE	0xa8
-
-/* vintr_get_valid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_GET_VALID
- * ARG0:	device handle
- * ARG1:	device ino
- * RET0:	status
- * RET1:	valid state
- */
-#define HV_FAST_VINTR_GET_VALID		0xa9
-
-/* vintr_set_valid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_SET_VALID
- * ARG0:	device handle
- * ARG1:	device ino
- * ARG2:	valid state
- * RET0:	status
- */
-#define HV_FAST_VINTR_SET_VALID		0xaa
-
-/* vintr_get_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_GET_STATE
- * ARG0:	device handle
- * ARG1:	device ino
- * RET0:	status
- * RET1:	state
- */
-#define HV_FAST_VINTR_GET_STATE		0xab
-
-/* vintr_set_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_SET_STATE
- * ARG0:	device handle
- * ARG1:	device ino
- * ARG2:	state
- * RET0:	status
- */
-#define HV_FAST_VINTR_SET_STATE		0xac
-
-/* vintr_get_target()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_GET_TARGET
- * ARG0:	device handle
- * ARG1:	device ino
- * RET0:	status
- * RET1:	cpuid
- */
-#define HV_FAST_VINTR_GET_TARGET	0xad
-
-/* vintr_set_target()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_VINTR_SET_TARGET
- * ARG0:	device handle
- * ARG1:	device ino
- * ARG2:	cpuid
- * RET0:	status
- */
-#define HV_FAST_VINTR_SET_TARGET	0xae
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_vintr_get_cookie(unsigned long dev_handle,
-					    unsigned long dev_ino,
-					    unsigned long *cookie);
-extern unsigned long sun4v_vintr_set_cookie(unsigned long dev_handle,
-					    unsigned long dev_ino,
-					    unsigned long cookie);
-extern unsigned long sun4v_vintr_get_valid(unsigned long dev_handle,
-					   unsigned long dev_ino,
-					   unsigned long *valid);
-extern unsigned long sun4v_vintr_set_valid(unsigned long dev_handle,
-					   unsigned long dev_ino,
-					   unsigned long valid);
-extern unsigned long sun4v_vintr_get_state(unsigned long dev_handle,
-					   unsigned long dev_ino,
-					   unsigned long *state);
-extern unsigned long sun4v_vintr_set_state(unsigned long dev_handle,
-					   unsigned long dev_ino,
-					   unsigned long state);
-extern unsigned long sun4v_vintr_get_target(unsigned long dev_handle,
-					    unsigned long dev_ino,
-					    unsigned long *cpuid);
-extern unsigned long sun4v_vintr_set_target(unsigned long dev_handle,
-					    unsigned long dev_ino,
-					    unsigned long cpuid);
-#endif
-
-/* PCI IO services.
- *
- * See the terminology descriptions in the device interrupt services
- * section above as those apply here too.  Here are terminology
- * definitions specific to these PCI IO services:
- *
- *	tsbnum		TSB number.  Indentifies which io-tsb is used.
- *			For this version of the specification, tsbnum
- *			must be zero.
- *
- *	tsbindex	TSB index.  Identifies which entry in the TSB
- *			is used.  The first entry is zero.
- *
- *	tsbid		A 64-bit aligned data structure which contains
- *			a tsbnum and a tsbindex.  Bits 63:32 contain the
- *			tsbnum and bits 31:00 contain the tsbindex.
- *
- *			Use the HV_PCI_TSBID() macro to construct such
- * 			values.
- *
- *	io_attributes	IO attributes for IOMMU mappings.  One of more
- *			of the attritbute bits are stores in a 64-bit
- *			value.  The values are defined below.
- *
- *	r_addr		64-bit real address
- *
- *	pci_device	PCI device address.  A PCI device address identifies
- *			a specific device on a specific PCI bus segment.
- *			A PCI device address ia a 32-bit unsigned integer
- *			with the following format:
- *
- *				00000000.bbbbbbbb.dddddfff.00000000
- *
- *			Use the HV_PCI_DEVICE_BUILD() macro to construct
- *			such values.
- *
- *	pci_config_offset
- *			PCI configureation space offset.  For conventional
- *			PCI a value between 0 and 255.  For extended
- *			configuration space, a value between 0 and 4095.
- *
- *			Note: For PCI configuration space accesses, the offset
- *			      must be aligned to the access size.
- *
- *	error_flag	A return value which specifies if the action succeeded
- *			or failed.  0 means no error, non-0 means some error
- *			occurred while performing the service.
- *
- *	io_sync_direction
- *			Direction definition for pci_dma_sync(), defined
- *			below in HV_PCI_SYNC_*.
- *
- *	io_page_list	A list of io_page_addresses, an io_page_address is
- *			a real address.
- *
- *	io_page_list_p	A pointer to an io_page_list.
- *
- *	"size based byte swap" - Some functions do size based byte swapping
- *				 which allows sw to access pointers and
- *				 counters in native form when the processor
- *				 operates in a different endianness than the
- *				 IO bus.  Size-based byte swapping converts a
- *				 multi-byte field between big-endian and
- *				 little-endian format.
- */
-
-#define HV_PCI_MAP_ATTR_READ		0x01
-#define HV_PCI_MAP_ATTR_WRITE		0x02
-
-#define HV_PCI_DEVICE_BUILD(b,d,f)	\
-	((((b) & 0xff) << 16) | \
-	 (((d) & 0x1f) << 11) | \
-	 (((f) & 0x07) <<  8))
-
-#define HV_PCI_TSBID(__tsb_num, __tsb_index) \
-	((((u64)(__tsb_num)) << 32UL) | ((u64)(__tsb_index)))
-
-#define HV_PCI_SYNC_FOR_DEVICE		0x01
-#define HV_PCI_SYNC_FOR_CPU		0x02
-
-/* pci_iommu_map()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_IOMMU_MAP
- * ARG0:	devhandle
- * ARG1:	tsbid
- * ARG2:	#ttes
- * ARG3:	io_attributes
- * ARG4:	io_page_list_p
- * RET0:	status
- * RET1:	#ttes mapped
- * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex/io_attributes
- *		EBADALIGN	Improperly aligned real address
- *		ENORADDR	Invalid real address
- *
- * Create IOMMU mappings in the sun4v device defined by the given
- * devhandle.  The mappings are created in the TSB defined by the
- * tsbnum component of the given tsbid.  The first mapping is created
- * in the TSB i ndex defined by the tsbindex component of the given tsbid.
- * The call creates up to #ttes mappings, the first one at tsbnum, tsbindex,
- * the second at tsbnum, tsbindex + 1, etc.
- *
- * All mappings are created with the attributes defined by the io_attributes
- * argument.  The page mapping addresses are described in the io_page_list
- * defined by the given io_page_list_p, which is a pointer to the io_page_list.
- * The first entry in the io_page_list is the address for the first iotte, the
- * 2nd for the 2nd iotte, and so on.
- *
- * Each io_page_address in the io_page_list must be appropriately aligned.
- * #ttes must be greater than zero.  For this version of the spec, the tsbnum
- * component of the given tsbid must be zero.
- *
- * Returns the actual number of mappings creates, which may be less than
- * or equal to the argument #ttes.  If the function returns a value which
- * is less than the #ttes, the caller may continus to call the function with
- * an updated tsbid, #ttes, io_page_list_p arguments until all pages are
- * mapped.
- *
- * Note: This function does not imply an iotte cache flush.  The guest must
- *       demap an entry before re-mapping it.
- */
-#define HV_FAST_PCI_IOMMU_MAP		0xb0
-
-/* pci_iommu_demap()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_IOMMU_DEMAP
- * ARG0:	devhandle
- * ARG1:	tsbid
- * ARG2:	#ttes
- * RET0:	status
- * RET1:	#ttes demapped
- * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex
- *
- * Demap and flush IOMMU mappings in the device defined by the given
- * devhandle.  Demaps up to #ttes entries in the TSB defined by the tsbnum
- * component of the given tsbid, starting at the TSB index defined by the
- * tsbindex component of the given tsbid.
- *
- * For this version of the spec, the tsbnum of the given tsbid must be zero.
- * #ttes must be greater than zero.
- *
- * Returns the actual number of ttes demapped, which may be less than or equal
- * to the argument #ttes.  If #ttes demapped is less than #ttes, the caller
- * may continue to call this function with updated tsbid and #ttes arguments
- * until all pages are demapped.
- *
- * Note: Entries do not have to be mapped to be demapped.  A demap of an
- *       unmapped page will flush the entry from the tte cache.
- */
-#define HV_FAST_PCI_IOMMU_DEMAP		0xb1
-
-/* pci_iommu_getmap()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_IOMMU_GETMAP
- * ARG0:	devhandle
- * ARG1:	tsbid
- * RET0:	status
- * RET1:	io_attributes
- * RET2:	real address
- * ERRORS:	EINVAL		Invalid devhandle/tsbnum/tsbindex
- *		ENOMAP		Mapping is not valid, no translation exists
- *
- * Read and return the mapping in the device described by the given devhandle
- * and tsbid.  If successful, the io_attributes shall be returned in RET1
- * and the page address of the mapping shall be returned in RET2.
- *
- * For this version of the spec, the tsbnum component of the given tsbid
- * must be zero.
- */
-#define HV_FAST_PCI_IOMMU_GETMAP	0xb2
-
-/* pci_iommu_getbypass()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_IOMMU_GETBYPASS
- * ARG0:	devhandle
- * ARG1:	real address
- * ARG2:	io_attributes
- * RET0:	status
- * RET1:	io_addr
- * ERRORS:	EINVAL		Invalid devhandle/io_attributes
- *		ENORADDR	Invalid real address
- *		ENOTSUPPORTED	Function not supported in this implementation.
- *
- * Create a "special" mapping in the device described by the given devhandle,
- * for the given real address and attributes.  Return the IO address in RET1
- * if successful.
- */
-#define HV_FAST_PCI_IOMMU_GETBYPASS	0xb3
-
-/* pci_config_get()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_CONFIG_GET
- * ARG0:	devhandle
- * ARG1:	pci_device
- * ARG2:	pci_config_offset
- * ARG3:	size
- * RET0:	status
- * RET1:	error_flag
- * RET2:	data
- * ERRORS:	EINVAL		Invalid devhandle/pci_device/offset/size
- *		EBADALIGN	pci_config_offset not size aligned
- *		ENOACCESS	Access to this offset is not permitted
- *
- * Read PCI configuration space for the adapter described by the given
- * devhandle.  Read size (1, 2, or 4) bytes of data from the given
- * pci_device, at pci_config_offset from the beginning of the device's
- * configuration space.  If there was no error, RET1 is set to zero and
- * RET2 is set to the data read.  Insignificant bits in RET2 are not
- * guarenteed to have any specific value and therefore must be ignored.
- *
- * The data returned in RET2 is size based byte swapped.
- *
- * If an error occurs during the read, set RET1 to a non-zero value.  The
- * given pci_config_offset must be 'size' aligned.
- */
-#define HV_FAST_PCI_CONFIG_GET		0xb4
-
-/* pci_config_put()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_CONFIG_PUT
- * ARG0:	devhandle
- * ARG1:	pci_device
- * ARG2:	pci_config_offset
- * ARG3:	size
- * ARG4:	data
- * RET0:	status
- * RET1:	error_flag
- * ERRORS:	EINVAL		Invalid devhandle/pci_device/offset/size
- *		EBADALIGN	pci_config_offset not size aligned
- *		ENOACCESS	Access to this offset is not permitted
- *
- * Write PCI configuration space for the adapter described by the given
- * devhandle.  Write size (1, 2, or 4) bytes of data in a single operation,
- * at pci_config_offset from the beginning of the device's configuration
- * space.  The data argument contains the data to be written to configuration
- * space.  Prior to writing, the data is size based byte swapped.
- *
- * If an error occurs during the write access, do not generate an error
- * report, do set RET1 to a non-zero value.  Otherwise RET1 is zero.
- * The given pci_config_offset must be 'size' aligned.
- *
- * This function is permitted to read from offset zero in the configuration
- * space described by the given pci_device if necessary to ensure that the
- * write access to config space completes.
- */
-#define HV_FAST_PCI_CONFIG_PUT		0xb5
-
-/* pci_peek()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_PEEK
- * ARG0:	devhandle
- * ARG1:	real address
- * ARG2:	size
- * RET0:	status
- * RET1:	error_flag
- * RET2:	data
- * ERRORS:	EINVAL		Invalid devhandle or size
- *		EBADALIGN	Improperly aligned real address
- *		ENORADDR	Bad real address
- *		ENOACCESS	Guest access prohibited
- *
- * Attempt to read the IO address given by the given devhandle, real address,
- * and size.  Size must be 1, 2, 4, or 8.  The read is performed as a single
- * access operation using the given size.  If an error occurs when reading
- * from the given location, do not generate an error report, but return a
- * non-zero value in RET1.  If the read was successful, return zero in RET1
- * and return the actual data read in RET2.  The data returned is size based
- * byte swapped.
- *
- * Non-significant bits in RET2 are not guarenteed to have any specific value
- * and therefore must be ignored.  If RET1 is returned as non-zero, the data 
- * value is not guarenteed to have any specific value and should be ignored.
- *
- * The caller must have permission to read from the given devhandle, real
- * address, which must be an IO address.  The argument real address must be a
- * size aligned address.
- *
- * The hypervisor implementation of this function must block access to any
- * IO address that the guest does not have explicit permission to access.
- */
-#define HV_FAST_PCI_PEEK		0xb6
-
-/* pci_poke()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_POKE
- * ARG0:	devhandle
- * ARG1:	real address
- * ARG2:	size
- * ARG3:	data
- * ARG4:	pci_device
- * RET0:	status
- * RET1:	error_flag
- * ERRORS:	EINVAL		Invalid devhandle, size, or pci_device
- *		EBADALIGN	Improperly aligned real address
- *		ENORADDR	Bad real address
- *		ENOACCESS	Guest access prohibited
- *		ENOTSUPPORTED	Function is not supported by implementation
- *
- * Attempt to write data to the IO address given by the given devhandle,
- * real address, and size.  Size must be 1, 2, 4, or 8.  The write is
- * performed as a single access operation using the given size. Prior to
- * writing the data is size based swapped.
- *
- * If an error occurs when writing to the given location, do not generate an
- * error report, but return a non-zero value in RET1.  If the write was
- * successful, return zero in RET1.
- *
- * pci_device describes the configuration address of the device being
- * written to.  The implementation may safely read from offset 0 with
- * the configuration space of the device described by devhandle and
- * pci_device in order to guarantee that the write portion of the operation
- * completes
- *
- * Any error that occurs due to the read shall be reported using the normal
- * error reporting mechanisms .. the read error is not suppressed.
- *
- * The caller must have permission to write to the given devhandle, real
- * address, which must be an IO address.  The argument real address must be a
- * size aligned address.  The caller must have permission to read from
- * the given devhandle, pci_device cofiguration space offset 0.
- *
- * The hypervisor implementation of this function must block access to any
- * IO address that the guest does not have explicit permission to access.
- */
-#define HV_FAST_PCI_POKE		0xb7
-
-/* pci_dma_sync()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_DMA_SYNC
- * ARG0:	devhandle
- * ARG1:	real address
- * ARG2:	size
- * ARG3:	io_sync_direction
- * RET0:	status
- * RET1:	#synced
- * ERRORS:	EINVAL		Invalid devhandle or io_sync_direction
- *		ENORADDR	Bad real address
- *
- * Synchronize a memory region described by the given real address and size,
- * for the device defined by the given devhandle using the direction(s)
- * defined by the given io_sync_direction.  The argument size is the size of
- * the memory region in bytes.
- *
- * Return the actual number of bytes synchronized in the return value #synced,
- * which may be less than or equal to the argument size.  If the return
- * value #synced is less than size, the caller must continue to call this
- * function with updated real address and size arguments until the entire
- * memory region is synchronized.
- */
-#define HV_FAST_PCI_DMA_SYNC		0xb8
-
-/* PCI MSI services.  */
-
-#define HV_MSITYPE_MSI32		0x00
-#define HV_MSITYPE_MSI64		0x01
-
-#define HV_MSIQSTATE_IDLE		0x00
-#define HV_MSIQSTATE_ERROR		0x01
-
-#define HV_MSIQ_INVALID			0x00
-#define HV_MSIQ_VALID			0x01
-
-#define HV_MSISTATE_IDLE		0x00
-#define HV_MSISTATE_DELIVERED		0x01
-
-#define HV_MSIVALID_INVALID		0x00
-#define HV_MSIVALID_VALID		0x01
-
-#define HV_PCIE_MSGTYPE_PME_MSG		0x18
-#define HV_PCIE_MSGTYPE_PME_ACK_MSG	0x1b
-#define HV_PCIE_MSGTYPE_CORR_MSG	0x30
-#define HV_PCIE_MSGTYPE_NONFATAL_MSG	0x31
-#define HV_PCIE_MSGTYPE_FATAL_MSG	0x33
-
-#define HV_MSG_INVALID			0x00
-#define HV_MSG_VALID			0x01
-
-/* pci_msiq_conf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_CONF
- * ARG0:	devhandle
- * ARG1:	msiqid
- * ARG2:	real address
- * ARG3:	number of entries
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle, msiqid or nentries
- *		EBADALIGN	Improperly aligned real address
- *		ENORADDR	Bad real address
- *
- * Configure the MSI queue given by the devhandle and msiqid arguments,
- * and to be placed at the given real address and be of the given
- * number of entries.  The real address must be aligned exactly to match
- * the queue size.  Each queue entry is 64-bytes long, so f.e. a 32 entry
- * queue must be aligned on a 2048 byte real address boundary.  The MSI-EQ
- * Head and Tail are initialized so that the MSI-EQ is 'empty'.
- *
- * Implementation Note: Certain implementations have fixed sized queues.  In
- *                      that case, number of entries must contain the correct
- *                      value.
- */
-#define HV_FAST_PCI_MSIQ_CONF		0xc0
-
-/* pci_msiq_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_INFO
- * ARG0:	devhandle
- * ARG1:	msiqid
- * RET0:	status
- * RET1:	real address
- * RET2:	number of entries
- * ERRORS:	EINVAL		Invalid devhandle or msiqid
- *
- * Return the configuration information for the MSI queue described
- * by the given devhandle and msiqid.  The base address of the queue
- * is returned in ARG1 and the number of entries is returned in ARG2.
- * If the queue is unconfigured, the real address is undefined and the
- * number of entries will be returned as zero.
- */
-#define HV_FAST_PCI_MSIQ_INFO		0xc1
-
-/* pci_msiq_getvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_GETVALID
- * ARG0:	devhandle
- * ARG1:	msiqid
- * RET0:	status
- * RET1:	msiqvalid	(HV_MSIQ_VALID or HV_MSIQ_INVALID)
- * ERRORS:	EINVAL		Invalid devhandle or msiqid
- *
- * Get the valid state of the MSI-EQ described by the given devhandle and
- * msiqid.
- */
-#define HV_FAST_PCI_MSIQ_GETVALID	0xc2
-
-/* pci_msiq_setvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_SETVALID
- * ARG0:	devhandle
- * ARG1:	msiqid
- * ARG2:	msiqvalid	(HV_MSIQ_VALID or HV_MSIQ_INVALID)
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqvalid
- *				value or MSI EQ is uninitialized
- *
- * Set the valid state of the MSI-EQ described by the given devhandle and
- * msiqid to the given msiqvalid.
- */
-#define HV_FAST_PCI_MSIQ_SETVALID	0xc3
-
-/* pci_msiq_getstate()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_GETSTATE
- * ARG0:	devhandle
- * ARG1:	msiqid
- * RET0:	status
- * RET1:	msiqstate	(HV_MSIQSTATE_IDLE or HV_MSIQSTATE_ERROR)
- * ERRORS:	EINVAL		Invalid devhandle or msiqid
- *
- * Get the state of the MSI-EQ described by the given devhandle and
- * msiqid.
- */
-#define HV_FAST_PCI_MSIQ_GETSTATE	0xc4
-
-/* pci_msiq_getvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_GETVALID
- * ARG0:	devhandle
- * ARG1:	msiqid
- * ARG2:	msiqstate	(HV_MSIQSTATE_IDLE or HV_MSIQSTATE_ERROR)
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqstate
- *				value or MSI EQ is uninitialized
- *
- * Set the state of the MSI-EQ described by the given devhandle and
- * msiqid to the given msiqvalid.
- */
-#define HV_FAST_PCI_MSIQ_SETSTATE	0xc5
-
-/* pci_msiq_gethead()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_GETHEAD
- * ARG0:	devhandle
- * ARG1:	msiqid
- * RET0:	status
- * RET1:	msiqhead
- * ERRORS:	EINVAL		Invalid devhandle or msiqid
- *
- * Get the current MSI EQ queue head for the MSI-EQ described by the
- * given devhandle and msiqid.
- */
-#define HV_FAST_PCI_MSIQ_GETHEAD	0xc6
-
-/* pci_msiq_sethead()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_SETHEAD
- * ARG0:	devhandle
- * ARG1:	msiqid
- * ARG2:	msiqhead
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msiqid or msiqhead,
- *				or MSI EQ is uninitialized
- *
- * Set the current MSI EQ queue head for the MSI-EQ described by the
- * given devhandle and msiqid.
- */
-#define HV_FAST_PCI_MSIQ_SETHEAD	0xc7
-
-/* pci_msiq_gettail()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSIQ_GETTAIL
- * ARG0:	devhandle
- * ARG1:	msiqid
- * RET0:	status
- * RET1:	msiqtail
- * ERRORS:	EINVAL		Invalid devhandle or msiqid
- *
- * Get the current MSI EQ queue tail for the MSI-EQ described by the
- * given devhandle and msiqid.
- */
-#define HV_FAST_PCI_MSIQ_GETTAIL	0xc8
-
-/* pci_msi_getvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_GETVALID
- * ARG0:	devhandle
- * ARG1:	msinum
- * RET0:	status
- * RET1:	msivalidstate
- * ERRORS:	EINVAL		Invalid devhandle or msinum
- *
- * Get the current valid/enabled state for the MSI defined by the
- * given devhandle and msinum.
- */
-#define HV_FAST_PCI_MSI_GETVALID	0xc9
-
-/* pci_msi_setvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_SETVALID
- * ARG0:	devhandle
- * ARG1:	msinum
- * ARG2:	msivalidstate
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msinum or msivalidstate
- *
- * Set the current valid/enabled state for the MSI defined by the
- * given devhandle and msinum.
- */
-#define HV_FAST_PCI_MSI_SETVALID	0xca
-
-/* pci_msi_getmsiq()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_GETMSIQ
- * ARG0:	devhandle
- * ARG1:	msinum
- * RET0:	status
- * RET1:	msiqid
- * ERRORS:	EINVAL		Invalid devhandle or msinum or MSI is unbound
- *
- * Get the MSI EQ that the MSI defined by the given devhandle and
- * msinum is bound to.
- */
-#define HV_FAST_PCI_MSI_GETMSIQ		0xcb
-
-/* pci_msi_setmsiq()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_SETMSIQ
- * ARG0:	devhandle
- * ARG1:	msinum
- * ARG2:	msitype
- * ARG3:	msiqid
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msinum or msiqid
- *
- * Set the MSI EQ that the MSI defined by the given devhandle and
- * msinum is bound to.
- */
-#define HV_FAST_PCI_MSI_SETMSIQ		0xcc
-
-/* pci_msi_getstate()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_GETSTATE
- * ARG0:	devhandle
- * ARG1:	msinum
- * RET0:	status
- * RET1:	msistate
- * ERRORS:	EINVAL		Invalid devhandle or msinum
- *
- * Get the state of the MSI defined by the given devhandle and msinum.
- * If not initialized, return HV_MSISTATE_IDLE.
- */
-#define HV_FAST_PCI_MSI_GETSTATE	0xcd
-
-/* pci_msi_setstate()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSI_SETSTATE
- * ARG0:	devhandle
- * ARG1:	msinum
- * ARG2:	msistate
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msinum or msistate
- *
- * Set the state of the MSI defined by the given devhandle and msinum.
- */
-#define HV_FAST_PCI_MSI_SETSTATE	0xce
-
-/* pci_msg_getmsiq()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSG_GETMSIQ
- * ARG0:	devhandle
- * ARG1:	msgtype
- * RET0:	status
- * RET1:	msiqid
- * ERRORS:	EINVAL		Invalid devhandle or msgtype
- *
- * Get the MSI EQ of the MSG defined by the given devhandle and msgtype.
- */
-#define HV_FAST_PCI_MSG_GETMSIQ		0xd0
-
-/* pci_msg_setmsiq()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSG_SETMSIQ
- * ARG0:	devhandle
- * ARG1:	msgtype
- * ARG2:	msiqid
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle, msgtype, or msiqid
- *
- * Set the MSI EQ of the MSG defined by the given devhandle and msgtype.
- */
-#define HV_FAST_PCI_MSG_SETMSIQ		0xd1
-
-/* pci_msg_getvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSG_GETVALID
- * ARG0:	devhandle
- * ARG1:	msgtype
- * RET0:	status
- * RET1:	msgvalidstate
- * ERRORS:	EINVAL		Invalid devhandle or msgtype
- *
- * Get the valid/enabled state of the MSG defined by the given
- * devhandle and msgtype.
- */
-#define HV_FAST_PCI_MSG_GETVALID	0xd2
-
-/* pci_msg_setvalid()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_PCI_MSG_SETVALID
- * ARG0:	devhandle
- * ARG1:	msgtype
- * ARG2:	msgvalidstate
- * RET0:	status
- * ERRORS:	EINVAL		Invalid devhandle or msgtype or msgvalidstate
- *
- * Set the valid/enabled state of the MSG defined by the given
- * devhandle and msgtype.
- */
-#define HV_FAST_PCI_MSG_SETVALID	0xd3
-
-/* Logical Domain Channel services.  */
-
-#define LDC_CHANNEL_DOWN		0
-#define LDC_CHANNEL_UP			1
-#define LDC_CHANNEL_RESETTING		2
-
-/* ldc_tx_qconf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_TX_QCONF
- * ARG0:	channel ID
- * ARG1:	real address base of queue
- * ARG2:	num entries in queue
- * RET0:	status
- *
- * Configure transmit queue for the LDC endpoint specified by the
- * given channel ID, to be placed at the given real address, and
- * be of the given num entries.  Num entries must be a power of two.
- * The real address base of the queue must be aligned on the queue
- * size.  Each queue entry is 64-bytes, so for example, a 32 entry
- * queue must be aligned on a 2048 byte real address boundary.
- *
- * Upon configuration of a valid transmit queue the head and tail
- * pointers are set to a hypervisor specific identical value indicating
- * that the queue initially is empty.
- *
- * The endpoint's transmit queue is un-configured if num entries is zero.
- *
- * The maximum number of entries for each queue for a specific cpu may be
- * determined from the machine description.  A transmit queue may be
- * specified even in the event that the LDC is down (peer endpoint has no
- * receive queue specified).  Transmission will begin as soon as the peer
- * endpoint defines a receive queue.
- *
- * It is recommended that a guest wait for a transmit queue to empty prior
- * to reconfiguring it, or un-configuring it.  Re or un-configuring of a
- * non-empty transmit queue behaves exactly as defined above, however it
- * is undefined as to how many of the pending entries in the original queue
- * will be delivered prior to the re-configuration taking effect.
- * Furthermore, as the queue configuration causes a reset of the head and
- * tail pointers there is no way for a guest to determine how many entries
- * have been sent after the configuration operation.
- */
-#define HV_FAST_LDC_TX_QCONF		0xe0
-
-/* ldc_tx_qinfo()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_TX_QINFO
- * ARG0:	channel ID
- * RET0:	status
- * RET1:	real address base of queue
- * RET2:	num entries in queue
- *
- * Return the configuration info for the transmit queue of LDC endpoint
- * defined by the given channel ID.  The real address is the currently
- * defined real address base of the defined queue, and num entries is the
- * size of the queue in terms of number of entries.
- *
- * If the specified channel ID is a valid endpoint number, but no transmit
- * queue has been defined this service will return success, but with num
- * entries set to zero and the real address will have an undefined value.
- */
-#define HV_FAST_LDC_TX_QINFO		0xe1
-
-/* ldc_tx_get_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_TX_GET_STATE
- * ARG0:	channel ID
- * RET0:	status
- * RET1:	head offset
- * RET2:	tail offset
- * RET3:	channel state
- *
- * Return the transmit state, and the head and tail queue pointers, for
- * the transmit queue of the LDC endpoint defined by the given channel ID.
- * The head and tail values are the byte offset of the head and tail
- * positions of the transmit queue for the specified endpoint.
- */
-#define HV_FAST_LDC_TX_GET_STATE	0xe2
-
-/* ldc_tx_set_qtail()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_TX_SET_QTAIL
- * ARG0:	channel ID
- * ARG1:	tail offset
- * RET0:	status
- *
- * Update the tail pointer for the transmit queue associated with the LDC
- * endpoint defined by the given channel ID.  The tail offset specified
- * must be aligned on a 64 byte boundary, and calculated so as to increase
- * the number of pending entries on the transmit queue.  Any attempt to
- * decrease the number of pending transmit queue entires is considered
- * an invalid tail offset and will result in an EINVAL error.
- *
- * Since the tail of the transmit queue may not be moved backwards, the
- * transmit queue may be flushed by configuring a new transmit queue,
- * whereupon the hypervisor will configure the initial transmit head and
- * tail pointers to be equal.
- */
-#define HV_FAST_LDC_TX_SET_QTAIL	0xe3
-
-/* ldc_rx_qconf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_RX_QCONF
- * ARG0:	channel ID
- * ARG1:	real address base of queue
- * ARG2:	num entries in queue
- * RET0:	status
- *
- * Configure receive queue for the LDC endpoint specified by the
- * given channel ID, to be placed at the given real address, and
- * be of the given num entries.  Num entries must be a power of two.
- * The real address base of the queue must be aligned on the queue
- * size.  Each queue entry is 64-bytes, so for example, a 32 entry
- * queue must be aligned on a 2048 byte real address boundary.
- *
- * The endpoint's transmit queue is un-configured if num entries is zero.
- *
- * If a valid receive queue is specified for a local endpoint the LDC is
- * in the up state for the purpose of transmission to this endpoint.
- *
- * The maximum number of entries for each queue for a specific cpu may be
- * determined from the machine description.
- *
- * As receive queue configuration causes a reset of the queue's head and
- * tail pointers there is no way for a gues to determine how many entries
- * have been received between a preceeding ldc_get_rx_state() API call
- * and the completion of the configuration operation.  It should be noted
- * that datagram delivery is not guarenteed via domain channels anyway,
- * and therefore any higher protocol should be resilient to datagram
- * loss if necessary.  However, to overcome this specific race potential
- * it is recommended, for example, that a higher level protocol be employed
- * to ensure either retransmission, or ensure that no datagrams are pending
- * on the peer endpoint's transmit queue prior to the configuration process.
- */
-#define HV_FAST_LDC_RX_QCONF		0xe4
-
-/* ldc_rx_qinfo()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_RX_QINFO
- * ARG0:	channel ID
- * RET0:	status
- * RET1:	real address base of queue
- * RET2:	num entries in queue
- *
- * Return the configuration info for the receive queue of LDC endpoint
- * defined by the given channel ID.  The real address is the currently
- * defined real address base of the defined queue, and num entries is the
- * size of the queue in terms of number of entries.
- *
- * If the specified channel ID is a valid endpoint number, but no receive
- * queue has been defined this service will return success, but with num
- * entries set to zero and the real address will have an undefined value.
- */
-#define HV_FAST_LDC_RX_QINFO		0xe5
-
-/* ldc_rx_get_state()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_RX_GET_STATE
- * ARG0:	channel ID
- * RET0:	status
- * RET1:	head offset
- * RET2:	tail offset
- * RET3:	channel state
- *
- * Return the receive state, and the head and tail queue pointers, for
- * the receive queue of the LDC endpoint defined by the given channel ID.
- * The head and tail values are the byte offset of the head and tail
- * positions of the receive queue for the specified endpoint.
- */
-#define HV_FAST_LDC_RX_GET_STATE	0xe6
-
-/* ldc_rx_set_qhead()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_RX_SET_QHEAD
- * ARG0:	channel ID
- * ARG1:	head offset
- * RET0:	status
- *
- * Update the head pointer for the receive queue associated with the LDC
- * endpoint defined by the given channel ID.  The head offset specified
- * must be aligned on a 64 byte boundary, and calculated so as to decrease
- * the number of pending entries on the receive queue.  Any attempt to
- * increase the number of pending receive queue entires is considered
- * an invalid head offset and will result in an EINVAL error.
- *
- * The receive queue may be flushed by setting the head offset equal
- * to the current tail offset.
- */
-#define HV_FAST_LDC_RX_SET_QHEAD	0xe7
-
-/* LDC Map Table Entry.  Each slot is defined by a translation table
- * entry, as specified by the LDC_MTE_* bits below, and a 64-bit
- * hypervisor invalidation cookie.
- */
-#define LDC_MTE_PADDR	0x0fffffffffffe000 /* pa[55:13]          */
-#define LDC_MTE_COPY_W	0x0000000000000400 /* copy write access  */
-#define LDC_MTE_COPY_R	0x0000000000000200 /* copy read access   */
-#define LDC_MTE_IOMMU_W	0x0000000000000100 /* IOMMU write access */
-#define LDC_MTE_IOMMU_R	0x0000000000000080 /* IOMMU read access  */
-#define LDC_MTE_EXEC	0x0000000000000040 /* execute            */
-#define LDC_MTE_WRITE	0x0000000000000020 /* read               */
-#define LDC_MTE_READ	0x0000000000000010 /* write              */
-#define LDC_MTE_SZALL	0x000000000000000f /* page size bits     */
-#define LDC_MTE_SZ16GB	0x0000000000000007 /* 16GB page          */
-#define LDC_MTE_SZ2GB	0x0000000000000006 /* 2GB page           */
-#define LDC_MTE_SZ256MB	0x0000000000000005 /* 256MB page         */
-#define LDC_MTE_SZ32MB	0x0000000000000004 /* 32MB page          */
-#define LDC_MTE_SZ4MB	0x0000000000000003 /* 4MB page           */
-#define LDC_MTE_SZ512K	0x0000000000000002 /* 512K page          */
-#define LDC_MTE_SZ64K	0x0000000000000001 /* 64K page           */
-#define LDC_MTE_SZ8K	0x0000000000000000 /* 8K page            */
-
-#ifndef __ASSEMBLY__
-struct ldc_mtable_entry {
-	unsigned long	mte;
-	unsigned long	cookie;
-};
-#endif
-
-/* ldc_set_map_table()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_SET_MAP_TABLE
- * ARG0:	channel ID
- * ARG1:	table real address
- * ARG2:	num entries
- * RET0:	status
- *
- * Register the MTE table at the given table real address, with the
- * specified num entries, for the LDC indicated by the given channel
- * ID.
- */
-#define HV_FAST_LDC_SET_MAP_TABLE	0xea
-
-/* ldc_get_map_table()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_GET_MAP_TABLE
- * ARG0:	channel ID
- * RET0:	status
- * RET1:	table real address
- * RET2:	num entries
- *
- * Return the configuration of the current mapping table registered
- * for the given channel ID.
- */
-#define HV_FAST_LDC_GET_MAP_TABLE	0xeb
-
-#define LDC_COPY_IN	0
-#define LDC_COPY_OUT	1
-
-/* ldc_copy()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_COPY
- * ARG0:	channel ID
- * ARG1:	LDC_COPY_* direction code
- * ARG2:	target real address
- * ARG3:	local real address
- * ARG4:	length in bytes
- * RET0:	status
- * RET1:	actual length in bytes
- */
-#define HV_FAST_LDC_COPY		0xec
-
-#define LDC_MEM_READ	1
-#define LDC_MEM_WRITE	2
-#define LDC_MEM_EXEC	4
-
-/* ldc_mapin()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_MAPIN
- * ARG0:	channel ID
- * ARG1:	cookie
- * RET0:	status
- * RET1:	real address
- * RET2:	LDC_MEM_* permissions
- */
-#define HV_FAST_LDC_MAPIN		0xed
-
-/* ldc_unmap()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_UNMAP
- * ARG0:	real address
- * RET0:	status
- */
-#define HV_FAST_LDC_UNMAP		0xee
-
-/* ldc_revoke()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_LDC_REVOKE
- * ARG0:	channel ID
- * ARG1:	cookie
- * ARG2:	ldc_mtable_entry cookie
- * RET0:	status
- */
-#define HV_FAST_LDC_REVOKE		0xef
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_ldc_tx_qconf(unsigned long channel,
-					unsigned long ra,
-					unsigned long num_entries);
-extern unsigned long sun4v_ldc_tx_qinfo(unsigned long channel,
-					unsigned long *ra,
-					unsigned long *num_entries);
-extern unsigned long sun4v_ldc_tx_get_state(unsigned long channel,
-					    unsigned long *head_off,
-					    unsigned long *tail_off,
-					    unsigned long *chan_state);
-extern unsigned long sun4v_ldc_tx_set_qtail(unsigned long channel,
-					    unsigned long tail_off);
-extern unsigned long sun4v_ldc_rx_qconf(unsigned long channel,
-					unsigned long ra,
-					unsigned long num_entries);
-extern unsigned long sun4v_ldc_rx_qinfo(unsigned long channel,
-					unsigned long *ra,
-					unsigned long *num_entries);
-extern unsigned long sun4v_ldc_rx_get_state(unsigned long channel,
-					    unsigned long *head_off,
-					    unsigned long *tail_off,
-					    unsigned long *chan_state);
-extern unsigned long sun4v_ldc_rx_set_qhead(unsigned long channel,
-					    unsigned long head_off);
-extern unsigned long sun4v_ldc_set_map_table(unsigned long channel,
-					     unsigned long ra,
-					     unsigned long num_entries);
-extern unsigned long sun4v_ldc_get_map_table(unsigned long channel,
-					     unsigned long *ra,
-					     unsigned long *num_entries);
-extern unsigned long sun4v_ldc_copy(unsigned long channel,
-				    unsigned long dir_code,
-				    unsigned long tgt_raddr,
-				    unsigned long lcl_raddr,
-				    unsigned long len,
-				    unsigned long *actual_len);
-extern unsigned long sun4v_ldc_mapin(unsigned long channel,
-				     unsigned long cookie,
-				     unsigned long *ra,
-				     unsigned long *perm);
-extern unsigned long sun4v_ldc_unmap(unsigned long ra);
-extern unsigned long sun4v_ldc_revoke(unsigned long channel,
-				      unsigned long cookie,
-				      unsigned long mte_cookie);
-#endif
-
-/* Performance counter services.  */
-
-#define HV_PERF_JBUS_PERF_CTRL_REG	0x00
-#define HV_PERF_JBUS_PERF_CNT_REG	0x01
-#define HV_PERF_DRAM_PERF_CTRL_REG_0	0x02
-#define HV_PERF_DRAM_PERF_CNT_REG_0	0x03
-#define HV_PERF_DRAM_PERF_CTRL_REG_1	0x04
-#define HV_PERF_DRAM_PERF_CNT_REG_1	0x05
-#define HV_PERF_DRAM_PERF_CTRL_REG_2	0x06
-#define HV_PERF_DRAM_PERF_CNT_REG_2	0x07
-#define HV_PERF_DRAM_PERF_CTRL_REG_3	0x08
-#define HV_PERF_DRAM_PERF_CNT_REG_3	0x09
-
-/* get_perfreg()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_GET_PERFREG
- * ARG0:	performance reg number
- * RET0:	status
- * RET1:	performance reg value
- * ERRORS:	EINVAL		Invalid performance register number
- *		ENOACCESS	No access allowed to performance counters
- *
- * Read the value of the given DRAM/JBUS performance counter/control register.
- */
-#define HV_FAST_GET_PERFREG		0x100
-
-/* set_perfreg()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_SET_PERFREG
- * ARG0:	performance reg number
- * ARG1:	performance reg value
- * RET0:	status
- * ERRORS:	EINVAL		Invalid performance register number
- *		ENOACCESS	No access allowed to performance counters
- *
- * Write the given performance reg value to the given DRAM/JBUS
- * performance counter/control register.
- */
-#define HV_FAST_SET_PERFREG		0x101
-
-/* MMU statistics services.
- *
- * The hypervisor maintains MMU statistics and privileged code provides
- * a buffer where these statistics can be collected.  It is continually
- * updated once configured.  The layout is as follows:
- */
-#ifndef __ASSEMBLY__
-struct hv_mmu_statistics {
-	unsigned long immu_tsb_hits_ctx0_8k_tte;
-	unsigned long immu_tsb_ticks_ctx0_8k_tte;
-	unsigned long immu_tsb_hits_ctx0_64k_tte;
-	unsigned long immu_tsb_ticks_ctx0_64k_tte;
-	unsigned long __reserved1[2];
-	unsigned long immu_tsb_hits_ctx0_4mb_tte;
-	unsigned long immu_tsb_ticks_ctx0_4mb_tte;
-	unsigned long __reserved2[2];
-	unsigned long immu_tsb_hits_ctx0_256mb_tte;
-	unsigned long immu_tsb_ticks_ctx0_256mb_tte;
-	unsigned long __reserved3[4];
-	unsigned long immu_tsb_hits_ctxnon0_8k_tte;
-	unsigned long immu_tsb_ticks_ctxnon0_8k_tte;
-	unsigned long immu_tsb_hits_ctxnon0_64k_tte;
-	unsigned long immu_tsb_ticks_ctxnon0_64k_tte;
-	unsigned long __reserved4[2];
-	unsigned long immu_tsb_hits_ctxnon0_4mb_tte;
-	unsigned long immu_tsb_ticks_ctxnon0_4mb_tte;
-	unsigned long __reserved5[2];
-	unsigned long immu_tsb_hits_ctxnon0_256mb_tte;
-	unsigned long immu_tsb_ticks_ctxnon0_256mb_tte;
-	unsigned long __reserved6[4];
-	unsigned long dmmu_tsb_hits_ctx0_8k_tte;
-	unsigned long dmmu_tsb_ticks_ctx0_8k_tte;
-	unsigned long dmmu_tsb_hits_ctx0_64k_tte;
-	unsigned long dmmu_tsb_ticks_ctx0_64k_tte;
-	unsigned long __reserved7[2];
-	unsigned long dmmu_tsb_hits_ctx0_4mb_tte;
-	unsigned long dmmu_tsb_ticks_ctx0_4mb_tte;
-	unsigned long __reserved8[2];
-	unsigned long dmmu_tsb_hits_ctx0_256mb_tte;
-	unsigned long dmmu_tsb_ticks_ctx0_256mb_tte;
-	unsigned long __reserved9[4];
-	unsigned long dmmu_tsb_hits_ctxnon0_8k_tte;
-	unsigned long dmmu_tsb_ticks_ctxnon0_8k_tte;
-	unsigned long dmmu_tsb_hits_ctxnon0_64k_tte;
-	unsigned long dmmu_tsb_ticks_ctxnon0_64k_tte;
-	unsigned long __reserved10[2];
-	unsigned long dmmu_tsb_hits_ctxnon0_4mb_tte;
-	unsigned long dmmu_tsb_ticks_ctxnon0_4mb_tte;
-	unsigned long __reserved11[2];
-	unsigned long dmmu_tsb_hits_ctxnon0_256mb_tte;
-	unsigned long dmmu_tsb_ticks_ctxnon0_256mb_tte;
-	unsigned long __reserved12[4];
-};
-#endif
-
-/* mmustat_conf()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMUSTAT_CONF
- * ARG0:	real address
- * RET0:	status
- * RET1:	real address
- * ERRORS:	ENORADDR	Invalid real address
- *		EBADALIGN	Real address not aligned on 64-byte boundary
- *		EBADTRAP	API not supported on this processor
- *
- * Enable MMU statistic gathering using the buffer at the given real
- * address on the current virtual CPU.  The new buffer real address
- * is given in ARG1, and the previously specified buffer real address
- * is returned in RET1, or is returned as zero for the first invocation.
- *
- * If the passed in real address argument is zero, this will disable
- * MMU statistic collection on the current virtual CPU.  If an error is
- * returned then no statistics are collected.
- *
- * The buffer contents should be initialized to all zeros before being
- * given to the hypervisor or else the statistics will be meaningless.
- */
-#define HV_FAST_MMUSTAT_CONF		0x102
-
-/* mmustat_info()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_MMUSTAT_INFO
- * RET0:	status
- * RET1:	real address
- * ERRORS:	EBADTRAP	API not supported on this processor
- *
- * Return the current state and real address of the currently configured
- * MMU statistics buffer on the current virtual CPU.
- */
-#define HV_FAST_MMUSTAT_INFO		0x103
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_mmustat_conf(unsigned long ra, unsigned long *orig_ra);
-extern unsigned long sun4v_mmustat_info(unsigned long *ra);
-#endif
-
-/* NCS crypto services  */
-
-/* ncs_request() sub-function numbers */
-#define HV_NCS_QCONF			0x01
-#define HV_NCS_QTAIL_UPDATE		0x02
-
-#ifndef __ASSEMBLY__
-struct hv_ncs_queue_entry {
-	/* MAU Control Register */
-	unsigned long	mau_control;
-#define MAU_CONTROL_INV_PARITY	0x0000000000002000
-#define MAU_CONTROL_STRAND	0x0000000000001800
-#define MAU_CONTROL_BUSY	0x0000000000000400
-#define MAU_CONTROL_INT		0x0000000000000200
-#define MAU_CONTROL_OP		0x00000000000001c0
-#define MAU_CONTROL_OP_SHIFT	6
-#define MAU_OP_LOAD_MA_MEMORY	0x0
-#define MAU_OP_STORE_MA_MEMORY	0x1
-#define MAU_OP_MODULAR_MULT	0x2
-#define MAU_OP_MODULAR_REDUCE	0x3
-#define MAU_OP_MODULAR_EXP_LOOP	0x4
-#define MAU_CONTROL_LEN		0x000000000000003f
-#define MAU_CONTROL_LEN_SHIFT	0
-
-	/* Real address of bytes to load or store bytes
-	 * into/out-of the MAU.
-	 */
-	unsigned long	mau_mpa;
-
-	/* Modular Arithmetic MA Offset Register.  */
-	unsigned long	mau_ma;
-
-	/* Modular Arithmetic N Prime Register.  */
-	unsigned long	mau_np;
-};
-
-struct hv_ncs_qconf_arg {
-	unsigned long	mid;      /* MAU ID, 1 per core on Niagara */
-	unsigned long	base;     /* Real address base of queue */
-	unsigned long	end;	  /* Real address end of queue */
-	unsigned long	num_ents; /* Number of entries in queue */
-};
-
-struct hv_ncs_qtail_update_arg {
-	unsigned long	mid;      /* MAU ID, 1 per core on Niagara */
-	unsigned long	tail;     /* New tail index to use */
-	unsigned long	syncflag; /* only SYNCFLAG_SYNC is implemented */
-#define HV_NCS_SYNCFLAG_SYNC	0x00
-#define HV_NCS_SYNCFLAG_ASYNC	0x01
-};
-#endif
-
-/* ncs_request()
- * TRAP:	HV_FAST_TRAP
- * FUNCTION:	HV_FAST_NCS_REQUEST
- * ARG0:	NCS sub-function
- * ARG1:	sub-function argument real address
- * ARG2:	size in bytes of sub-function argument
- * RET0:	status
- *
- * The MAU chip of the Niagara processor is not directly accessible
- * to privileged code, instead it is programmed indirectly via this
- * hypervisor API.
- *
- * The interfaces defines a queue of MAU operations to perform.
- * Privileged code registers a queue with the hypervisor by invoking
- * this HVAPI with the HV_NCS_QCONF sub-function, which defines the
- * base, end, and number of entries of the queue.  Each queue entry
- * contains a MAU register struct block.
- *
- * The privileged code then proceeds to add entries to the queue and
- * then invoke the HV_NCS_QTAIL_UPDATE sub-function.  Since only
- * synchronous operations are supported by the current hypervisor,
- * HV_NCS_QTAIL_UPDATE will run all the pending queue entries to
- * completion and return HV_EOK, or return an error code.
- *
- * The real address of the sub-function argument must be aligned on at
- * least an 8-byte boundary.
- *
- * The tail argument of HV_NCS_QTAIL_UPDATE is an index, not a byte
- * offset, into the queue and must be less than or equal the 'num_ents'
- * argument given in the HV_NCS_QCONF call.
- */
-#define HV_FAST_NCS_REQUEST		0x110
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_ncs_request(unsigned long request,
-				       unsigned long arg_ra,
-				       unsigned long arg_size);
-#endif
-
-#define HV_FAST_FIRE_GET_PERFREG	0x120
-#define HV_FAST_FIRE_SET_PERFREG	0x121
-
-/* Function numbers for HV_CORE_TRAP.  */
-#define HV_CORE_SET_VER			0x00
-#define HV_CORE_PUTCHAR			0x01
-#define HV_CORE_EXIT			0x02
-#define HV_CORE_GET_VER			0x03
-
-/* Hypervisor API groups for use with HV_CORE_SET_VER and
- * HV_CORE_GET_VER.
- */
-#define HV_GRP_SUN4V			0x0000
-#define HV_GRP_CORE			0x0001
-#define HV_GRP_INTR			0x0002
-#define HV_GRP_SOFT_STATE		0x0003
-#define HV_GRP_PCI			0x0100
-#define HV_GRP_LDOM			0x0101
-#define HV_GRP_SVC_CHAN			0x0102
-#define HV_GRP_NCS			0x0103
-#define HV_GRP_NIAG_PERF		0x0200
-#define HV_GRP_FIRE_PERF		0x0201
-#define HV_GRP_DIAG			0x0300
-
-#ifndef __ASSEMBLY__
-extern unsigned long sun4v_get_version(unsigned long group,
-				       unsigned long *major,
-				       unsigned long *minor);
-extern unsigned long sun4v_set_version(unsigned long group,
-				       unsigned long major,
-				       unsigned long minor,
-				       unsigned long *actual_minor);
-
-extern int sun4v_hvapi_register(unsigned long group, unsigned long major,
-				unsigned long *minor);
-extern void sun4v_hvapi_unregister(unsigned long group);
-extern int sun4v_hvapi_get(unsigned long group,
-			   unsigned long *major,
-			   unsigned long *minor);
-extern void sun4v_hvapi_init(void);
-#endif
-
-#endif /* !(_SPARC64_HYPERVISOR_H) */
+#include <asm-sparc/hypervisor.h>
diff --git a/include/asm-sparc64/ide.h b/include/asm-sparc64/ide.h
index 1282676da1c..7125317a428 100644
--- a/include/asm-sparc64/ide.h
+++ b/include/asm-sparc64/ide.h
@@ -1,118 +1 @@
-/*
- * ide.h: Ultra/PCI specific IDE glue.
- *
- * Copyright (C) 1997  David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1998  Eddie C. Dost   (ecd@skynet.be)
- */
-
-#ifndef _SPARC64_IDE_H
-#define _SPARC64_IDE_H
-
-#ifdef __KERNEL__
-
-#include <asm/pgalloc.h>
-#include <asm/io.h>
-#include <asm/spitfire.h>
-#include <asm/cacheflush.h>
-#include <asm/page.h>
-
-#ifndef MAX_HWIFS
-# ifdef CONFIG_BLK_DEV_IDEPCI
-#define MAX_HWIFS	10
-# else
-#define MAX_HWIFS	2
-# endif
-#endif
-
-#define __ide_insl(data_reg, buffer, wcount) \
-	__ide_insw(data_reg, buffer, (wcount)<<1)
-#define __ide_outsl(data_reg, buffer, wcount) \
-	__ide_outsw(data_reg, buffer, (wcount)<<1)
-
-/* On sparc64, I/O ports and MMIO registers are accessed identically.  */
-#define __ide_mm_insw	__ide_insw
-#define __ide_mm_insl	__ide_insl
-#define __ide_mm_outsw	__ide_outsw
-#define __ide_mm_outsl	__ide_outsl
-
-static inline unsigned int inw_be(void __iomem *addr)
-{
-	unsigned int ret;
-
-	__asm__ __volatile__("lduha [%1] %2, %0"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline void __ide_insw(void __iomem *port, void *dst, u32 count)
-{
-#ifdef DCACHE_ALIASING_POSSIBLE
-	unsigned long end = (unsigned long)dst + (count << 1);
-#endif
-	u16 *ps = dst;
-	u32 *pi;
-
-	if(((u64)ps) & 0x2) {
-		*ps++ = inw_be(port);
-		count--;
-	}
-	pi = (u32 *)ps;
-	while(count >= 2) {
-		u32 w;
-
-		w  = inw_be(port) << 16;
-		w |= inw_be(port);
-		*pi++ = w;
-		count -= 2;
-	}
-	ps = (u16 *)pi;
-	if(count)
-		*ps++ = inw_be(port);
-
-#ifdef DCACHE_ALIASING_POSSIBLE
-	__flush_dcache_range((unsigned long)dst, end);
-#endif
-}
-
-static inline void outw_be(unsigned short w, void __iomem *addr)
-{
-	__asm__ __volatile__("stha %0, [%1] %2"
-			     : /* no outputs */
-			     : "r" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void __ide_outsw(void __iomem *port, void *src, u32 count)
-{
-#ifdef DCACHE_ALIASING_POSSIBLE
-	unsigned long end = (unsigned long)src + (count << 1);
-#endif
-	const u16 *ps = src;
-	const u32 *pi;
-
-	if(((u64)src) & 0x2) {
-		outw_be(*ps++, port);
-		count--;
-	}
-	pi = (const u32 *)ps;
-	while(count >= 2) {
-		u32 w;
-
-		w = *pi++;
-		outw_be((w >> 16), port);
-		outw_be(w, port);
-		count -= 2;
-	}
-	ps = (const u16 *)pi;
-	if(count)
-		outw_be(*ps, port);
-
-#ifdef DCACHE_ALIASING_POSSIBLE
-	__flush_dcache_range((unsigned long)src, end);
-#endif
-}
-
-#endif /* __KERNEL__ */
-
-#endif /* _SPARC64_IDE_H */
+#include <asm-sparc/ide.h>
diff --git a/include/asm-sparc64/idprom.h b/include/asm-sparc64/idprom.h
index a363fa0a112..c22f9c30bc7 100644
--- a/include/asm-sparc64/idprom.h
+++ b/include/asm-sparc64/idprom.h
@@ -1,25 +1 @@
-/*
- * idprom.h: Macros and defines for idprom routines
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC64_IDPROM_H
-#define _SPARC64_IDPROM_H
-
-#include <linux/types.h>
-
-struct idprom {
-	u8		id_format;	/* Format identifier (always 0x01) */
-	u8		id_machtype;	/* Machine type */
-	u8		id_ethaddr[6];	/* Hardware ethernet address */
-	s32		id_date;	/* Date of manufacture */
-	u32		id_sernum:24;	/* Unique serial number */
-	u8		id_cksum;	/* Checksum - xor of the data bytes */
-	u8		reserved[16];
-};
-
-extern struct idprom *idprom;
-extern void idprom_init(void);
-
-#endif /* !(_SPARC_IDPROM_H) */
+#include <asm-sparc/idprom.h>
diff --git a/include/asm-sparc64/intr_queue.h b/include/asm-sparc64/intr_queue.h
index 206077dedc2..f7225015b3d 100644
--- a/include/asm-sparc64/intr_queue.h
+++ b/include/asm-sparc64/intr_queue.h
@@ -1,15 +1 @@
-#ifndef _SPARC64_INTR_QUEUE_H
-#define _SPARC64_INTR_QUEUE_H
-
-/* Sun4v interrupt queue registers, accessed via ASI_QUEUE.  */
-
-#define INTRQ_CPU_MONDO_HEAD	  0x3c0 /* CPU mondo head	          */
-#define INTRQ_CPU_MONDO_TAIL	  0x3c8 /* CPU mondo tail	          */
-#define INTRQ_DEVICE_MONDO_HEAD	  0x3d0 /* Device mondo head	          */
-#define INTRQ_DEVICE_MONDO_TAIL	  0x3d8 /* Device mondo tail	          */
-#define INTRQ_RESUM_MONDO_HEAD	  0x3e0 /* Resumable error mondo head     */
-#define INTRQ_RESUM_MONDO_TAIL	  0x3e8 /* Resumable error mondo tail     */
-#define INTRQ_NONRESUM_MONDO_HEAD 0x3f0 /* Non-resumable error mondo head */
-#define INTRQ_NONRESUM_MONDO_TAIL 0x3f8 /* Non-resumable error mondo head */
-
-#endif /* !(_SPARC64_INTR_QUEUE_H) */
+#include <asm-sparc/intr_queue.h>
diff --git a/include/asm-sparc64/io.h b/include/asm-sparc64/io.h
index 3158960f3eb..25ff258dfd3 100644
--- a/include/asm-sparc64/io.h
+++ b/include/asm-sparc64/io.h
@@ -1,511 +1 @@
-#ifndef __SPARC64_IO_H
-#define __SPARC64_IO_H
-
-#include <linux/kernel.h>
-#include <linux/compiler.h>
-#include <linux/types.h>
-
-#include <asm/page.h>      /* IO address mapping routines need this */
-#include <asm/system.h>
-#include <asm/asi.h>
-
-/* PC crapola... */
-#define __SLOW_DOWN_IO	do { } while (0)
-#define SLOW_DOWN_IO	do { } while (0)
-
-/* BIO layer definitions. */
-extern unsigned long kern_base, kern_size;
-#define page_to_phys(page)	(page_to_pfn(page) << PAGE_SHIFT)
-#define BIO_VMERGE_BOUNDARY	8192
-
-static inline u8 _inb(unsigned long addr)
-{
-	u8 ret;
-
-	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_inb */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u16 _inw(unsigned long addr)
-{
-	u16 ret;
-
-	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_inw */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u32 _inl(unsigned long addr)
-{
-	u32 ret;
-
-	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_inl */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline void _outb(u8 b, unsigned long addr)
-{
-	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_outb */"
-			     : /* no outputs */
-			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-static inline void _outw(u16 w, unsigned long addr)
-{
-	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_outw */"
-			     : /* no outputs */
-			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-static inline void _outl(u32 l, unsigned long addr)
-{
-	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_outl */"
-			     : /* no outputs */
-			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-#define inb(__addr)		(_inb((unsigned long)(__addr)))
-#define inw(__addr)		(_inw((unsigned long)(__addr)))
-#define inl(__addr)		(_inl((unsigned long)(__addr)))
-#define outb(__b, __addr)	(_outb((u8)(__b), (unsigned long)(__addr)))
-#define outw(__w, __addr)	(_outw((u16)(__w), (unsigned long)(__addr)))
-#define outl(__l, __addr)	(_outl((u32)(__l), (unsigned long)(__addr)))
-
-#define inb_p(__addr) 		inb(__addr)
-#define outb_p(__b, __addr)	outb(__b, __addr)
-#define inw_p(__addr)		inw(__addr)
-#define outw_p(__w, __addr)	outw(__w, __addr)
-#define inl_p(__addr)		inl(__addr)
-#define outl_p(__l, __addr)	outl(__l, __addr)
-
-extern void outsb(unsigned long, const void *, unsigned long);
-extern void outsw(unsigned long, const void *, unsigned long);
-extern void outsl(unsigned long, const void *, unsigned long);
-extern void insb(unsigned long, void *, unsigned long);
-extern void insw(unsigned long, void *, unsigned long);
-extern void insl(unsigned long, void *, unsigned long);
-
-static inline void ioread8_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insb((unsigned long __force)port, buf, count);
-}
-static inline void ioread16_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insw((unsigned long __force)port, buf, count);
-}
-
-static inline void ioread32_rep(void __iomem *port, void *buf, unsigned long count)
-{
-	insl((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite8_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsb((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite16_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsw((unsigned long __force)port, buf, count);
-}
-
-static inline void iowrite32_rep(void __iomem *port, const void *buf, unsigned long count)
-{
-	outsl((unsigned long __force)port, buf, count);
-}
-
-/* Memory functions, same as I/O accesses on Ultra. */
-static inline u8 _readb(const volatile void __iomem *addr)
-{	u8 ret;
-
-	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_readb */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-	return ret;
-}
-
-static inline u16 _readw(const volatile void __iomem *addr)
-{	u16 ret;
-
-	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_readw */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u32 _readl(const volatile void __iomem *addr)
-{	u32 ret;
-
-	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_readl */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u64 _readq(const volatile void __iomem *addr)
-{	u64 ret;
-
-	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* pci_readq */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-
-	return ret;
-}
-
-static inline void _writeb(u8 b, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_writeb */"
-			     : /* no outputs */
-			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-static inline void _writew(u16 w, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_writew */"
-			     : /* no outputs */
-			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-static inline void _writel(u32 l, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_writel */"
-			     : /* no outputs */
-			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-static inline void _writeq(u64 q, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* pci_writeq */"
-			     : /* no outputs */
-			     : "Jr" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E_L)
-			     : "memory");
-}
-
-#define readb(__addr)		_readb(__addr)
-#define readw(__addr)		_readw(__addr)
-#define readl(__addr)		_readl(__addr)
-#define readq(__addr)		_readq(__addr)
-#define readb_relaxed(__addr)	_readb(__addr)
-#define readw_relaxed(__addr)	_readw(__addr)
-#define readl_relaxed(__addr)	_readl(__addr)
-#define readq_relaxed(__addr)	_readq(__addr)
-#define writeb(__b, __addr)	_writeb(__b, __addr)
-#define writew(__w, __addr)	_writew(__w, __addr)
-#define writel(__l, __addr)	_writel(__l, __addr)
-#define writeq(__q, __addr)	_writeq(__q, __addr)
-
-/* Now versions without byte-swapping. */
-static inline u8 _raw_readb(unsigned long addr)
-{
-	u8 ret;
-
-	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* pci_raw_readb */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline u16 _raw_readw(unsigned long addr)
-{
-	u16 ret;
-
-	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* pci_raw_readw */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline u32 _raw_readl(unsigned long addr)
-{
-	u32 ret;
-
-	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* pci_raw_readl */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline u64 _raw_readq(unsigned long addr)
-{
-	u64 ret;
-
-	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* pci_raw_readq */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline void _raw_writeb(u8 b, unsigned long addr)
-{
-	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* pci_raw_writeb */"
-			     : /* no outputs */
-			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _raw_writew(u16 w, unsigned long addr)
-{
-	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* pci_raw_writew */"
-			     : /* no outputs */
-			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _raw_writel(u32 l, unsigned long addr)
-{
-	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* pci_raw_writel */"
-			     : /* no outputs */
-			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _raw_writeq(u64 q, unsigned long addr)
-{
-	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* pci_raw_writeq */"
-			     : /* no outputs */
-			     : "Jr" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-#define __raw_readb(__addr)		(_raw_readb((unsigned long)(__addr)))
-#define __raw_readw(__addr)		(_raw_readw((unsigned long)(__addr)))
-#define __raw_readl(__addr)		(_raw_readl((unsigned long)(__addr)))
-#define __raw_readq(__addr)		(_raw_readq((unsigned long)(__addr)))
-#define __raw_writeb(__b, __addr)	(_raw_writeb((u8)(__b), (unsigned long)(__addr)))
-#define __raw_writew(__w, __addr)	(_raw_writew((u16)(__w), (unsigned long)(__addr)))
-#define __raw_writel(__l, __addr)	(_raw_writel((u32)(__l), (unsigned long)(__addr)))
-#define __raw_writeq(__q, __addr)	(_raw_writeq((u64)(__q), (unsigned long)(__addr)))
-
-/* Valid I/O Space regions are anywhere, because each PCI bus supported
- * can live in an arbitrary area of the physical address range.
- */
-#define IO_SPACE_LIMIT 0xffffffffffffffffUL
-
-/* Now, SBUS variants, only difference from PCI is that we do
- * not use little-endian ASIs.
- */
-static inline u8 _sbus_readb(const volatile void __iomem *addr)
-{
-	u8 ret;
-
-	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* sbus_readb */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u16 _sbus_readw(const volatile void __iomem *addr)
-{
-	u16 ret;
-
-	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* sbus_readw */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u32 _sbus_readl(const volatile void __iomem *addr)
-{
-	u32 ret;
-
-	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* sbus_readl */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-
-	return ret;
-}
-
-static inline u64 _sbus_readq(const volatile void __iomem *addr)
-{
-	u64 ret;
-
-	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* sbus_readq */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-
-	return ret;
-}
-
-static inline void _sbus_writeb(u8 b, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stba\t%r0, [%1] %2\t/* sbus_writeb */"
-			     : /* no outputs */
-			     : "Jr" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-}
-
-static inline void _sbus_writew(u16 w, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stha\t%r0, [%1] %2\t/* sbus_writew */"
-			     : /* no outputs */
-			     : "Jr" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-}
-
-static inline void _sbus_writel(u32 l, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stwa\t%r0, [%1] %2\t/* sbus_writel */"
-			     : /* no outputs */
-			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-}
-
-static inline void _sbus_writeq(u64 l, volatile void __iomem *addr)
-{
-	__asm__ __volatile__("stxa\t%r0, [%1] %2\t/* sbus_writeq */"
-			     : /* no outputs */
-			     : "Jr" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E)
-			     : "memory");
-}
-
-#define sbus_readb(__addr)		_sbus_readb(__addr)
-#define sbus_readw(__addr)		_sbus_readw(__addr)
-#define sbus_readl(__addr)		_sbus_readl(__addr)
-#define sbus_readq(__addr)		_sbus_readq(__addr)
-#define sbus_writeb(__b, __addr)	_sbus_writeb(__b, __addr)
-#define sbus_writew(__w, __addr)	_sbus_writew(__w, __addr)
-#define sbus_writel(__l, __addr)	_sbus_writel(__l, __addr)
-#define sbus_writeq(__l, __addr)	_sbus_writeq(__l, __addr)
-
-static inline void _sbus_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
-{
-	while(n--) {
-		sbus_writeb(c, dst);
-		dst++;
-	}
-}
-
-#define sbus_memset_io(d,c,sz)	_sbus_memset_io(d,c,sz)
-
-static inline void
-_memset_io(volatile void __iomem *dst, int c, __kernel_size_t n)
-{
-	volatile void __iomem *d = dst;
-
-	while (n--) {
-		writeb(c, d);
-		d++;
-	}
-}
-
-#define memset_io(d,c,sz)	_memset_io(d,c,sz)
-
-static inline void
-_memcpy_fromio(void *dst, const volatile void __iomem *src, __kernel_size_t n)
-{
-	char *d = dst;
-
-	while (n--) {
-		char tmp = readb(src);
-		*d++ = tmp;
-		src++;
-	}
-}
-
-#define memcpy_fromio(d,s,sz)	_memcpy_fromio(d,s,sz)
-
-static inline void 
-_memcpy_toio(volatile void __iomem *dst, const void *src, __kernel_size_t n)
-{
-	const char *s = src;
-	volatile void __iomem *d = dst;
-
-	while (n--) {
-		char tmp = *s++;
-		writeb(tmp, d);
-		d++;
-	}
-}
-
-#define memcpy_toio(d,s,sz)	_memcpy_toio(d,s,sz)
-
-#define mmiowb()
-
-#ifdef __KERNEL__
-
-/* On sparc64 we have the whole physical IO address space accessible
- * using physically addressed loads and stores, so this does nothing.
- */
-static inline void __iomem *ioremap(unsigned long offset, unsigned long size)
-{
-	return (void __iomem *)offset;
-}
-
-#define ioremap_nocache(X,Y)		ioremap((X),(Y))
-
-static inline void iounmap(volatile void __iomem *addr)
-{
-}
-
-#define ioread8(X)			readb(X)
-#define ioread16(X)			readw(X)
-#define ioread32(X)			readl(X)
-#define iowrite8(val,X)			writeb(val,X)
-#define iowrite16(val,X)		writew(val,X)
-#define iowrite32(val,X)		writel(val,X)
-
-/* Create a virtual mapping cookie for an IO port range */
-extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
-extern void ioport_unmap(void __iomem *);
-
-/* Create a virtual mapping cookie for a PCI BAR (memory or IO) */
-struct pci_dev;
-extern void __iomem *pci_iomap(struct pci_dev *dev, int bar, unsigned long max);
-extern void pci_iounmap(struct pci_dev *dev, void __iomem *);
-
-/* Similarly for SBUS. */
-#define sbus_ioremap(__res, __offset, __size, __name) \
-({	unsigned long __ret; \
-	__ret  = (__res)->start + (((__res)->flags & 0x1ffUL) << 32UL); \
-	__ret += (unsigned long) (__offset); \
-	if (! request_region((__ret), (__size), (__name))) \
-		__ret = 0UL; \
-	(void __iomem *) __ret; \
-})
-
-#define sbus_iounmap(__addr, __size)	\
-	release_region((unsigned long)(__addr), (__size))
-
-/*
- * Convert a physical pointer to a virtual kernel pointer for /dev/mem
- * access
- */
-#define xlate_dev_mem_ptr(p)	__va(p)
-
-/*
- * Convert a virtual cached pointer to an uncached pointer
- */
-#define xlate_dev_kmem_ptr(p)	p
-
-#endif
-
-#endif /* !(__SPARC64_IO_H) */
+#include <asm-sparc/io.h>
diff --git a/include/asm-sparc64/ioctl.h b/include/asm-sparc64/ioctl.h
index d634c21b472..18fc5623ff5 100644
--- a/include/asm-sparc64/ioctl.h
+++ b/include/asm-sparc64/ioctl.h
@@ -1,67 +1 @@
-#ifndef _SPARC64_IOCTL_H
-#define _SPARC64_IOCTL_H
-
-/*
- * Our DIR and SIZE overlap in order to simulteneously provide
- * a non-zero _IOC_NONE (for binary compatibility) and
- * 14 bits of size as on i386. Here's the layout:
- *
- *   0xE0000000   DIR
- *   0x80000000     DIR = WRITE
- *   0x40000000     DIR = READ
- *   0x20000000     DIR = NONE
- *   0x3FFF0000   SIZE (overlaps NONE bit)
- *   0x0000FF00   TYPE
- *   0x000000FF   NR (CMD)
- */
-
-#define _IOC_NRBITS      8
-#define _IOC_TYPEBITS    8
-#define _IOC_SIZEBITS   13	/* Actually 14, see below. */
-#define _IOC_DIRBITS     3
-
-#define _IOC_NRMASK      ((1 << _IOC_NRBITS)-1)
-#define _IOC_TYPEMASK    ((1 << _IOC_TYPEBITS)-1)
-#define _IOC_SIZEMASK    ((1 << _IOC_SIZEBITS)-1)
-#define _IOC_XSIZEMASK   ((1 << (_IOC_SIZEBITS+1))-1)
-#define _IOC_DIRMASK     ((1 << _IOC_DIRBITS)-1)
-
-#define _IOC_NRSHIFT     0
-#define _IOC_TYPESHIFT   (_IOC_NRSHIFT + _IOC_NRBITS)
-#define _IOC_SIZESHIFT   (_IOC_TYPESHIFT + _IOC_TYPEBITS)
-#define _IOC_DIRSHIFT    (_IOC_SIZESHIFT + _IOC_SIZEBITS)
-
-#define _IOC_NONE        1U
-#define _IOC_READ        2U
-#define _IOC_WRITE       4U
-
-#define _IOC(dir,type,nr,size) \
-        (((dir)  << _IOC_DIRSHIFT) | \
-         ((type) << _IOC_TYPESHIFT) | \
-         ((nr)   << _IOC_NRSHIFT) | \
-         ((size) << _IOC_SIZESHIFT))
-
-#define _IO(type,nr)        _IOC(_IOC_NONE,(type),(nr),0)
-#define _IOR(type,nr,size)  _IOC(_IOC_READ,(type),(nr),sizeof(size))
-#define _IOW(type,nr,size)  _IOC(_IOC_WRITE,(type),(nr),sizeof(size))
-#define _IOWR(type,nr,size) _IOC(_IOC_READ|_IOC_WRITE,(type),(nr),sizeof(size))
-
-/* Used to decode ioctl numbers in drivers despite the leading underscore... */
-#define _IOC_DIR(nr)    \
- ( (((((nr) >> _IOC_DIRSHIFT) & _IOC_DIRMASK) & (_IOC_WRITE|_IOC_READ)) != 0)?   \
-                            (((nr) >> _IOC_DIRSHIFT) & (_IOC_WRITE|_IOC_READ)):  \
-                            (((nr) >> _IOC_DIRSHIFT) & _IOC_DIRMASK) )
-#define _IOC_TYPE(nr)       (((nr) >> _IOC_TYPESHIFT) & _IOC_TYPEMASK)
-#define _IOC_NR(nr)         (((nr) >> _IOC_NRSHIFT) & _IOC_NRMASK)
-#define _IOC_SIZE(nr)   \
- ((((((nr) >> _IOC_DIRSHIFT) & _IOC_DIRMASK) & (_IOC_WRITE|_IOC_READ)) == 0)?    \
-                         0: (((nr) >> _IOC_SIZESHIFT) & _IOC_XSIZEMASK))
-
-/* ...and for the PCMCIA and sound. */
-#define IOC_IN          (_IOC_WRITE << _IOC_DIRSHIFT)
-#define IOC_OUT         (_IOC_READ << _IOC_DIRSHIFT)
-#define IOC_INOUT       ((_IOC_WRITE|_IOC_READ) << _IOC_DIRSHIFT)
-#define IOCSIZE_MASK    (_IOC_XSIZEMASK << _IOC_SIZESHIFT)
-#define IOCSIZE_SHIFT   (_IOC_SIZESHIFT)
-
-#endif /* !(_SPARC64_IOCTL_H) */
+#include <asm-sparc/ioctl.h>
diff --git a/include/asm-sparc64/ioctls.h b/include/asm-sparc64/ioctls.h
index 94d1b75e512..dcd5540ec10 100644
--- a/include/asm-sparc64/ioctls.h
+++ b/include/asm-sparc64/ioctls.h
@@ -1,136 +1 @@
-#ifndef _ASM_SPARC64_IOCTLS_H
-#define _ASM_SPARC64_IOCTLS_H
-
-#include <asm/ioctl.h>
-
-/* Big T */
-#define TCGETA		_IOR('T', 1, struct termio)
-#define TCSETA		_IOW('T', 2, struct termio)
-#define TCSETAW		_IOW('T', 3, struct termio)
-#define TCSETAF		_IOW('T', 4, struct termio)
-#define TCSBRK		_IO('T', 5)
-#define TCXONC		_IO('T', 6)
-#define TCFLSH		_IO('T', 7)
-#define TCGETS		_IOR('T', 8, struct termios)
-#define TCSETS		_IOW('T', 9, struct termios)
-#define TCSETSW		_IOW('T', 10, struct termios)
-#define TCSETSF		_IOW('T', 11, struct termios)
-#define TCGETS2		_IOR('T', 12, struct termios2)
-#define TCSETS2		_IOW('T', 13, struct termios2)
-#define TCSETSW2	_IOW('T', 14, struct termios2)
-#define TCSETSF2	_IOW('T', 15, struct termios2)
-
-/* Note that all the ioctls that are not available in Linux have a 
- * double underscore on the front to: a) avoid some programs to
- * think we support some ioctls under Linux (autoconfiguration stuff)
- */
-/* Little t */
-#define TIOCGETD	_IOR('t', 0, int)
-#define TIOCSETD	_IOW('t', 1, int)
-#define __TIOCHPCL        _IO('t', 2) /* SunOS Specific */
-#define __TIOCMODG        _IOR('t', 3, int) /* SunOS Specific */
-#define __TIOCMODS        _IOW('t', 4, int) /* SunOS Specific */
-#define __TIOCGETP        _IOR('t', 8, struct sgttyb) /* SunOS Specific */
-#define __TIOCSETP        _IOW('t', 9, struct sgttyb) /* SunOS Specific */
-#define __TIOCSETN        _IOW('t', 10, struct sgttyb) /* SunOS Specific */
-#define TIOCEXCL	_IO('t', 13)
-#define TIOCNXCL	_IO('t', 14)
-#define __TIOCFLUSH       _IOW('t', 16, int) /* SunOS Specific */
-#define __TIOCSETC        _IOW('t', 17, struct tchars) /* SunOS Specific */
-#define __TIOCGETC        _IOR('t', 18, struct tchars) /* SunOS Specific */
-#define __TIOCTCNTL       _IOW('t', 32, int) /* SunOS Specific */
-#define __TIOCSIGNAL      _IOW('t', 33, int) /* SunOS Specific */
-#define __TIOCSETX        _IOW('t', 34, int) /* SunOS Specific */
-#define __TIOCGETX        _IOR('t', 35, int) /* SunOS Specific */
-#define TIOCCONS	_IO('t', 36)
-#define TIOCGSOFTCAR	_IOR('t', 100, int)
-#define TIOCSSOFTCAR	_IOW('t', 101, int)
-#define __TIOCUCNTL       _IOW('t', 102, int) /* SunOS Specific */
-#define TIOCSWINSZ	_IOW('t', 103, struct winsize)
-#define TIOCGWINSZ	_IOR('t', 104, struct winsize)
-#define __TIOCREMOTE      _IOW('t', 105, int) /* SunOS Specific */
-#define TIOCMGET	_IOR('t', 106, int)
-#define TIOCMBIC	_IOW('t', 107, int)
-#define TIOCMBIS	_IOW('t', 108, int)
-#define TIOCMSET	_IOW('t', 109, int)
-#define TIOCSTART       _IO('t', 110)
-#define TIOCSTOP        _IO('t', 111)
-#define TIOCPKT		_IOW('t', 112, int)
-#define TIOCNOTTY	_IO('t', 113)
-#define TIOCSTI		_IOW('t', 114, char)
-#define TIOCOUTQ	_IOR('t', 115, int)
-#define __TIOCGLTC        _IOR('t', 116, struct ltchars) /* SunOS Specific */
-#define __TIOCSLTC        _IOW('t', 117, struct ltchars) /* SunOS Specific */
-/* 118 is the non-posix setpgrp tty ioctl */
-/* 119 is the non-posix getpgrp tty ioctl */
-#define __TIOCCDTR        _IO('t', 120) /* SunOS Specific */
-#define __TIOCSDTR        _IO('t', 121) /* SunOS Specific */
-#define TIOCCBRK        _IO('t', 122)
-#define TIOCSBRK        _IO('t', 123)
-#define __TIOCLGET        _IOW('t', 124, int) /* SunOS Specific */
-#define __TIOCLSET        _IOW('t', 125, int) /* SunOS Specific */
-#define __TIOCLBIC        _IOW('t', 126, int) /* SunOS Specific */
-#define __TIOCLBIS        _IOW('t', 127, int) /* SunOS Specific */
-#define __TIOCISPACE      _IOR('t', 128, int) /* SunOS Specific */
-#define __TIOCISIZE       _IOR('t', 129, int) /* SunOS Specific */
-#define TIOCSPGRP	_IOW('t', 130, int)
-#define TIOCGPGRP	_IOR('t', 131, int)
-#define TIOCSCTTY	_IO('t', 132)
-#define TIOCGSID	_IOR('t', 133, int)
-/* Get minor device of a pty master's FD -- Solaris equiv is ISPTM */
-#define TIOCGPTN	_IOR('t', 134, unsigned int) /* Get Pty Number */
-#define TIOCSPTLCK	_IOW('t', 135, int) /* Lock/unlock PTY */
-
-/* Little f */
-#define FIOCLEX		_IO('f', 1)
-#define FIONCLEX	_IO('f', 2)
-#define FIOASYNC	_IOW('f', 125, int)
-#define FIONBIO		_IOW('f', 126, int)
-#define FIONREAD	_IOR('f', 127, int)
-#define TIOCINQ		FIONREAD
-#define FIOQSIZE	_IOR('f', 128, loff_t)
-
-/* SCARY Rutgers local SunOS kernel hackery, perhaps I will support it
- * someday.  This is completely bogus, I know...
- */
-#define __TCGETSTAT       _IO('T', 200) /* Rutgers specific */
-#define __TCSETSTAT       _IO('T', 201) /* Rutgers specific */
-
-/* Linux specific, no SunOS equivalent. */
-#define TIOCLINUX	0x541C
-#define TIOCGSERIAL	0x541E
-#define TIOCSSERIAL	0x541F
-#define TCSBRKP		0x5425
-#define TIOCSERCONFIG	0x5453
-#define TIOCSERGWILD	0x5454
-#define TIOCSERSWILD	0x5455
-#define TIOCGLCKTRMIOS	0x5456
-#define TIOCSLCKTRMIOS	0x5457
-#define TIOCSERGSTRUCT	0x5458 /* For debugging only */
-#define TIOCSERGETLSR   0x5459 /* Get line status register */
-#define TIOCSERGETMULTI 0x545A /* Get multiport config  */
-#define TIOCSERSETMULTI 0x545B /* Set multiport config */
-#define TIOCMIWAIT	0x545C /* Wait for change on serial input line(s) */
-#define TIOCGICOUNT	0x545D /* Read serial port inline interrupt counts */
-
-/* Kernel definitions */
-#ifdef __KERNEL__
-#define TIOCGETC __TIOCGETC
-#define TIOCGETP __TIOCGETP
-#define TIOCGLTC __TIOCGLTC
-#define TIOCSLTC __TIOCSLTC
-#define TIOCSETP __TIOCSETP
-#define TIOCSETN __TIOCSETN
-#define TIOCSETC __TIOCSETC
-#endif
-
-/* Used for packet mode */
-#define TIOCPKT_DATA		 0
-#define TIOCPKT_FLUSHREAD	 1
-#define TIOCPKT_FLUSHWRITE	 2
-#define TIOCPKT_STOP		 4
-#define TIOCPKT_START		 8
-#define TIOCPKT_NOSTOP		16
-#define TIOCPKT_DOSTOP		32
-
-#endif /* !(_ASM_SPARC64_IOCTLS_H) */
+#include <asm-sparc/ioctls.h>
diff --git a/include/asm-sparc64/iommu.h b/include/asm-sparc64/iommu.h
index d7b9afcba08..76252bb85e9 100644
--- a/include/asm-sparc64/iommu.h
+++ b/include/asm-sparc64/iommu.h
@@ -1,62 +1 @@
-/* iommu.h: Definitions for the sun5 IOMMU.
- *
- * Copyright (C) 1996, 1999, 2007 David S. Miller (davem@davemloft.net)
- */
-#ifndef _SPARC64_IOMMU_H
-#define _SPARC64_IOMMU_H
-
-/* The format of an iopte in the page tables. */
-#define IOPTE_VALID   0x8000000000000000UL
-#define IOPTE_64K     0x2000000000000000UL
-#define IOPTE_STBUF   0x1000000000000000UL
-#define IOPTE_INTRA   0x0800000000000000UL
-#define IOPTE_CONTEXT 0x07ff800000000000UL
-#define IOPTE_PAGE    0x00007fffffffe000UL
-#define IOPTE_CACHE   0x0000000000000010UL
-#define IOPTE_WRITE   0x0000000000000002UL
-
-#define IOMMU_NUM_CTXS	4096
-
-struct iommu_arena {
-	unsigned long	*map;
-	unsigned int	hint;
-	unsigned int	limit;
-};
-
-struct iommu {
-	spinlock_t		lock;
-	struct iommu_arena	arena;
-	void			(*flush_all)(struct iommu *);
-	iopte_t			*page_table;
-	u32			page_table_map_base;
-	unsigned long		iommu_control;
-	unsigned long		iommu_tsbbase;
-	unsigned long		iommu_flush;
-	unsigned long		iommu_flushinv;
-	unsigned long		iommu_tags;
-	unsigned long		iommu_ctxflush;
-	unsigned long		write_complete_reg;
-	unsigned long		dummy_page;
-	unsigned long		dummy_page_pa;
-	unsigned long		ctx_lowest_free;
-	DECLARE_BITMAP(ctx_bitmap, IOMMU_NUM_CTXS);
-	u32			dma_addr_mask;
-};
-
-struct strbuf {
-	int			strbuf_enabled;
-	unsigned long		strbuf_control;
-	unsigned long		strbuf_pflush;
-	unsigned long		strbuf_fsync;
-	unsigned long		strbuf_ctxflush;
-	unsigned long		strbuf_ctxmatch_base;
-	unsigned long		strbuf_flushflag_pa;
-	volatile unsigned long *strbuf_flushflag;
-	volatile unsigned long	__flushflag_buf[(64+(64-1)) / sizeof(long)];
-};
-
-extern int iommu_table_init(struct iommu *iommu, int tsbsize,
-			    u32 dma_offset, u32 dma_addr_mask,
-			    int numa_node);
-
-#endif /* !(_SPARC64_IOMMU_H) */
+#include <asm-sparc/iommu.h>
diff --git a/include/asm-sparc64/ipcbuf.h b/include/asm-sparc64/ipcbuf.h
index 9c5bf1bc423..41dfaf1149b 100644
--- a/include/asm-sparc64/ipcbuf.h
+++ b/include/asm-sparc64/ipcbuf.h
@@ -1,28 +1 @@
-#ifndef _SPARC64_IPCBUF_H
-#define _SPARC64_IPCBUF_H
-
-/* 
- * The ipc64_perm structure for sparc64 architecture.
- * Note extra padding because this structure is passed back and forth
- * between kernel and user space.
- *
- * Pad space is left for:
- * - 32-bit seq
- * - 2 miscellaneous 64-bit values
- */
-
-struct ipc64_perm
-{
-	__kernel_key_t	key;
-	__kernel_uid_t	uid;
-	__kernel_gid_t	gid;
-	__kernel_uid_t	cuid;
-	__kernel_gid_t	cgid;
-	__kernel_mode_t	mode; 
-	unsigned short	__pad1;
-	unsigned short	seq;
-	unsigned long	__unused1;
-	unsigned long	__unused2;
-};
-
-#endif /* _SPARC64_IPCBUF_H */
+#include <asm-sparc/ipcbuf.h>
diff --git a/include/asm-sparc64/irq.h b/include/asm-sparc64/irq.h
index 0bb9bf53174..b2102e65947 100644
--- a/include/asm-sparc64/irq.h
+++ b/include/asm-sparc64/irq.h
@@ -1,93 +1 @@
-/* irq.h: IRQ registers on the 64-bit Sparc.
- *
- * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
- * Copyright (C) 1998 Jakub Jelinek (jj@ultra.linux.cz)
- */
-
-#ifndef _SPARC64_IRQ_H
-#define _SPARC64_IRQ_H
-
-#include <linux/linkage.h>
-#include <linux/kernel.h>
-#include <linux/errno.h>
-#include <linux/interrupt.h>
-#include <asm/pil.h>
-#include <asm/ptrace.h>
-
-/* IMAP/ICLR register defines */
-#define IMAP_VALID		0x80000000UL	/* IRQ Enabled		*/
-#define IMAP_TID_UPA		0x7c000000UL	/* UPA TargetID		*/
-#define IMAP_TID_JBUS		0x7c000000UL	/* JBUS TargetID	*/
-#define IMAP_TID_SHIFT		26
-#define IMAP_AID_SAFARI		0x7c000000UL	/* Safari AgentID	*/
-#define IMAP_AID_SHIFT		26
-#define IMAP_NID_SAFARI		0x03e00000UL	/* Safari NodeID	*/
-#define IMAP_NID_SHIFT		21
-#define IMAP_IGN		0x000007c0UL	/* IRQ Group Number	*/
-#define IMAP_INO		0x0000003fUL	/* IRQ Number		*/
-#define IMAP_INR		0x000007ffUL	/* Full interrupt number*/
-
-#define ICLR_IDLE		0x00000000UL	/* Idle state		*/
-#define ICLR_TRANSMIT		0x00000001UL	/* Transmit state	*/
-#define ICLR_PENDING		0x00000003UL	/* Pending state	*/
-
-/* The largest number of unique interrupt sources we support.
- * If this needs to ever be larger than 255, you need to change
- * the type of ino_bucket->virt_irq as appropriate.
- *
- * ino_bucket->virt_irq allocation is made during {sun4v_,}build_irq().
- */
-#define NR_IRQS    255
-
-extern void irq_install_pre_handler(int virt_irq,
-				    void (*func)(unsigned int, void *, void *),
-				    void *arg1, void *arg2);
-#define irq_canonicalize(irq)	(irq)
-extern unsigned int build_irq(int inofixup, unsigned long iclr, unsigned long imap);
-extern unsigned int sun4v_build_irq(u32 devhandle, unsigned int devino);
-extern unsigned int sun4v_build_virq(u32 devhandle, unsigned int devino);
-extern unsigned int sun4v_build_msi(u32 devhandle, unsigned int *virt_irq_p,
-				    unsigned int msi_devino_start,
-				    unsigned int msi_devino_end);
-extern void sun4v_destroy_msi(unsigned int virt_irq);
-extern unsigned int sun4u_build_msi(u32 portid, unsigned int *virt_irq_p,
-				    unsigned int msi_devino_start,
-				    unsigned int msi_devino_end,
-				    unsigned long imap_base,
-				    unsigned long iclr_base);
-extern void sun4u_destroy_msi(unsigned int virt_irq);
-extern unsigned int sbus_build_irq(void *sbus, unsigned int ino);
-
-extern unsigned char virt_irq_alloc(unsigned int dev_handle,
-				    unsigned int dev_ino);
-#ifdef CONFIG_PCI_MSI
-extern void virt_irq_free(unsigned int virt_irq);
-#endif
-
-extern void __init init_IRQ(void);
-extern void fixup_irqs(void);
-
-static inline void set_softint(unsigned long bits)
-{
-	__asm__ __volatile__("wr	%0, 0x0, %%set_softint"
-			     : /* No outputs */
-			     : "r" (bits));
-}
-
-static inline void clear_softint(unsigned long bits)
-{
-	__asm__ __volatile__("wr	%0, 0x0, %%clear_softint"
-			     : /* No outputs */
-			     : "r" (bits));
-}
-
-static inline unsigned long get_softint(void)
-{
-	unsigned long retval;
-
-	__asm__ __volatile__("rd	%%softint, %0"
-			     : "=r" (retval));
-	return retval;
-}
-
-#endif
+#include <asm-sparc/irq.h>
diff --git a/include/asm-sparc64/irq_regs.h b/include/asm-sparc64/irq_regs.h
index 3dd9c0b7027..1e2b8a1e745 100644
--- a/include/asm-sparc64/irq_regs.h
+++ b/include/asm-sparc64/irq_regs.h
@@ -1 +1 @@
-#include <asm-generic/irq_regs.h>
+#include <asm-sparc/irq_regs.h>
diff --git a/include/asm-sparc64/irqflags.h b/include/asm-sparc64/irqflags.h
index 024fc54d068..27b091fc3fa 100644
--- a/include/asm-sparc64/irqflags.h
+++ b/include/asm-sparc64/irqflags.h
@@ -1,89 +1 @@
-/*
- * include/asm-sparc64/irqflags.h
- *
- * IRQ flags handling
- *
- * This file gets included from lowlevel asm headers too, to provide
- * wrapped versions of the local_irq_*() APIs, based on the
- * raw_local_irq_*() functions from the lowlevel headers.
- */
-#ifndef _ASM_IRQFLAGS_H
-#define _ASM_IRQFLAGS_H
-
-#ifndef __ASSEMBLY__
-
-static inline unsigned long __raw_local_save_flags(void)
-{
-	unsigned long flags;
-
-	__asm__ __volatile__(
-		"rdpr	%%pil, %0"
-		: "=r" (flags)
-	);
-
-	return flags;
-}
-
-#define raw_local_save_flags(flags) \
-		do { (flags) = __raw_local_save_flags(); } while (0)
-
-static inline void raw_local_irq_restore(unsigned long flags)
-{
-	__asm__ __volatile__(
-		"wrpr	%0, %%pil"
-		: /* no output */
-		: "r" (flags)
-		: "memory"
-	);
-}
-
-static inline void raw_local_irq_disable(void)
-{
-	__asm__ __volatile__(
-		"wrpr	15, %%pil"
-		: /* no outputs */
-		: /* no inputs */
-		: "memory"
-	);
-}
-
-static inline void raw_local_irq_enable(void)
-{
-	__asm__ __volatile__(
-		"wrpr	0, %%pil"
-		: /* no outputs */
-		: /* no inputs */
-		: "memory"
-	);
-}
-
-static inline int raw_irqs_disabled_flags(unsigned long flags)
-{
-	return (flags > 0);
-}
-
-static inline int raw_irqs_disabled(void)
-{
-	unsigned long flags = __raw_local_save_flags();
-
-	return raw_irqs_disabled_flags(flags);
-}
-
-/*
- * For spinlocks, etc:
- */
-static inline unsigned long __raw_local_irq_save(void)
-{
-	unsigned long flags = __raw_local_save_flags();
-
-	raw_local_irq_disable();
-
-	return flags;
-}
-
-#define raw_local_irq_save(flags) \
-		do { (flags) = __raw_local_irq_save(); } while (0)
-
-#endif /* (__ASSEMBLY__) */
-
-#endif /* !(_ASM_IRQFLAGS_H) */
+#include <asm-sparc/irqflags.h>
diff --git a/include/asm-sparc64/kdebug.h b/include/asm-sparc64/kdebug.h
index f905b773235..78cfd5d2749 100644
--- a/include/asm-sparc64/kdebug.h
+++ b/include/asm-sparc64/kdebug.h
@@ -1,19 +1 @@
-#ifndef _SPARC64_KDEBUG_H
-#define _SPARC64_KDEBUG_H
-
-struct pt_regs;
-
-extern void bad_trap(struct pt_regs *, long);
-
-/* Grossly misnamed. */
-enum die_val {
-	DIE_OOPS = 1,
-	DIE_DEBUG,	/* ta 0x70 */
-	DIE_DEBUG_2,	/* ta 0x71 */
-	DIE_DIE,
-	DIE_TRAP,
-	DIE_TRAP_TL1,
-	DIE_CALL,
-};
-
-#endif
+#include <asm-sparc/kdebug.h>
diff --git a/include/asm-sparc64/kmap_types.h b/include/asm-sparc64/kmap_types.h
index 34c1d3d9a3b..276530cf539 100644
--- a/include/asm-sparc64/kmap_types.h
+++ b/include/asm-sparc64/kmap_types.h
@@ -1,25 +1 @@
-#ifndef _ASM_KMAP_TYPES_H
-#define _ASM_KMAP_TYPES_H
-
-/* Dummy header just to define km_type.  None of this
- * is actually used on sparc64.  -DaveM
- */
-
-enum km_type {
-	KM_BOUNCE_READ,
-	KM_SKB_SUNRPC_DATA,
-	KM_SKB_DATA_SOFTIRQ,
-	KM_USER0,
-	KM_USER1,
-	KM_BIO_SRC_IRQ,
-	KM_BIO_DST_IRQ,
-	KM_PTE0,
-	KM_PTE1,
-	KM_IRQ0,
-	KM_IRQ1,
-	KM_SOFTIRQ0,
-	KM_SOFTIRQ1,
-	KM_TYPE_NR
-};
-
-#endif
+#include <asm-sparc/kmap_types.h>
diff --git a/include/asm-sparc64/kprobes.h b/include/asm-sparc64/kprobes.h
index 5879d71afda..c55e43e4d2a 100644
--- a/include/asm-sparc64/kprobes.h
+++ b/include/asm-sparc64/kprobes.h
@@ -1,49 +1 @@
-#ifndef _SPARC64_KPROBES_H
-#define _SPARC64_KPROBES_H
-
-#include <linux/types.h>
-#include <linux/percpu.h>
-
-typedef u32 kprobe_opcode_t;
-
-#define BREAKPOINT_INSTRUCTION   0x91d02070 /* ta 0x70 */
-#define BREAKPOINT_INSTRUCTION_2 0x91d02071 /* ta 0x71 */
-#define MAX_INSN_SIZE 2
-
-#define kretprobe_blacklist_size 0
-
-#define arch_remove_kprobe(p)	do {} while (0)
-
-#define flush_insn_slot(p)		\
-do { 	flushi(&(p)->ainsn.insn[0]);	\
-	flushi(&(p)->ainsn.insn[1]);	\
-} while (0)
-
-void kretprobe_trampoline(void);
-
-/* Architecture specific copy of original instruction*/
-struct arch_specific_insn {
-	/* copy of the original instruction */
-	kprobe_opcode_t insn[MAX_INSN_SIZE];
-};
-
-struct prev_kprobe {
-	struct kprobe *kp;
-	unsigned long status;
-	unsigned long orig_tnpc;
-	unsigned long orig_tstate_pil;
-};
-
-/* per-cpu kprobe control block */
-struct kprobe_ctlblk {
-	unsigned long kprobe_status;
-	unsigned long kprobe_orig_tnpc;
-	unsigned long kprobe_orig_tstate_pil;
-	struct pt_regs jprobe_saved_regs;
-	struct prev_kprobe prev_kprobe;
-};
-
-extern int kprobe_exceptions_notify(struct notifier_block *self,
-				    unsigned long val, void *data);
-extern int kprobe_fault_handler(struct pt_regs *regs, int trapnr);
-#endif /* _SPARC64_KPROBES_H */
+#include <asm-sparc/kprobes.h>
diff --git a/include/asm-sparc64/kvm.h b/include/asm-sparc64/kvm.h
index 380537a77bf..53564ad86b1 100644
--- a/include/asm-sparc64/kvm.h
+++ b/include/asm-sparc64/kvm.h
@@ -1,6 +1 @@
-#ifndef __LINUX_KVM_SPARC64_H
-#define __LINUX_KVM_SPARC64_H
-
-/* sparc64 does not support KVM */
-
-#endif
+#include <asm-sparc/kvm.h>
diff --git a/include/asm-sparc64/ldc.h b/include/asm-sparc64/ldc.h
index bdb524a7b81..40f3f231c45 100644
--- a/include/asm-sparc64/ldc.h
+++ b/include/asm-sparc64/ldc.h
@@ -1,138 +1 @@
-#ifndef _SPARC64_LDC_H
-#define _SPARC64_LDC_H
-
-#include <asm/hypervisor.h>
-
-extern int ldom_domaining_enabled;
-extern void ldom_set_var(const char *var, const char *value);
-extern void ldom_reboot(const char *boot_command);
-extern void ldom_power_off(void);
-
-/* The event handler will be evoked when link state changes
- * or data becomes available on the receive side.
- *
- * For non-RAW links, if the LDC_EVENT_RESET event arrives the
- * driver should reset all of it's internal state and reinvoke
- * ldc_connect() to try and bring the link up again.
- *
- * For RAW links, ldc_connect() is not used.  Instead the driver
- * just waits for the LDC_EVENT_UP event.
- */
-struct ldc_channel_config {
-	void (*event)(void *arg, int event);
-
-	u32			mtu;
-	unsigned int		rx_irq;
-	unsigned int		tx_irq;
-	u8			mode;
-#define LDC_MODE_RAW		0x00
-#define LDC_MODE_UNRELIABLE	0x01
-#define LDC_MODE_RESERVED	0x02
-#define LDC_MODE_STREAM		0x03
-
-	u8			debug;
-#define LDC_DEBUG_HS		0x01
-#define LDC_DEBUG_STATE		0x02
-#define LDC_DEBUG_RX		0x04
-#define LDC_DEBUG_TX		0x08
-#define LDC_DEBUG_DATA		0x10
-};
-
-#define LDC_EVENT_RESET		0x01
-#define LDC_EVENT_UP		0x02
-#define LDC_EVENT_DATA_READY	0x04
-
-#define LDC_STATE_INVALID	0x00
-#define LDC_STATE_INIT		0x01
-#define LDC_STATE_BOUND		0x02
-#define LDC_STATE_READY		0x03
-#define LDC_STATE_CONNECTED	0x04
-
-struct ldc_channel;
-
-/* Allocate state for a channel.  */
-extern struct ldc_channel *ldc_alloc(unsigned long id,
-				     const struct ldc_channel_config *cfgp,
-				     void *event_arg);
-
-/* Shut down and free state for a channel.  */
-extern void ldc_free(struct ldc_channel *lp);
-
-/* Register TX and RX queues of the link with the hypervisor.  */
-extern int ldc_bind(struct ldc_channel *lp, const char *name);
-
-/* For non-RAW protocols we need to complete a handshake before
- * communication can proceed.  ldc_connect() does that, if the
- * handshake completes successfully, an LDC_EVENT_UP event will
- * be sent up to the driver.
- */
-extern int ldc_connect(struct ldc_channel *lp);
-extern int ldc_disconnect(struct ldc_channel *lp);
-
-extern int ldc_state(struct ldc_channel *lp);
-
-/* Read and write operations.  Only valid when the link is up.  */
-extern int ldc_write(struct ldc_channel *lp, const void *buf,
-		     unsigned int size);
-extern int ldc_read(struct ldc_channel *lp, void *buf, unsigned int size);
-
-#define LDC_MAP_SHADOW	0x01
-#define LDC_MAP_DIRECT	0x02
-#define LDC_MAP_IO	0x04
-#define LDC_MAP_R	0x08
-#define LDC_MAP_W	0x10
-#define LDC_MAP_X	0x20
-#define LDC_MAP_RW	(LDC_MAP_R | LDC_MAP_W)
-#define LDC_MAP_RWX	(LDC_MAP_R | LDC_MAP_W | LDC_MAP_X)
-#define LDC_MAP_ALL	0x03f
-
-struct ldc_trans_cookie {
-	u64			cookie_addr;
-	u64			cookie_size;
-};
-
-struct scatterlist;
-extern int ldc_map_sg(struct ldc_channel *lp,
-		      struct scatterlist *sg, int num_sg,
-		      struct ldc_trans_cookie *cookies, int ncookies,
-		      unsigned int map_perm);
-
-extern int ldc_map_single(struct ldc_channel *lp,
-			  void *buf, unsigned int len,
-			  struct ldc_trans_cookie *cookies, int ncookies,
-			  unsigned int map_perm);
-
-extern void ldc_unmap(struct ldc_channel *lp, struct ldc_trans_cookie *cookies,
-		      int ncookies);
-
-extern int ldc_copy(struct ldc_channel *lp, int copy_dir,
-		    void *buf, unsigned int len, unsigned long offset,
-		    struct ldc_trans_cookie *cookies, int ncookies);
-
-static inline int ldc_get_dring_entry(struct ldc_channel *lp,
-				      void *buf, unsigned int len,
-				      unsigned long offset,
-				      struct ldc_trans_cookie *cookies,
-				      int ncookies)
-{
-	return ldc_copy(lp, LDC_COPY_IN, buf, len, offset, cookies, ncookies);
-}
-
-static inline int ldc_put_dring_entry(struct ldc_channel *lp,
-				      void *buf, unsigned int len,
-				      unsigned long offset,
-				      struct ldc_trans_cookie *cookies,
-				      int ncookies)
-{
-	return ldc_copy(lp, LDC_COPY_OUT, buf, len, offset, cookies, ncookies);
-}
-
-extern void *ldc_alloc_exp_dring(struct ldc_channel *lp, unsigned int len,
-				 struct ldc_trans_cookie *cookies,
-				 int *ncookies, unsigned int map_perm);
-
-extern void ldc_free_exp_dring(struct ldc_channel *lp, void *buf,
-			       unsigned int len,
-			       struct ldc_trans_cookie *cookies, int ncookies);
-
-#endif /* _SPARC64_LDC_H */
+#include <asm-sparc/ldc.h>
diff --git a/include/asm-sparc64/linkage.h b/include/asm-sparc64/linkage.h
index 291c2d01c44..3ea4fd13f19 100644
--- a/include/asm-sparc64/linkage.h
+++ b/include/asm-sparc64/linkage.h
@@ -1,6 +1 @@
-#ifndef __ASM_LINKAGE_H
-#define __ASM_LINKAGE_H
-
-/* Nothing to see here... */
-
-#endif
+#include <asm-sparc/linkage.h>
diff --git a/include/asm-sparc64/lmb.h b/include/asm-sparc64/lmb.h
index 6a352cbcf52..3d04981701e 100644
--- a/include/asm-sparc64/lmb.h
+++ b/include/asm-sparc64/lmb.h
@@ -1,10 +1 @@
-#ifndef _SPARC64_LMB_H
-#define _SPARC64_LMB_H
-
-#include <asm/oplib.h>
-
-#define LMB_DBG(fmt...) prom_printf(fmt)
-
-#define LMB_REAL_LIMIT	0
-
-#endif /* !(_SPARC64_LMB_H) */
+#include <asm-sparc/lmb.h>
diff --git a/include/asm-sparc64/lsu.h b/include/asm-sparc64/lsu.h
index 7190f8de90a..4e3d8b128a5 100644
--- a/include/asm-sparc64/lsu.h
+++ b/include/asm-sparc64/lsu.h
@@ -1,19 +1 @@
-#ifndef _SPARC64_LSU_H
-#define _SPARC64_LSU_H
-
-#include <linux/const.h>
-
-/* LSU Control Register */
-#define LSU_CONTROL_PM _AC(0x000001fe00000000,UL) /* Phys-watchpoint byte mask*/
-#define LSU_CONTROL_VM _AC(0x00000001fe000000,UL) /* Virt-watchpoint byte mask*/
-#define LSU_CONTROL_PR _AC(0x0000000001000000,UL) /* Phys-rd watchpoint enable*/
-#define LSU_CONTROL_PW _AC(0x0000000000800000,UL) /* Phys-wr watchpoint enable*/
-#define LSU_CONTROL_VR _AC(0x0000000000400000,UL) /* Virt-rd watchpoint enable*/
-#define LSU_CONTROL_VW _AC(0x0000000000200000,UL) /* Virt-wr watchpoint enable*/
-#define LSU_CONTROL_FM _AC(0x00000000000ffff0,UL) /* Parity mask enables.     */
-#define LSU_CONTROL_DM _AC(0x0000000000000008,UL) /* Data MMU enable.         */
-#define LSU_CONTROL_IM _AC(0x0000000000000004,UL) /* Instruction MMU enable.  */
-#define LSU_CONTROL_DC _AC(0x0000000000000002,UL) /* Data cache enable.       */
-#define LSU_CONTROL_IC _AC(0x0000000000000001,UL) /* Instruction cache enable.*/
-
-#endif /* !(_SPARC64_LSU_H) */
+#include <asm-sparc/lsu.h>
diff --git a/include/asm-sparc64/mc146818rtc.h b/include/asm-sparc64/mc146818rtc.h
index e9c0fcc25c6..97842e6ed1c 100644
--- a/include/asm-sparc64/mc146818rtc.h
+++ b/include/asm-sparc64/mc146818rtc.h
@@ -1,34 +1 @@
-/*
- * Machine dependent access functions for RTC registers.
- */
-#ifndef __ASM_SPARC64_MC146818RTC_H
-#define __ASM_SPARC64_MC146818RTC_H
-
-#include <asm/io.h>
-
-#ifndef RTC_PORT
-#ifdef CONFIG_PCI
-extern unsigned long ds1287_regs;
-#else
-#define ds1287_regs (0UL)
-#endif
-#define RTC_PORT(x)	(ds1287_regs + (x))
-#define RTC_ALWAYS_BCD	0
-#endif
-
-/*
- * The yet supported machines all access the RTC index register via
- * an ISA port access but the way to access the date register differs ...
- */
-#define CMOS_READ(addr) ({ \
-outb_p((addr),RTC_PORT(0)); \
-inb_p(RTC_PORT(1)); \
-})
-#define CMOS_WRITE(val, addr) ({ \
-outb_p((addr),RTC_PORT(0)); \
-outb_p((val),RTC_PORT(1)); \
-})
-
-#define RTC_IRQ 8
-
-#endif /* __ASM_SPARC64_MC146818RTC_H */
+#include <asm-sparc/mc146818rtc.h>
diff --git a/include/asm-sparc64/mdesc.h b/include/asm-sparc64/mdesc.h
index 1acc7272e53..165a1934728 100644
--- a/include/asm-sparc64/mdesc.h
+++ b/include/asm-sparc64/mdesc.h
@@ -1,78 +1 @@
-#ifndef _SPARC64_MDESC_H
-#define _SPARC64_MDESC_H
-
-#include <linux/types.h>
-#include <linux/cpumask.h>
-#include <asm/prom.h>
-
-struct mdesc_handle;
-
-/* Machine description operations are to be surrounded by grab and
- * release calls.  The mdesc_handle returned from the grab is
- * the first argument to all of the operational calls that work
- * on mdescs.
- */
-extern struct mdesc_handle *mdesc_grab(void);
-extern void mdesc_release(struct mdesc_handle *);
-
-#define MDESC_NODE_NULL		(~(u64)0)
-
-extern u64 mdesc_node_by_name(struct mdesc_handle *handle,
-			      u64 from_node, const char *name);
-#define mdesc_for_each_node_by_name(__hdl, __node, __name) \
-	for (__node = mdesc_node_by_name(__hdl, MDESC_NODE_NULL, __name); \
-	     (__node) != MDESC_NODE_NULL; \
-	     __node = mdesc_node_by_name(__hdl, __node, __name))
-
-/* Access to property values returned from mdesc_get_property() are
- * only valid inside of a mdesc_grab()/mdesc_release() sequence.
- * Once mdesc_release() is called, the memory backed up by these
- * pointers may reference freed up memory.
- *
- * Therefore callers must make copies of any property values
- * they need.
- *
- * These same rules apply to mdesc_node_name().
- */
-extern const void *mdesc_get_property(struct mdesc_handle *handle,
-				      u64 node, const char *name, int *lenp);
-extern const char *mdesc_node_name(struct mdesc_handle *hp, u64 node);
-
-/* MD arc iteration, the standard sequence is:
- *
- *	unsigned long arc;
- *	mdesc_for_each_arc(arc, handle, node, MDESC_ARC_TYPE_{FWD,BACK}) {
- *		unsigned long target = mdesc_arc_target(handle, arc);
- *		...
- *	}
- */
-
-#define MDESC_ARC_TYPE_FWD	"fwd"
-#define MDESC_ARC_TYPE_BACK	"back"
-
-extern u64 mdesc_next_arc(struct mdesc_handle *handle, u64 from,
-			  const char *arc_type);
-#define mdesc_for_each_arc(__arc, __hdl, __node, __type) \
-	for (__arc = mdesc_next_arc(__hdl, __node, __type); \
-	     (__arc) != MDESC_NODE_NULL; \
-	     __arc = mdesc_next_arc(__hdl, __arc, __type))
-
-extern u64 mdesc_arc_target(struct mdesc_handle *hp, u64 arc);
-
-extern void mdesc_update(void);
-
-struct mdesc_notifier_client {
-	void (*add)(struct mdesc_handle *handle, u64 node);
-	void (*remove)(struct mdesc_handle *handle, u64 node);
-
-	const char			*node_name;
-	struct mdesc_notifier_client	*next;
-};
-
-extern void mdesc_register_notifier(struct mdesc_notifier_client *client);
-
-extern void mdesc_fill_in_cpu_data(cpumask_t mask);
-
-extern void sun4v_mdesc_init(void);
-
-#endif
+#include <asm-sparc/mdesc.h>
diff --git a/include/asm-sparc64/mman.h b/include/asm-sparc64/mman.h
index d2ae67cd1bd..17ddb1724f5 100644
--- a/include/asm-sparc64/mman.h
+++ b/include/asm-sparc64/mman.h
@@ -1,31 +1 @@
-#ifndef __SPARC64_MMAN_H__
-#define __SPARC64_MMAN_H__
-
-#include <asm-generic/mman.h>
-
-/* SunOS'ified... */
-
-#define MAP_RENAME      MAP_ANONYMOUS   /* In SunOS terminology */
-#define MAP_NORESERVE   0x40            /* don't reserve swap pages */
-#define MAP_INHERIT     0x80            /* SunOS doesn't do this, but... */
-#define MAP_LOCKED      0x100           /* lock the mapping */
-#define _MAP_NEW        0x80000000      /* Binary compatibility is fun... */
-
-#define MAP_GROWSDOWN	0x0200		/* stack-like segment */
-#define MAP_DENYWRITE	0x0800		/* ETXTBSY */
-#define MAP_EXECUTABLE	0x1000		/* mark it as an executable */
-
-#define MCL_CURRENT     0x2000          /* lock all currently mapped pages */
-#define MCL_FUTURE      0x4000          /* lock all additions to address space */
-
-#define MAP_POPULATE	0x8000		/* populate (prefault) pagetables */
-#define MAP_NONBLOCK	0x10000		/* do not block on IO */
-
-#ifdef __KERNEL__
-#ifndef __ASSEMBLY__
-#define arch_mmap_check(addr,len,flags)	sparc64_mmap_check(addr,len)
-int sparc64_mmap_check(unsigned long addr, unsigned long len);
-#endif
-#endif
-
-#endif /* __SPARC64_MMAN_H__ */
+#include <asm-sparc/mman.h>
diff --git a/include/asm-sparc64/mmu.h b/include/asm-sparc64/mmu.h
index 8abc58f0f9d..e677a64d8db 100644
--- a/include/asm-sparc64/mmu.h
+++ b/include/asm-sparc64/mmu.h
@@ -1,127 +1 @@
-#ifndef __MMU_H
-#define __MMU_H
-
-#include <linux/const.h>
-#include <asm/page.h>
-#include <asm/hypervisor.h>
-
-#define CTX_NR_BITS		13
-
-#define TAG_CONTEXT_BITS	((_AC(1,UL) << CTX_NR_BITS) - _AC(1,UL))
-
-/* UltraSPARC-III+ and later have a feature whereby you can
- * select what page size the various Data-TLB instances in the
- * chip.  In order to gracefully support this, we put the version
- * field in a spot outside of the areas of the context register
- * where this parameter is specified.
- */
-#define CTX_VERSION_SHIFT	22
-#define CTX_VERSION_MASK	((~0UL) << CTX_VERSION_SHIFT)
-
-#define CTX_PGSZ_8KB		_AC(0x0,UL)
-#define CTX_PGSZ_64KB		_AC(0x1,UL)
-#define CTX_PGSZ_512KB		_AC(0x2,UL)
-#define CTX_PGSZ_4MB		_AC(0x3,UL)
-#define CTX_PGSZ_BITS		_AC(0x7,UL)
-#define CTX_PGSZ0_NUC_SHIFT	61
-#define CTX_PGSZ1_NUC_SHIFT	58
-#define CTX_PGSZ0_SHIFT		16
-#define CTX_PGSZ1_SHIFT		19
-#define CTX_PGSZ_MASK		((CTX_PGSZ_BITS << CTX_PGSZ0_SHIFT) | \
-				 (CTX_PGSZ_BITS << CTX_PGSZ1_SHIFT))
-
-#if defined(CONFIG_SPARC64_PAGE_SIZE_8KB)
-#define CTX_PGSZ_BASE	CTX_PGSZ_8KB
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_64KB)
-#define CTX_PGSZ_BASE	CTX_PGSZ_64KB
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_512KB)
-#define CTX_PGSZ_BASE	CTX_PGSZ_512KB
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_4MB)
-#define CTX_PGSZ_BASE	CTX_PGSZ_4MB
-#else
-#error No page size specified in kernel configuration
-#endif
-
-#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
-#define CTX_PGSZ_HUGE		CTX_PGSZ_4MB
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
-#define CTX_PGSZ_HUGE		CTX_PGSZ_512KB
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
-#define CTX_PGSZ_HUGE		CTX_PGSZ_64KB
-#endif
-
-#define CTX_PGSZ_KERN	CTX_PGSZ_4MB
-
-/* Thus, when running on UltraSPARC-III+ and later, we use the following
- * PRIMARY_CONTEXT register values for the kernel context.
- */
-#define CTX_CHEETAH_PLUS_NUC \
-	((CTX_PGSZ_KERN << CTX_PGSZ0_NUC_SHIFT) | \
-	 (CTX_PGSZ_BASE << CTX_PGSZ1_NUC_SHIFT))
-
-#define CTX_CHEETAH_PLUS_CTX0 \
-	((CTX_PGSZ_KERN << CTX_PGSZ0_SHIFT) | \
-	 (CTX_PGSZ_BASE << CTX_PGSZ1_SHIFT))
-
-/* If you want "the TLB context number" use CTX_NR_MASK.  If you
- * want "the bits I program into the context registers" use
- * CTX_HW_MASK.
- */
-#define CTX_NR_MASK		TAG_CONTEXT_BITS
-#define CTX_HW_MASK		(CTX_NR_MASK | CTX_PGSZ_MASK)
-
-#define CTX_FIRST_VERSION	((_AC(1,UL) << CTX_VERSION_SHIFT) + _AC(1,UL))
-#define CTX_VALID(__ctx)	\
-	 (!(((__ctx.sparc64_ctx_val) ^ tlb_context_cache) & CTX_VERSION_MASK))
-#define CTX_HWBITS(__ctx)	((__ctx.sparc64_ctx_val) & CTX_HW_MASK)
-#define CTX_NRBITS(__ctx)	((__ctx.sparc64_ctx_val) & CTX_NR_MASK)
-
-#ifndef __ASSEMBLY__
-
-#define TSB_ENTRY_ALIGNMENT	16
-
-struct tsb {
-	unsigned long tag;
-	unsigned long pte;
-} __attribute__((aligned(TSB_ENTRY_ALIGNMENT)));
-
-extern void __tsb_insert(unsigned long ent, unsigned long tag, unsigned long pte);
-extern void tsb_flush(unsigned long ent, unsigned long tag);
-extern void tsb_init(struct tsb *tsb, unsigned long size);
-
-struct tsb_config {
-	struct tsb		*tsb;
-	unsigned long		tsb_rss_limit;
-	unsigned long		tsb_nentries;
-	unsigned long		tsb_reg_val;
-	unsigned long		tsb_map_vaddr;
-	unsigned long		tsb_map_pte;
-};
-
-#define MM_TSB_BASE	0
-
-#ifdef CONFIG_HUGETLB_PAGE
-#define MM_TSB_HUGE	1
-#define MM_NUM_TSBS	2
-#else
-#define MM_NUM_TSBS	1
-#endif
-
-typedef struct {
-	spinlock_t		lock;
-	unsigned long		sparc64_ctx_val;
-	unsigned long		huge_pte_count;
-	struct tsb_config	tsb_block[MM_NUM_TSBS];
-	struct hv_tsb_descr	tsb_descr[MM_NUM_TSBS];
-} mm_context_t;
-
-#endif /* !__ASSEMBLY__ */
-
-#define TSB_CONFIG_TSB		0x00
-#define TSB_CONFIG_RSS_LIMIT	0x08
-#define TSB_CONFIG_NENTRIES	0x10
-#define TSB_CONFIG_REG_VAL	0x18
-#define TSB_CONFIG_MAP_VADDR	0x20
-#define TSB_CONFIG_MAP_PTE	0x28
-
-#endif /* __MMU_H */
+#include <asm-sparc/mmu.h>
diff --git a/include/asm-sparc64/mmu_context.h b/include/asm-sparc64/mmu_context.h
index 5693ab48260..877fee94bd4 100644
--- a/include/asm-sparc64/mmu_context.h
+++ b/include/asm-sparc64/mmu_context.h
@@ -1,155 +1 @@
-#ifndef __SPARC64_MMU_CONTEXT_H
-#define __SPARC64_MMU_CONTEXT_H
-
-/* Derived heavily from Linus's Alpha/AXP ASN code... */
-
-#ifndef __ASSEMBLY__
-
-#include <linux/spinlock.h>
-#include <asm/system.h>
-#include <asm/spitfire.h>
-#include <asm-generic/mm_hooks.h>
-
-static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
-{
-}
-
-extern spinlock_t ctx_alloc_lock;
-extern unsigned long tlb_context_cache;
-extern unsigned long mmu_context_bmap[];
-
-extern void get_new_mmu_context(struct mm_struct *mm);
-#ifdef CONFIG_SMP
-extern void smp_new_mmu_context_version(void);
-#else
-#define smp_new_mmu_context_version() do { } while (0)
-#endif
-
-extern int init_new_context(struct task_struct *tsk, struct mm_struct *mm);
-extern void destroy_context(struct mm_struct *mm);
-
-extern void __tsb_context_switch(unsigned long pgd_pa,
-				 struct tsb_config *tsb_base,
-				 struct tsb_config *tsb_huge,
-				 unsigned long tsb_descr_pa);
-
-static inline void tsb_context_switch(struct mm_struct *mm)
-{
-	__tsb_context_switch(__pa(mm->pgd),
-			     &mm->context.tsb_block[0],
-#ifdef CONFIG_HUGETLB_PAGE
-			     (mm->context.tsb_block[1].tsb ?
-			      &mm->context.tsb_block[1] :
-			      NULL)
-#else
-			     NULL
-#endif
-			     , __pa(&mm->context.tsb_descr[0]));
-}
-
-extern void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long mm_rss);
-#ifdef CONFIG_SMP
-extern void smp_tsb_sync(struct mm_struct *mm);
-#else
-#define smp_tsb_sync(__mm) do { } while (0)
-#endif
-
-/* Set MMU context in the actual hardware. */
-#define load_secondary_context(__mm) \
-	__asm__ __volatile__( \
-	"\n661:	stxa		%0, [%1] %2\n" \
-	"	.section	.sun4v_1insn_patch, \"ax\"\n" \
-	"	.word		661b\n" \
-	"	stxa		%0, [%1] %3\n" \
-	"	.previous\n" \
-	"	flush		%%g6\n" \
-	: /* No outputs */ \
-	: "r" (CTX_HWBITS((__mm)->context)), \
-	  "r" (SECONDARY_CONTEXT), "i" (ASI_DMMU), "i" (ASI_MMU))
-
-extern void __flush_tlb_mm(unsigned long, unsigned long);
-
-/* Switch the current MM context.  Interrupts are disabled.  */
-static inline void switch_mm(struct mm_struct *old_mm, struct mm_struct *mm, struct task_struct *tsk)
-{
-	unsigned long ctx_valid, flags;
-	int cpu;
-
-	if (unlikely(mm == &init_mm))
-		return;
-
-	spin_lock_irqsave(&mm->context.lock, flags);
-	ctx_valid = CTX_VALID(mm->context);
-	if (!ctx_valid)
-		get_new_mmu_context(mm);
-
-	/* We have to be extremely careful here or else we will miss
-	 * a TSB grow if we switch back and forth between a kernel
-	 * thread and an address space which has it's TSB size increased
-	 * on another processor.
-	 *
-	 * It is possible to play some games in order to optimize the
-	 * switch, but the safest thing to do is to unconditionally
-	 * perform the secondary context load and the TSB context switch.
-	 *
-	 * For reference the bad case is, for address space "A":
-	 *
-	 *		CPU 0			CPU 1
-	 *	run address space A
-	 *	set cpu0's bits in cpu_vm_mask
-	 *	switch to kernel thread, borrow
-	 *	address space A via entry_lazy_tlb
-	 *					run address space A
-	 *					set cpu1's bit in cpu_vm_mask
-	 *					flush_tlb_pending()
-	 *					reset cpu_vm_mask to just cpu1
-	 *					TSB grow
-	 *	run address space A
-	 *	context was valid, so skip
-	 *	TSB context switch
-	 *
-	 * At that point cpu0 continues to use a stale TSB, the one from
-	 * before the TSB grow performed on cpu1.  cpu1 did not cross-call
-	 * cpu0 to update it's TSB because at that point the cpu_vm_mask
-	 * only had cpu1 set in it.
-	 */
-	load_secondary_context(mm);
-	tsb_context_switch(mm);
-
-	/* Any time a processor runs a context on an address space
-	 * for the first time, we must flush that context out of the
-	 * local TLB.
-	 */
-	cpu = smp_processor_id();
-	if (!ctx_valid || !cpu_isset(cpu, mm->cpu_vm_mask)) {
-		cpu_set(cpu, mm->cpu_vm_mask);
-		__flush_tlb_mm(CTX_HWBITS(mm->context),
-			       SECONDARY_CONTEXT);
-	}
-	spin_unlock_irqrestore(&mm->context.lock, flags);
-}
-
-#define deactivate_mm(tsk,mm)	do { } while (0)
-
-/* Activate a new MM instance for the current task. */
-static inline void activate_mm(struct mm_struct *active_mm, struct mm_struct *mm)
-{
-	unsigned long flags;
-	int cpu;
-
-	spin_lock_irqsave(&mm->context.lock, flags);
-	if (!CTX_VALID(mm->context))
-		get_new_mmu_context(mm);
-	cpu = smp_processor_id();
-	if (!cpu_isset(cpu, mm->cpu_vm_mask))
-		cpu_set(cpu, mm->cpu_vm_mask);
-
-	load_secondary_context(mm);
-	__flush_tlb_mm(CTX_HWBITS(mm->context), SECONDARY_CONTEXT);
-	tsb_context_switch(mm);
-	spin_unlock_irqrestore(&mm->context.lock, flags);
-}
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC64_MMU_CONTEXT_H) */
+#include <asm-sparc/mmu_context.h>
diff --git a/include/asm-sparc64/mmzone.h b/include/asm-sparc64/mmzone.h
index ebf5986c12e..43a710f7892 100644
--- a/include/asm-sparc64/mmzone.h
+++ b/include/asm-sparc64/mmzone.h
@@ -1,17 +1 @@
-#ifndef _SPARC64_MMZONE_H
-#define _SPARC64_MMZONE_H
-
-#ifdef CONFIG_NEED_MULTIPLE_NODES
-
-extern struct pglist_data *node_data[];
-
-#define NODE_DATA(nid)		(node_data[nid])
-#define node_start_pfn(nid)	(NODE_DATA(nid)->node_start_pfn)
-#define node_end_pfn(nid)	(NODE_DATA(nid)->node_end_pfn)
-
-extern int numa_cpu_lookup_table[];
-extern cpumask_t numa_cpumask_lookup_table[];
-
-#endif /* CONFIG_NEED_MULTIPLE_NODES */
-
-#endif /* _SPARC64_MMZONE_H */
+#include <asm-sparc/mmzone.h>
diff --git a/include/asm-sparc64/module.h b/include/asm-sparc64/module.h
index 3d77ba46578..a9606db55e4 100644
--- a/include/asm-sparc64/module.h
+++ b/include/asm-sparc64/module.h
@@ -1,7 +1 @@
-#ifndef _ASM_SPARC64_MODULE_H
-#define _ASM_SPARC64_MODULE_H
-struct mod_arch_specific { };
-#define Elf_Shdr Elf64_Shdr
-#define Elf_Sym Elf64_Sym
-#define Elf_Ehdr Elf64_Ehdr
-#endif /* _ASM_SPARC64_MODULE_H */
+#include <asm-sparc/module.h>
diff --git a/include/asm-sparc64/mostek.h b/include/asm-sparc64/mostek.h
index c5652de2ace..95a752f7e87 100644
--- a/include/asm-sparc64/mostek.h
+++ b/include/asm-sparc64/mostek.h
@@ -1,143 +1 @@
-/* mostek.h:  Describes the various Mostek time of day clock registers.
- *
- * Copyright (C) 1995 David S. Miller (davem@davemloft.net)
- * Copyright (C) 1996 Thomas K. Dyas (tdyas@eden.rutgers.edu)
- */
-
-#ifndef _SPARC64_MOSTEK_H
-#define _SPARC64_MOSTEK_H
-
-#include <asm/idprom.h>
-
-/*       M48T02 Register Map (adapted from Sun NVRAM/Hostid FAQ)
- *
- *                             Data
- * Address                                                 Function
- *        Bit 7 Bit 6 Bit 5 Bit 4Bit 3 Bit 2 Bit 1 Bit 0
- *   7ff  -     -     -     -    -     -     -     -       Year 00-99
- *   7fe  0     0     0     -    -     -     -     -      Month 01-12
- *   7fd  0     0     -     -    -     -     -     -       Date 01-31
- *   7fc  0     FT    0     0    0     -     -     -        Day 01-07
- *   7fb  KS    0     -     -    -     -     -     -      Hours 00-23
- *   7fa  0     -     -     -    -     -     -     -    Minutes 00-59
- *   7f9  ST    -     -     -    -     -     -     -    Seconds 00-59
- *   7f8  W     R     S     -    -     -     -     -    Control
- *
- *   * ST is STOP BIT
- *   * W is WRITE BIT
- *   * R is READ BIT
- *   * S is SIGN BIT
- *   * FT is FREQ TEST BIT
- *   * KS is KICK START BIT
- */
-
-/* The Mostek 48t02 real time clock and NVRAM chip. The registers
- * other than the control register are in binary coded decimal. Some
- * control bits also live outside the control register.
- *
- * We now deal with physical addresses for I/O to the chip. -DaveM
- */
-static inline u8 mostek_read(void __iomem *addr)
-{
-	u8 ret;
-
-	__asm__ __volatile__("lduba	[%1] %2, %0"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-	return ret;
-}
-
-static inline void mostek_write(void __iomem *addr, u8 val)
-{
-	__asm__ __volatile__("stba	%0, [%1] %2"
-			     : /* no outputs */
-			     : "r" (val), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-#define MOSTEK_EEPROM		0x0000UL
-#define MOSTEK_IDPROM		0x07d8UL
-#define MOSTEK_CREG		0x07f8UL
-#define MOSTEK_SEC		0x07f9UL
-#define MOSTEK_MIN		0x07faUL
-#define MOSTEK_HOUR		0x07fbUL
-#define MOSTEK_DOW		0x07fcUL
-#define MOSTEK_DOM		0x07fdUL
-#define MOSTEK_MONTH		0x07feUL
-#define MOSTEK_YEAR		0x07ffUL
-
-extern spinlock_t mostek_lock;
-extern void __iomem *mstk48t02_regs;
-
-/* Control register values. */
-#define	MSTK_CREG_WRITE	0x80	/* Must set this before placing values. */
-#define	MSTK_CREG_READ	0x40	/* Stop updates to allow a clean read. */
-#define	MSTK_CREG_SIGN	0x20	/* Slow/speed clock in calibration mode. */
-
-/* Control bits that live in the other registers. */
-#define	MSTK_STOP	0x80	/* Stop the clock oscillator. (sec) */
-#define	MSTK_KICK_START	0x80	/* Kick start the clock chip. (hour) */
-#define MSTK_FREQ_TEST	0x40	/* Frequency test mode. (day) */
-
-#define MSTK_YEAR_ZERO       1968   /* If year reg has zero, it is 1968. */
-#define MSTK_CVT_YEAR(yr)  ((yr) + MSTK_YEAR_ZERO)
-
-/* Masks that define how much space each value takes up. */
-#define	MSTK_SEC_MASK	0x7f
-#define	MSTK_MIN_MASK	0x7f
-#define	MSTK_HOUR_MASK	0x3f
-#define	MSTK_DOW_MASK	0x07
-#define	MSTK_DOM_MASK	0x3f
-#define	MSTK_MONTH_MASK	0x1f
-#define	MSTK_YEAR_MASK	0xffU
-
-/* Binary coded decimal conversion macros. */
-#define MSTK_REGVAL_TO_DECIMAL(x)  (((x) & 0x0F) + 0x0A * ((x) >> 0x04))
-#define MSTK_DECIMAL_TO_REGVAL(x)  ((((x) / 0x0A) << 0x04) + ((x) % 0x0A))
-
-/* Generic register set and get macros for internal use. */
-#define MSTK_GET(regs,name)	\
-	(MSTK_REGVAL_TO_DECIMAL(mostek_read(regs + MOSTEK_ ## name) & MSTK_ ## name ## _MASK))
-#define MSTK_SET(regs,name,value) \
-do {	u8 __val = mostek_read(regs + MOSTEK_ ## name); \
-	__val &= ~(MSTK_ ## name ## _MASK); \
-	__val |= (MSTK_DECIMAL_TO_REGVAL(value) & \
-		  (MSTK_ ## name ## _MASK)); \
-	mostek_write(regs + MOSTEK_ ## name, __val); \
-} while(0)
-
-/* Macros to make register access easier on our fingers. These give you
- * the decimal value of the register requested if applicable. You pass
- * the a pointer to a 'struct mostek48t02'.
- */
-#define	MSTK_REG_CREG(regs)	(mostek_read((regs) + MOSTEK_CREG))
-#define	MSTK_REG_SEC(regs)	MSTK_GET(regs,SEC)
-#define	MSTK_REG_MIN(regs)	MSTK_GET(regs,MIN)
-#define	MSTK_REG_HOUR(regs)	MSTK_GET(regs,HOUR)
-#define	MSTK_REG_DOW(regs)	MSTK_GET(regs,DOW)
-#define	MSTK_REG_DOM(regs)	MSTK_GET(regs,DOM)
-#define	MSTK_REG_MONTH(regs)	MSTK_GET(regs,MONTH)
-#define	MSTK_REG_YEAR(regs)	MSTK_GET(regs,YEAR)
-
-#define	MSTK_SET_REG_SEC(regs,value)	MSTK_SET(regs,SEC,value)
-#define	MSTK_SET_REG_MIN(regs,value)	MSTK_SET(regs,MIN,value)
-#define	MSTK_SET_REG_HOUR(regs,value)	MSTK_SET(regs,HOUR,value)
-#define	MSTK_SET_REG_DOW(regs,value)	MSTK_SET(regs,DOW,value)
-#define	MSTK_SET_REG_DOM(regs,value)	MSTK_SET(regs,DOM,value)
-#define	MSTK_SET_REG_MONTH(regs,value)	MSTK_SET(regs,MONTH,value)
-#define	MSTK_SET_REG_YEAR(regs,value)	MSTK_SET(regs,YEAR,value)
-
-
-/* The Mostek 48t08 clock chip. Found on Sun4m's I think. It has the
- * same (basically) layout of the 48t02 chip except for the extra
- * NVRAM on board (8 KB against the 48t02's 2 KB).
- */
-#define MOSTEK_48T08_OFFSET	0x0000UL	/* Lower NVRAM portions */
-#define MOSTEK_48T08_48T02	0x1800UL	/* Offset to 48T02 chip */
-
-/* SUN5 systems usually have 48t59 model clock chipsets.  But we keep the older
- * clock chip definitions around just in case.
- */
-#define MOSTEK_48T59_OFFSET	0x0000UL	/* Lower NVRAM portions */
-#define MOSTEK_48T59_48T02	0x1800UL	/* Offset to 48T02 chip */
-
-#endif /* !(_SPARC64_MOSTEK_H) */
+#include <asm-sparc/mostek.h>
diff --git a/include/asm-sparc64/msgbuf.h b/include/asm-sparc64/msgbuf.h
index 55c101bd0e7..5b33cc9d9bf 100644
--- a/include/asm-sparc64/msgbuf.h
+++ b/include/asm-sparc64/msgbuf.h
@@ -1,27 +1 @@
-#ifndef _SPARC64_MSGBUF_H
-#define _SPARC64_MSGBUF_H
-
-/* 
- * The msqid64_ds structure for sparc64 architecture.
- * Note extra padding because this structure is passed back and forth
- * between kernel and user space.
- *
- * Pad space is left for:
- * - 2 miscellaneous 64-bit values
- */
-
-struct msqid64_ds {
-	struct ipc64_perm msg_perm;
-	__kernel_time_t msg_stime;	/* last msgsnd time */
-	__kernel_time_t msg_rtime;	/* last msgrcv time */
-	__kernel_time_t msg_ctime;	/* last change time */
-	unsigned long  msg_cbytes;	/* current number of bytes on queue */
-	unsigned long  msg_qnum;	/* number of messages in queue */
-	unsigned long  msg_qbytes;	/* max number of bytes on queue */
-	__kernel_pid_t msg_lspid;	/* pid of last msgsnd */
-	__kernel_pid_t msg_lrpid;	/* last receive pid */
-	unsigned long  __unused1;
-	unsigned long  __unused2;
-};
-
-#endif /* _SPARC64_MSGBUF_H */
+#include <asm-sparc/msgbuf.h>
diff --git a/include/asm-sparc64/mutex.h b/include/asm-sparc64/mutex.h
index 458c1f7fbc1..c0c0f8f260d 100644
--- a/include/asm-sparc64/mutex.h
+++ b/include/asm-sparc64/mutex.h
@@ -1,9 +1 @@
-/*
- * Pull in the generic implementation for the mutex fastpath.
- *
- * TODO: implement optimized primitives instead, or leave the generic
- * implementation in place, or pick the atomic_xchg() based generic
- * implementation. (see asm-generic/mutex-xchg.h for details)
- */
-
-#include <asm-generic/mutex-dec.h>
+#include <asm-sparc/mutex.h>
diff --git a/include/asm-sparc64/namei.h b/include/asm-sparc64/namei.h
index cbc1b4c0689..1344a910ba2 100644
--- a/include/asm-sparc64/namei.h
+++ b/include/asm-sparc64/namei.h
@@ -1,13 +1 @@
-/*
- * linux/include/asm-sparc64/namei.h
- *
- * Routines to handle famous /usr/gnemul/s*.
- * Included from linux/fs/namei.c
- */
-
-#ifndef __SPARC64_NAMEI_H
-#define __SPARC64_NAMEI_H
-
-#define __emul_prefix() NULL
-
-#endif /* __SPARC64_NAMEI_H */
+#include <asm-sparc/namei.h>
diff --git a/include/asm-sparc64/ns87303.h b/include/asm-sparc64/ns87303.h
index 686defe6aaa..5f369d4df3d 100644
--- a/include/asm-sparc64/ns87303.h
+++ b/include/asm-sparc64/ns87303.h
@@ -1,118 +1 @@
-/* ns87303.h: Configuration Register Description for the
- *            National Semiconductor PC87303 (SuperIO).
- *
- * Copyright (C) 1997  Eddie C. Dost  (ecd@skynet.be)
- */
-
-#ifndef _SPARC_NS87303_H
-#define _SPARC_NS87303_H 1
-
-/*
- * Control Register Index Values
- */
-#define FER	0x00
-#define FAR	0x01
-#define PTR	0x02
-#define FCR	0x03
-#define PCR	0x04
-#define KRR	0x05
-#define PMC	0x06
-#define TUP	0x07
-#define SID	0x08
-#define ASC	0x09
-#define CS0CF0	0x0a
-#define CS0CF1	0x0b
-#define CS1CF0	0x0c
-#define CS1CF1	0x0d
-
-/* Function Enable Register (FER) bits */
-#define FER_EDM		0x10	/* Encoded Drive and Motor pin information   */
-
-/* Function Address Register (FAR) bits */
-#define FAR_LPT_MASK	0x03
-#define FAR_LPTB	0x00
-#define FAR_LPTA	0x01
-#define FAR_LPTC	0x02
-
-/* Power and Test Register (PTR) bits */
-#define PTR_LPTB_IRQ7	0x08
-#define PTR_LEVEL_IRQ	0x80	/* When not ECP/EPP: Use level IRQ           */
-#define PTR_LPT_REG_DIR	0x80	/* When ECP/EPP: LPT CTR controlls direction */
-				/*               of the parallel port	     */
-
-/* Function Control Register (FCR) bits */
-#define FCR_LDE		0x10	/* Logical Drive Exchange                    */
-#define FCR_ZWS_ENA	0x20	/* Enable short host read/write in ECP/EPP   */
-
-/* Printer Control Register (PCR) bits */
-#define PCR_EPP_ENABLE	0x01
-#define PCR_EPP_IEEE	0x02	/* Enable EPP Version 1.9 (IEEE 1284)        */
-#define PCR_ECP_ENABLE	0x04
-#define PCR_ECP_CLK_ENA	0x08	/* If 0 ECP Clock is stopped on Power down   */
-#define PCR_IRQ_POLAR	0x20	/* If 0 IRQ is level high or negative pulse, */
-				/* if 1 polarity is inverted                 */
-#define PCR_IRQ_ODRAIN	0x40	/* If 1, IRQ is open drain                   */
-
-/* Tape UARTs and Parallel Port Config Register (TUP) bits */
-#define TUP_EPP_TIMO	0x02	/* Enable EPP timeout IRQ                    */
-
-/* Advanced SuperIO Config Register (ASC) bits */
-#define ASC_LPT_IRQ7	0x01	/* Always use IRQ7 for LPT                  */
-#define ASC_DRV2_SEL	0x02	/* Logical Drive Exchange controlled by TDR  */
-
-#define FER_RESERVED	0x00
-#define FAR_RESERVED	0x00
-#define PTR_RESERVED	0x73
-#define FCR_RESERVED	0xc4
-#define PCR_RESERVED	0x10
-#define KRR_RESERVED	0x00
-#define PMC_RESERVED	0x98
-#define TUP_RESERVED	0xfb
-#define SIP_RESERVED	0x00
-#define ASC_RESERVED	0x18
-#define CS0CF0_RESERVED	0x00
-#define CS0CF1_RESERVED	0x08
-#define CS1CF0_RESERVED	0x00
-#define CS1CF1_RESERVED	0x08
-
-#ifdef __KERNEL__
-
-#include <linux/spinlock.h>
-
-#include <asm/system.h>
-#include <asm/io.h>
-
-extern spinlock_t ns87303_lock;
-
-static inline int ns87303_modify(unsigned long port, unsigned int index,
-				     unsigned char clr, unsigned char set)
-{
-	static unsigned char reserved[] = {
-		FER_RESERVED, FAR_RESERVED, PTR_RESERVED, FCR_RESERVED,
-		PCR_RESERVED, KRR_RESERVED, PMC_RESERVED, TUP_RESERVED,
-		SIP_RESERVED, ASC_RESERVED, CS0CF0_RESERVED, CS0CF1_RESERVED,
-		CS1CF0_RESERVED, CS1CF1_RESERVED
-	};
-	unsigned long flags;
-	unsigned char value;
-
-	if (index > 0x0d)
-		return -EINVAL;
-
-	spin_lock_irqsave(&ns87303_lock, flags);
-
-	outb(index, port);
-	value = inb(port + 1);
-	value &= ~(reserved[index] | clr);
-	value |= set;
-	outb(value, port + 1);
-	outb(value, port + 1);
-
-	spin_unlock_irqrestore(&ns87303_lock, flags);
-
-	return 0;
-}
-
-#endif /* __KERNEL__ */
-
-#endif /* !(_SPARC_NS87303_H) */
+#include <asm-sparc/ns87303.h>
diff --git a/include/asm-sparc64/of_platform.h b/include/asm-sparc64/of_platform.h
index 78aa032b674..f7c427b8bc6 100644
--- a/include/asm-sparc64/of_platform.h
+++ b/include/asm-sparc64/of_platform.h
@@ -1,25 +1 @@
-#ifndef _ASM_SPARC64_OF_PLATFORM_H
-#define _ASM_SPARC64_OF_PLATFORM_H
-/*
- *    Copyright (C) 2006 Benjamin Herrenschmidt, IBM Corp.
- *			 <benh@kernel.crashing.org>
- *    Modified for Sparc by merging parts of asm-sparc/of_device.h
- *		by Stephen Rothwell
- *
- *  This program is free software; you can redistribute it and/or
- *  modify it under the terms of the GNU General Public License
- *  as published by the Free Software Foundation; either version
- *  2 of the License, or (at your option) any later version.
- *
- */
-
-/* This is just here during the transition */
-#include <linux/of_platform.h>
-
-extern struct bus_type isa_bus_type;
-extern struct bus_type ebus_bus_type;
-extern struct bus_type sbus_bus_type;
-
-#define of_bus_type	of_platform_bus_type	/* for compatibility */
-
-#endif	/* _ASM_SPARC64_OF_PLATFORM_H */
+#include <asm-sparc/of_platform.h>
diff --git a/include/asm-sparc64/openprom.h b/include/asm-sparc64/openprom.h
index b69e4a8c917..acf4b234fae 100644
--- a/include/asm-sparc64/openprom.h
+++ b/include/asm-sparc64/openprom.h
@@ -1,280 +1 @@
-#ifndef __SPARC64_OPENPROM_H
-#define __SPARC64_OPENPROM_H
-
-/* openprom.h:  Prom structures and defines for access to the OPENBOOT
- *              prom routines and data areas.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __ASSEMBLY__
-/* V0 prom device operations. */
-struct linux_dev_v0_funcs {
-	int (*v0_devopen)(char *device_str);
-	int (*v0_devclose)(int dev_desc);
-	int (*v0_rdblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
-	int (*v0_wrblkdev)(int dev_desc, int num_blks, int blk_st, char *buf);
-	int (*v0_wrnetdev)(int dev_desc, int num_bytes, char *buf);
-	int (*v0_rdnetdev)(int dev_desc, int num_bytes, char *buf);
-	int (*v0_rdchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
-	int (*v0_wrchardev)(int dev_desc, int num_bytes, int dummy, char *buf);
-	int (*v0_seekdev)(int dev_desc, long logical_offst, int from);
-};
-
-/* V2 and later prom device operations. */
-struct linux_dev_v2_funcs {
-	int (*v2_inst2pkg)(int d);	/* Convert ihandle to phandle */
-	char * (*v2_dumb_mem_alloc)(char *va, unsigned sz);
-	void (*v2_dumb_mem_free)(char *va, unsigned sz);
-
-	/* To map devices into virtual I/O space. */
-	char * (*v2_dumb_mmap)(char *virta, int which_io, unsigned paddr, unsigned sz);
-	void (*v2_dumb_munmap)(char *virta, unsigned size);
-
-	int (*v2_dev_open)(char *devpath);
-	void (*v2_dev_close)(int d);
-	int (*v2_dev_read)(int d, char *buf, int nbytes);
-	int (*v2_dev_write)(int d, char *buf, int nbytes);
-	int (*v2_dev_seek)(int d, int hi, int lo);
-
-	/* Never issued (multistage load support) */
-	void (*v2_wheee2)(void);
-	void (*v2_wheee3)(void);
-};
-
-struct linux_mlist_v0 {
-	struct linux_mlist_v0 *theres_more;
-	unsigned start_adr;
-	unsigned num_bytes;
-};
-
-struct linux_mem_v0 {
-	struct linux_mlist_v0 **v0_totphys;
-	struct linux_mlist_v0 **v0_prommap;
-	struct linux_mlist_v0 **v0_available; /* What we can use */
-};
-
-/* Arguments sent to the kernel from the boot prompt. */
-struct linux_arguments_v0 {
-	char *argv[8];
-	char args[100];
-	char boot_dev[2];
-	int boot_dev_ctrl;
-	int boot_dev_unit;
-	int dev_partition;
-	char *kernel_file_name;
-	void *aieee1;           /* XXX */
-};
-
-/* V2 and up boot things. */
-struct linux_bootargs_v2 {
-	char **bootpath;
-	char **bootargs;
-	int *fd_stdin;
-	int *fd_stdout;
-};
-
-/* The top level PROM vector. */
-struct linux_romvec {
-	/* Version numbers. */
-	unsigned int pv_magic_cookie;
-	unsigned int pv_romvers;
-	unsigned int pv_plugin_revision;
-	unsigned int pv_printrev;
-
-	/* Version 0 memory descriptors. */
-	struct linux_mem_v0 pv_v0mem;
-
-	/* Node operations. */
-	struct linux_nodeops *pv_nodeops;
-
-	char **pv_bootstr;
-	struct linux_dev_v0_funcs pv_v0devops;
-
-	char *pv_stdin;
-	char *pv_stdout;
-#define	PROMDEV_KBD	0		/* input from keyboard */
-#define	PROMDEV_SCREEN	0		/* output to screen */
-#define	PROMDEV_TTYA	1		/* in/out to ttya */
-#define	PROMDEV_TTYB	2		/* in/out to ttyb */
-
-	/* Blocking getchar/putchar.  NOT REENTRANT! (grr) */
-	int (*pv_getchar)(void);
-	void (*pv_putchar)(int ch);
-
-	/* Non-blocking variants. */
-	int (*pv_nbgetchar)(void);
-	int (*pv_nbputchar)(int ch);
-
-	void (*pv_putstr)(char *str, int len);
-
-	/* Miscellany. */
-	void (*pv_reboot)(char *bootstr);
-	void (*pv_printf)(__const__ char *fmt, ...);
-	void (*pv_abort)(void);
-	__volatile__ int *pv_ticks;
-	void (*pv_halt)(void);
-	void (**pv_synchook)(void);
-
-	/* Evaluate a forth string, not different proto for V0 and V2->up. */
-	union {
-		void (*v0_eval)(int len, char *str);
-		void (*v2_eval)(char *str);
-	} pv_fortheval;
-
-	struct linux_arguments_v0 **pv_v0bootargs;
-
-	/* Get ether address. */
-	unsigned int (*pv_enaddr)(int d, char *enaddr);
-
-	struct linux_bootargs_v2 pv_v2bootargs;
-	struct linux_dev_v2_funcs pv_v2devops;
-
-	int filler[15];
-
-	/* This one is sun4c/sun4 only. */
-	void (*pv_setctxt)(int ctxt, char *va, int pmeg);
-
-	/* Prom version 3 Multiprocessor routines. This stuff is crazy.
-	 * No joke. Calling these when there is only one cpu probably
-	 * crashes the machine, have to test this. :-)
-	 */
-
-	/* v3_cpustart() will start the cpu 'whichcpu' in mmu-context
-	 * 'thiscontext' executing at address 'prog_counter'
-	 */
-	int (*v3_cpustart)(unsigned int whichcpu, int ctxtbl_ptr,
-			   int thiscontext, char *prog_counter);
-
-	/* v3_cpustop() will cause cpu 'whichcpu' to stop executing
-	 * until a resume cpu call is made.
-	 */
-	int (*v3_cpustop)(unsigned int whichcpu);
-
-	/* v3_cpuidle() will idle cpu 'whichcpu' until a stop or
-	 * resume cpu call is made.
-	 */
-	int (*v3_cpuidle)(unsigned int whichcpu);
-
-	/* v3_cpuresume() will resume processor 'whichcpu' executing
-	 * starting with whatever 'pc' and 'npc' were left at the
-	 * last 'idle' or 'stop' call.
-	 */
-	int (*v3_cpuresume)(unsigned int whichcpu);
-};
-
-/* Routines for traversing the prom device tree. */
-struct linux_nodeops {
-	int (*no_nextnode)(int node);
-	int (*no_child)(int node);
-	int (*no_proplen)(int node, char *name);
-	int (*no_getprop)(int node, char *name, char *val);
-	int (*no_setprop)(int node, char *name, char *val, int len);
-	char * (*no_nextprop)(int node, char *name);
-};
-
-/* More fun PROM structures for device probing. */
-#define PROMREG_MAX     24
-#define PROMVADDR_MAX   16
-#define PROMINTR_MAX    32
-
-struct linux_prom_registers {
-	unsigned which_io;	/* hi part of physical address			*/
-	unsigned phys_addr;	/* The physical address of this register	*/
-	int reg_size;		/* How many bytes does this register take up?	*/
-};
-
-struct linux_prom64_registers {
-	unsigned long phys_addr;
-	unsigned long reg_size;
-};
-
-struct linux_prom_irqs {
-	int pri;    /* IRQ priority */
-	int vector; /* This is foobar, what does it do? */
-};
-
-/* Element of the "ranges" vector */
-struct linux_prom_ranges {
-	unsigned int ot_child_space;
-	unsigned int ot_child_base;		/* Bus feels this */
-	unsigned int ot_parent_space;
-	unsigned int ot_parent_base;		/* CPU looks from here */
-	unsigned int or_size;
-};
-
-struct linux_prom64_ranges {
-	unsigned long ot_child_base;		/* Bus feels this */
-	unsigned long ot_parent_base;		/* CPU looks from here */
-	unsigned long or_size;
-};
-
-/* Ranges and reg properties are a bit different for PCI. */
-struct linux_prom_pci_registers {
-	unsigned int phys_hi;
-	unsigned int phys_mid;
-	unsigned int phys_lo;
-
-	unsigned int size_hi;
-	unsigned int size_lo;
-};
-
-struct linux_prom_pci_ranges {
-	unsigned int child_phys_hi;	/* Only certain bits are encoded here. */
-	unsigned int child_phys_mid;
-	unsigned int child_phys_lo;
-
-	unsigned int parent_phys_hi;
-	unsigned int parent_phys_lo;
-
-	unsigned int size_hi;
-	unsigned int size_lo;
-};
-
-struct linux_prom_pci_intmap {
-	unsigned int phys_hi;
-	unsigned int phys_mid;
-	unsigned int phys_lo;
-
-	unsigned int interrupt;
-
-	int          cnode;
-	unsigned int cinterrupt;
-};
-
-struct linux_prom_pci_intmask {
-	unsigned int phys_hi;
-	unsigned int phys_mid;
-	unsigned int phys_lo;
-	unsigned int interrupt;
-};
-
-struct linux_prom_ebus_ranges {
-	unsigned int child_phys_hi;
-	unsigned int child_phys_lo;
-
-	unsigned int parent_phys_hi;
-	unsigned int parent_phys_mid;
-	unsigned int parent_phys_lo;
-
-	unsigned int size;
-};
-
-struct linux_prom_ebus_intmap {
-	unsigned int phys_hi;
-	unsigned int phys_lo;
-
-	unsigned int interrupt;
-
-	int          cnode;
-	unsigned int cinterrupt;
-};
-
-struct linux_prom_ebus_intmask {
-	unsigned int phys_hi;
-	unsigned int phys_lo;
-	unsigned int interrupt;
-};
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC64_OPENPROM_H) */
+#include <asm-sparc/openprom.h>
diff --git a/include/asm-sparc64/openpromio.h b/include/asm-sparc64/openpromio.h
index 847ce2326ad..122fabda21f 100644
--- a/include/asm-sparc64/openpromio.h
+++ b/include/asm-sparc64/openpromio.h
@@ -1,69 +1 @@
-#ifndef	_SPARC64_OPENPROMIO_H
-#define	_SPARC64_OPENPROMIO_H
-
-#include <linux/compiler.h>
-#include <linux/ioctl.h>
-#include <linux/types.h>
-
-/*
- * SunOS and Solaris /dev/openprom definitions. The ioctl values
- * were chosen to be exactly equal to the SunOS equivalents.
- */
-
-struct openpromio
-{
-	u_int	oprom_size;		/* Actual size of the oprom_array. */
-	char	oprom_array[1];		/* Holds property names and values. */
-};
-
-#define	OPROMMAXPARAM	4096		/* Maximum size of oprom_array. */
-
-#define	OPROMGETOPT		0x20004F01
-#define	OPROMSETOPT		0x20004F02
-#define	OPROMNXTOPT		0x20004F03
-#define	OPROMSETOPT2		0x20004F04
-#define	OPROMNEXT		0x20004F05
-#define	OPROMCHILD		0x20004F06
-#define	OPROMGETPROP		0x20004F07
-#define	OPROMNXTPROP		0x20004F08
-#define	OPROMU2P		0x20004F09
-#define	OPROMGETCONS		0x20004F0A
-#define	OPROMGETFBNAME		0x20004F0B
-#define	OPROMGETBOOTARGS	0x20004F0C
-/* Linux extensions */				/* Arguments in oprom_array: */
-#define OPROMSETCUR		0x20004FF0	/* int node - Sets current node */
-#define OPROMPCI2NODE		0x20004FF1	/* int pci_bus, pci_devfn - Sets current node to PCI device's node */
-#define OPROMPATH2NODE		0x20004FF2	/* char path[] - Set current node from fully qualified PROM path */
-
-/*
- * Return values from OPROMGETCONS:
- */
-
-#define OPROMCONS_NOT_WSCONS    0
-#define OPROMCONS_STDIN_IS_KBD  0x1     /* stdin device is kbd */
-#define OPROMCONS_STDOUT_IS_FB  0x2     /* stdout is a framebuffer */
-#define OPROMCONS_OPENPROM      0x4     /* supports openboot */
-
-
-/*
- *  NetBSD/OpenBSD /dev/openprom definitions.
- */
-
-struct opiocdesc
-{
-	int	op_nodeid;		/* PROM Node ID (value-result) */
-	int	op_namelen;		/* Length of op_name. */
-	char	__user *op_name;	/* Pointer to the property name. */
-	int	op_buflen;		/* Length of op_buf (value-result) */
-	char	__user *op_buf;		/* Pointer to buffer. */
-};
-
-#define	OPIOCGET	_IOWR('O', 1, struct opiocdesc)
-#define	OPIOCSET	_IOW('O', 2, struct opiocdesc)
-#define	OPIOCNEXTPROP	_IOWR('O', 3, struct opiocdesc)
-#define	OPIOCGETOPTNODE	_IOR('O', 4, int)
-#define	OPIOCGETNEXT	_IOWR('O', 5, int)
-#define	OPIOCGETCHILD	_IOWR('O', 6, int)
-
-#endif /* _SPARC64_OPENPROMIO_H */
-
+#include <asm-sparc/openpromio.h>
diff --git a/include/asm-sparc64/oplib.h b/include/asm-sparc64/oplib.h
index 55c5bb27e4d..d93e44e6351 100644
--- a/include/asm-sparc64/oplib.h
+++ b/include/asm-sparc64/oplib.h
@@ -1,322 +1 @@
-/* oplib.h:  Describes the interface and available routines in the
- *           Linux Prom library.
- *
- * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
- * Copyright (C) 1996 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-
-#ifndef __SPARC64_OPLIB_H
-#define __SPARC64_OPLIB_H
-
-#include <asm/openprom.h>
-
-/* OBP version string. */
-extern char prom_version[];
-
-/* Root node of the prom device tree, this stays constant after
- * initialization is complete.
- */
-extern int prom_root_node;
-
-/* PROM stdin and stdout */
-extern int prom_stdin, prom_stdout;
-
-/* /chosen node of the prom device tree, this stays constant after
- * initialization is complete.
- */
-extern int prom_chosen_node;
-
-/* Helper values and strings in arch/sparc64/kernel/head.S */
-extern const char prom_peer_name[];
-extern const char prom_compatible_name[];
-extern const char prom_root_compatible[];
-extern const char prom_cpu_compatible[];
-extern const char prom_finddev_name[];
-extern const char prom_chosen_path[];
-extern const char prom_cpu_path[];
-extern const char prom_getprop_name[];
-extern const char prom_mmu_name[];
-extern const char prom_callmethod_name[];
-extern const char prom_translate_name[];
-extern const char prom_map_name[];
-extern const char prom_unmap_name[];
-extern int prom_mmu_ihandle_cache;
-extern unsigned int prom_boot_mapped_pc;
-extern unsigned int prom_boot_mapping_mode;
-extern unsigned long prom_boot_mapping_phys_high, prom_boot_mapping_phys_low;
-
-struct linux_mlist_p1275 {
-	struct linux_mlist_p1275 *theres_more;
-	unsigned long start_adr;
-	unsigned long num_bytes;
-};
-
-struct linux_mem_p1275 {
-	struct linux_mlist_p1275 **p1275_totphys;
-	struct linux_mlist_p1275 **p1275_prommap;
-	struct linux_mlist_p1275 **p1275_available; /* What we can use */
-};
-
-/* The functions... */
-
-/* You must call prom_init() before using any of the library services,
- * preferably as early as possible.  Pass it the romvec pointer.
- */
-extern void prom_init(void *cif_handler, void *cif_stack);
-
-/* Boot argument acquisition, returns the boot command line string. */
-extern char *prom_getbootargs(void);
-
-/* Device utilities. */
-
-/* Device operations. */
-
-/* Open the device described by the passed string.  Note, that the format
- * of the string is different on V0 vs. V2->higher proms.  The caller must
- * know what he/she is doing!  Returns the device descriptor, an int.
- */
-extern int prom_devopen(const char *device_string);
-
-/* Close a previously opened device described by the passed integer
- * descriptor.
- */
-extern int prom_devclose(int device_handle);
-
-/* Do a seek operation on the device described by the passed integer
- * descriptor.
- */
-extern void prom_seek(int device_handle, unsigned int seek_hival,
-		      unsigned int seek_lowval);
-
-/* Miscellaneous routines, don't really fit in any category per se. */
-
-/* Reboot the machine with the command line passed. */
-extern void prom_reboot(const char *boot_command);
-
-/* Evaluate the forth string passed. */
-extern void prom_feval(const char *forth_string);
-
-/* Enter the prom, with possibility of continuation with the 'go'
- * command in newer proms.
- */
-extern void prom_cmdline(void);
-
-/* Enter the prom, with no chance of continuation for the stand-alone
- * which calls this.
- */
-extern void prom_halt(void) __attribute__ ((noreturn));
-
-/* Halt and power-off the machine. */
-extern void prom_halt_power_off(void) __attribute__ ((noreturn));
-
-/* Set the PROM 'sync' callback function to the passed function pointer.
- * When the user gives the 'sync' command at the prom prompt while the
- * kernel is still active, the prom will call this routine.
- *
- */
-typedef int (*callback_func_t)(long *cmd);
-extern void prom_setcallback(callback_func_t func_ptr);
-
-/* Acquire the IDPROM of the root node in the prom device tree.  This
- * gets passed a buffer where you would like it stuffed.  The return value
- * is the format type of this idprom or 0xff on error.
- */
-extern unsigned char prom_get_idprom(char *idp_buffer, int idpbuf_size);
-
-/* Character operations to/from the console.... */
-
-/* Non-blocking get character from console. */
-extern int prom_nbgetchar(void);
-
-/* Non-blocking put character to console. */
-extern int prom_nbputchar(char character);
-
-/* Blocking get character from console. */
-extern char prom_getchar(void);
-
-/* Blocking put character to console. */
-extern void prom_putchar(char character);
-
-/* Prom's internal routines, don't use in kernel/boot code. */
-extern void prom_printf(const char *fmt, ...);
-extern void prom_write(const char *buf, unsigned int len);
-
-/* Multiprocessor operations... */
-#ifdef CONFIG_SMP
-/* Start the CPU with the given device tree node at the passed program
- * counter with the given arg passed in via register %o0.
- */
-extern void prom_startcpu(int cpunode, unsigned long pc, unsigned long arg);
-
-/* Start the CPU with the given cpu ID at the passed program
- * counter with the given arg passed in via register %o0.
- */
-extern void prom_startcpu_cpuid(int cpuid, unsigned long pc, unsigned long arg);
-
-/* Stop the CPU with the given cpu ID.  */
-extern void prom_stopcpu_cpuid(int cpuid);
-
-/* Stop the current CPU. */
-extern void prom_stopself(void);
-
-/* Idle the current CPU. */
-extern void prom_idleself(void);
-
-/* Resume the CPU with the passed device tree node. */
-extern void prom_resumecpu(int cpunode);
-#endif
-
-/* Power management interfaces. */
-
-/* Put the current CPU to sleep. */
-extern void prom_sleepself(void);
-
-/* Put the entire system to sleep. */
-extern int prom_sleepsystem(void);
-
-/* Initiate a wakeup event. */
-extern int prom_wakeupsystem(void);
-
-/* MMU and memory related OBP interfaces. */
-
-/* Get unique string identifying SIMM at given physical address. */
-extern int prom_getunumber(int syndrome_code,
-			   unsigned long phys_addr,
-			   char *buf, int buflen);
-
-/* Retain physical memory to the caller across soft resets. */
-extern unsigned long prom_retain(const char *name,
-				 unsigned long pa_low, unsigned long pa_high,
-				 long size, long align);
-
-/* Load explicit I/D TLB entries into the calling processor. */
-extern long prom_itlb_load(unsigned long index,
-			   unsigned long tte_data,
-			   unsigned long vaddr);
-
-extern long prom_dtlb_load(unsigned long index,
-			   unsigned long tte_data,
-			   unsigned long vaddr);
-
-/* Map/Unmap client program address ranges.  First the format of
- * the mapping mode argument.
- */
-#define PROM_MAP_WRITE	0x0001 /* Writable */
-#define PROM_MAP_READ	0x0002 /* Readable - sw */
-#define PROM_MAP_EXEC	0x0004 /* Executable - sw */
-#define PROM_MAP_LOCKED	0x0010 /* Locked, use i/dtlb load calls for this instead */
-#define PROM_MAP_CACHED	0x0020 /* Cacheable in both L1 and L2 caches */
-#define PROM_MAP_SE	0x0040 /* Side-Effects */
-#define PROM_MAP_GLOB	0x0080 /* Global */
-#define PROM_MAP_IE	0x0100 /* Invert-Endianness */
-#define PROM_MAP_DEFAULT (PROM_MAP_WRITE | PROM_MAP_READ | PROM_MAP_EXEC | PROM_MAP_CACHED)
-
-extern int prom_map(int mode, unsigned long size,
-		    unsigned long vaddr, unsigned long paddr);
-extern void prom_unmap(unsigned long size, unsigned long vaddr);
-
-
-/* PROM device tree traversal functions... */
-
-#ifdef PROMLIB_INTERNAL
-
-/* Internal version of prom_getchild. */
-extern int __prom_getchild(int parent_node);
-
-/* Internal version of prom_getsibling. */
-extern int __prom_getsibling(int node);
-
-#endif
-
-/* Get the child node of the given node, or zero if no child exists. */
-extern int prom_getchild(int parent_node);
-
-/* Get the next sibling node of the given node, or zero if no further
- * siblings exist.
- */
-extern int prom_getsibling(int node);
-
-/* Get the length, at the passed node, of the given property type.
- * Returns -1 on error (ie. no such property at this node).
- */
-extern int prom_getproplen(int thisnode, const char *property);
-
-/* Fetch the requested property using the given buffer.  Returns
- * the number of bytes the prom put into your buffer or -1 on error.
- */
-extern int prom_getproperty(int thisnode, const char *property,
-			    char *prop_buffer, int propbuf_size);
-
-/* Acquire an integer property. */
-extern int prom_getint(int node, const char *property);
-
-/* Acquire an integer property, with a default value. */
-extern int prom_getintdefault(int node, const char *property, int defval);
-
-/* Acquire a boolean property, 0=FALSE 1=TRUE. */
-extern int prom_getbool(int node, const char *prop);
-
-/* Acquire a string property, null string on error. */
-extern void prom_getstring(int node, const char *prop, char *buf, int bufsize);
-
-/* Does the passed node have the given "name"? YES=1 NO=0 */
-extern int prom_nodematch(int thisnode, const char *name);
-
-/* Search all siblings starting at the passed node for "name" matching
- * the given string.  Returns the node on success, zero on failure.
- */
-extern int prom_searchsiblings(int node_start, const char *name);
-
-/* Return the first property type, as a string, for the given node.
- * Returns a null string on error. Buffer should be at least 32B long.
- */
-extern char *prom_firstprop(int node, char *buffer);
-
-/* Returns the next property after the passed property for the given
- * node.  Returns null string on failure. Buffer should be at least 32B long.
- */
-extern char *prom_nextprop(int node, const char *prev_property, char *buffer);
-
-/* Returns 1 if the specified node has given property. */
-extern int prom_node_has_property(int node, const char *property);
-
-/* Returns phandle of the path specified */
-extern int prom_finddevice(const char *name);
-
-/* Set the indicated property at the given node with the passed value.
- * Returns the number of bytes of your value that the prom took.
- */
-extern int prom_setprop(int node, const char *prop_name, char *prop_value,
-			int value_size);
-			
-extern int prom_pathtoinode(const char *path);
-extern int prom_inst2pkg(int);
-extern int prom_service_exists(const char *service_name);
-extern void prom_sun4v_guest_soft_state(void);
-
-extern int prom_ihandle2path(int handle, char *buffer, int bufsize);
-
-/* Client interface level routines. */
-extern long p1275_cmd(const char *, long, ...);
-
-#if 0
-#define P1275_SIZE(x) ((((long)((x) / 32)) << 32) | (x))
-#else
-#define P1275_SIZE(x) x
-#endif
-
-/* We support at most 16 input and 1 output argument */
-#define P1275_ARG_NUMBER		0
-#define P1275_ARG_IN_STRING		1
-#define P1275_ARG_OUT_BUF		2
-#define P1275_ARG_OUT_32B		3
-#define P1275_ARG_IN_FUNCTION		4
-#define P1275_ARG_IN_BUF		5
-#define P1275_ARG_IN_64B		6
-
-#define P1275_IN(x) ((x) & 0xf)
-#define P1275_OUT(x) (((x) << 4) & 0xf0)
-#define P1275_INOUT(i,o) (P1275_IN(i)|P1275_OUT(o))
-#define P1275_ARG(n,x) ((x) << ((n)*3 + 8))
-
-#endif /* !(__SPARC64_OPLIB_H) */
+#include <asm-sparc/oplib.h>
diff --git a/include/asm-sparc64/page.h b/include/asm-sparc64/page.h
index 93f0881b766..f46c1fb5302 100644
--- a/include/asm-sparc64/page.h
+++ b/include/asm-sparc64/page.h
@@ -1,142 +1 @@
-#ifndef _SPARC64_PAGE_H
-#define _SPARC64_PAGE_H
-
-#include <linux/const.h>
-
-#if defined(CONFIG_SPARC64_PAGE_SIZE_8KB)
-#define PAGE_SHIFT   13
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_64KB)
-#define PAGE_SHIFT   16
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_512KB)
-#define PAGE_SHIFT   19
-#elif defined(CONFIG_SPARC64_PAGE_SIZE_4MB)
-#define PAGE_SHIFT   22
-#else
-#error No page size specified in kernel configuration
-#endif
-
-#define PAGE_SIZE    (_AC(1,UL) << PAGE_SHIFT)
-#define PAGE_MASK    (~(PAGE_SIZE-1))
-
-/* Flushing for D-cache alias handling is only needed if
- * the page size is smaller than 16K.
- */
-#if PAGE_SHIFT < 14
-#define DCACHE_ALIASING_POSSIBLE
-#endif
-
-#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
-#define HPAGE_SHIFT		22
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
-#define HPAGE_SHIFT		19
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
-#define HPAGE_SHIFT		16
-#endif
-
-#ifdef CONFIG_HUGETLB_PAGE
-#define HPAGE_SIZE		(_AC(1,UL) << HPAGE_SHIFT)
-#define HPAGE_MASK		(~(HPAGE_SIZE - 1UL))
-#define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
-#define HAVE_ARCH_HUGETLB_UNMAPPED_AREA
-#endif
-
-#ifndef __ASSEMBLY__
-
-extern void _clear_page(void *page);
-#define clear_page(X)	_clear_page((void *)(X))
-struct page;
-extern void clear_user_page(void *addr, unsigned long vaddr, struct page *page);
-#define copy_page(X,Y)	memcpy((void *)(X), (void *)(Y), PAGE_SIZE)
-extern void copy_user_page(void *to, void *from, unsigned long vaddr, struct page *topage);
-
-/* Unlike sparc32, sparc64's parameter passing API is more
- * sane in that structures which as small enough are passed
- * in registers instead of on the stack.  Thus, setting
- * STRICT_MM_TYPECHECKS does not generate worse code so
- * let's enable it to get the type checking.
- */
-
-#define STRICT_MM_TYPECHECKS
-
-#ifdef STRICT_MM_TYPECHECKS
-/* These are used to make use of C type-checking.. */
-typedef struct { unsigned long pte; } pte_t;
-typedef struct { unsigned long iopte; } iopte_t;
-typedef struct { unsigned int pmd; } pmd_t;
-typedef struct { unsigned int pgd; } pgd_t;
-typedef struct { unsigned long pgprot; } pgprot_t;
-
-#define pte_val(x)	((x).pte)
-#define iopte_val(x)	((x).iopte)
-#define pmd_val(x)      ((x).pmd)
-#define pgd_val(x)	((x).pgd)
-#define pgprot_val(x)	((x).pgprot)
-
-#define __pte(x)	((pte_t) { (x) } )
-#define __iopte(x)	((iopte_t) { (x) } )
-#define __pmd(x)        ((pmd_t) { (x) } )
-#define __pgd(x)	((pgd_t) { (x) } )
-#define __pgprot(x)	((pgprot_t) { (x) } )
-
-#else
-/* .. while these make it easier on the compiler */
-typedef unsigned long pte_t;
-typedef unsigned long iopte_t;
-typedef unsigned int pmd_t;
-typedef unsigned int pgd_t;
-typedef unsigned long pgprot_t;
-
-#define pte_val(x)	(x)
-#define iopte_val(x)	(x)
-#define pmd_val(x)      (x)
-#define pgd_val(x)	(x)
-#define pgprot_val(x)	(x)
-
-#define __pte(x)	(x)
-#define __iopte(x)	(x)
-#define __pmd(x)        (x)
-#define __pgd(x)	(x)
-#define __pgprot(x)	(x)
-
-#endif /* (STRICT_MM_TYPECHECKS) */
-
-typedef struct page *pgtable_t;
-
-#define TASK_UNMAPPED_BASE	(test_thread_flag(TIF_32BIT) ? \
-				 (_AC(0x0000000070000000,UL)) : \
-				 (_AC(0xfffff80000000000,UL) + (1UL << 32UL)))
-
-#include <asm-generic/memory_model.h>
-
-#endif /* !(__ASSEMBLY__) */
-
-/* to align the pointer to the (next) page boundary */
-#define PAGE_ALIGN(addr)	(((addr)+PAGE_SIZE-1)&PAGE_MASK)
-
-/* We used to stick this into a hard-coded global register (%g4)
- * but that does not make sense anymore.
- */
-#define PAGE_OFFSET		_AC(0xFFFFF80000000000,UL)
-
-#ifndef __ASSEMBLY__
-
-#define __pa(x)			((unsigned long)(x) - PAGE_OFFSET)
-#define __va(x)			((void *)((unsigned long) (x) + PAGE_OFFSET))
-
-#define pfn_to_kaddr(pfn)	__va((pfn) << PAGE_SHIFT)
-
-#define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr)>>PAGE_SHIFT)
-
-#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
-
-#define virt_to_phys __pa
-#define phys_to_virt __va
-
-#endif /* !(__ASSEMBLY__) */
-
-#define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \
-				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
-
-#include <asm-generic/page.h>
-
-#endif /* _SPARC64_PAGE_H */
+#include <asm-sparc/page.h>
diff --git a/include/asm-sparc64/param.h b/include/asm-sparc64/param.h
index f0125cf5a9d..40c6dc11082 100644
--- a/include/asm-sparc64/param.h
+++ b/include/asm-sparc64/param.h
@@ -1,23 +1 @@
-#ifndef _ASMSPARC64_PARAM_H
-#define _ASMSPARC64_PARAM_H
-
-
-#ifdef __KERNEL__
-# define HZ		CONFIG_HZ	/* Internal kernel timer frequency */
-# define USER_HZ	100	/* .. some user interfaces are in "ticks" */
-# define CLOCKS_PER_SEC (USER_HZ)
-#endif
-
-#ifndef HZ
-#define HZ 100
-#endif
-
-#define EXEC_PAGESIZE	8192    /* Thanks for sun4's we carry baggage... */
-
-#ifndef NOGROUP
-#define NOGROUP		(-1)
-#endif
-
-#define MAXHOSTNAMELEN	64	/* max length of hostname */
-
-#endif /* _ASMSPARC64_PARAM_H */
+#include <asm-sparc/param.h>
diff --git a/include/asm-sparc64/parport.h b/include/asm-sparc64/parport.h
index e9555b246c8..b4e4ca812eb 100644
--- a/include/asm-sparc64/parport.h
+++ b/include/asm-sparc64/parport.h
@@ -1,246 +1 @@
-/* parport.h: sparc64 specific parport initialization and dma.
- *
- * Copyright (C) 1999  Eddie C. Dost  (ecd@skynet.be)
- */
-
-#ifndef _ASM_SPARC64_PARPORT_H
-#define _ASM_SPARC64_PARPORT_H 1
-
-#include <asm/ebus.h>
-#include <asm/ns87303.h>
-#include <asm/of_device.h>
-#include <asm/prom.h>
-
-#define PARPORT_PC_MAX_PORTS	PARPORT_MAX
-
-/*
- * While sparc64 doesn't have an ISA DMA API, we provide something that looks
- * close enough to make parport_pc happy
- */
-#define HAS_DMA
-
-static DEFINE_SPINLOCK(dma_spin_lock);
-
-#define claim_dma_lock() \
-({	unsigned long flags; \
-	spin_lock_irqsave(&dma_spin_lock, flags); \
-	flags; \
-})
-
-#define release_dma_lock(__flags) \
-	spin_unlock_irqrestore(&dma_spin_lock, __flags);
-
-static struct sparc_ebus_info {
-	struct ebus_dma_info info;
-	unsigned int addr;
-	unsigned int count;
-	int lock;
-
-	struct parport *port;
-} sparc_ebus_dmas[PARPORT_PC_MAX_PORTS];
-
-static DECLARE_BITMAP(dma_slot_map, PARPORT_PC_MAX_PORTS);
-
-static inline int request_dma(unsigned int dmanr, const char *device_id)
-{
-	if (dmanr >= PARPORT_PC_MAX_PORTS)
-		return -EINVAL;
-	if (xchg(&sparc_ebus_dmas[dmanr].lock, 1) != 0)
-		return -EBUSY;
-	return 0;
-}
-
-static inline void free_dma(unsigned int dmanr)
-{
-	if (dmanr >= PARPORT_PC_MAX_PORTS) {
-		printk(KERN_WARNING "Trying to free DMA%d\n", dmanr);
-		return;
-	}
-	if (xchg(&sparc_ebus_dmas[dmanr].lock, 0) == 0) {
-		printk(KERN_WARNING "Trying to free free DMA%d\n", dmanr);
-		return;
-	}	
-}
-
-static inline void enable_dma(unsigned int dmanr)
-{
-	ebus_dma_enable(&sparc_ebus_dmas[dmanr].info, 1);
-
-	if (ebus_dma_request(&sparc_ebus_dmas[dmanr].info,
-			     sparc_ebus_dmas[dmanr].addr,
-			     sparc_ebus_dmas[dmanr].count))
-		BUG();
-}
-
-static inline void disable_dma(unsigned int dmanr)
-{
-	ebus_dma_enable(&sparc_ebus_dmas[dmanr].info, 0);
-}
-
-static inline void clear_dma_ff(unsigned int dmanr)
-{
-	/* nothing */
-}
-
-static inline void set_dma_mode(unsigned int dmanr, char mode)
-{
-	ebus_dma_prepare(&sparc_ebus_dmas[dmanr].info, (mode != DMA_MODE_WRITE));
-}
-
-static inline void set_dma_addr(unsigned int dmanr, unsigned int addr)
-{
-	sparc_ebus_dmas[dmanr].addr = addr;
-}
-
-static inline void set_dma_count(unsigned int dmanr, unsigned int count)
-{
-	sparc_ebus_dmas[dmanr].count = count;
-}
-
-static inline unsigned int get_dma_residue(unsigned int dmanr)
-{
-	return ebus_dma_residue(&sparc_ebus_dmas[dmanr].info);
-}
-
-static int __devinit ecpp_probe(struct of_device *op, const struct of_device_id *match)
-{
-	unsigned long base = op->resource[0].start;
-	unsigned long config = op->resource[1].start;
-	unsigned long d_base = op->resource[2].start;
-	unsigned long d_len;
-	struct device_node *parent;
-	struct parport *p;
-	int slot, err;
-
-	parent = op->node->parent;
-	if (!strcmp(parent->name, "dma")) {
-		p = parport_pc_probe_port(base, base + 0x400,
-					  op->irqs[0], PARPORT_DMA_NOFIFO,
-					  op->dev.parent->parent);
-		if (!p)
-			return -ENOMEM;
-		dev_set_drvdata(&op->dev, p);
-		return 0;
-	}
-
-	for (slot = 0; slot < PARPORT_PC_MAX_PORTS; slot++) {
-		if (!test_and_set_bit(slot, dma_slot_map))
-			break;
-	}
-	err = -ENODEV;
-	if (slot >= PARPORT_PC_MAX_PORTS)
-		goto out_err;
-
-	spin_lock_init(&sparc_ebus_dmas[slot].info.lock);
-
-	d_len = (op->resource[2].end - d_base) + 1UL;
-	sparc_ebus_dmas[slot].info.regs =
-		of_ioremap(&op->resource[2], 0, d_len, "ECPP DMA");
-
-	if (!sparc_ebus_dmas[slot].info.regs)
-		goto out_clear_map;
-
-	sparc_ebus_dmas[slot].info.flags = 0;
-	sparc_ebus_dmas[slot].info.callback = NULL;
-	sparc_ebus_dmas[slot].info.client_cookie = NULL;
-	sparc_ebus_dmas[slot].info.irq = 0xdeadbeef;
-	strcpy(sparc_ebus_dmas[slot].info.name, "parport");
-	if (ebus_dma_register(&sparc_ebus_dmas[slot].info))
-		goto out_unmap_regs;
-
-	ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 1);
-
-	/* Configure IRQ to Push Pull, Level Low */
-	/* Enable ECP, set bit 2 of the CTR first */
-	outb(0x04, base + 0x02);
-	ns87303_modify(config, PCR,
-		       PCR_EPP_ENABLE |
-		       PCR_IRQ_ODRAIN,
-		       PCR_ECP_ENABLE |
-		       PCR_ECP_CLK_ENA |
-		       PCR_IRQ_POLAR);
-
-	/* CTR bit 5 controls direction of port */
-	ns87303_modify(config, PTR,
-		       0, PTR_LPT_REG_DIR);
-
-	p = parport_pc_probe_port(base, base + 0x400,
-				  op->irqs[0],
-				  slot,
-				  op->dev.parent);
-	err = -ENOMEM;
-	if (!p)
-		goto out_disable_irq;
-
-	dev_set_drvdata(&op->dev, p);
-
-	return 0;
-
-out_disable_irq:
-	ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 0);
-	ebus_dma_unregister(&sparc_ebus_dmas[slot].info);
-
-out_unmap_regs:
-	of_iounmap(&op->resource[2], sparc_ebus_dmas[slot].info.regs, d_len);
-
-out_clear_map:
-	clear_bit(slot, dma_slot_map);
-
-out_err:
-	return err;
-}
-
-static int __devexit ecpp_remove(struct of_device *op)
-{
-	struct parport *p = dev_get_drvdata(&op->dev);
-	int slot = p->dma;
-
-	parport_pc_unregister_port(p);
-
-	if (slot != PARPORT_DMA_NOFIFO) {
-		unsigned long d_base = op->resource[2].start;
-		unsigned long d_len;
-
-		d_len = (op->resource[2].end - d_base) + 1UL;
-
-		ebus_dma_irq_enable(&sparc_ebus_dmas[slot].info, 0);
-		ebus_dma_unregister(&sparc_ebus_dmas[slot].info);
-		of_iounmap(&op->resource[2],
-			   sparc_ebus_dmas[slot].info.regs,
-			   d_len);
-		clear_bit(slot, dma_slot_map);
-	}
-
-	return 0;
-}
-
-static struct of_device_id ecpp_match[] = {
-	{
-		.name = "ecpp",
-	},
-	{
-		.name = "parallel",
-		.compatible = "ecpp",
-	},
-	{
-		.name = "parallel",
-		.compatible = "ns87317-ecpp",
-	},
-	{},
-};
-
-static struct of_platform_driver ecpp_driver = {
-	.name			= "ecpp",
-	.match_table		= ecpp_match,
-	.probe			= ecpp_probe,
-	.remove			= __devexit_p(ecpp_remove),
-};
-
-static int parport_pc_find_nonpci_ports(int autoirq, int autodma)
-{
-	of_register_driver(&ecpp_driver, &of_bus_type);
-
-	return 0;
-}
-
-#endif /* !(_ASM_SPARC64_PARPORT_H */
+#include <asm-sparc/parport.h>
diff --git a/include/asm-sparc64/pci.h b/include/asm-sparc64/pci.h
index f59f2571295..da54c4d1f39 100644
--- a/include/asm-sparc64/pci.h
+++ b/include/asm-sparc64/pci.h
@@ -1,209 +1 @@
-#ifndef __SPARC64_PCI_H
-#define __SPARC64_PCI_H
-
-#ifdef __KERNEL__
-
-#include <linux/dma-mapping.h>
-
-/* Can be used to override the logic in pci_scan_bus for skipping
- * already-configured bus numbers - to be used for buggy BIOSes
- * or architectures with incomplete PCI setup by the loader.
- */
-#define pcibios_assign_all_busses()	0
-#define pcibios_scan_all_fns(a, b)	0
-
-#define PCIBIOS_MIN_IO		0UL
-#define PCIBIOS_MIN_MEM		0UL
-
-#define PCI_IRQ_NONE		0xffffffff
-
-#define PCI_CACHE_LINE_BYTES	64
-
-static inline void pcibios_set_master(struct pci_dev *dev)
-{
-	/* No special bus mastering setup handling */
-}
-
-static inline void pcibios_penalize_isa_irq(int irq, int active)
-{
-	/* We don't do dynamic PCI IRQ allocation */
-}
-
-/* The PCI address space does not equal the physical memory
- * address space.  The networking and block device layers use
- * this boolean for bounce buffer decisions.
- */
-#define PCI_DMA_BUS_IS_PHYS	(0)
-
-static inline void *pci_alloc_consistent(struct pci_dev *pdev, size_t size,
-					 dma_addr_t *dma_handle)
-{
-	return dma_alloc_coherent(&pdev->dev, size, dma_handle, GFP_ATOMIC);
-}
-
-static inline void pci_free_consistent(struct pci_dev *pdev, size_t size,
-				       void *vaddr, dma_addr_t dma_handle)
-{
-	return dma_free_coherent(&pdev->dev, size, vaddr, dma_handle);
-}
-
-static inline dma_addr_t pci_map_single(struct pci_dev *pdev, void *ptr,
-					size_t size, int direction)
-{
-	return dma_map_single(&pdev->dev, ptr, size,
-			      (enum dma_data_direction) direction);
-}
-
-static inline void pci_unmap_single(struct pci_dev *pdev, dma_addr_t dma_addr,
-				    size_t size, int direction)
-{
-	dma_unmap_single(&pdev->dev, dma_addr, size,
-			 (enum dma_data_direction) direction);
-}
-
-#define pci_map_page(dev, page, off, size, dir) \
-	pci_map_single(dev, (page_address(page) + (off)), size, dir)
-#define pci_unmap_page(dev,addr,sz,dir) \
-	pci_unmap_single(dev,addr,sz,dir)
-
-/* pci_unmap_{single,page} is not a nop, thus... */
-#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
-	dma_addr_t ADDR_NAME;
-#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
-	__u32 LEN_NAME;
-#define pci_unmap_addr(PTR, ADDR_NAME)			\
-	((PTR)->ADDR_NAME)
-#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
-	(((PTR)->ADDR_NAME) = (VAL))
-#define pci_unmap_len(PTR, LEN_NAME)			\
-	((PTR)->LEN_NAME)
-#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
-	(((PTR)->LEN_NAME) = (VAL))
-
-static inline int pci_map_sg(struct pci_dev *pdev, struct scatterlist *sg,
-			     int nents, int direction)
-{
-	return dma_map_sg(&pdev->dev, sg, nents,
-			  (enum dma_data_direction) direction);
-}
-
-static inline void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sg,
-				int nents, int direction)
-{
-	dma_unmap_sg(&pdev->dev, sg, nents,
-		     (enum dma_data_direction) direction);
-}
-
-static inline void pci_dma_sync_single_for_cpu(struct pci_dev *pdev,
-					       dma_addr_t dma_handle,
-					       size_t size, int direction)
-{
-	dma_sync_single_for_cpu(&pdev->dev, dma_handle, size,
-				(enum dma_data_direction) direction);
-}
-
-static inline void pci_dma_sync_single_for_device(struct pci_dev *pdev,
-						  dma_addr_t dma_handle,
-						  size_t size, int direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-static inline void pci_dma_sync_sg_for_cpu(struct pci_dev *pdev,
-					   struct scatterlist *sg,
-					   int nents, int direction)
-{
-	dma_sync_sg_for_cpu(&pdev->dev, sg, nents,
-			    (enum dma_data_direction) direction);
-}
-
-static inline void pci_dma_sync_sg_for_device(struct pci_dev *pdev,
-					      struct scatterlist *sg,
-					      int nelems, int direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-/* Return whether the given PCI device DMA address mask can
- * be supported properly.  For example, if your device can
- * only drive the low 24-bits during PCI bus mastering, then
- * you would pass 0x00ffffff as the mask to this function.
- */
-extern int pci_dma_supported(struct pci_dev *hwdev, u64 mask);
-
-/* PCI IOMMU mapping bypass support. */
-
-/* PCI 64-bit addressing works for all slots on all controller
- * types on sparc64.  However, it requires that the device
- * can drive enough of the 64 bits.
- */
-#define PCI64_REQUIRED_MASK	(~(dma64_addr_t)0)
-#define PCI64_ADDR_BASE		0xfffc000000000000UL
-
-static inline int pci_dma_mapping_error(dma_addr_t dma_addr)
-{
-	return dma_mapping_error(dma_addr);
-}
-
-#ifdef CONFIG_PCI
-static inline void pci_dma_burst_advice(struct pci_dev *pdev,
-					enum pci_dma_burst_strategy *strat,
-					unsigned long *strategy_parameter)
-{
-	unsigned long cacheline_size;
-	u8 byte;
-
-	pci_read_config_byte(pdev, PCI_CACHE_LINE_SIZE, &byte);
-	if (byte == 0)
-		cacheline_size = 1024;
-	else
-		cacheline_size = (int) byte * 4;
-
-	*strat = PCI_DMA_BURST_BOUNDARY;
-	*strategy_parameter = cacheline_size;
-}
-#endif
-
-/* Return the index of the PCI controller for device PDEV. */
-
-extern int pci_domain_nr(struct pci_bus *bus);
-static inline int pci_proc_domain(struct pci_bus *bus)
-{
-	return 1;
-}
-
-/* Platform support for /proc/bus/pci/X/Y mmap()s. */
-
-#define HAVE_PCI_MMAP
-#define HAVE_ARCH_PCI_GET_UNMAPPED_AREA
-#define get_pci_unmapped_area get_fb_unmapped_area
-
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			       enum pci_mmap_state mmap_state,
-			       int write_combine);
-
-extern void
-pcibios_resource_to_bus(struct pci_dev *dev, struct pci_bus_region *region,
-			struct resource *res);
-
-extern void
-pcibios_bus_to_resource(struct pci_dev *dev, struct resource *res,
-			struct pci_bus_region *region);
-
-extern struct resource *pcibios_select_root(struct pci_dev *, struct resource *);
-
-static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
-{
-	return PCI_IRQ_NONE;
-}
-
-struct device_node;
-extern struct device_node *pci_device_to_OF_node(struct pci_dev *pdev);
-
-#define HAVE_ARCH_PCI_RESOURCE_TO_USER
-extern void pci_resource_to_user(const struct pci_dev *dev, int bar,
-				 const struct resource *rsrc,
-				 resource_size_t *start, resource_size_t *end);
-#endif /* __KERNEL__ */
-
-#endif /* __SPARC64_PCI_H */
+#include <asm-sparc/pci.h>
diff --git a/include/asm-sparc64/percpu.h b/include/asm-sparc64/percpu.h
index bee64593023..292729bb350 100644
--- a/include/asm-sparc64/percpu.h
+++ b/include/asm-sparc64/percpu.h
@@ -1,28 +1 @@
-#ifndef __ARCH_SPARC64_PERCPU__
-#define __ARCH_SPARC64_PERCPU__
-
-#include <linux/compiler.h>
-
-register unsigned long __local_per_cpu_offset asm("g5");
-
-#ifdef CONFIG_SMP
-
-extern void real_setup_per_cpu_areas(void);
-
-extern unsigned long __per_cpu_base;
-extern unsigned long __per_cpu_shift;
-#define __per_cpu_offset(__cpu) \
-	(__per_cpu_base + ((unsigned long)(__cpu) << __per_cpu_shift))
-#define per_cpu_offset(x) (__per_cpu_offset(x))
-
-#define __my_cpu_offset __local_per_cpu_offset
-
-#else /* ! SMP */
-
-#define real_setup_per_cpu_areas()		do { } while (0)
-
-#endif	/* SMP */
-
-#include <asm-generic/percpu.h>
-
-#endif /* __ARCH_SPARC64_PERCPU__ */
+#include <asm-sparc/percpu.h>
diff --git a/include/asm-sparc64/perfctr.h b/include/asm-sparc64/perfctr.h
index 836873002b7..52073a9f8e3 100644
--- a/include/asm-sparc64/perfctr.h
+++ b/include/asm-sparc64/perfctr.h
@@ -1,173 +1 @@
-/*----------------------------------------
-  PERFORMANCE INSTRUMENTATION  
-  Guillaume Thouvenin           08/10/98
-  David S. Miller               10/06/98
-  ---------------------------------------*/
-#ifndef PERF_COUNTER_API
-#define PERF_COUNTER_API
-
-/* sys_perfctr() interface.  First arg is operation code
- * from enumeration below.  The meaning of further arguments
- * are determined by the operation code.
- *
- * int sys_perfctr(int opcode, unsigned long arg0,
- *                 unsigned long arg1, unsigned long arg2)
- *
- * Pointers which are passed by the user are pointers to 64-bit
- * integers.
- *
- * Once enabled, performance counter state is retained until the
- * process either exits or performs an exec.  That is, performance
- * counters remain enabled for fork/clone children.
- */
-enum perfctr_opcode {
-	/* Enable UltraSparc performance counters, ARG0 is pointer
-	 * to 64-bit accumulator for D0 counter in PIC, ARG1 is pointer
-	 * to 64-bit accumulator for D1 counter.  ARG2 is a pointer to
-	 * the initial PCR register value to use.
-	 */
-	PERFCTR_ON,
-
-	/* Disable UltraSparc performance counters.  The PCR is written
-	 * with zero and the user counter accumulator pointers and
-	 * working PCR register value are forgotten.
-	 */
-	PERFCTR_OFF,
-
-	/* Add current D0 and D1 PIC values into user pointers given
-	 * in PERFCTR_ON operation.  The PIC is cleared before returning.
-	 */
-	PERFCTR_READ,
-
-	/* Clear the PIC register. */
-	PERFCTR_CLRPIC,
-
-	/* Begin using a new PCR value, the pointer to which is passed
-	 * in ARG0.  The PIC is also cleared after the new PCR value is
-	 * written.
-	 */
-	PERFCTR_SETPCR,
-
-	/* Store in pointer given in ARG0 the current PCR register value
-	 * being used.
-	 */
-	PERFCTR_GETPCR
-};
-
-/* I don't want the kernel's namespace to be polluted with this
- * stuff when this file is included.  --DaveM
- */
-#ifndef __KERNEL__
-
-#define  PRIV 0x00000001
-#define  SYS  0x00000002
-#define  USR  0x00000004
-
-/* Pic.S0 Selection Bit Field Encoding, Ultra-I/II  */
-#define  CYCLE_CNT            0x00000000
-#define  INSTR_CNT            0x00000010
-#define  DISPATCH0_IC_MISS    0x00000020
-#define  DISPATCH0_STOREBUF   0x00000030
-#define  IC_REF               0x00000080
-#define  DC_RD                0x00000090
-#define  DC_WR                0x000000A0
-#define  LOAD_USE             0x000000B0
-#define  EC_REF               0x000000C0
-#define  EC_WRITE_HIT_RDO     0x000000D0
-#define  EC_SNOOP_INV         0x000000E0
-#define  EC_RD_HIT            0x000000F0
-
-/* Pic.S0 Selection Bit Field Encoding, Ultra-III  */
-#define  US3_CYCLE_CNT	      	0x00000000
-#define  US3_INSTR_CNT	      	0x00000010
-#define  US3_DISPATCH0_IC_MISS	0x00000020
-#define  US3_DISPATCH0_BR_TGT	0x00000030
-#define  US3_DISPATCH0_2ND_BR	0x00000040
-#define  US3_RSTALL_STOREQ	0x00000050
-#define  US3_RSTALL_IU_USE	0x00000060
-#define  US3_IC_REF		0x00000080
-#define  US3_DC_RD		0x00000090
-#define  US3_DC_WR		0x000000a0
-#define  US3_EC_REF		0x000000c0
-#define  US3_EC_WR_HIT_RTO	0x000000d0
-#define  US3_EC_SNOOP_INV	0x000000e0
-#define  US3_EC_RD_MISS		0x000000f0
-#define  US3_PC_PORT0_RD	0x00000100
-#define  US3_SI_SNOOP		0x00000110
-#define  US3_SI_CIQ_FLOW	0x00000120
-#define  US3_SI_OWNED		0x00000130
-#define  US3_SW_COUNT_0		0x00000140
-#define  US3_IU_BR_MISS_TAKEN	0x00000150
-#define  US3_IU_BR_COUNT_TAKEN	0x00000160
-#define  US3_DISP_RS_MISPRED	0x00000170
-#define  US3_FA_PIPE_COMPL	0x00000180
-#define  US3_MC_READS_0		0x00000200
-#define  US3_MC_READS_1		0x00000210
-#define  US3_MC_READS_2		0x00000220
-#define  US3_MC_READS_3		0x00000230
-#define  US3_MC_STALLS_0	0x00000240
-#define  US3_MC_STALLS_2	0x00000250
-
-/* Pic.S1 Selection Bit Field Encoding, Ultra-I/II  */
-#define  CYCLE_CNT_D1         0x00000000
-#define  INSTR_CNT_D1         0x00000800
-#define  DISPATCH0_IC_MISPRED 0x00001000
-#define  DISPATCH0_FP_USE     0x00001800
-#define  IC_HIT               0x00004000
-#define  DC_RD_HIT            0x00004800
-#define  DC_WR_HIT            0x00005000
-#define  LOAD_USE_RAW         0x00005800
-#define  EC_HIT               0x00006000
-#define  EC_WB                0x00006800
-#define  EC_SNOOP_CB          0x00007000
-#define  EC_IT_HIT            0x00007800
-
-/* Pic.S1 Selection Bit Field Encoding, Ultra-III  */
-#define  US3_CYCLE_CNT_D1	0x00000000
-#define  US3_INSTR_CNT_D1	0x00000800
-#define  US3_DISPATCH0_MISPRED	0x00001000
-#define  US3_IC_MISS_CANCELLED	0x00001800
-#define  US3_RE_ENDIAN_MISS	0x00002000
-#define  US3_RE_FPU_BYPASS	0x00002800
-#define  US3_RE_DC_MISS		0x00003000
-#define  US3_RE_EC_MISS		0x00003800
-#define  US3_IC_MISS		0x00004000
-#define  US3_DC_RD_MISS		0x00004800
-#define  US3_DC_WR_MISS		0x00005000
-#define  US3_RSTALL_FP_USE	0x00005800
-#define  US3_EC_MISSES		0x00006000
-#define  US3_EC_WB		0x00006800
-#define  US3_EC_SNOOP_CB	0x00007000
-#define  US3_EC_IC_MISS		0x00007800
-#define  US3_RE_PC_MISS		0x00008000
-#define  US3_ITLB_MISS		0x00008800
-#define  US3_DTLB_MISS		0x00009000
-#define  US3_WC_MISS		0x00009800
-#define  US3_WC_SNOOP_CB	0x0000a000
-#define  US3_WC_SCRUBBED	0x0000a800
-#define  US3_WC_WB_WO_READ	0x0000b000
-#define  US3_PC_SOFT_HIT	0x0000c000
-#define  US3_PC_SNOOP_INV	0x0000c800
-#define  US3_PC_HARD_HIT	0x0000d000
-#define  US3_PC_PORT1_RD	0x0000d800
-#define  US3_SW_COUNT_1		0x0000e000
-#define  US3_IU_STAT_BR_MIS_UNTAKEN	0x0000e800
-#define  US3_IU_STAT_BR_COUNT_UNTAKEN	0x0000f000
-#define  US3_PC_MS_MISSES	0x0000f800
-#define  US3_MC_WRITES_0	0x00010800
-#define  US3_MC_WRITES_1	0x00011000
-#define  US3_MC_WRITES_2	0x00011800
-#define  US3_MC_WRITES_3	0x00012000
-#define  US3_MC_STALLS_1	0x00012800
-#define  US3_MC_STALLS_3	0x00013000
-#define  US3_RE_RAW_MISS	0x00013800
-#define  US3_FM_PIPE_COMPLETION	0x00014000
-
-struct vcounter_struct {
-  unsigned long long vcnt0;
-  unsigned long long vcnt1;
-};
-
-#endif /* !(__KERNEL__) */
-
-#endif /* !(PERF_COUNTER_API) */
+#include <asm-sparc/perfctr.h>
diff --git a/include/asm-sparc64/pgalloc.h b/include/asm-sparc64/pgalloc.h
index 326de104d01..bec31641011 100644
--- a/include/asm-sparc64/pgalloc.h
+++ b/include/asm-sparc64/pgalloc.h
@@ -1,81 +1 @@
-#ifndef _SPARC64_PGALLOC_H
-#define _SPARC64_PGALLOC_H
-
-#include <linux/kernel.h>
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/slab.h>
-#include <linux/quicklist.h>
-
-#include <asm/spitfire.h>
-#include <asm/cpudata.h>
-#include <asm/cacheflush.h>
-#include <asm/page.h>
-
-/* Page table allocation/freeing. */
-
-static inline pgd_t *pgd_alloc(struct mm_struct *mm)
-{
-	return quicklist_alloc(0, GFP_KERNEL, NULL);
-}
-
-static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
-{
-	quicklist_free(0, NULL, pgd);
-}
-
-#define pud_populate(MM, PUD, PMD)	pud_set(PUD, PMD)
-
-static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long addr)
-{
-	return quicklist_alloc(0, GFP_KERNEL, NULL);
-}
-
-static inline void pmd_free(struct mm_struct *mm, pmd_t *pmd)
-{
-	quicklist_free(0, NULL, pmd);
-}
-
-static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
-					  unsigned long address)
-{
-	return quicklist_alloc(0, GFP_KERNEL, NULL);
-}
-
-static inline pgtable_t pte_alloc_one(struct mm_struct *mm,
-					unsigned long address)
-{
-	struct page *page;
-	void *pg;
-
-	pg = quicklist_alloc(0, GFP_KERNEL, NULL);
-	if (!pg)
-		return NULL;
-	page = virt_to_page(pg);
-	pgtable_page_ctor(page);
-	return page;
-}
-		
-static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
-{
-	quicklist_free(0, NULL, pte);
-}
-
-static inline void pte_free(struct mm_struct *mm, pgtable_t ptepage)
-{
-	pgtable_page_dtor(ptepage);
-	quicklist_free_page(0, NULL, ptepage);
-}
-
-
-#define pmd_populate_kernel(MM, PMD, PTE)	pmd_set(PMD, PTE)
-#define pmd_populate(MM,PMD,PTE_PAGE)		\
-	pmd_populate_kernel(MM,PMD,page_address(PTE_PAGE))
-#define pmd_pgtable(pmd) pmd_page(pmd)
-
-static inline void check_pgt_cache(void)
-{
-	quicklist_trim(0, NULL, 25, 16);
-}
-
-#endif /* _SPARC64_PGALLOC_H */
+#include <asm-sparc/pgalloc.h>
diff --git a/include/asm-sparc64/pgtable.h b/include/asm-sparc64/pgtable.h
index b87017747b5..9decbd99aef 100644
--- a/include/asm-sparc64/pgtable.h
+++ b/include/asm-sparc64/pgtable.h
@@ -1,781 +1 @@
-/*
- * pgtable.h: SpitFire page table operations.
- *
- * Copyright 1996,1997 David S. Miller (davem@caip.rutgers.edu)
- * Copyright 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-
-#ifndef _SPARC64_PGTABLE_H
-#define _SPARC64_PGTABLE_H
-
-/* This file contains the functions and defines necessary to modify and use
- * the SpitFire page tables.
- */
-
-#include <asm-generic/pgtable-nopud.h>
-
-#include <linux/compiler.h>
-#include <linux/const.h>
-#include <asm/types.h>
-#include <asm/spitfire.h>
-#include <asm/asi.h>
-#include <asm/system.h>
-#include <asm/page.h>
-#include <asm/processor.h>
-
-/* The kernel image occupies 0x4000000 to 0x6000000 (4MB --> 96MB).
- * The page copy blockops can use 0x6000000 to 0x8000000.
- * The TSB is mapped in the 0x8000000 to 0xa000000 range.
- * The PROM resides in an area spanning 0xf0000000 to 0x100000000.
- * The vmalloc area spans 0x100000000 to 0x200000000.
- * Since modules need to be in the lowest 32-bits of the address space,
- * we place them right before the OBP area from 0x10000000 to 0xf0000000.
- * There is a single static kernel PMD which maps from 0x0 to address
- * 0x400000000.
- */
-#define	TLBTEMP_BASE		_AC(0x0000000006000000,UL)
-#define	TSBMAP_BASE		_AC(0x0000000008000000,UL)
-#define MODULES_VADDR		_AC(0x0000000010000000,UL)
-#define MODULES_LEN		_AC(0x00000000e0000000,UL)
-#define MODULES_END		_AC(0x00000000f0000000,UL)
-#define LOW_OBP_ADDRESS		_AC(0x00000000f0000000,UL)
-#define HI_OBP_ADDRESS		_AC(0x0000000100000000,UL)
-#define VMALLOC_START		_AC(0x0000000100000000,UL)
-#define VMALLOC_END		_AC(0x0000000200000000,UL)
-#define VMEMMAP_BASE		_AC(0x0000000200000000,UL)
-
-#define vmemmap			((struct page *)VMEMMAP_BASE)
-
-/* XXX All of this needs to be rethought so we can take advantage
- * XXX cheetah's full 64-bit virtual address space, ie. no more hole
- * XXX in the middle like on spitfire. -DaveM
- */
-/*
- * Given a virtual address, the lowest PAGE_SHIFT bits determine offset
- * into the page; the next higher PAGE_SHIFT-3 bits determine the pte#
- * in the proper pagetable (the -3 is from the 8 byte ptes, and each page
- * table is a single page long). The next higher PMD_BITS determine pmd# 
- * in the proper pmdtable (where we must have PMD_BITS <= (PAGE_SHIFT-2) 
- * since the pmd entries are 4 bytes, and each pmd page is a single page 
- * long). Finally, the higher few bits determine pgde#.
- */
-
-/* PMD_SHIFT determines the size of the area a second-level page
- * table can map
- */
-#define PMD_SHIFT	(PAGE_SHIFT + (PAGE_SHIFT-3))
-#define PMD_SIZE	(_AC(1,UL) << PMD_SHIFT)
-#define PMD_MASK	(~(PMD_SIZE-1))
-#define PMD_BITS	(PAGE_SHIFT - 2)
-
-/* PGDIR_SHIFT determines what a third-level page table entry can map */
-#define PGDIR_SHIFT	(PAGE_SHIFT + (PAGE_SHIFT-3) + PMD_BITS)
-#define PGDIR_SIZE	(_AC(1,UL) << PGDIR_SHIFT)
-#define PGDIR_MASK	(~(PGDIR_SIZE-1))
-#define PGDIR_BITS	(PAGE_SHIFT - 2)
-
-#ifndef __ASSEMBLY__
-
-#include <linux/sched.h>
-
-/* Entries per page directory level. */
-#define PTRS_PER_PTE	(1UL << (PAGE_SHIFT-3))
-#define PTRS_PER_PMD	(1UL << PMD_BITS)
-#define PTRS_PER_PGD	(1UL << PGDIR_BITS)
-
-/* Kernel has a separate 44bit address space. */
-#define FIRST_USER_ADDRESS	0
-
-#define pte_ERROR(e)	__builtin_trap()
-#define pmd_ERROR(e)	__builtin_trap()
-#define pgd_ERROR(e)	__builtin_trap()
-
-#endif /* !(__ASSEMBLY__) */
-
-/* PTE bits which are the same in SUN4U and SUN4V format.  */
-#define _PAGE_VALID	  _AC(0x8000000000000000,UL) /* Valid TTE            */
-#define _PAGE_R	  	  _AC(0x8000000000000000,UL) /* Keep ref bit uptodate*/
-
-/* SUN4U pte bits... */
-#define _PAGE_SZ4MB_4U	  _AC(0x6000000000000000,UL) /* 4MB Page             */
-#define _PAGE_SZ512K_4U	  _AC(0x4000000000000000,UL) /* 512K Page            */
-#define _PAGE_SZ64K_4U	  _AC(0x2000000000000000,UL) /* 64K Page             */
-#define _PAGE_SZ8K_4U	  _AC(0x0000000000000000,UL) /* 8K Page              */
-#define _PAGE_NFO_4U	  _AC(0x1000000000000000,UL) /* No Fault Only        */
-#define _PAGE_IE_4U	  _AC(0x0800000000000000,UL) /* Invert Endianness    */
-#define _PAGE_SOFT2_4U	  _AC(0x07FC000000000000,UL) /* Software bits, set 2 */
-#define _PAGE_RES1_4U	  _AC(0x0002000000000000,UL) /* Reserved             */
-#define _PAGE_SZ32MB_4U	  _AC(0x0001000000000000,UL) /* (Panther) 32MB page  */
-#define _PAGE_SZ256MB_4U  _AC(0x2001000000000000,UL) /* (Panther) 256MB page */
-#define _PAGE_SZALL_4U	  _AC(0x6001000000000000,UL) /* All pgsz bits        */
-#define _PAGE_SN_4U	  _AC(0x0000800000000000,UL) /* (Cheetah) Snoop      */
-#define _PAGE_RES2_4U	  _AC(0x0000780000000000,UL) /* Reserved             */
-#define _PAGE_PADDR_4U	  _AC(0x000007FFFFFFE000,UL) /* (Cheetah) pa[42:13]  */
-#define _PAGE_SOFT_4U	  _AC(0x0000000000001F80,UL) /* Software bits:       */
-#define _PAGE_EXEC_4U	  _AC(0x0000000000001000,UL) /* Executable SW bit    */
-#define _PAGE_MODIFIED_4U _AC(0x0000000000000800,UL) /* Modified (dirty)     */
-#define _PAGE_FILE_4U	  _AC(0x0000000000000800,UL) /* Pagecache page       */
-#define _PAGE_ACCESSED_4U _AC(0x0000000000000400,UL) /* Accessed (ref'd)     */
-#define _PAGE_READ_4U	  _AC(0x0000000000000200,UL) /* Readable SW Bit      */
-#define _PAGE_WRITE_4U	  _AC(0x0000000000000100,UL) /* Writable SW Bit      */
-#define _PAGE_PRESENT_4U  _AC(0x0000000000000080,UL) /* Present              */
-#define _PAGE_L_4U	  _AC(0x0000000000000040,UL) /* Locked TTE           */
-#define _PAGE_CP_4U	  _AC(0x0000000000000020,UL) /* Cacheable in P-Cache */
-#define _PAGE_CV_4U	  _AC(0x0000000000000010,UL) /* Cacheable in V-Cache */
-#define _PAGE_E_4U	  _AC(0x0000000000000008,UL) /* side-Effect          */
-#define _PAGE_P_4U	  _AC(0x0000000000000004,UL) /* Privileged Page      */
-#define _PAGE_W_4U	  _AC(0x0000000000000002,UL) /* Writable             */
-
-/* SUN4V pte bits... */
-#define _PAGE_NFO_4V	  _AC(0x4000000000000000,UL) /* No Fault Only        */
-#define _PAGE_SOFT2_4V	  _AC(0x3F00000000000000,UL) /* Software bits, set 2 */
-#define _PAGE_MODIFIED_4V _AC(0x2000000000000000,UL) /* Modified (dirty)     */
-#define _PAGE_ACCESSED_4V _AC(0x1000000000000000,UL) /* Accessed (ref'd)     */
-#define _PAGE_READ_4V	  _AC(0x0800000000000000,UL) /* Readable SW Bit      */
-#define _PAGE_WRITE_4V	  _AC(0x0400000000000000,UL) /* Writable SW Bit      */
-#define _PAGE_PADDR_4V	  _AC(0x00FFFFFFFFFFE000,UL) /* paddr[55:13]         */
-#define _PAGE_IE_4V	  _AC(0x0000000000001000,UL) /* Invert Endianness    */
-#define _PAGE_E_4V	  _AC(0x0000000000000800,UL) /* side-Effect          */
-#define _PAGE_CP_4V	  _AC(0x0000000000000400,UL) /* Cacheable in P-Cache */
-#define _PAGE_CV_4V	  _AC(0x0000000000000200,UL) /* Cacheable in V-Cache */
-#define _PAGE_P_4V	  _AC(0x0000000000000100,UL) /* Privileged Page      */
-#define _PAGE_EXEC_4V	  _AC(0x0000000000000080,UL) /* Executable Page      */
-#define _PAGE_W_4V	  _AC(0x0000000000000040,UL) /* Writable             */
-#define _PAGE_SOFT_4V	  _AC(0x0000000000000030,UL) /* Software bits        */
-#define _PAGE_FILE_4V	  _AC(0x0000000000000020,UL) /* Pagecache page       */
-#define _PAGE_PRESENT_4V  _AC(0x0000000000000010,UL) /* Present              */
-#define _PAGE_RESV_4V	  _AC(0x0000000000000008,UL) /* Reserved             */
-#define _PAGE_SZ16GB_4V	  _AC(0x0000000000000007,UL) /* 16GB Page            */
-#define _PAGE_SZ2GB_4V	  _AC(0x0000000000000006,UL) /* 2GB Page             */
-#define _PAGE_SZ256MB_4V  _AC(0x0000000000000005,UL) /* 256MB Page           */
-#define _PAGE_SZ32MB_4V	  _AC(0x0000000000000004,UL) /* 32MB Page            */
-#define _PAGE_SZ4MB_4V	  _AC(0x0000000000000003,UL) /* 4MB Page             */
-#define _PAGE_SZ512K_4V	  _AC(0x0000000000000002,UL) /* 512K Page            */
-#define _PAGE_SZ64K_4V	  _AC(0x0000000000000001,UL) /* 64K Page             */
-#define _PAGE_SZ8K_4V	  _AC(0x0000000000000000,UL) /* 8K Page              */
-#define _PAGE_SZALL_4V	  _AC(0x0000000000000007,UL) /* All pgsz bits        */
-
-#if PAGE_SHIFT == 13
-#define _PAGE_SZBITS_4U	_PAGE_SZ8K_4U
-#define _PAGE_SZBITS_4V	_PAGE_SZ8K_4V
-#elif PAGE_SHIFT == 16
-#define _PAGE_SZBITS_4U	_PAGE_SZ64K_4U
-#define _PAGE_SZBITS_4V	_PAGE_SZ64K_4V
-#elif PAGE_SHIFT == 19
-#define _PAGE_SZBITS_4U	_PAGE_SZ512K_4U
-#define _PAGE_SZBITS_4V	_PAGE_SZ512K_4V
-#elif PAGE_SHIFT == 22
-#define _PAGE_SZBITS_4U	_PAGE_SZ4MB_4U
-#define _PAGE_SZBITS_4V	_PAGE_SZ4MB_4V
-#else
-#error Wrong PAGE_SHIFT specified
-#endif
-
-#if defined(CONFIG_HUGETLB_PAGE_SIZE_4MB)
-#define _PAGE_SZHUGE_4U	_PAGE_SZ4MB_4U
-#define _PAGE_SZHUGE_4V	_PAGE_SZ4MB_4V
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_512K)
-#define _PAGE_SZHUGE_4U	_PAGE_SZ512K_4U
-#define _PAGE_SZHUGE_4V	_PAGE_SZ512K_4V
-#elif defined(CONFIG_HUGETLB_PAGE_SIZE_64K)
-#define _PAGE_SZHUGE_4U	_PAGE_SZ64K_4U
-#define _PAGE_SZHUGE_4V	_PAGE_SZ64K_4V
-#endif
-
-/* These are actually filled in at boot time by sun4{u,v}_pgprot_init() */
-#define __P000	__pgprot(0)
-#define __P001	__pgprot(0)
-#define __P010	__pgprot(0)
-#define __P011	__pgprot(0)
-#define __P100	__pgprot(0)
-#define __P101	__pgprot(0)
-#define __P110	__pgprot(0)
-#define __P111	__pgprot(0)
-
-#define __S000	__pgprot(0)
-#define __S001	__pgprot(0)
-#define __S010	__pgprot(0)
-#define __S011	__pgprot(0)
-#define __S100	__pgprot(0)
-#define __S101	__pgprot(0)
-#define __S110	__pgprot(0)
-#define __S111	__pgprot(0)
-
-#ifndef __ASSEMBLY__
-
-extern pte_t mk_pte_io(unsigned long, pgprot_t, int, unsigned long);
-
-extern unsigned long pte_sz_bits(unsigned long size);
-
-extern pgprot_t PAGE_KERNEL;
-extern pgprot_t PAGE_KERNEL_LOCKED;
-extern pgprot_t PAGE_COPY;
-extern pgprot_t PAGE_SHARED;
-
-/* XXX This uglyness is for the atyfb driver's sparc mmap() support. XXX */
-extern unsigned long _PAGE_IE;
-extern unsigned long _PAGE_E;
-extern unsigned long _PAGE_CACHE;
-
-extern unsigned long pg_iobits;
-extern unsigned long _PAGE_ALL_SZ_BITS;
-extern unsigned long _PAGE_SZBITS;
-
-extern struct page *mem_map_zero;
-#define ZERO_PAGE(vaddr)	(mem_map_zero)
-
-/* PFNs are real physical page numbers.  However, mem_map only begins to record
- * per-page information starting at pfn_base.  This is to handle systems where
- * the first physical page in the machine is at some huge physical address,
- * such as 4GB.   This is common on a partitioned E10000, for example.
- */
-static inline pte_t pfn_pte(unsigned long pfn, pgprot_t prot)
-{
-	unsigned long paddr = pfn << PAGE_SHIFT;
-	unsigned long sz_bits;
-
-	sz_bits = 0UL;
-	if (_PAGE_SZBITS_4U != 0UL || _PAGE_SZBITS_4V != 0UL) {
-		__asm__ __volatile__(
-		"\n661:	sethi		%%uhi(%1), %0\n"
-		"	sllx		%0, 32, %0\n"
-		"	.section	.sun4v_2insn_patch, \"ax\"\n"
-		"	.word		661b\n"
-		"	mov		%2, %0\n"
-		"	nop\n"
-		"	.previous\n"
-		: "=r" (sz_bits)
-		: "i" (_PAGE_SZBITS_4U), "i" (_PAGE_SZBITS_4V));
-	}
-	return __pte(paddr | sz_bits | pgprot_val(prot));
-}
-#define mk_pte(page, pgprot)	pfn_pte(page_to_pfn(page), (pgprot))
-
-/* This one can be done with two shifts.  */
-static inline unsigned long pte_pfn(pte_t pte)
-{
-	unsigned long ret;
-
-	__asm__ __volatile__(
-	"\n661:	sllx		%1, %2, %0\n"
-	"	srlx		%0, %3, %0\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sllx		%1, %4, %0\n"
-	"	srlx		%0, %5, %0\n"
-	"	.previous\n"
-	: "=r" (ret)
-	: "r" (pte_val(pte)),
-	  "i" (21), "i" (21 + PAGE_SHIFT),
-	  "i" (8), "i" (8 + PAGE_SHIFT));
-
-	return ret;
-}
-#define pte_page(x) pfn_to_page(pte_pfn(x))
-
-static inline pte_t pte_modify(pte_t pte, pgprot_t prot)
-{
-	unsigned long mask, tmp;
-
-	/* SUN4U: 0x600307ffffffecb8 (negated == 0x9ffcf80000001347)
-	 * SUN4V: 0x30ffffffffffee17 (negated == 0xcf000000000011e8)
-	 *
-	 * Even if we use negation tricks the result is still a 6
-	 * instruction sequence, so don't try to play fancy and just
-	 * do the most straightforward implementation.
-	 *
-	 * Note: We encode this into 3 sun4v 2-insn patch sequences.
-	 */
-
-	__asm__ __volatile__(
-	"\n661:	sethi		%%uhi(%2), %1\n"
-	"	sethi		%%hi(%2), %0\n"
-	"\n662:	or		%1, %%ulo(%2), %1\n"
-	"	or		%0, %%lo(%2), %0\n"
-	"\n663:	sllx		%1, 32, %1\n"
-	"	or		%0, %1, %0\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%3), %1\n"
-	"	sethi		%%hi(%3), %0\n"
-	"	.word		662b\n"
-	"	or		%1, %%ulo(%3), %1\n"
-	"	or		%0, %%lo(%3), %0\n"
-	"	.word		663b\n"
-	"	sllx		%1, 32, %1\n"
-	"	or		%0, %1, %0\n"
-	"	.previous\n"
-	: "=r" (mask), "=r" (tmp)
-	: "i" (_PAGE_PADDR_4U | _PAGE_MODIFIED_4U | _PAGE_ACCESSED_4U |
-	       _PAGE_CP_4U | _PAGE_CV_4U | _PAGE_E_4U | _PAGE_PRESENT_4U |
-	       _PAGE_SZBITS_4U),
-	  "i" (_PAGE_PADDR_4V | _PAGE_MODIFIED_4V | _PAGE_ACCESSED_4V |
-	       _PAGE_CP_4V | _PAGE_CV_4V | _PAGE_E_4V | _PAGE_PRESENT_4V |
-	       _PAGE_SZBITS_4V));
-
-	return __pte((pte_val(pte) & mask) | (pgprot_val(prot) & ~mask));
-}
-
-static inline pte_t pgoff_to_pte(unsigned long off)
-{
-	off <<= PAGE_SHIFT;
-
-	__asm__ __volatile__(
-	"\n661:	or		%0, %2, %0\n"
-	"	.section	.sun4v_1insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	or		%0, %3, %0\n"
-	"	.previous\n"
-	: "=r" (off)
-	: "0" (off), "i" (_PAGE_FILE_4U), "i" (_PAGE_FILE_4V));
-
-	return __pte(off);
-}
-
-static inline pgprot_t pgprot_noncached(pgprot_t prot)
-{
-	unsigned long val = pgprot_val(prot);
-
-	__asm__ __volatile__(
-	"\n661:	andn		%0, %2, %0\n"
-	"	or		%0, %3, %0\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	andn		%0, %4, %0\n"
-	"	or		%0, %5, %0\n"
-	"	.previous\n"
-	: "=r" (val)
-	: "0" (val), "i" (_PAGE_CP_4U | _PAGE_CV_4U), "i" (_PAGE_E_4U),
-	             "i" (_PAGE_CP_4V | _PAGE_CV_4V), "i" (_PAGE_E_4V));
-
-	return __pgprot(val);
-}
-/* Various pieces of code check for platform support by ifdef testing
- * on "pgprot_noncached".  That's broken and should be fixed, but for
- * now...
- */
-#define pgprot_noncached pgprot_noncached
-
-#ifdef CONFIG_HUGETLB_PAGE
-static inline pte_t pte_mkhuge(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	sethi		%%uhi(%1), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	mov		%2, %0\n"
-	"	nop\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_SZHUGE_4U), "i" (_PAGE_SZHUGE_4V));
-
-	return __pte(pte_val(pte) | mask);
-}
-#endif
-
-static inline pte_t pte_mkdirty(pte_t pte)
-{
-	unsigned long val = pte_val(pte), tmp;
-
-	__asm__ __volatile__(
-	"\n661:	or		%0, %3, %0\n"
-	"	nop\n"
-	"\n662:	nop\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%4), %1\n"
-	"	sllx		%1, 32, %1\n"
-	"	.word		662b\n"
-	"	or		%1, %%lo(%4), %1\n"
-	"	or		%0, %1, %0\n"
-	"	.previous\n"
-	: "=r" (val), "=r" (tmp)
-	: "0" (val), "i" (_PAGE_MODIFIED_4U | _PAGE_W_4U),
-	  "i" (_PAGE_MODIFIED_4V | _PAGE_W_4V));
-
-	return __pte(val);
-}
-
-static inline pte_t pte_mkclean(pte_t pte)
-{
-	unsigned long val = pte_val(pte), tmp;
-
-	__asm__ __volatile__(
-	"\n661:	andn		%0, %3, %0\n"
-	"	nop\n"
-	"\n662:	nop\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%4), %1\n"
-	"	sllx		%1, 32, %1\n"
-	"	.word		662b\n"
-	"	or		%1, %%lo(%4), %1\n"
-	"	andn		%0, %1, %0\n"
-	"	.previous\n"
-	: "=r" (val), "=r" (tmp)
-	: "0" (val), "i" (_PAGE_MODIFIED_4U | _PAGE_W_4U),
-	  "i" (_PAGE_MODIFIED_4V | _PAGE_W_4V));
-
-	return __pte(val);
-}
-
-static inline pte_t pte_mkwrite(pte_t pte)
-{
-	unsigned long val = pte_val(pte), mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_WRITE_4U), "i" (_PAGE_WRITE_4V));
-
-	return __pte(val | mask);
-}
-
-static inline pte_t pte_wrprotect(pte_t pte)
-{
-	unsigned long val = pte_val(pte), tmp;
-
-	__asm__ __volatile__(
-	"\n661:	andn		%0, %3, %0\n"
-	"	nop\n"
-	"\n662:	nop\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%4), %1\n"
-	"	sllx		%1, 32, %1\n"
-	"	.word		662b\n"
-	"	or		%1, %%lo(%4), %1\n"
-	"	andn		%0, %1, %0\n"
-	"	.previous\n"
-	: "=r" (val), "=r" (tmp)
-	: "0" (val), "i" (_PAGE_WRITE_4U | _PAGE_W_4U),
-	  "i" (_PAGE_WRITE_4V | _PAGE_W_4V));
-
-	return __pte(val);
-}
-
-static inline pte_t pte_mkold(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
-
-	mask |= _PAGE_R;
-
-	return __pte(pte_val(pte) & ~mask);
-}
-
-static inline pte_t pte_mkyoung(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
-
-	mask |= _PAGE_R;
-
-	return __pte(pte_val(pte) | mask);
-}
-
-static inline pte_t pte_mkspecial(pte_t pte)
-{
-	return pte;
-}
-
-static inline unsigned long pte_young(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_ACCESSED_4U), "i" (_PAGE_ACCESSED_4V));
-
-	return (pte_val(pte) & mask);
-}
-
-static inline unsigned long pte_dirty(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_MODIFIED_4U), "i" (_PAGE_MODIFIED_4V));
-
-	return (pte_val(pte) & mask);
-}
-
-static inline unsigned long pte_write(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	mov		%1, %0\n"
-	"	nop\n"
-	"	.section	.sun4v_2insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	sethi		%%uhi(%2), %0\n"
-	"	sllx		%0, 32, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_WRITE_4U), "i" (_PAGE_WRITE_4V));
-
-	return (pte_val(pte) & mask);
-}
-
-static inline unsigned long pte_exec(pte_t pte)
-{
-	unsigned long mask;
-
-	__asm__ __volatile__(
-	"\n661:	sethi		%%hi(%1), %0\n"
-	"	.section	.sun4v_1insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	mov		%2, %0\n"
-	"	.previous\n"
-	: "=r" (mask)
-	: "i" (_PAGE_EXEC_4U), "i" (_PAGE_EXEC_4V));
-
-	return (pte_val(pte) & mask);
-}
-
-static inline unsigned long pte_file(pte_t pte)
-{
-	unsigned long val = pte_val(pte);
-
-	__asm__ __volatile__(
-	"\n661:	and		%0, %2, %0\n"
-	"	.section	.sun4v_1insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	and		%0, %3, %0\n"
-	"	.previous\n"
-	: "=r" (val)
-	: "0" (val), "i" (_PAGE_FILE_4U), "i" (_PAGE_FILE_4V));
-
-	return val;
-}
-
-static inline unsigned long pte_present(pte_t pte)
-{
-	unsigned long val = pte_val(pte);
-
-	__asm__ __volatile__(
-	"\n661:	and		%0, %2, %0\n"
-	"	.section	.sun4v_1insn_patch, \"ax\"\n"
-	"	.word		661b\n"
-	"	and		%0, %3, %0\n"
-	"	.previous\n"
-	: "=r" (val)
-	: "0" (val), "i" (_PAGE_PRESENT_4U), "i" (_PAGE_PRESENT_4V));
-
-	return val;
-}
-
-static inline int pte_special(pte_t pte)
-{
-	return 0;
-}
-
-#define pmd_set(pmdp, ptep)	\
-	(pmd_val(*(pmdp)) = (__pa((unsigned long) (ptep)) >> 11UL))
-#define pud_set(pudp, pmdp)	\
-	(pud_val(*(pudp)) = (__pa((unsigned long) (pmdp)) >> 11UL))
-#define __pmd_page(pmd)		\
-	((unsigned long) __va((((unsigned long)pmd_val(pmd))<<11UL)))
-#define pmd_page(pmd) 			virt_to_page((void *)__pmd_page(pmd))
-#define pud_page_vaddr(pud)		\
-	((unsigned long) __va((((unsigned long)pud_val(pud))<<11UL)))
-#define pud_page(pud) 			virt_to_page((void *)pud_page_vaddr(pud))
-#define pmd_none(pmd)			(!pmd_val(pmd))
-#define pmd_bad(pmd)			(0)
-#define pmd_present(pmd)		(pmd_val(pmd) != 0U)
-#define pmd_clear(pmdp)			(pmd_val(*(pmdp)) = 0U)
-#define pud_none(pud)			(!pud_val(pud))
-#define pud_bad(pud)			(0)
-#define pud_present(pud)		(pud_val(pud) != 0U)
-#define pud_clear(pudp)			(pud_val(*(pudp)) = 0U)
-
-/* Same in both SUN4V and SUN4U.  */
-#define pte_none(pte) 			(!pte_val(pte))
-
-/* to find an entry in a page-table-directory. */
-#define pgd_index(address)	(((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD - 1))
-#define pgd_offset(mm, address)	((mm)->pgd + pgd_index(address))
-
-/* to find an entry in a kernel page-table-directory */
-#define pgd_offset_k(address) pgd_offset(&init_mm, address)
-
-/* Find an entry in the second-level page table.. */
-#define pmd_offset(pudp, address)	\
-	((pmd_t *) pud_page_vaddr(*(pudp)) + \
-	 (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1)))
-
-/* Find an entry in the third-level page table.. */
-#define pte_index(dir, address)	\
-	((pte_t *) __pmd_page(*(dir)) + \
-	 ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
-#define pte_offset_kernel		pte_index
-#define pte_offset_map			pte_index
-#define pte_offset_map_nested		pte_index
-#define pte_unmap(pte)			do { } while (0)
-#define pte_unmap_nested(pte)		do { } while (0)
-
-/* Actual page table PTE updates.  */
-extern void tlb_batch_add(struct mm_struct *mm, unsigned long vaddr, pte_t *ptep, pte_t orig);
-
-static inline void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte)
-{
-	pte_t orig = *ptep;
-
-	*ptep = pte;
-
-	/* It is more efficient to let flush_tlb_kernel_range()
-	 * handle init_mm tlb flushes.
-	 *
-	 * SUN4V NOTE: _PAGE_VALID is the same value in both the SUN4U
-	 *             and SUN4V pte layout, so this inline test is fine.
-	 */
-	if (likely(mm != &init_mm) && (pte_val(orig) & _PAGE_VALID))
-		tlb_batch_add(mm, addr, ptep, orig);
-}
-
-#define pte_clear(mm,addr,ptep)		\
-	set_pte_at((mm), (addr), (ptep), __pte(0UL))
-
-#ifdef DCACHE_ALIASING_POSSIBLE
-#define __HAVE_ARCH_MOVE_PTE
-#define move_pte(pte, prot, old_addr, new_addr)				\
-({									\
- 	pte_t newpte = (pte);						\
-	if (tlb_type != hypervisor && pte_present(pte)) {		\
-		unsigned long this_pfn = pte_pfn(pte);			\
-									\
-		if (pfn_valid(this_pfn) &&				\
-		    (((old_addr) ^ (new_addr)) & (1 << 13)))		\
-			flush_dcache_page_all(current->mm,		\
-					      pfn_to_page(this_pfn));	\
-	}								\
-	newpte;								\
-})
-#endif
-
-extern pgd_t swapper_pg_dir[2048];
-extern pmd_t swapper_low_pmd_dir[2048];
-
-extern void paging_init(void);
-extern unsigned long find_ecache_flush_span(unsigned long size);
-
-/* These do nothing with the way I have things setup. */
-#define mmu_lockarea(vaddr, len)		(vaddr)
-#define mmu_unlockarea(vaddr, len)		do { } while(0)
-
-struct vm_area_struct;
-extern void update_mmu_cache(struct vm_area_struct *, unsigned long, pte_t);
-
-/* Encode and de-code a swap entry */
-#define __swp_type(entry)	(((entry).val >> PAGE_SHIFT) & 0xffUL)
-#define __swp_offset(entry)	((entry).val >> (PAGE_SHIFT + 8UL))
-#define __swp_entry(type, offset)	\
-	( (swp_entry_t) \
-	  { \
-		(((long)(type) << PAGE_SHIFT) | \
-                 ((long)(offset) << (PAGE_SHIFT + 8UL))) \
-	  } )
-#define __pte_to_swp_entry(pte)		((swp_entry_t) { pte_val(pte) })
-#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
-
-/* File offset in PTE support. */
-extern unsigned long pte_file(pte_t);
-#define pte_to_pgoff(pte)	(pte_val(pte) >> PAGE_SHIFT)
-extern pte_t pgoff_to_pte(unsigned long);
-#define PTE_FILE_MAX_BITS	(64UL - PAGE_SHIFT - 1UL)
-
-extern unsigned long *sparc64_valid_addr_bitmap;
-
-/* Needs to be defined here and not in linux/mm.h, as it is arch dependent */
-#define kern_addr_valid(addr)	\
-	(test_bit(__pa((unsigned long)(addr))>>22, sparc64_valid_addr_bitmap))
-
-extern int page_in_phys_avail(unsigned long paddr);
-
-extern int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
-			       unsigned long pfn,
-			       unsigned long size, pgprot_t prot);
-
-/*
- * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
- * its high 4 bits.  These macros/functions put it there or get it from there.
- */
-#define MK_IOSPACE_PFN(space, pfn)	(pfn | (space << (BITS_PER_LONG - 4)))
-#define GET_IOSPACE(pfn)		(pfn >> (BITS_PER_LONG - 4))
-#define GET_PFN(pfn)			(pfn & 0x0fffffffffffffffUL)
-
-#include <asm-generic/pgtable.h>
-
-/* We provide our own get_unmapped_area to cope with VA holes and
- * SHM area cache aliasing for userland.
- */
-#define HAVE_ARCH_UNMAPPED_AREA
-#define HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
-
-/* We provide a special get_unmapped_area for framebuffer mmaps to try and use
- * the largest alignment possible such that larget PTEs can be used.
- */
-extern unsigned long get_fb_unmapped_area(struct file *filp, unsigned long,
-					  unsigned long, unsigned long,
-					  unsigned long);
-#define HAVE_ARCH_FB_UNMAPPED_AREA
-
-extern void pgtable_cache_init(void);
-extern void sun4v_register_fault_status(void);
-extern void sun4v_ktsb_register(void);
-extern void __init cheetah_ecache_flush_init(void);
-extern void sun4v_patch_tlb_handlers(void);
-
-extern unsigned long cmdline_memory_size;
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(_SPARC64_PGTABLE_H) */
+#include <asm-sparc/pgtable.h>
diff --git a/include/asm-sparc64/pil.h b/include/asm-sparc64/pil.h
index eaac842d88c..d805f33f1e0 100644
--- a/include/asm-sparc64/pil.h
+++ b/include/asm-sparc64/pil.h
@@ -1,21 +1 @@
-#ifndef _SPARC64_PIL_H
-#define _SPARC64_PIL_H
-
-/* To avoid some locking problems, we hard allocate certain PILs
- * for SMP cross call messages that must do a etrap/rtrap.
- *
- * A local_irq_disable() does not block the cross call delivery, so
- * when SMP locking is an issue we reschedule the event into a PIL
- * interrupt which is blocked by local_irq_disable().
- *
- * In fact any XCALL which has to etrap/rtrap has a problem because
- * it is difficult to prevent rtrap from running BH's, and that would
- * need to be done if the XCALL arrived while %pil==15.
- */
-#define PIL_SMP_CALL_FUNC	1
-#define PIL_SMP_RECEIVE_SIGNAL	2
-#define PIL_SMP_CAPTURE		3
-#define PIL_SMP_CTX_NEW_VERSION	4
-#define PIL_DEVICE_IRQ		5
-
-#endif /* !(_SPARC64_PIL_H) */
+#include <asm-sparc/pil.h>
diff --git a/include/asm-sparc64/poll.h b/include/asm-sparc64/poll.h
index ebeeb3816c4..8e2f31b4641 100644
--- a/include/asm-sparc64/poll.h
+++ b/include/asm-sparc64/poll.h
@@ -1,12 +1 @@
-#ifndef __SPARC64_POLL_H
-#define __SPARC64_POLL_H
-
-#define POLLWRNORM	POLLOUT
-#define POLLWRBAND	256
-#define POLLMSG		512
-#define POLLREMOVE	1024
-#define POLLRDHUP       2048
-
-#include <asm-generic/poll.h>
-
-#endif
+#include <asm-sparc/poll.h>
diff --git a/include/asm-sparc64/posix_types.h b/include/asm-sparc64/posix_types.h
index 4eaaa019663..8cee9920023 100644
--- a/include/asm-sparc64/posix_types.h
+++ b/include/asm-sparc64/posix_types.h
@@ -1,122 +1 @@
-#ifndef __ARCH_SPARC64_POSIX_TYPES_H
-#define __ARCH_SPARC64_POSIX_TYPES_H
-
-/*
- * This file is generally used by user-level software, so you need to
- * be a little careful about namespace pollution etc.  Also, we cannot
- * assume GCC is being used.
- */
-
-typedef unsigned long          __kernel_size_t;
-typedef long                   __kernel_ssize_t;
-typedef long                   __kernel_ptrdiff_t;
-typedef long                   __kernel_time_t;
-typedef long                   __kernel_clock_t;
-typedef int                    __kernel_pid_t;
-typedef int                    __kernel_ipc_pid_t;
-typedef unsigned int           __kernel_uid_t;
-typedef unsigned int           __kernel_gid_t;
-typedef unsigned long          __kernel_ino_t;
-typedef unsigned int           __kernel_mode_t;
-typedef unsigned short         __kernel_umode_t;
-typedef unsigned int           __kernel_nlink_t;
-typedef int                    __kernel_daddr_t;
-typedef long                   __kernel_off_t;
-typedef char *                 __kernel_caddr_t;
-typedef unsigned short	       __kernel_uid16_t;
-typedef unsigned short	       __kernel_gid16_t;
-typedef int                    __kernel_clockid_t;
-typedef int                    __kernel_timer_t;
-
-typedef unsigned short 	       __kernel_old_uid_t;
-typedef unsigned short         __kernel_old_gid_t;
-typedef __kernel_uid_t	       __kernel_uid32_t;
-typedef __kernel_gid_t	       __kernel_gid32_t;
-
-typedef unsigned int	       __kernel_old_dev_t;
-
-/* Note this piece of asymmetry from the v9 ABI.  */
-typedef int		       __kernel_suseconds_t;
-
-#ifdef __GNUC__
-typedef long long              __kernel_loff_t;
-#endif
-
-typedef struct {
-	int	val[2];
-} __kernel_fsid_t;
-
-#if defined(__KERNEL__)
-
-#undef __FD_SET
-static inline void __FD_SET(unsigned long fd, __kernel_fd_set *fdsetp)
-{
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	fdsetp->fds_bits[_tmp] |= (1UL<<_rem);
-}
-
-#undef __FD_CLR
-static inline void __FD_CLR(unsigned long fd, __kernel_fd_set *fdsetp)
-{
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	fdsetp->fds_bits[_tmp] &= ~(1UL<<_rem);
-}
-
-#undef __FD_ISSET
-static inline int __FD_ISSET(unsigned long fd, __const__ __kernel_fd_set *p)
-{ 
-	unsigned long _tmp = fd / __NFDBITS;
-	unsigned long _rem = fd % __NFDBITS;
-	return (p->fds_bits[_tmp] & (1UL<<_rem)) != 0;
-}
-
-/*
- * This will unroll the loop for the normal constant cases (8 or 32 longs,
- * for 256 and 1024-bit fd_sets respectively)
- */
-#undef __FD_ZERO
-static inline void __FD_ZERO(__kernel_fd_set *p)
-{
-	unsigned long *tmp = p->fds_bits;
-	int i;
-
-	if (__builtin_constant_p(__FDSET_LONGS)) {
-		switch (__FDSET_LONGS) {
-			case 32:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
-			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
-			  tmp[16] = 0; tmp[17] = 0; tmp[18] = 0; tmp[19] = 0;
-			  tmp[20] = 0; tmp[21] = 0; tmp[22] = 0; tmp[23] = 0;
-			  tmp[24] = 0; tmp[25] = 0; tmp[26] = 0; tmp[27] = 0;
-			  tmp[28] = 0; tmp[29] = 0; tmp[30] = 0; tmp[31] = 0;
-			  return;
-			case 16:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
-			  tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;
-			  return;
-			case 8:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  tmp[ 4] = 0; tmp[ 5] = 0; tmp[ 6] = 0; tmp[ 7] = 0;
-			  return;
-			case 4:
-			  tmp[ 0] = 0; tmp[ 1] = 0; tmp[ 2] = 0; tmp[ 3] = 0;
-			  return;
-		}
-	}
-	i = __FDSET_LONGS;
-	while (i) {
-		i--;
-		*tmp = 0;
-		tmp++;
-	}
-}
-
-#endif /* defined(__KERNEL__) */
-
-#endif /* !(__ARCH_SPARC64_POSIX_TYPES_H) */
+#include <asm-sparc/posix_types.h>
diff --git a/include/asm-sparc64/processor.h b/include/asm-sparc64/processor.h
index 26b4e525576..21de6cc182e 100644
--- a/include/asm-sparc64/processor.h
+++ b/include/asm-sparc64/processor.h
@@ -1,237 +1 @@
-/*
- * include/asm-sparc64/processor.h
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __ASM_SPARC64_PROCESSOR_H
-#define __ASM_SPARC64_PROCESSOR_H
-
-/*
- * Sparc64 implementation of macro that returns current
- * instruction pointer ("program counter").
- */
-#define current_text_addr() ({ void *pc; __asm__("rd %%pc, %0" : "=r" (pc)); pc; })
-
-#include <asm/asi.h>
-#include <asm/pstate.h>
-#include <asm/ptrace.h>
-#include <asm/page.h>
-
-/* The sparc has no problems with write protection */
-#define wp_works_ok 1
-#define wp_works_ok__is_a_macro /* for versions in ksyms.c */
-
-/*
- * User lives in his very own context, and cannot reference us. Note
- * that TASK_SIZE is a misnomer, it really gives maximum user virtual 
- * address that the kernel will allocate out.
- *
- * XXX No longer using virtual page tables, kill this upper limit...
- */
-#define VA_BITS		44
-#ifndef __ASSEMBLY__
-#define VPTE_SIZE	(1UL << (VA_BITS - PAGE_SHIFT + 3))
-#else
-#define VPTE_SIZE	(1 << (VA_BITS - PAGE_SHIFT + 3))
-#endif
-
-#define TASK_SIZE	((unsigned long)-VPTE_SIZE)
-#define TASK_SIZE_OF(tsk) \
-	(test_tsk_thread_flag(tsk,TIF_32BIT) ? \
-	 (1UL << 32UL) : TASK_SIZE)
-#ifdef __KERNEL__
-
-#define STACK_TOP32	((1UL << 32UL) - PAGE_SIZE)
-#define STACK_TOP64	(0x0000080000000000UL - (1UL << 32UL))
-
-#define STACK_TOP	(test_thread_flag(TIF_32BIT) ? \
-			 STACK_TOP32 : STACK_TOP64)
-
-#define STACK_TOP_MAX	STACK_TOP64
-
-#endif
-
-#ifndef __ASSEMBLY__
-
-typedef struct {
-	unsigned char seg;
-} mm_segment_t;
-
-/* The Sparc processor specific thread struct. */
-/* XXX This should die, everything can go into thread_info now. */
-struct thread_struct {
-#ifdef CONFIG_DEBUG_SPINLOCK
-	/* How many spinlocks held by this thread.
-	 * Used with spin lock debugging to catch tasks
-	 * sleeping illegally with locks held.
-	 */
-	int smp_lock_count;
-	unsigned int smp_lock_pc;
-#else
-	int dummy; /* f'in gcc bug... */
-#endif
-};
-
-#endif /* !(__ASSEMBLY__) */
-
-#ifndef CONFIG_DEBUG_SPINLOCK
-#define INIT_THREAD  {			\
-	0,				\
-}
-#else /* CONFIG_DEBUG_SPINLOCK */
-#define INIT_THREAD  {					\
-/* smp_lock_count, smp_lock_pc, */			\
-   0,		   0,					\
-}
-#endif /* !(CONFIG_DEBUG_SPINLOCK) */
-
-#ifndef __ASSEMBLY__
-
-#include <linux/types.h>
-
-/* Return saved PC of a blocked thread. */
-struct task_struct;
-extern unsigned long thread_saved_pc(struct task_struct *);
-
-/* On Uniprocessor, even in RMO processes see TSO semantics */
-#ifdef CONFIG_SMP
-#define TSTATE_INITIAL_MM	TSTATE_TSO
-#else
-#define TSTATE_INITIAL_MM	TSTATE_RMO
-#endif
-
-/* Do necessary setup to start up a newly executed thread. */
-#define start_thread(regs, pc, sp) \
-do { \
-	unsigned long __asi = ASI_PNF; \
-	regs->tstate = (regs->tstate & (TSTATE_CWP)) | (TSTATE_INITIAL_MM|TSTATE_IE) | (__asi << 24UL); \
-	regs->tpc = ((pc & (~3)) - 4); \
-	regs->tnpc = regs->tpc + 4; \
-	regs->y = 0; \
-	set_thread_wstate(1 << 3); \
-	if (current_thread_info()->utraps) { \
-		if (*(current_thread_info()->utraps) < 2) \
-			kfree(current_thread_info()->utraps); \
-		else \
-			(*(current_thread_info()->utraps))--; \
-		current_thread_info()->utraps = NULL; \
-	} \
-	__asm__ __volatile__( \
-	"stx		%%g0, [%0 + %2 + 0x00]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x08]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x10]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x18]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x20]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x28]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x30]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x38]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x40]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x48]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x50]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x58]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x60]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x68]\n\t" \
-	"stx		%1,   [%0 + %2 + 0x70]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x78]\n\t" \
-	"wrpr		%%g0, (1 << 3), %%wstate\n\t" \
-	: \
-	: "r" (regs), "r" (sp - sizeof(struct reg_window) - STACK_BIAS), \
-	  "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))); \
-} while (0)
-
-#define start_thread32(regs, pc, sp) \
-do { \
-	unsigned long __asi = ASI_PNF; \
-	pc &= 0x00000000ffffffffUL; \
-	sp &= 0x00000000ffffffffUL; \
-	regs->tstate = (regs->tstate & (TSTATE_CWP))|(TSTATE_INITIAL_MM|TSTATE_IE|TSTATE_AM) | (__asi << 24UL); \
-	regs->tpc = ((pc & (~3)) - 4); \
-	regs->tnpc = regs->tpc + 4; \
-	regs->y = 0; \
-	set_thread_wstate(2 << 3); \
-	if (current_thread_info()->utraps) { \
-		if (*(current_thread_info()->utraps) < 2) \
-			kfree(current_thread_info()->utraps); \
-		else \
-			(*(current_thread_info()->utraps))--; \
-		current_thread_info()->utraps = NULL; \
-	} \
-	__asm__ __volatile__( \
-	"stx		%%g0, [%0 + %2 + 0x00]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x08]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x10]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x18]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x20]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x28]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x30]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x38]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x40]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x48]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x50]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x58]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x60]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x68]\n\t" \
-	"stx		%1,   [%0 + %2 + 0x70]\n\t" \
-	"stx		%%g0, [%0 + %2 + 0x78]\n\t" \
-	"wrpr		%%g0, (2 << 3), %%wstate\n\t" \
-	: \
-	: "r" (regs), "r" (sp - sizeof(struct reg_window32)), \
-	  "i" ((const unsigned long)(&((struct pt_regs *)0)->u_regs[0]))); \
-} while (0)
-
-/* Free all resources held by a thread. */
-#define release_thread(tsk)		do { } while (0)
-
-/* Prepare to copy thread state - unlazy all lazy status */
-#define prepare_to_copy(tsk)	do { } while (0)
-
-extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
-
-extern unsigned long get_wchan(struct task_struct *task);
-
-#define task_pt_regs(tsk) (task_thread_info(tsk)->kregs)
-#define KSTK_EIP(tsk)  (task_pt_regs(tsk)->tpc)
-#define KSTK_ESP(tsk)  (task_pt_regs(tsk)->u_regs[UREG_FP])
-
-#define cpu_relax()	barrier()
-
-/* Prefetch support.  This is tuned for UltraSPARC-III and later.
- * UltraSPARC-I will treat these as nops, and UltraSPARC-II has
- * a shallower prefetch queue than later chips.
- */
-#define ARCH_HAS_PREFETCH
-#define ARCH_HAS_PREFETCHW
-#define ARCH_HAS_SPINLOCK_PREFETCH
-
-static inline void prefetch(const void *x)
-{
-	/* We do not use the read prefetch mnemonic because that
-	 * prefetches into the prefetch-cache which only is accessible
-	 * by floating point operations in UltraSPARC-III and later.
-	 * By contrast, "#one_write" prefetches into the L2 cache
-	 * in shared state.
-	 */
-	__asm__ __volatile__("prefetch [%0], #one_write"
-			     : /* no outputs */
-			     : "r" (x));
-}
-
-static inline void prefetchw(const void *x)
-{
-	/* The most optimal prefetch to use for writes is
-	 * "#n_writes".  This brings the cacheline into the
-	 * L2 cache in "owned" state.
-	 */
-	__asm__ __volatile__("prefetch [%0], #n_writes"
-			     : /* no outputs */
-			     : "r" (x));
-}
-
-#define spin_lock_prefetch(x)	prefetchw(x)
-
-#define HAVE_ARCH_PICK_MMAP_LAYOUT
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__ASM_SPARC64_PROCESSOR_H) */
+#include <asm-sparc/processor.h>
diff --git a/include/asm-sparc64/psrcompat.h b/include/asm-sparc64/psrcompat.h
index 44b6327dbbf..587846f4835 100644
--- a/include/asm-sparc64/psrcompat.h
+++ b/include/asm-sparc64/psrcompat.h
@@ -1,45 +1 @@
-#ifndef _SPARC64_PSRCOMPAT_H
-#define _SPARC64_PSRCOMPAT_H
-
-#include <asm/pstate.h>
-
-/* Old 32-bit PSR fields for the compatibility conversion code. */
-#define PSR_CWP     0x0000001f         /* current window pointer     */
-#define PSR_ET      0x00000020         /* enable traps field         */
-#define PSR_PS      0x00000040         /* previous privilege level   */
-#define PSR_S       0x00000080         /* current privilege level    */
-#define PSR_PIL     0x00000f00         /* processor interrupt level  */
-#define PSR_EF      0x00001000         /* enable floating point      */
-#define PSR_EC      0x00002000         /* enable co-processor        */
-#define PSR_SYSCALL 0x00004000         /* inside of a syscall        */
-#define PSR_LE      0x00008000         /* SuperSparcII little-endian */
-#define PSR_ICC     0x00f00000         /* integer condition codes    */
-#define PSR_C       0x00100000         /* carry bit                  */
-#define PSR_V       0x00200000         /* overflow bit               */
-#define PSR_Z       0x00400000         /* zero bit                   */
-#define PSR_N       0x00800000         /* negative bit               */
-#define PSR_VERS    0x0f000000         /* cpu-version field          */
-#define PSR_IMPL    0xf0000000         /* cpu-implementation field   */
-
-#define PSR_V8PLUS  0xff000000         /* fake impl/ver, meaning a 64bit CPU is present */
-#define PSR_XCC	    0x000f0000         /* if PSR_V8PLUS, this is %xcc */
-
-static inline unsigned int tstate_to_psr(unsigned long tstate)
-{
-	return ((tstate & TSTATE_CWP)			|
-		PSR_S					|
-		((tstate & TSTATE_ICC) >> 12)		|
-		((tstate & TSTATE_XCC) >> 20)		|
-		((tstate & TSTATE_SYSCALL) ? PSR_SYSCALL : 0) |
-		PSR_V8PLUS);
-}
-
-static inline unsigned long psr_to_tstate_icc(unsigned int psr)
-{
-	unsigned long tstate = ((unsigned long)(psr & PSR_ICC)) << 12;
-	if ((psr & (PSR_VERS|PSR_IMPL)) == PSR_V8PLUS)
-		tstate |= ((unsigned long)(psr & PSR_XCC)) << 20;
-	return tstate;
-}
-
-#endif /* !(_SPARC64_PSRCOMPAT_H) */
+#include <asm-sparc/psrcompat.h>
diff --git a/include/asm-sparc64/pstate.h b/include/asm-sparc64/pstate.h
index a26a53777bb..3ccf0be2536 100644
--- a/include/asm-sparc64/pstate.h
+++ b/include/asm-sparc64/pstate.h
@@ -1,91 +1 @@
-#ifndef _SPARC64_PSTATE_H
-#define _SPARC64_PSTATE_H
-
-#include <linux/const.h>
-
-/* The V9 PSTATE Register (with SpitFire extensions).
- *
- * -----------------------------------------------------------------------
- * | Resv | IG | MG | CLE | TLE |  MM  | RED | PEF | AM | PRIV | IE | AG |
- * -----------------------------------------------------------------------
- *  63  12  11   10    9     8    7   6   5     4     3     2     1    0
- */
-#define PSTATE_IG   _AC(0x0000000000000800,UL) /* Interrupt Globals.	*/
-#define PSTATE_MG   _AC(0x0000000000000400,UL) /* MMU Globals.		*/
-#define PSTATE_CLE  _AC(0x0000000000000200,UL) /* Current Little Endian.*/
-#define PSTATE_TLE  _AC(0x0000000000000100,UL) /* Trap Little Endian.	*/
-#define PSTATE_MM   _AC(0x00000000000000c0,UL) /* Memory Model.		*/
-#define PSTATE_TSO  _AC(0x0000000000000000,UL) /* MM: TotalStoreOrder	*/
-#define PSTATE_PSO  _AC(0x0000000000000040,UL) /* MM: PartialStoreOrder	*/
-#define PSTATE_RMO  _AC(0x0000000000000080,UL) /* MM: RelaxedMemoryOrder*/
-#define PSTATE_RED  _AC(0x0000000000000020,UL) /* Reset Error Debug.	*/
-#define PSTATE_PEF  _AC(0x0000000000000010,UL) /* Floating Point Enable.*/
-#define PSTATE_AM   _AC(0x0000000000000008,UL) /* Address Mask.		*/
-#define PSTATE_PRIV _AC(0x0000000000000004,UL) /* Privilege.		*/
-#define PSTATE_IE   _AC(0x0000000000000002,UL) /* Interrupt Enable.	*/
-#define PSTATE_AG   _AC(0x0000000000000001,UL) /* Alternate Globals.	*/
-
-/* The V9 TSTATE Register (with SpitFire and Linux extensions).
- *
- * ---------------------------------------------------------------------
- * |  Resv |  GL  |  CCR  |  ASI  |  %pil  |  PSTATE  |  Resv  |  CWP  |
- * ---------------------------------------------------------------------
- *  63   43 42  40 39   32 31   24 23    20 19       8 7      5 4     0
- */
-#define TSTATE_GL	_AC(0x0000070000000000,UL) /* Global reg level  */
-#define TSTATE_CCR	_AC(0x000000ff00000000,UL) /* Condition Codes.	*/
-#define TSTATE_XCC	_AC(0x000000f000000000,UL) /* Condition Codes.	*/
-#define TSTATE_XNEG	_AC(0x0000008000000000,UL) /* %xcc Negative.	*/
-#define TSTATE_XZERO	_AC(0x0000004000000000,UL) /* %xcc Zero.	*/
-#define TSTATE_XOVFL	_AC(0x0000002000000000,UL) /* %xcc Overflow.	*/
-#define TSTATE_XCARRY	_AC(0x0000001000000000,UL) /* %xcc Carry.	*/
-#define TSTATE_ICC	_AC(0x0000000f00000000,UL) /* Condition Codes.	*/
-#define TSTATE_INEG	_AC(0x0000000800000000,UL) /* %icc Negative.	*/
-#define TSTATE_IZERO	_AC(0x0000000400000000,UL) /* %icc Zero.	*/
-#define TSTATE_IOVFL	_AC(0x0000000200000000,UL) /* %icc Overflow.	*/
-#define TSTATE_ICARRY	_AC(0x0000000100000000,UL) /* %icc Carry.	*/
-#define TSTATE_ASI	_AC(0x00000000ff000000,UL) /* AddrSpace ID.	*/
-#define TSTATE_PIL	_AC(0x0000000000f00000,UL) /* %pil (Linux traps)*/
-#define TSTATE_PSTATE	_AC(0x00000000000fff00,UL) /* PSTATE.		*/
-#define TSTATE_IG	_AC(0x0000000000080000,UL) /* Interrupt Globals.*/
-#define TSTATE_MG	_AC(0x0000000000040000,UL) /* MMU Globals.	*/
-#define TSTATE_CLE	_AC(0x0000000000020000,UL) /* CurrLittleEndian.	*/
-#define TSTATE_TLE	_AC(0x0000000000010000,UL) /* TrapLittleEndian.	*/
-#define TSTATE_MM	_AC(0x000000000000c000,UL) /* Memory Model.	*/
-#define TSTATE_TSO	_AC(0x0000000000000000,UL) /* MM: TSO		*/
-#define TSTATE_PSO	_AC(0x0000000000004000,UL) /* MM: PSO		*/
-#define TSTATE_RMO	_AC(0x0000000000008000,UL) /* MM: RMO		*/
-#define TSTATE_RED	_AC(0x0000000000002000,UL) /* Reset Error Debug.*/
-#define TSTATE_PEF	_AC(0x0000000000001000,UL) /* FPU Enable.	*/
-#define TSTATE_AM	_AC(0x0000000000000800,UL) /* Address Mask.	*/
-#define TSTATE_PRIV	_AC(0x0000000000000400,UL) /* Privilege.	*/
-#define TSTATE_IE	_AC(0x0000000000000200,UL) /* Interrupt Enable.	*/
-#define TSTATE_AG	_AC(0x0000000000000100,UL) /* Alternate Globals.*/
-#define TSTATE_SYSCALL	_AC(0x0000000000000020,UL) /* in syscall trap   */
-#define TSTATE_CWP	_AC(0x000000000000001f,UL) /* Curr Win-Pointer.	*/
-
-/* Floating-Point Registers State Register.
- *
- * --------------------------------
- * |  Resv  |  FEF  |  DU  |  DL  |
- * --------------------------------
- *  63     3    2       1      0
- */
-#define FPRS_FEF	_AC(0x0000000000000004,UL) /* FPU Enable.	*/
-#define FPRS_DU		_AC(0x0000000000000002,UL) /* Dirty Upper.	*/
-#define FPRS_DL		_AC(0x0000000000000001,UL) /* Dirty Lower.	*/
-
-/* Version Register.
- *
- * ------------------------------------------------------
- * | MANUF | IMPL | MASK | Resv | MAXTL | Resv | MAXWIN |
- * ------------------------------------------------------
- *  63   48 47  32 31  24 23  16 15    8 7    5 4      0
- */
-#define VERS_MANUF	_AC(0xffff000000000000,UL) /* Manufacturer.	*/
-#define VERS_IMPL	_AC(0x0000ffff00000000,UL) /* Implementation.	*/
-#define VERS_MASK	_AC(0x00000000ff000000,UL) /* Mask Set Revision.*/
-#define VERS_MAXTL	_AC(0x000000000000ff00,UL) /* Max Trap Level.	*/
-#define VERS_MAXWIN	_AC(0x000000000000001f,UL) /* Max RegWindow Idx.*/
-
-#endif /* !(_SPARC64_PSTATE_H) */
+#include <asm-sparc/pstate.h>
diff --git a/include/asm-sparc64/ptrace.h b/include/asm-sparc64/ptrace.h
index b163da79bb6..1a55b9fb3b0 100644
--- a/include/asm-sparc64/ptrace.h
+++ b/include/asm-sparc64/ptrace.h
@@ -1,346 +1 @@
-#ifndef _SPARC64_PTRACE_H
-#define _SPARC64_PTRACE_H
-
-#include <asm/pstate.h>
-
-/* This struct defines the way the registers are stored on the 
- * stack during a system call and basically all traps.
- */
-
-/* This magic value must have the low 9 bits clear,
- * as that is where we encode the %tt value, see below.
- */
-#define PT_REGS_MAGIC 0x57ac6c00
-
-#ifndef __ASSEMBLY__
-
-#include <linux/types.h>
-
-struct pt_regs {
-	unsigned long u_regs[16]; /* globals and ins */
-	unsigned long tstate;
-	unsigned long tpc;
-	unsigned long tnpc;
-	unsigned int y;
-
-	/* We encode a magic number, PT_REGS_MAGIC, along
-	 * with the %tt (trap type) register value at trap
-	 * entry time.  The magic number allows us to identify
-	 * accurately a trap stack frame in the stack
-	 * unwinder, and the %tt value allows us to test
-	 * things like "in a system call" etc. for an arbitray
-	 * process.
-	 *
-	 * The PT_REGS_MAGIC is choosen such that it can be
-	 * loaded completely using just a sethi instruction.
-	 */
-	unsigned int magic;
-};
-
-static inline int pt_regs_trap_type(struct pt_regs *regs)
-{
-	return regs->magic & 0x1ff;
-}
-
-static inline bool pt_regs_is_syscall(struct pt_regs *regs)
-{
-	return (regs->tstate & TSTATE_SYSCALL);
-}
-
-static inline bool pt_regs_clear_syscall(struct pt_regs *regs)
-{
-	return (regs->tstate &= ~TSTATE_SYSCALL);
-}
-
-struct pt_regs32 {
-	unsigned int psr;
-	unsigned int pc;
-	unsigned int npc;
-	unsigned int y;
-	unsigned int u_regs[16]; /* globals and ins */
-};
-
-#define UREG_G0        0
-#define UREG_G1        1
-#define UREG_G2        2
-#define UREG_G3        3
-#define UREG_G4        4
-#define UREG_G5        5
-#define UREG_G6        6
-#define UREG_G7        7
-#define UREG_I0        8
-#define UREG_I1        9
-#define UREG_I2        10
-#define UREG_I3        11
-#define UREG_I4        12
-#define UREG_I5        13
-#define UREG_I6        14
-#define UREG_I7        15
-#define UREG_FP        UREG_I6
-#define UREG_RETPC     UREG_I7
-
-/* A V9 register window */
-struct reg_window {
-	unsigned long locals[8];
-	unsigned long ins[8];
-};
-
-/* A 32-bit register window. */
-struct reg_window32 {
-	unsigned int locals[8];
-	unsigned int ins[8];
-};
-
-/* A V9 Sparc stack frame */
-struct sparc_stackf {
-	unsigned long locals[8];
-        unsigned long ins[6];
-	struct sparc_stackf *fp;
-	unsigned long callers_pc;
-	char *structptr;
-	unsigned long xargs[6];
-	unsigned long xxargs[1];
-};	
-
-/* A 32-bit Sparc stack frame */
-struct sparc_stackf32 {
-	unsigned int locals[8];
-        unsigned int ins[6];
-	unsigned int fp;
-	unsigned int callers_pc;
-	unsigned int structptr;
-	unsigned int xargs[6];
-	unsigned int xxargs[1];
-};	
-
-struct sparc_trapf {
-	unsigned long locals[8];
-	unsigned long ins[8];
-	unsigned long _unused;
-	struct pt_regs *regs;
-};
-
-#define TRACEREG_SZ	sizeof(struct pt_regs)
-#define STACKFRAME_SZ	sizeof(struct sparc_stackf)
-
-#define TRACEREG32_SZ	sizeof(struct pt_regs32)
-#define STACKFRAME32_SZ	sizeof(struct sparc_stackf32)
-
-#ifdef __KERNEL__
-
-struct global_reg_snapshot {
-	unsigned long		tstate;
-	unsigned long		tpc;
-	unsigned long		tnpc;
-	unsigned long		o7;
-	unsigned long		i7;
-	struct thread_info	*thread;
-	unsigned long		pad1;
-	unsigned long		pad2;
-};
-
-#define __ARCH_WANT_COMPAT_SYS_PTRACE
-
-#define force_successful_syscall_return()	    \
-do {	current_thread_info()->syscall_noerror = 1; \
-} while (0)
-#define user_mode(regs) (!((regs)->tstate & TSTATE_PRIV))
-#define instruction_pointer(regs) ((regs)->tpc)
-#define regs_return_value(regs) ((regs)->u_regs[UREG_I0])
-#ifdef CONFIG_SMP
-extern unsigned long profile_pc(struct pt_regs *);
-#else
-#define profile_pc(regs) instruction_pointer(regs)
-#endif
-extern void show_regs(struct pt_regs *);
-extern void __show_regs(struct pt_regs *);
-#endif
-
-#else /* __ASSEMBLY__ */
-/* For assembly code. */
-#define TRACEREG_SZ		0xa0
-#define STACKFRAME_SZ		0xc0
-
-#define TRACEREG32_SZ		0x50
-#define STACKFRAME32_SZ		0x60
-#endif
-
-#ifdef __KERNEL__
-#define STACK_BIAS		2047
-#endif
-
-/* These are for pt_regs. */
-#define PT_V9_G0     0x00
-#define PT_V9_G1     0x08
-#define PT_V9_G2     0x10
-#define PT_V9_G3     0x18
-#define PT_V9_G4     0x20
-#define PT_V9_G5     0x28
-#define PT_V9_G6     0x30
-#define PT_V9_G7     0x38
-#define PT_V9_I0     0x40
-#define PT_V9_I1     0x48
-#define PT_V9_I2     0x50
-#define PT_V9_I3     0x58
-#define PT_V9_I4     0x60
-#define PT_V9_I5     0x68
-#define PT_V9_I6     0x70
-#define PT_V9_FP     PT_V9_I6
-#define PT_V9_I7     0x78
-#define PT_V9_TSTATE 0x80
-#define PT_V9_TPC    0x88
-#define PT_V9_TNPC   0x90
-#define PT_V9_Y      0x98
-#define PT_V9_MAGIC  0x9c
-#define PT_TSTATE	PT_V9_TSTATE
-#define PT_TPC		PT_V9_TPC
-#define PT_TNPC		PT_V9_TNPC
-
-/* These for pt_regs32. */
-#define PT_PSR    0x0
-#define PT_PC     0x4
-#define PT_NPC    0x8
-#define PT_Y      0xc
-#define PT_G0     0x10
-#define PT_WIM    PT_G0
-#define PT_G1     0x14
-#define PT_G2     0x18
-#define PT_G3     0x1c
-#define PT_G4     0x20
-#define PT_G5     0x24
-#define PT_G6     0x28
-#define PT_G7     0x2c
-#define PT_I0     0x30
-#define PT_I1     0x34
-#define PT_I2     0x38
-#define PT_I3     0x3c
-#define PT_I4     0x40
-#define PT_I5     0x44
-#define PT_I6     0x48
-#define PT_FP     PT_I6
-#define PT_I7     0x4c
-
-/* Reg_window offsets */
-#define RW_V9_L0     0x00
-#define RW_V9_L1     0x08
-#define RW_V9_L2     0x10
-#define RW_V9_L3     0x18
-#define RW_V9_L4     0x20
-#define RW_V9_L5     0x28
-#define RW_V9_L6     0x30
-#define RW_V9_L7     0x38
-#define RW_V9_I0     0x40
-#define RW_V9_I1     0x48
-#define RW_V9_I2     0x50
-#define RW_V9_I3     0x58
-#define RW_V9_I4     0x60
-#define RW_V9_I5     0x68
-#define RW_V9_I6     0x70
-#define RW_V9_I7     0x78
-
-#define RW_L0     0x00
-#define RW_L1     0x04
-#define RW_L2     0x08
-#define RW_L3     0x0c
-#define RW_L4     0x10
-#define RW_L5     0x14
-#define RW_L6     0x18
-#define RW_L7     0x1c
-#define RW_I0     0x20
-#define RW_I1     0x24
-#define RW_I2     0x28
-#define RW_I3     0x2c
-#define RW_I4     0x30
-#define RW_I5     0x34
-#define RW_I6     0x38
-#define RW_I7     0x3c
-
-/* Stack_frame offsets */
-#define SF_V9_L0     0x00
-#define SF_V9_L1     0x08
-#define SF_V9_L2     0x10
-#define SF_V9_L3     0x18
-#define SF_V9_L4     0x20
-#define SF_V9_L5     0x28
-#define SF_V9_L6     0x30
-#define SF_V9_L7     0x38
-#define SF_V9_I0     0x40
-#define SF_V9_I1     0x48
-#define SF_V9_I2     0x50
-#define SF_V9_I3     0x58
-#define SF_V9_I4     0x60
-#define SF_V9_I5     0x68
-#define SF_V9_FP     0x70
-#define SF_V9_PC     0x78
-#define SF_V9_RETP   0x80
-#define SF_V9_XARG0  0x88
-#define SF_V9_XARG1  0x90
-#define SF_V9_XARG2  0x98
-#define SF_V9_XARG3  0xa0
-#define SF_V9_XARG4  0xa8
-#define SF_V9_XARG5  0xb0
-#define SF_V9_XXARG  0xb8
-
-#define SF_L0     0x00
-#define SF_L1     0x04
-#define SF_L2     0x08
-#define SF_L3     0x0c
-#define SF_L4     0x10
-#define SF_L5     0x14
-#define SF_L6     0x18
-#define SF_L7     0x1c
-#define SF_I0     0x20
-#define SF_I1     0x24
-#define SF_I2     0x28
-#define SF_I3     0x2c
-#define SF_I4     0x30
-#define SF_I5     0x34
-#define SF_FP     0x38
-#define SF_PC     0x3c
-#define SF_RETP   0x40
-#define SF_XARG0  0x44
-#define SF_XARG1  0x48
-#define SF_XARG2  0x4c
-#define SF_XARG3  0x50
-#define SF_XARG4  0x54
-#define SF_XARG5  0x58
-#define SF_XXARG  0x5c
-
-#ifdef __KERNEL__
-
-/* global_reg_snapshot offsets */
-#define GR_SNAP_TSTATE	0x00
-#define GR_SNAP_TPC	0x08
-#define GR_SNAP_TNPC	0x10
-#define GR_SNAP_O7	0x18
-#define GR_SNAP_I7	0x20
-#define GR_SNAP_THREAD	0x28
-#define GR_SNAP_PAD1	0x30
-#define GR_SNAP_PAD2	0x38
-
-#endif  /*  __KERNEL__  */
-
-/* Stuff for the ptrace system call */
-#define PTRACE_SPARC_DETACH       11
-#define PTRACE_GETREGS            12
-#define PTRACE_SETREGS            13
-#define PTRACE_GETFPREGS          14
-#define PTRACE_SETFPREGS          15
-#define PTRACE_READDATA           16
-#define PTRACE_WRITEDATA          17
-#define PTRACE_READTEXT           18
-#define PTRACE_WRITETEXT          19
-#define PTRACE_GETFPAREGS         20
-#define PTRACE_SETFPAREGS         21
-
-/* There are for debugging 64-bit processes, either from a 32 or 64 bit
- * parent.  Thus their complements are for debugging 32-bit processes only.
- */
-
-#define PTRACE_GETREGS64	  22
-#define PTRACE_SETREGS64	  23
-/* PTRACE_SYSCALL is 24 */
-#define PTRACE_GETFPREGS64	  25
-#define PTRACE_SETFPREGS64	  26
-
-#endif /* !(_SPARC64_PTRACE_H) */
+#include <asm-sparc/ptrace.h>
diff --git a/include/asm-sparc64/reboot.h b/include/asm-sparc64/reboot.h
index 3f3f43f5be5..0d72eb811cc 100644
--- a/include/asm-sparc64/reboot.h
+++ b/include/asm-sparc64/reboot.h
@@ -1,6 +1 @@
-#ifndef _SPARC64_REBOOT_H
-#define _SPARC64_REBOOT_H
-
-extern void machine_alt_power_off(void);
-
-#endif /* _SPARC64_REBOOT_H */
+#include <asm-sparc/reboot.h>
diff --git a/include/asm-sparc64/reg.h b/include/asm-sparc64/reg.h
index 77aa4804a60..495bab27da0 100644
--- a/include/asm-sparc64/reg.h
+++ b/include/asm-sparc64/reg.h
@@ -1,56 +1 @@
-/*
- * linux/asm-sparc64/reg.h
- * Layout of the registers as expected by gdb on the Sparc
- * we should replace the user.h definitions with those in
- * this file, we don't even use the other 
- * -miguel
- *
- * The names of the structures, constants and aliases in this file
- * have the same names as the sunos ones, some programs rely on these
- * names (gdb for example).
- *
- */
-
-#ifndef __SPARC64_REG_H
-#define __SPARC64_REG_H
-
-struct regs {
-        unsigned long r_g1;
-        unsigned long r_g2;
-        unsigned long r_g3;
-        unsigned long r_g4;
-        unsigned long r_g5;
-        unsigned long r_g6;
-        unsigned long r_g7;
-        unsigned long r_o0;
-        unsigned long r_o1;
-        unsigned long r_o2;
-        unsigned long r_o3;
-        unsigned long r_o4;
-        unsigned long r_o5;
-        unsigned long r_o6;
-        unsigned long r_o7;
-        unsigned long __pad;
-        unsigned long r_tstate;
-        unsigned long r_tpc;
-        unsigned long r_tnpc;
-        unsigned int  r_y;
-        unsigned int  r_fprs;
-};
-
-#define FPU_REGS_TYPE unsigned int
-#define FPU_FSR_TYPE unsigned long
-
-struct fp_status {
-        unsigned long fpu_fr[32];
-        unsigned long Fpu_fsr;
-};
-
-struct fpu {
-	struct fp_status f_fpstatus;
-};
-
-#define fpu_regs  f_fpstatus.fpu_fr
-#define fpu_fsr   f_fpstatus.Fpu_fsr
-
-#endif /* __SPARC64_REG_H */
+#include <asm-sparc/reg.h>
diff --git a/include/asm-sparc64/resource.h b/include/asm-sparc64/resource.h
index 4f08fb5e4ca..46e3bc0de47 100644
--- a/include/asm-sparc64/resource.h
+++ b/include/asm-sparc64/resource.h
@@ -1,19 +1 @@
-/*
- * resource.h: Resource definitions.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef _SPARC64_RESOURCE_H
-#define _SPARC64_RESOURCE_H
-
-/*
- * These two resource limit IDs have a Sparc/Linux-specific ordering,
- * the rest comes from the generic header:
- */
-#define RLIMIT_NOFILE		6	/* max number of open files */
-#define RLIMIT_NPROC		7	/* max number of processes */
-
-#include <asm-generic/resource.h>
-
-#endif /* !(_SPARC64_RESOURCE_H) */
+#include <asm-sparc/resource.h>
diff --git a/include/asm-sparc64/rtc.h b/include/asm-sparc64/rtc.h
index f9ecb1fe2ec..e49a9685aea 100644
--- a/include/asm-sparc64/rtc.h
+++ b/include/asm-sparc64/rtc.h
@@ -1,26 +1 @@
-/*
- * rtc.h: Definitions for access to the Mostek real time clock
- *
- * Copyright (C) 1996 Thomas K. Dyas (tdyas@eden.rutgers.edu)
- */
-
-#ifndef _RTC_H
-#define _RTC_H
-
-#include <linux/ioctl.h>
-
-struct rtc_time
-{
-	int	sec;	/* Seconds (0-59) */
-	int	min;	/* Minutes (0-59) */
-	int	hour;	/* Hour (0-23) */
-	int	dow;	/* Day of the week (1-7) */
-	int	dom;	/* Day of the month (1-31) */
-	int	month;	/* Month of year (1-12) */
-	int	year;	/* Year (0-99) */
-};
-
-#define RTCGET _IOR('p', 20, struct rtc_time)
-#define RTCSET _IOW('p', 21, struct rtc_time)
-
-#endif
+#include <asm-sparc/rtc.h>
diff --git a/include/asm-sparc64/rwsem-const.h b/include/asm-sparc64/rwsem-const.h
index a303c9d64d8..2a1de315c86 100644
--- a/include/asm-sparc64/rwsem-const.h
+++ b/include/asm-sparc64/rwsem-const.h
@@ -1,12 +1 @@
-/* rwsem-const.h: RW semaphore counter constants.  */
-#ifndef _SPARC64_RWSEM_CONST_H
-#define _SPARC64_RWSEM_CONST_H
-
-#define RWSEM_UNLOCKED_VALUE		0x00000000
-#define RWSEM_ACTIVE_BIAS		0x00000001
-#define RWSEM_ACTIVE_MASK		0x0000ffff
-#define RWSEM_WAITING_BIAS		0xffff0000
-#define RWSEM_ACTIVE_READ_BIAS		RWSEM_ACTIVE_BIAS
-#define RWSEM_ACTIVE_WRITE_BIAS		(RWSEM_WAITING_BIAS + RWSEM_ACTIVE_BIAS)
-
-#endif /* _SPARC64_RWSEM_CONST_H */
+#include <asm-sparc/rwsem-const.h>
diff --git a/include/asm-sparc64/rwsem.h b/include/asm-sparc64/rwsem.h
index 1dc129ac2fe..6943c56ed08 100644
--- a/include/asm-sparc64/rwsem.h
+++ b/include/asm-sparc64/rwsem.h
@@ -1,84 +1 @@
-/*
- * rwsem.h: R/W semaphores implemented using CAS
- *
- * Written by David S. Miller (davem@redhat.com), 2001.
- * Derived from asm-i386/rwsem.h
- */
-#ifndef _SPARC64_RWSEM_H
-#define _SPARC64_RWSEM_H
-
-#ifndef _LINUX_RWSEM_H
-#error "please don't include asm/rwsem.h directly, use linux/rwsem.h instead"
-#endif
-
-#ifdef __KERNEL__
-
-#include <linux/list.h>
-#include <linux/spinlock.h>
-#include <asm/rwsem-const.h>
-
-struct rwsem_waiter;
-
-struct rw_semaphore {
-	signed int count;
-	spinlock_t		wait_lock;
-	struct list_head	wait_list;
-#ifdef CONFIG_DEBUG_LOCK_ALLOC
-	struct lockdep_map	dep_map;
-#endif
-};
-
-#ifdef CONFIG_DEBUG_LOCK_ALLOC
-# define __RWSEM_DEP_MAP_INIT(lockname) , .dep_map = { .name = #lockname }
-#else
-# define __RWSEM_DEP_MAP_INIT(lockname)
-#endif
-
-#define __RWSEM_INITIALIZER(name) \
-{ RWSEM_UNLOCKED_VALUE, SPIN_LOCK_UNLOCKED, LIST_HEAD_INIT((name).wait_list) \
-  __RWSEM_DEP_MAP_INIT(name) }
-
-#define DECLARE_RWSEM(name) \
-	struct rw_semaphore name = __RWSEM_INITIALIZER(name)
-
-extern void __init_rwsem(struct rw_semaphore *sem, const char *name,
-			 struct lock_class_key *key);
-
-#define init_rwsem(sem)						\
-do {								\
-	static struct lock_class_key __key;			\
-								\
-	__init_rwsem((sem), #sem, &__key);			\
-} while (0)
-
-extern void __down_read(struct rw_semaphore *sem);
-extern int __down_read_trylock(struct rw_semaphore *sem);
-extern void __down_write(struct rw_semaphore *sem);
-extern int __down_write_trylock(struct rw_semaphore *sem);
-extern void __up_read(struct rw_semaphore *sem);
-extern void __up_write(struct rw_semaphore *sem);
-extern void __downgrade_write(struct rw_semaphore *sem);
-
-static inline void __down_write_nested(struct rw_semaphore *sem, int subclass)
-{
-	__down_write(sem);
-}
-
-static inline int rwsem_atomic_update(int delta, struct rw_semaphore *sem)
-{
-	return atomic_add_return(delta, (atomic_t *)(&sem->count));
-}
-
-static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem)
-{
-	atomic_add(delta, (atomic_t *)(&sem->count));
-}
-
-static inline int rwsem_is_locked(struct rw_semaphore *sem)
-{
-	return (sem->count != 0);
-}
-
-#endif /* __KERNEL__ */
-
-#endif /* _SPARC64_RWSEM_H */
+#include <asm-sparc/rwsem.h>
diff --git a/include/asm-sparc64/sbus.h b/include/asm-sparc64/sbus.h
index 24a04a55cf8..0cab0e89b87 100644
--- a/include/asm-sparc64/sbus.h
+++ b/include/asm-sparc64/sbus.h
@@ -1,190 +1 @@
-/* sbus.h: Defines for the Sun SBus.
- *
- * Copyright (C) 1996, 1999, 2007 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC64_SBUS_H
-#define _SPARC64_SBUS_H
-
-#include <linux/dma-mapping.h>
-#include <linux/ioport.h>
-
-#include <asm/oplib.h>
-#include <asm/prom.h>
-#include <asm/of_device.h>
-#include <asm/iommu.h>
-#include <asm/scatterlist.h>
-
-/* We scan which devices are on the SBus using the PROM node device
- * tree.  SBus devices are described in two different ways.  You can
- * either get an absolute address at which to access the device, or
- * you can get a SBus 'slot' number and an offset within that slot.
- */
-
-/* The base address at which to calculate device OBIO addresses. */
-#define SUN_SBUS_BVADDR        0x00000000
-#define SBUS_OFF_MASK          0x0fffffff
-
-/* These routines are used to calculate device address from slot
- * numbers + offsets, and vice versa.
- */
-
-static inline unsigned long sbus_devaddr(int slotnum, unsigned long offset)
-{
-  return (unsigned long) (SUN_SBUS_BVADDR+((slotnum)<<28)+(offset));
-}
-
-static inline int sbus_dev_slot(unsigned long dev_addr)
-{
-  return (int) (((dev_addr)-SUN_SBUS_BVADDR)>>28);
-}
-
-struct sbus_bus;
-
-/* Linux SBUS device tables */
-struct sbus_dev {
-	struct of_device	ofdev;
-	struct sbus_bus		*bus;
-	struct sbus_dev		*next;
-	struct sbus_dev		*child;
-	struct sbus_dev		*parent;
-	int prom_node;	
-	char prom_name[64];
-	int slot;
-
-	struct resource resource[PROMREG_MAX];
-
-	struct linux_prom_registers reg_addrs[PROMREG_MAX];
-	int num_registers;
-
-	struct linux_prom_ranges device_ranges[PROMREG_MAX];
-	int num_device_ranges;
-
-	unsigned int irqs[4];
-	int num_irqs;
-};
-#define to_sbus_device(d) container_of(d, struct sbus_dev, ofdev.dev)
-
-/* This struct describes the SBus(s) found on this machine. */
-struct sbus_bus {
-	struct of_device	ofdev;
-	struct sbus_dev		*devices;	/* Tree of SBUS devices	*/
-	struct sbus_bus		*next;		/* Next SBUS in system	*/
-	int			prom_node;      /* OBP node of SBUS	*/
-	char			prom_name[64];	/* Usually "sbus" or "sbi" */
-	int			clock_freq;
-
-	struct linux_prom_ranges sbus_ranges[PROMREG_MAX];
-	int num_sbus_ranges;
-
-	int portid;
-};
-#define to_sbus(d) container_of(d, struct sbus_bus, ofdev.dev)
-
-extern struct sbus_bus *sbus_root;
-
-/* Device probing routines could find these handy */
-#define for_each_sbus(bus) \
-        for((bus) = sbus_root; (bus); (bus)=(bus)->next)
-
-#define for_each_sbusdev(device, bus) \
-        for((device) = (bus)->devices; (device); (device)=(device)->next)
-        
-#define for_all_sbusdev(device, bus) \
-	for ((bus) = sbus_root; (bus); (bus) = (bus)->next) \
-		for ((device) = (bus)->devices; (device); (device) = (device)->next)
-
-/* Driver DVMA interfaces. */
-#define sbus_can_dma_64bit(sdev)	(1)
-#define sbus_can_burst64(sdev)		(1)
-extern void sbus_set_sbus64(struct sbus_dev *, int);
-extern void sbus_fill_device_irq(struct sbus_dev *);
-
-static inline void *sbus_alloc_consistent(struct sbus_dev *sdev , size_t size,
-					  dma_addr_t *dma_handle)
-{
-	return dma_alloc_coherent(&sdev->ofdev.dev, size,
-				  dma_handle, GFP_ATOMIC);
-}
-
-static inline void sbus_free_consistent(struct sbus_dev *sdev, size_t size,
-					void *vaddr, dma_addr_t dma_handle)
-{
-	return dma_free_coherent(&sdev->ofdev.dev, size, vaddr, dma_handle);
-}
-
-#define SBUS_DMA_BIDIRECTIONAL	DMA_BIDIRECTIONAL
-#define SBUS_DMA_TODEVICE	DMA_TO_DEVICE
-#define SBUS_DMA_FROMDEVICE	DMA_FROM_DEVICE
-#define	SBUS_DMA_NONE		DMA_NONE
-
-/* All the rest use streaming mode mappings. */
-static inline dma_addr_t sbus_map_single(struct sbus_dev *sdev, void *ptr,
-					 size_t size, int direction)
-{
-	return dma_map_single(&sdev->ofdev.dev, ptr, size,
-			      (enum dma_data_direction) direction);
-}
-
-static inline void sbus_unmap_single(struct sbus_dev *sdev,
-				     dma_addr_t dma_addr, size_t size,
-				     int direction)
-{
-	dma_unmap_single(&sdev->ofdev.dev, dma_addr, size,
-			 (enum dma_data_direction) direction);
-}
-
-static inline int sbus_map_sg(struct sbus_dev *sdev, struct scatterlist *sg,
-			      int nents, int direction)
-{
-	return dma_map_sg(&sdev->ofdev.dev, sg, nents,
-			  (enum dma_data_direction) direction);
-}
-
-static inline void sbus_unmap_sg(struct sbus_dev *sdev, struct scatterlist *sg,
-				 int nents, int direction)
-{
-	dma_unmap_sg(&sdev->ofdev.dev, sg, nents,
-		     (enum dma_data_direction) direction);
-}
-
-/* Finally, allow explicit synchronization of streamable mappings. */
-static inline void sbus_dma_sync_single_for_cpu(struct sbus_dev *sdev,
-						dma_addr_t dma_handle,
-						size_t size, int direction)
-{
-	dma_sync_single_for_cpu(&sdev->ofdev.dev, dma_handle, size,
-				(enum dma_data_direction) direction);
-}
-#define sbus_dma_sync_single sbus_dma_sync_single_for_cpu
-
-static inline void sbus_dma_sync_single_for_device(struct sbus_dev *sdev,
-						   dma_addr_t dma_handle,
-						   size_t size, int direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-static inline void sbus_dma_sync_sg_for_cpu(struct sbus_dev *sdev,
-					    struct scatterlist *sg,
-					    int nents, int direction)
-{
-	dma_sync_sg_for_cpu(&sdev->ofdev.dev, sg, nents,
-			    (enum dma_data_direction) direction);
-}
-#define sbus_dma_sync_sg sbus_dma_sync_sg_for_cpu
-
-static inline void sbus_dma_sync_sg_for_device(struct sbus_dev *sdev,
-					       struct scatterlist *sg,
-					       int nents, int direction)
-{
-	/* No flushing needed to sync cpu writes to the device.  */
-}
-
-extern void sbus_arch_bus_ranges_init(struct device_node *, struct sbus_bus *);
-extern void sbus_setup_iommu(struct sbus_bus *, struct device_node *);
-extern void sbus_setup_arch_props(struct sbus_bus *, struct device_node *);
-extern int sbus_arch_preinit(void);
-extern void sbus_arch_postinit(void);
-
-#endif /* !(_SPARC64_SBUS_H) */
+#include <asm-sparc/sbus.h>
diff --git a/include/asm-sparc64/scatterlist.h b/include/asm-sparc64/scatterlist.h
index 81bd058f938..b7fef95953c 100644
--- a/include/asm-sparc64/scatterlist.h
+++ b/include/asm-sparc64/scatterlist.h
@@ -1,27 +1 @@
-#ifndef _SPARC64_SCATTERLIST_H
-#define _SPARC64_SCATTERLIST_H
-
-#include <asm/page.h>
-#include <asm/types.h>
-
-struct scatterlist {
-#ifdef CONFIG_DEBUG_SG
-	unsigned long	sg_magic;
-#endif
-	unsigned long	page_link;
-	unsigned int	offset;
-
-	unsigned int	length;
-
-	dma_addr_t	dma_address;
-	__u32		dma_length;
-};
-
-#define sg_dma_address(sg)	((sg)->dma_address)
-#define sg_dma_len(sg)     	((sg)->dma_length)
-
-#define ISA_DMA_THRESHOLD	(~0UL)
-
-#define ARCH_HAS_SG_CHAIN
-
-#endif /* !(_SPARC64_SCATTERLIST_H) */
+#include <asm-sparc/scatterlist.h>
diff --git a/include/asm-sparc64/scratchpad.h b/include/asm-sparc64/scratchpad.h
index 5e8b01fb334..23675f6a915 100644
--- a/include/asm-sparc64/scratchpad.h
+++ b/include/asm-sparc64/scratchpad.h
@@ -1,14 +1 @@
-#ifndef _SPARC64_SCRATCHPAD_H
-#define _SPARC64_SCRATCHPAD_H
-
-/* Sun4v scratchpad registers, accessed via ASI_SCRATCHPAD.  */
-
-#define SCRATCHPAD_MMU_MISS	0x00 /* Shared with OBP - set by OBP	    */
-#define SCRATCHPAD_CPUID	0x08 /* Shared with OBP - set by hypervisor */
-#define SCRATCHPAD_UTSBREG1	0x10
-#define SCRATCHPAD_UTSBREG2	0x18
-	/* 0x20 and 0x28, hypervisor only... */
-#define SCRATCHPAD_UNUSED1	0x30
-#define SCRATCHPAD_UNUSED2	0x38 /* Reserved for OBP		    */
-
-#endif /* !(_SPARC64_SCRATCHPAD_H) */
+#include <asm-sparc/scratchpad.h>
diff --git a/include/asm-sparc64/seccomp.h b/include/asm-sparc64/seccomp.h
index 7fcd9968192..f22f02a08a6 100644
--- a/include/asm-sparc64/seccomp.h
+++ b/include/asm-sparc64/seccomp.h
@@ -1,21 +1 @@
-#ifndef _ASM_SECCOMP_H
-
-#include <linux/thread_info.h> /* already defines TIF_32BIT */
-
-#ifndef TIF_32BIT
-#error "unexpected TIF_32BIT on sparc64"
-#endif
-
-#include <linux/unistd.h>
-
-#define __NR_seccomp_read __NR_read
-#define __NR_seccomp_write __NR_write
-#define __NR_seccomp_exit __NR_exit
-#define __NR_seccomp_sigreturn __NR_rt_sigreturn
-
-#define __NR_seccomp_read_32 __NR_read
-#define __NR_seccomp_write_32 __NR_write
-#define __NR_seccomp_exit_32 __NR_exit
-#define __NR_seccomp_sigreturn_32 __NR_sigreturn
-
-#endif /* _ASM_SECCOMP_H */
+#include <asm-sparc/seccomp.h>
diff --git a/include/asm-sparc64/sections.h b/include/asm-sparc64/sections.h
index 3f4b9fdc28d..721496f8b2b 100644
--- a/include/asm-sparc64/sections.h
+++ b/include/asm-sparc64/sections.h
@@ -1,9 +1 @@
-#ifndef _SPARC64_SECTIONS_H
-#define _SPARC64_SECTIONS_H
-
-/* nothing to see, move along */
-#include <asm-generic/sections.h>
-
-extern char _start[];
-
-#endif
+#include <asm-sparc/sections.h>
diff --git a/include/asm-sparc64/semaphore.h b/include/asm-sparc64/semaphore.h
index d9b2034ed1d..39362afde5f 100644
--- a/include/asm-sparc64/semaphore.h
+++ b/include/asm-sparc64/semaphore.h
@@ -1 +1 @@
-#include <linux/semaphore.h>
+#include <asm-sparc/semaphore.h>
diff --git a/include/asm-sparc64/sembuf.h b/include/asm-sparc64/sembuf.h
index 99f04e4e288..c55b9521413 100644
--- a/include/asm-sparc64/sembuf.h
+++ b/include/asm-sparc64/sembuf.h
@@ -1,22 +1 @@
-#ifndef _SPARC64_SEMBUF_H
-#define _SPARC64_SEMBUF_H
-
-/* 
- * The semid64_ds structure for sparc64 architecture.
- * Note extra padding because this structure is passed back and forth
- * between kernel and user space.
- *
- * Pad space is left for:
- * - 2 miscellaneous 64-bit values
- */
-
-struct semid64_ds {
-	struct ipc64_perm sem_perm;		/* permissions .. see ipc.h */
-	__kernel_time_t	sem_otime;		/* last semop time */
-	__kernel_time_t	sem_ctime;		/* last change time */
-	unsigned long	sem_nsems;		/* no. of semaphores in array */
-	unsigned long	__unused1;
-	unsigned long	__unused2;
-};
-
-#endif /* _SPARC64_SEMBUF_H */
+#include <asm-sparc/sembuf.h>
diff --git a/include/asm-sparc64/setup.h b/include/asm-sparc64/setup.h
index 5053df3cec4..7143d06b2c5 100644
--- a/include/asm-sparc64/setup.h
+++ b/include/asm-sparc64/setup.h
@@ -1,10 +1 @@
-/*
- *	Just a place holder. 
- */
-
-#ifndef _SPARC64_SETUP_H
-#define _SPARC64_SETUP_H
-
-#define COMMAND_LINE_SIZE	2048
-
-#endif /* _SPARC64_SETUP_H */
+#include <asm-sparc/setup.h>
diff --git a/include/asm-sparc64/sfafsr.h b/include/asm-sparc64/sfafsr.h
index e96137b04a4..8036fc377a4 100644
--- a/include/asm-sparc64/sfafsr.h
+++ b/include/asm-sparc64/sfafsr.h
@@ -1,82 +1 @@
-#ifndef _SPARC64_SFAFSR_H
-#define _SPARC64_SFAFSR_H
-
-#include <linux/const.h>
-
-/* Spitfire Asynchronous Fault Status register, ASI=0x4C VA<63:0>=0x0 */
-
-#define SFAFSR_ME		(_AC(1,UL) << SFAFSR_ME_SHIFT)
-#define SFAFSR_ME_SHIFT		32
-#define SFAFSR_PRIV		(_AC(1,UL) << SFAFSR_PRIV_SHIFT)
-#define SFAFSR_PRIV_SHIFT	31
-#define SFAFSR_ISAP		(_AC(1,UL) << SFAFSR_ISAP_SHIFT)
-#define SFAFSR_ISAP_SHIFT	30
-#define SFAFSR_ETP		(_AC(1,UL) << SFAFSR_ETP_SHIFT)
-#define SFAFSR_ETP_SHIFT	29
-#define SFAFSR_IVUE		(_AC(1,UL) << SFAFSR_IVUE_SHIFT)
-#define SFAFSR_IVUE_SHIFT	28
-#define SFAFSR_TO		(_AC(1,UL) << SFAFSR_TO_SHIFT)
-#define SFAFSR_TO_SHIFT		27
-#define SFAFSR_BERR		(_AC(1,UL) << SFAFSR_BERR_SHIFT)
-#define SFAFSR_BERR_SHIFT	26
-#define SFAFSR_LDP		(_AC(1,UL) << SFAFSR_LDP_SHIFT)
-#define SFAFSR_LDP_SHIFT	25
-#define SFAFSR_CP		(_AC(1,UL) << SFAFSR_CP_SHIFT)
-#define SFAFSR_CP_SHIFT		24
-#define SFAFSR_WP		(_AC(1,UL) << SFAFSR_WP_SHIFT)
-#define SFAFSR_WP_SHIFT		23
-#define SFAFSR_EDP		(_AC(1,UL) << SFAFSR_EDP_SHIFT)
-#define SFAFSR_EDP_SHIFT	22
-#define SFAFSR_UE		(_AC(1,UL) << SFAFSR_UE_SHIFT)
-#define SFAFSR_UE_SHIFT		21
-#define SFAFSR_CE		(_AC(1,UL) << SFAFSR_CE_SHIFT)
-#define SFAFSR_CE_SHIFT		20
-#define SFAFSR_ETS		(_AC(0xf,UL) << SFAFSR_ETS_SHIFT)
-#define SFAFSR_ETS_SHIFT	16
-#define SFAFSR_PSYND		(_AC(0xffff,UL) << SFAFSR_PSYND_SHIFT)
-#define SFAFSR_PSYND_SHIFT	0
-
-/* UDB Error Register, ASI=0x7f VA<63:0>=0x0(High),0x18(Low) for read
- *                     ASI=0x77 VA<63:0>=0x0(High),0x18(Low) for write
- */
-
-#define UDBE_UE			(_AC(1,UL) << 9)
-#define UDBE_CE			(_AC(1,UL) << 8)
-#define UDBE_E_SYNDR		(_AC(0xff,UL) << 0)
-
-/* The trap handlers for asynchronous errors encode the AFSR and
- * other pieces of information into a 64-bit argument for C code
- * encoded as follows:
- *
- * -----------------------------------------------
- * |  UDB_H  |  UDB_L  | TL>1  |  TT  |   AFSR   |
- * -----------------------------------------------
- *  63     54 53     44    42   41  33 32       0
- *
- * The AFAR is passed in unchanged.
- */
-#define SFSTAT_UDBH_MASK	(_AC(0x3ff,UL) << SFSTAT_UDBH_SHIFT)
-#define SFSTAT_UDBH_SHIFT	54
-#define SFSTAT_UDBL_MASK	(_AC(0x3ff,UL) << SFSTAT_UDBH_SHIFT)
-#define SFSTAT_UDBL_SHIFT	44
-#define SFSTAT_TL_GT_ONE	(_AC(1,UL) << SFSTAT_TL_GT_ONE_SHIFT)
-#define SFSTAT_TL_GT_ONE_SHIFT	42
-#define SFSTAT_TRAP_TYPE	(_AC(0x1FF,UL) << SFSTAT_TRAP_TYPE_SHIFT)
-#define SFSTAT_TRAP_TYPE_SHIFT	33
-#define SFSTAT_AFSR_MASK	(_AC(0x1ffffffff,UL) << SFSTAT_AFSR_SHIFT)
-#define SFSTAT_AFSR_SHIFT	0
-
-/* ESTATE Error Enable Register, ASI=0x4b VA<63:0>=0x0 */
-#define ESTATE_ERR_CE		0x1 /* Correctable errors                    */
-#define ESTATE_ERR_NCE		0x2 /* TO, BERR, LDP, ETP, EDP, WP, UE, IVUE */
-#define ESTATE_ERR_ISAP		0x4 /* System address parity error           */
-#define ESTATE_ERR_ALL		(ESTATE_ERR_CE | \
-				 ESTATE_ERR_NCE | \
-				 ESTATE_ERR_ISAP)
-
-/* The various trap types that report using the above state. */
-#define TRAP_TYPE_IAE		0x09 /* Instruction Access Error             */
-#define TRAP_TYPE_DAE		0x32 /* Data Access Error                    */
-#define TRAP_TYPE_CEE		0x63 /* Correctable ECC Error                */
-
-#endif /* _SPARC64_SFAFSR_H */
+#include <asm-sparc/sfafsr.h>
diff --git a/include/asm-sparc64/sfp-machine.h b/include/asm-sparc64/sfp-machine.h
index c9331b02d9c..7bbc4fecdc7 100644
--- a/include/asm-sparc64/sfp-machine.h
+++ b/include/asm-sparc64/sfp-machine.h
@@ -1,93 +1 @@
-/* Machine-dependent software floating-point definitions.
-   Sparc64 kernel version.
-   Copyright (C) 1997,1998,1999 Free Software Foundation, Inc.
-   This file is part of the GNU C Library.
-   Contributed by Richard Henderson (rth@cygnus.com),
-		  Jakub Jelinek (jj@ultra.linux.cz) and
-		  David S. Miller (davem@redhat.com).
-
-   The GNU C Library is free software; you can redistribute it and/or
-   modify it under the terms of the GNU Library General Public License as
-   published by the Free Software Foundation; either version 2 of the
-   License, or (at your option) any later version.
-
-   The GNU C Library is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
-   Library General Public License for more details.
-
-   You should have received a copy of the GNU Library General Public
-   License along with the GNU C Library; see the file COPYING.LIB.  If
-   not, write to the Free Software Foundation, Inc.,
-   59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
-
-#ifndef _SFP_MACHINE_H
-#define _SFP_MACHINE_H
-   
-#define _FP_W_TYPE_SIZE		64
-#define _FP_W_TYPE		unsigned long
-#define _FP_WS_TYPE		signed long
-#define _FP_I_TYPE		long
-
-#define _FP_MUL_MEAT_S(R,X,Y)					\
-  _FP_MUL_MEAT_1_imm(_FP_WFRACBITS_S,R,X,Y)
-#define _FP_MUL_MEAT_D(R,X,Y)					\
-  _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
-#define _FP_MUL_MEAT_Q(R,X,Y)					\
-  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
-
-#define _FP_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_imm(S,R,X,Y,_FP_DIV_HELP_imm)
-#define _FP_DIV_MEAT_D(R,X,Y)	_FP_DIV_MEAT_1_udiv_norm(D,R,X,Y)
-#define _FP_DIV_MEAT_Q(R,X,Y)	_FP_DIV_MEAT_2_udiv(Q,R,X,Y)
-
-#define _FP_NANFRAC_S		((_FP_QNANBIT_S << 1) - 1)
-#define _FP_NANFRAC_D		((_FP_QNANBIT_D << 1) - 1)
-#define _FP_NANFRAC_Q		((_FP_QNANBIT_Q << 1) - 1), -1
-#define _FP_NANSIGN_S		0
-#define _FP_NANSIGN_D		0
-#define _FP_NANSIGN_Q		0
-
-#define _FP_KEEPNANFRACP 1
-
-/* If one NaN is signaling and the other is not,
- * we choose that one, otherwise we choose X.
- */
-/* For _Qp_* and _Q_*, this should prefer X, for
- * CPU instruction emulation this should prefer Y.
- * (see SPAMv9 B.2.2 section).
- */
-#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
-  do {								\
-    if ((_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs)		\
-	&& !(_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs))	\
-      {								\
-	R##_s = X##_s;						\
-	_FP_FRAC_COPY_##wc(R,X);				\
-      }								\
-    else							\
-      {								\
-	R##_s = Y##_s;						\
-	_FP_FRAC_COPY_##wc(R,Y);				\
-      }								\
-    R##_c = FP_CLS_NAN;						\
-  } while (0)
-
-/* Obtain the current rounding mode. */
-#ifndef FP_ROUNDMODE
-#define FP_ROUNDMODE	((current_thread_info()->xfsr[0] >> 30) & 0x3)
-#endif
-
-/* Exception flags. */
-#define FP_EX_INVALID		(1 << 4)
-#define FP_EX_OVERFLOW		(1 << 3)
-#define FP_EX_UNDERFLOW		(1 << 2)
-#define FP_EX_DIVZERO		(1 << 1)
-#define FP_EX_INEXACT		(1 << 0)
-
-#define FP_HANDLE_EXCEPTIONS return _fex
-
-#define FP_INHIBIT_RESULTS ((current_thread_info()->xfsr[0] >> 23) & _fex)
-
-#define FP_TRAPPING_EXCEPTIONS ((current_thread_info()->xfsr[0] >> 23) & 0x1f)
-
-#endif
+#include <asm-sparc/sfp-machine.h>
diff --git a/include/asm-sparc64/shmbuf.h b/include/asm-sparc64/shmbuf.h
index 61c2ef42eba..0c54a2d6868 100644
--- a/include/asm-sparc64/shmbuf.h
+++ b/include/asm-sparc64/shmbuf.h
@@ -1,38 +1 @@
-#ifndef _SPARC64_SHMBUF_H
-#define _SPARC64_SHMBUF_H
-
-/* 
- * The shmid64_ds structure for sparc64 architecture.
- * Note extra padding because this structure is passed back and forth
- * between kernel and user space.
- *
- * Pad space is left for:
- * - 2 miscellaneous 64-bit values
- */
-
-struct shmid64_ds {
-	struct ipc64_perm	shm_perm;	/* operation perms */
-	__kernel_time_t		shm_atime;	/* last attach time */
-	__kernel_time_t		shm_dtime;	/* last detach time */
-	__kernel_time_t		shm_ctime;	/* last change time */
-	size_t			shm_segsz;	/* size of segment (bytes) */
-	__kernel_pid_t		shm_cpid;	/* pid of creator */
-	__kernel_pid_t		shm_lpid;	/* pid of last operator */
-	unsigned long		shm_nattch;	/* no. of current attaches */
-	unsigned long		__unused1;
-	unsigned long		__unused2;
-};
-
-struct shminfo64 {
-	unsigned long	shmmax;
-	unsigned long	shmmin;
-	unsigned long	shmmni;
-	unsigned long	shmseg;
-	unsigned long	shmall;
-	unsigned long	__unused1;
-	unsigned long	__unused2;
-	unsigned long	__unused3;
-	unsigned long	__unused4;
-};
-
-#endif /* _SPARC64_SHMBUF_H */
+#include <asm-sparc/shmbuf.h>
diff --git a/include/asm-sparc64/shmparam.h b/include/asm-sparc64/shmparam.h
index 1ed0d6701a9..5fa3a9b05e7 100644
--- a/include/asm-sparc64/shmparam.h
+++ b/include/asm-sparc64/shmparam.h
@@ -1,10 +1 @@
-#ifndef _ASMSPARC64_SHMPARAM_H
-#define _ASMSPARC64_SHMPARAM_H
-
-#include <asm/spitfire.h>
-
-#define __ARCH_FORCE_SHMLBA	1
-/* attach addr a multiple of this */
-#define	SHMLBA	((PAGE_SIZE > L1DCACHE_SIZE) ? PAGE_SIZE : L1DCACHE_SIZE)
-
-#endif /* _ASMSPARC64_SHMPARAM_H */
+#include <asm-sparc/shmparam.h>
diff --git a/include/asm-sparc64/sigcontext.h b/include/asm-sparc64/sigcontext.h
index 1c868d680cf..5b16dcce44f 100644
--- a/include/asm-sparc64/sigcontext.h
+++ b/include/asm-sparc64/sigcontext.h
@@ -1,87 +1 @@
-#ifndef __SPARC64_SIGCONTEXT_H
-#define __SPARC64_SIGCONTEXT_H
-
-#ifdef __KERNEL__
-#include <asm/ptrace.h>
-#endif
-
-#ifndef __ASSEMBLY__
-
-#ifdef __KERNEL__
-
-#define __SUNOS_MAXWIN   31
-
-/* This is what SunOS does, so shall I unless we use new 32bit signals or rt signals. */
-struct sigcontext32 {
-	int sigc_onstack;      /* state to restore */
-	int sigc_mask;         /* sigmask to restore */
-	int sigc_sp;           /* stack pointer */
-	int sigc_pc;           /* program counter */
-	int sigc_npc;          /* next program counter */
-	int sigc_psr;          /* for condition codes etc */
-	int sigc_g1;           /* User uses these two registers */
-	int sigc_o0;           /* within the trampoline code. */
-
-	/* Now comes information regarding the users window set
-	 * at the time of the signal.
-	 */
-	int sigc_oswins;       /* outstanding windows */
-
-	/* stack ptrs for each regwin buf */
-	unsigned sigc_spbuf[__SUNOS_MAXWIN];
-
-	/* Windows to restore after signal */
-	struct reg_window32 sigc_wbuf[__SUNOS_MAXWIN];
-};
-
-#endif
-
-#ifdef __KERNEL__
-
-/* This is what we use for 32bit new non-rt signals. */
-
-typedef struct {
-	struct {
-		unsigned int psr;
-		unsigned int pc;
-		unsigned int npc;
-		unsigned int y;
-		unsigned int u_regs[16]; /* globals and ins */
-	}			si_regs;
-	int			si_mask;
-} __siginfo32_t;
-
-#endif
-
-typedef struct {
-	unsigned   int si_float_regs [64];
-	unsigned   long si_fsr;
-	unsigned   long si_gsr;
-	unsigned   long si_fprs;
-} __siginfo_fpu_t;
-
-/* This is what SunOS doesn't, so we have to write this alone
-   and do it properly. */
-struct sigcontext {
-	/* The size of this array has to match SI_MAX_SIZE from siginfo.h */
-	char			sigc_info[128];
-	struct {
-		unsigned long	u_regs[16]; /* globals and ins */
-		unsigned long	tstate;
-		unsigned long	tpc;
-		unsigned long	tnpc;
-		unsigned int	y;
-		unsigned int	fprs;
-	}			sigc_regs;
-	__siginfo_fpu_t *	sigc_fpu_save;
-	struct {
-		void	*	ss_sp;
-		int		ss_flags;
-		unsigned long	ss_size;
-	}			sigc_stack;
-	unsigned long		sigc_mask;
-};
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC64_SIGCONTEXT_H) */
+#include <asm-sparc/sigcontext.h>
diff --git a/include/asm-sparc64/siginfo.h b/include/asm-sparc64/siginfo.h
index c96e6c30f8b..8ffd6ebabc7 100644
--- a/include/asm-sparc64/siginfo.h
+++ b/include/asm-sparc64/siginfo.h
@@ -1,32 +1 @@
-#ifndef _SPARC64_SIGINFO_H
-#define _SPARC64_SIGINFO_H
-
-#define SI_PAD_SIZE32	((SI_MAX_SIZE/sizeof(int)) - 3)
-
-#define __ARCH_SI_PREAMBLE_SIZE	(4 * sizeof(int))
-#define __ARCH_SI_TRAPNO
-#define __ARCH_SI_BAND_T int
-
-#include <asm-generic/siginfo.h>
-
-#ifdef __KERNEL__
-
-#include <linux/compat.h>
-
-#ifdef CONFIG_COMPAT
-
-struct compat_siginfo;
-
-#endif /* CONFIG_COMPAT */
-
-#endif /* __KERNEL__ */
-
-#define SI_NOINFO	32767		/* no information in siginfo_t */
-
-/*
- * SIGEMT si_codes
- */
-#define EMT_TAGOVF	(__SI_FAULT|1)	/* tag overflow */
-#define NSIGEMT		1
-
-#endif
+#include <asm-sparc/siginfo.h>
diff --git a/include/asm-sparc64/signal.h b/include/asm-sparc64/signal.h
index 2a7c7934ac0..79705e5d49c 100644
--- a/include/asm-sparc64/signal.h
+++ b/include/asm-sparc64/signal.h
@@ -1,194 +1 @@
-#ifndef _ASMSPARC64_SIGNAL_H
-#define _ASMSPARC64_SIGNAL_H
-
-#include <asm/sigcontext.h>
-
-#ifdef __KERNEL__
-#ifndef __ASSEMBLY__
-#include <linux/personality.h>
-#include <linux/types.h>
-#endif
-#endif
-
-/* On the Sparc the signal handlers get passed a 'sub-signal' code
- * for certain signal types, which we document here.
- */
-#define SIGHUP		 1
-#define SIGINT		 2
-#define SIGQUIT		 3
-#define SIGILL		 4
-#define    SUBSIG_STACK       0
-#define    SUBSIG_ILLINST     2
-#define    SUBSIG_PRIVINST    3
-#define    SUBSIG_BADTRAP(t)  (0x80 + (t))
-
-#define SIGTRAP		 5
-#define SIGABRT		 6
-#define SIGIOT		 6
-
-#define SIGEMT           7
-#define    SUBSIG_TAG    10
-
-#define SIGFPE		 8
-#define    SUBSIG_FPDISABLED     0x400
-#define    SUBSIG_FPERROR        0x404
-#define    SUBSIG_FPINTOVFL      0x001
-#define    SUBSIG_FPSTSIG        0x002
-#define    SUBSIG_IDIVZERO       0x014
-#define    SUBSIG_FPINEXACT      0x0c4
-#define    SUBSIG_FPDIVZERO      0x0c8
-#define    SUBSIG_FPUNFLOW       0x0cc
-#define    SUBSIG_FPOPERROR      0x0d0
-#define    SUBSIG_FPOVFLOW       0x0d4
-
-#define SIGKILL		 9
-#define SIGBUS          10
-#define    SUBSIG_BUSTIMEOUT    1
-#define    SUBSIG_ALIGNMENT     2
-#define    SUBSIG_MISCERROR     5
-
-#define SIGSEGV		11
-#define    SUBSIG_NOMAPPING     3
-#define    SUBSIG_PROTECTION    4
-#define    SUBSIG_SEGERROR      5
-
-#define SIGSYS		12
-
-#define SIGPIPE		13
-#define SIGALRM		14
-#define SIGTERM		15
-#define SIGURG          16
-
-/* SunOS values which deviate from the Linux/i386 ones */
-#define SIGSTOP		17
-#define SIGTSTP		18
-#define SIGCONT		19
-#define SIGCHLD		20
-#define SIGTTIN		21
-#define SIGTTOU		22
-#define SIGIO		23
-#define SIGPOLL		SIGIO   /* SysV name for SIGIO */
-#define SIGXCPU		24
-#define SIGXFSZ		25
-#define SIGVTALRM	26
-#define SIGPROF		27
-#define SIGWINCH	28
-#define SIGLOST		29
-#define SIGPWR		SIGLOST
-#define SIGUSR1		30
-#define SIGUSR2		31
-
-/* Most things should be clean enough to redefine this at will, if care
-   is taken to make libc match.  */
-
-#define __OLD_NSIG	32
-#define __NEW_NSIG      64
-#define _NSIG_BPW     	64
-#define _NSIG_WORDS   	(__NEW_NSIG / _NSIG_BPW)
-
-#define SIGRTMIN       32
-#define SIGRTMAX       __NEW_NSIG
-
-#if defined(__KERNEL__) || defined(__WANT_POSIX1B_SIGNALS__)
-#define _NSIG			__NEW_NSIG
-#define __new_sigset_t		sigset_t
-#define __new_sigaction		sigaction
-#define __new_sigaction32	sigaction32
-#define __old_sigset_t		old_sigset_t
-#define __old_sigaction		old_sigaction
-#define __old_sigaction32	old_sigaction32
-#else
-#define _NSIG			__OLD_NSIG
-#define NSIG			_NSIG
-#define __old_sigset_t		sigset_t
-#define __old_sigaction		sigaction
-#define __old_sigaction32	sigaction32
-#endif
-
-#ifndef __ASSEMBLY__
-
-typedef unsigned long __old_sigset_t;            /* at least 32 bits */
-
-typedef struct {
-       unsigned long sig[_NSIG_WORDS];
-} __new_sigset_t;
-
-/* A SunOS sigstack */
-struct sigstack {
-	/* XXX 32-bit pointers pinhead XXX */
-	char *the_stack;
-	int   cur_status;
-};
-
-/* Sigvec flags */
-#define _SV_SSTACK    1u    /* This signal handler should use sig-stack */
-#define _SV_INTR      2u    /* Sig return should not restart system call */
-#define _SV_RESET     4u    /* Set handler to SIG_DFL upon taken signal */
-#define _SV_IGNCHILD  8u    /* Do not send SIGCHLD */
-
-/*
- * sa_flags values: SA_STACK is not currently supported, but will allow the
- * usage of signal stacks by using the (now obsolete) sa_restorer field in
- * the sigaction structure as a stack pointer. This is now possible due to
- * the changes in signal handling. LBT 010493.
- * SA_RESTART flag to get restarting signals (which were the default long ago)
- */
-#define SA_NOCLDSTOP	_SV_IGNCHILD
-#define SA_STACK	_SV_SSTACK
-#define SA_ONSTACK	_SV_SSTACK
-#define SA_RESTART	_SV_INTR
-#define SA_ONESHOT	_SV_RESET
-#define SA_NOMASK	0x20u
-#define SA_NOCLDWAIT    0x100u
-#define SA_SIGINFO      0x200u
-
-
-#define SIG_BLOCK          0x01	/* for blocking signals */
-#define SIG_UNBLOCK        0x02	/* for unblocking signals */
-#define SIG_SETMASK        0x04	/* for setting the signal mask */
-
-/* 
- * sigaltstack controls
- */
-#define SS_ONSTACK	1
-#define SS_DISABLE	2
-
-#define MINSIGSTKSZ	4096
-#define SIGSTKSZ	16384
-
-#include <asm-generic/signal.h>
-
-struct __new_sigaction {
-	__sighandler_t		sa_handler;
-	unsigned long		sa_flags;
-	__sigrestore_t 		sa_restorer;  /* not used by Linux/SPARC yet */
-	__new_sigset_t		sa_mask;
-};
-
-struct __old_sigaction {
-	__sighandler_t  	sa_handler;
-	__old_sigset_t  	sa_mask;
-	unsigned long   	sa_flags;
-	void 			(*sa_restorer)(void);     /* not used by Linux/SPARC yet */
-};
-
-typedef struct sigaltstack {
-	void			__user *ss_sp;
-	int			ss_flags;
-	size_t			ss_size;
-} stack_t;
-
-#ifdef __KERNEL__
-
-struct k_sigaction {
-	struct __new_sigaction 	sa;
-	void __user		*ka_restorer;
-};
-
-#define ptrace_signal_deliver(regs, cookie) do { } while (0)
-
-#endif /* !(__KERNEL__) */
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(_ASMSPARC64_SIGNAL_H) */
+#include <asm-sparc/signal.h>
diff --git a/include/asm-sparc64/smp.h b/include/asm-sparc64/smp.h
index cd0311b2e19..5095a2cbea5 100644
--- a/include/asm-sparc64/smp.h
+++ b/include/asm-sparc64/smp.h
@@ -1,64 +1 @@
-/* smp.h: Sparc64 specific SMP stuff.
- *
- * Copyright (C) 1996, 2008 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC64_SMP_H
-#define _SPARC64_SMP_H
-
-#include <linux/threads.h>
-#include <asm/asi.h>
-#include <asm/starfire.h>
-#include <asm/spitfire.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/cpumask.h>
-#include <linux/cache.h>
-
-#endif /* !(__ASSEMBLY__) */
-
-#ifdef CONFIG_SMP
-
-#ifndef __ASSEMBLY__
-
-/*
- *	Private routines/data
- */
- 
-#include <linux/bitops.h>
-#include <asm/atomic.h>
-#include <asm/percpu.h>
-
-DECLARE_PER_CPU(cpumask_t, cpu_sibling_map);
-extern cpumask_t cpu_core_map[NR_CPUS];
-extern int sparc64_multi_core;
-
-/*
- *	General functions that each host system must provide.
- */
-
-extern int hard_smp_processor_id(void);
-#define raw_smp_processor_id() (current_thread_info()->cpu)
-
-extern void smp_fill_in_sib_core_maps(void);
-extern void cpu_play_dead(void);
-
-extern void smp_fetch_global_regs(void);
-
-#ifdef CONFIG_HOTPLUG_CPU
-extern int __cpu_disable(void);
-extern void __cpu_die(unsigned int cpu);
-#endif
-
-#endif /* !(__ASSEMBLY__) */
-
-#else
-
-#define hard_smp_processor_id()		0
-#define smp_fill_in_sib_core_maps() do { } while (0)
-#define smp_fetch_global_regs() do { } while (0)
-
-#endif /* !(CONFIG_SMP) */
-
-#endif /* !(_SPARC64_SMP_H) */
+#include <asm-sparc/smp.h>
diff --git a/include/asm-sparc64/socket.h b/include/asm-sparc64/socket.h
index 5af688f5671..13e0d5d94bb 100644
--- a/include/asm-sparc64/socket.h
+++ b/include/asm-sparc64/socket.h
@@ -1,57 +1 @@
-#ifndef _ASM_SOCKET_H
-#define _ASM_SOCKET_H
-
-#include <asm/sockios.h>
-
-/* For setsockopt(2) */
-#define SOL_SOCKET	0xffff
-
-#define SO_DEBUG	0x0001
-#define SO_PASSCRED	0x0002
-#define SO_REUSEADDR	0x0004
-#define SO_KEEPALIVE	0x0008
-#define SO_DONTROUTE	0x0010
-#define SO_BROADCAST	0x0020
-#define SO_PEERCRED	0x0040
-#define SO_LINGER	0x0080
-#define SO_OOBINLINE	0x0100
-/* To add :#define SO_REUSEPORT 0x0200 */
-#define SO_BSDCOMPAT    0x0400
-#define SO_RCVLOWAT     0x0800
-#define SO_SNDLOWAT     0x1000
-#define SO_RCVTIMEO     0x2000
-#define SO_SNDTIMEO     0x4000
-#define SO_ACCEPTCONN	0x8000
-
-#define SO_SNDBUF	0x1001
-#define SO_RCVBUF	0x1002
-#define SO_SNDBUFFORCE	0x100a
-#define SO_RCVBUFFORCE	0x100b
-#define SO_ERROR	0x1007
-#define SO_TYPE		0x1008
-
-/* Linux specific, keep the same. */
-#define SO_NO_CHECK	0x000b
-#define SO_PRIORITY	0x000c
-
-#define SO_BINDTODEVICE 0x000d
-
-#define SO_ATTACH_FILTER	0x001a
-#define SO_DETACH_FILTER        0x001b
-
-#define SO_PEERNAME		0x001c
-#define SO_TIMESTAMP		0x001d
-#define SCM_TIMESTAMP		SO_TIMESTAMP
-
-#define SO_PEERSEC		0x001e
-#define SO_PASSSEC		0x001f
-#define SO_TIMESTAMPNS		0x0021
-#define SCM_TIMESTAMPNS		SO_TIMESTAMPNS
-
-/* Security levels - as per NRL IPv6 - don't actually do anything */
-#define SO_SECURITY_AUTHENTICATION		0x5001
-#define SO_SECURITY_ENCRYPTION_TRANSPORT	0x5002
-#define SO_SECURITY_ENCRYPTION_NETWORK		0x5004
-
-#define SO_MARK			0x0022
-#endif /* _ASM_SOCKET_H */
+#include <asm-sparc/socket.h>
diff --git a/include/asm-sparc64/sockios.h b/include/asm-sparc64/sockios.h
index c7d9900638d..2cb4b641482 100644
--- a/include/asm-sparc64/sockios.h
+++ b/include/asm-sparc64/sockios.h
@@ -1,14 +1 @@
-#ifndef _ASM_SPARC64_SOCKIOS_H
-#define _ASM_SPARC64_SOCKIOS_H
-
-/* Socket-level I/O control calls. */
-#define FIOSETOWN 	0x8901
-#define SIOCSPGRP	0x8902
-#define FIOGETOWN	0x8903
-#define SIOCGPGRP	0x8904
-#define SIOCATMARK	0x8905
-#define SIOCGSTAMP	0x8906		/* Get stamp (timeval) */
-#define SIOCGSTAMPNS	0x8907		/* Get stamp (timespec) */
-
-#endif /* !(_ASM_SPARC64_SOCKIOS_H) */
-
+#include <asm-sparc/sockios.h>
diff --git a/include/asm-sparc64/sparsemem.h b/include/asm-sparc64/sparsemem.h
index b99d4e4b6d2..e681f22a97a 100644
--- a/include/asm-sparc64/sparsemem.h
+++ b/include/asm-sparc64/sparsemem.h
@@ -1,12 +1 @@
-#ifndef _SPARC64_SPARSEMEM_H
-#define _SPARC64_SPARSEMEM_H
-
-#ifdef __KERNEL__
-
-#define SECTION_SIZE_BITS       30
-#define MAX_PHYSADDR_BITS       42
-#define MAX_PHYSMEM_BITS        42
-
-#endif /* !(__KERNEL__) */
-
-#endif /* !(_SPARC64_SPARSEMEM_H) */
+#include <asm-sparc/sparsemem.h>
diff --git a/include/asm-sparc64/spinlock.h b/include/asm-sparc64/spinlock.h
index 0006fe9f8c7..0115b8156eb 100644
--- a/include/asm-sparc64/spinlock.h
+++ b/include/asm-sparc64/spinlock.h
@@ -1,250 +1 @@
-/* spinlock.h: 64-bit Sparc spinlock support.
- *
- * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __SPARC64_SPINLOCK_H
-#define __SPARC64_SPINLOCK_H
-
-#include <linux/threads.h>	/* For NR_CPUS */
-
-#ifndef __ASSEMBLY__
-
-/* To get debugging spinlocks which detect and catch
- * deadlock situations, set CONFIG_DEBUG_SPINLOCK
- * and rebuild your kernel.
- */
-
-/* All of these locking primitives are expected to work properly
- * even in an RMO memory model, which currently is what the kernel
- * runs in.
- *
- * There is another issue.  Because we play games to save cycles
- * in the non-contention case, we need to be extra careful about
- * branch targets into the "spinning" code.  They live in their
- * own section, but the newer V9 branches have a shorter range
- * than the traditional 32-bit sparc branch variants.  The rule
- * is that the branches that go into and out of the spinner sections
- * must be pre-V9 branches.
- */
-
-#define __raw_spin_is_locked(lp)	((lp)->lock != 0)
-
-#define __raw_spin_unlock_wait(lp)	\
-	do {	rmb();			\
-	} while((lp)->lock)
-
-static inline void __raw_spin_lock(raw_spinlock_t *lock)
-{
-	unsigned long tmp;
-
-	__asm__ __volatile__(
-"1:	ldstub		[%1], %0\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	brnz,pn		%0, 2f\n"
-"	 nop\n"
-"	.subsection	2\n"
-"2:	ldub		[%1], %0\n"
-"	membar		#LoadLoad\n"
-"	brnz,pt		%0, 2b\n"
-"	 nop\n"
-"	ba,a,pt		%%xcc, 1b\n"
-"	.previous"
-	: "=&r" (tmp)
-	: "r" (lock)
-	: "memory");
-}
-
-static inline int __raw_spin_trylock(raw_spinlock_t *lock)
-{
-	unsigned long result;
-
-	__asm__ __volatile__(
-"	ldstub		[%1], %0\n"
-"	membar		#StoreLoad | #StoreStore"
-	: "=r" (result)
-	: "r" (lock)
-	: "memory");
-
-	return (result == 0UL);
-}
-
-static inline void __raw_spin_unlock(raw_spinlock_t *lock)
-{
-	__asm__ __volatile__(
-"	membar		#StoreStore | #LoadStore\n"
-"	stb		%%g0, [%0]"
-	: /* No outputs */
-	: "r" (lock)
-	: "memory");
-}
-
-static inline void __raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags)
-{
-	unsigned long tmp1, tmp2;
-
-	__asm__ __volatile__(
-"1:	ldstub		[%2], %0\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	brnz,pn		%0, 2f\n"
-"	 nop\n"
-"	.subsection	2\n"
-"2:	rdpr		%%pil, %1\n"
-"	wrpr		%3, %%pil\n"
-"3:	ldub		[%2], %0\n"
-"	membar		#LoadLoad\n"
-"	brnz,pt		%0, 3b\n"
-"	 nop\n"
-"	ba,pt		%%xcc, 1b\n"
-"	 wrpr		%1, %%pil\n"
-"	.previous"
-	: "=&r" (tmp1), "=&r" (tmp2)
-	: "r"(lock), "r"(flags)
-	: "memory");
-}
-
-/* Multi-reader locks, these are much saner than the 32-bit Sparc ones... */
-
-static void inline __read_lock(raw_rwlock_t *lock)
-{
-	unsigned long tmp1, tmp2;
-
-	__asm__ __volatile__ (
-"1:	ldsw		[%2], %0\n"
-"	brlz,pn		%0, 2f\n"
-"4:	 add		%0, 1, %1\n"
-"	cas		[%2], %0, %1\n"
-"	cmp		%0, %1\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	bne,pn		%%icc, 1b\n"
-"	 nop\n"
-"	.subsection	2\n"
-"2:	ldsw		[%2], %0\n"
-"	membar		#LoadLoad\n"
-"	brlz,pt		%0, 2b\n"
-"	 nop\n"
-"	ba,a,pt		%%xcc, 4b\n"
-"	.previous"
-	: "=&r" (tmp1), "=&r" (tmp2)
-	: "r" (lock)
-	: "memory");
-}
-
-static int inline __read_trylock(raw_rwlock_t *lock)
-{
-	int tmp1, tmp2;
-
-	__asm__ __volatile__ (
-"1:	ldsw		[%2], %0\n"
-"	brlz,a,pn	%0, 2f\n"
-"	 mov		0, %0\n"
-"	add		%0, 1, %1\n"
-"	cas		[%2], %0, %1\n"
-"	cmp		%0, %1\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	bne,pn		%%icc, 1b\n"
-"	 mov		1, %0\n"
-"2:"
-	: "=&r" (tmp1), "=&r" (tmp2)
-	: "r" (lock)
-	: "memory");
-
-	return tmp1;
-}
-
-static void inline __read_unlock(raw_rwlock_t *lock)
-{
-	unsigned long tmp1, tmp2;
-
-	__asm__ __volatile__(
-"	membar	#StoreLoad | #LoadLoad\n"
-"1:	lduw	[%2], %0\n"
-"	sub	%0, 1, %1\n"
-"	cas	[%2], %0, %1\n"
-"	cmp	%0, %1\n"
-"	bne,pn	%%xcc, 1b\n"
-"	 nop"
-	: "=&r" (tmp1), "=&r" (tmp2)
-	: "r" (lock)
-	: "memory");
-}
-
-static void inline __write_lock(raw_rwlock_t *lock)
-{
-	unsigned long mask, tmp1, tmp2;
-
-	mask = 0x80000000UL;
-
-	__asm__ __volatile__(
-"1:	lduw		[%2], %0\n"
-"	brnz,pn		%0, 2f\n"
-"4:	 or		%0, %3, %1\n"
-"	cas		[%2], %0, %1\n"
-"	cmp		%0, %1\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	bne,pn		%%icc, 1b\n"
-"	 nop\n"
-"	.subsection	2\n"
-"2:	lduw		[%2], %0\n"
-"	membar		#LoadLoad\n"
-"	brnz,pt		%0, 2b\n"
-"	 nop\n"
-"	ba,a,pt		%%xcc, 4b\n"
-"	.previous"
-	: "=&r" (tmp1), "=&r" (tmp2)
-	: "r" (lock), "r" (mask)
-	: "memory");
-}
-
-static void inline __write_unlock(raw_rwlock_t *lock)
-{
-	__asm__ __volatile__(
-"	membar		#LoadStore | #StoreStore\n"
-"	stw		%%g0, [%0]"
-	: /* no outputs */
-	: "r" (lock)
-	: "memory");
-}
-
-static int inline __write_trylock(raw_rwlock_t *lock)
-{
-	unsigned long mask, tmp1, tmp2, result;
-
-	mask = 0x80000000UL;
-
-	__asm__ __volatile__(
-"	mov		0, %2\n"
-"1:	lduw		[%3], %0\n"
-"	brnz,pn		%0, 2f\n"
-"	 or		%0, %4, %1\n"
-"	cas		[%3], %0, %1\n"
-"	cmp		%0, %1\n"
-"	membar		#StoreLoad | #StoreStore\n"
-"	bne,pn		%%icc, 1b\n"
-"	 nop\n"
-"	mov		1, %2\n"
-"2:"
-	: "=&r" (tmp1), "=&r" (tmp2), "=&r" (result)
-	: "r" (lock), "r" (mask)
-	: "memory");
-
-	return result;
-}
-
-#define __raw_read_lock(p)	__read_lock(p)
-#define __raw_read_trylock(p)	__read_trylock(p)
-#define __raw_read_unlock(p)	__read_unlock(p)
-#define __raw_write_lock(p)	__write_lock(p)
-#define __raw_write_unlock(p)	__write_unlock(p)
-#define __raw_write_trylock(p)	__write_trylock(p)
-
-#define __raw_read_can_lock(rw)		(!((rw)->lock & 0x80000000UL))
-#define __raw_write_can_lock(rw)	(!(rw)->lock)
-
-#define _raw_spin_relax(lock)	cpu_relax()
-#define _raw_read_relax(lock)	cpu_relax()
-#define _raw_write_relax(lock)	cpu_relax()
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(__SPARC64_SPINLOCK_H) */
+#include <asm-sparc/spinlock.h>
diff --git a/include/asm-sparc64/spinlock_types.h b/include/asm-sparc64/spinlock_types.h
index e128112a0d7..48d81c8734b 100644
--- a/include/asm-sparc64/spinlock_types.h
+++ b/include/asm-sparc64/spinlock_types.h
@@ -1,20 +1 @@
-#ifndef __SPARC64_SPINLOCK_TYPES_H
-#define __SPARC64_SPINLOCK_TYPES_H
-
-#ifndef __LINUX_SPINLOCK_TYPES_H
-# error "please don't include this file directly"
-#endif
-
-typedef struct {
-	volatile unsigned char lock;
-} raw_spinlock_t;
-
-#define __RAW_SPIN_LOCK_UNLOCKED	{ 0 }
-
-typedef struct {
-	volatile unsigned int lock;
-} raw_rwlock_t;
-
-#define __RAW_RW_LOCK_UNLOCKED		{ 0 }
-
-#endif
+#include <asm-sparc/spinlock_types.h>
diff --git a/include/asm-sparc64/spitfire.h b/include/asm-sparc64/spitfire.h
index 985ea7e3199..4430d2fbb0d 100644
--- a/include/asm-sparc64/spitfire.h
+++ b/include/asm-sparc64/spitfire.h
@@ -1,342 +1 @@
-/* spitfire.h: SpitFire/BlackBird/Cheetah inline MMU operations.
- *
- * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC64_SPITFIRE_H
-#define _SPARC64_SPITFIRE_H
-
-#include <asm/asi.h>
-
-/* The following register addresses are accessible via ASI_DMMU
- * and ASI_IMMU, that is there is a distinct and unique copy of
- * each these registers for each TLB.
- */
-#define TSB_TAG_TARGET		0x0000000000000000 /* All chips				*/
-#define TLB_SFSR		0x0000000000000018 /* All chips				*/
-#define TSB_REG			0x0000000000000028 /* All chips				*/
-#define TLB_TAG_ACCESS		0x0000000000000030 /* All chips				*/
-#define VIRT_WATCHPOINT		0x0000000000000038 /* All chips				*/
-#define PHYS_WATCHPOINT		0x0000000000000040 /* All chips				*/
-#define TSB_EXTENSION_P		0x0000000000000048 /* Ultra-III and later		*/
-#define TSB_EXTENSION_S		0x0000000000000050 /* Ultra-III and later, D-TLB only	*/
-#define TSB_EXTENSION_N		0x0000000000000058 /* Ultra-III and later		*/
-#define TLB_TAG_ACCESS_EXT	0x0000000000000060 /* Ultra-III+ and later		*/
-
-/* These registers only exist as one entity, and are accessed
- * via ASI_DMMU only.
- */
-#define PRIMARY_CONTEXT		0x0000000000000008
-#define SECONDARY_CONTEXT	0x0000000000000010
-#define DMMU_SFAR		0x0000000000000020
-#define VIRT_WATCHPOINT		0x0000000000000038
-#define PHYS_WATCHPOINT		0x0000000000000040
-
-#define SPITFIRE_HIGHEST_LOCKED_TLBENT	(64 - 1)
-#define CHEETAH_HIGHEST_LOCKED_TLBENT	(16 - 1)
-
-#define L1DCACHE_SIZE		0x4000
-
-#define SUN4V_CHIP_INVALID	0x00
-#define SUN4V_CHIP_NIAGARA1	0x01
-#define SUN4V_CHIP_NIAGARA2	0x02
-#define SUN4V_CHIP_UNKNOWN	0xff
-
-#ifndef __ASSEMBLY__
-
-enum ultra_tlb_layout {
-	spitfire = 0,
-	cheetah = 1,
-	cheetah_plus = 2,
-	hypervisor = 3,
-};
-
-extern enum ultra_tlb_layout tlb_type;
-
-extern int sun4v_chip_type;
-
-extern int cheetah_pcache_forced_on;
-extern void cheetah_enable_pcache(void);
-
-#define sparc64_highest_locked_tlbent()	\
-	(tlb_type == spitfire ? \
-	 SPITFIRE_HIGHEST_LOCKED_TLBENT : \
-	 CHEETAH_HIGHEST_LOCKED_TLBENT)
-
-extern int num_kernel_image_mappings;
-
-/* The data cache is write through, so this just invalidates the
- * specified line.
- */
-static inline void spitfire_put_dcache_tag(unsigned long addr, unsigned long tag)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (tag), "r" (addr), "i" (ASI_DCACHE_TAG));
-}
-
-/* The instruction cache lines are flushed with this, but note that
- * this does not flush the pipeline.  It is possible for a line to
- * get flushed but stale instructions to still be in the pipeline,
- * a flush instruction (to any address) is sufficient to handle
- * this issue after the line is invalidated.
- */
-static inline void spitfire_put_icache_tag(unsigned long addr, unsigned long tag)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (tag), "r" (addr), "i" (ASI_IC_TAG));
-}
-
-static inline unsigned long spitfire_get_dtlb_data(int entry)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" (entry << 3), "i" (ASI_DTLB_DATA_ACCESS));
-
-	/* Clear TTE diag bits. */
-	data &= ~0x0003fe0000000000UL;
-
-	return data;
-}
-
-static inline unsigned long spitfire_get_dtlb_tag(int entry)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" (entry << 3), "i" (ASI_DTLB_TAG_READ));
-	return tag;
-}
-
-static inline void spitfire_put_dtlb_data(int entry, unsigned long data)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data), "r" (entry << 3),
-			       "i" (ASI_DTLB_DATA_ACCESS));
-}
-
-static inline unsigned long spitfire_get_itlb_data(int entry)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" (entry << 3), "i" (ASI_ITLB_DATA_ACCESS));
-
-	/* Clear TTE diag bits. */
-	data &= ~0x0003fe0000000000UL;
-
-	return data;
-}
-
-static inline unsigned long spitfire_get_itlb_tag(int entry)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" (entry << 3), "i" (ASI_ITLB_TAG_READ));
-	return tag;
-}
-
-static inline void spitfire_put_itlb_data(int entry, unsigned long data)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data), "r" (entry << 3),
-			       "i" (ASI_ITLB_DATA_ACCESS));
-}
-
-static inline void spitfire_flush_dtlb_nucleus_page(unsigned long page)
-{
-	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (page | 0x20), "i" (ASI_DMMU_DEMAP));
-}
-
-static inline void spitfire_flush_itlb_nucleus_page(unsigned long page)
-{
-	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (page | 0x20), "i" (ASI_IMMU_DEMAP));
-}
-
-/* Cheetah has "all non-locked" tlb flushes. */
-static inline void cheetah_flush_dtlb_all(void)
-{
-	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (0x80), "i" (ASI_DMMU_DEMAP));
-}
-
-static inline void cheetah_flush_itlb_all(void)
-{
-	__asm__ __volatile__("stxa	%%g0, [%0] %1\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (0x80), "i" (ASI_IMMU_DEMAP));
-}
-
-/* Cheetah has a 4-tlb layout so direct access is a bit different.
- * The first two TLBs are fully assosciative, hold 16 entries, and are
- * used only for locked and >8K sized translations.  One exists for
- * data accesses and one for instruction accesses.
- *
- * The third TLB is for data accesses to 8K non-locked translations, is
- * 2 way assosciative, and holds 512 entries.  The fourth TLB is for
- * instruction accesses to 8K non-locked translations, is 2 way
- * assosciative, and holds 128 entries.
- *
- * Cheetah has some bug where bogus data can be returned from
- * ASI_{D,I}TLB_DATA_ACCESS loads, doing the load twice fixes
- * the problem for me. -DaveM
- */
-static inline unsigned long cheetah_get_ldtlb_data(int entry)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
-			     "ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" ((0 << 16) | (entry << 3)),
-			     "i" (ASI_DTLB_DATA_ACCESS));
-
-	return data;
-}
-
-static inline unsigned long cheetah_get_litlb_data(int entry)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
-			     "ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" ((0 << 16) | (entry << 3)),
-			     "i" (ASI_ITLB_DATA_ACCESS));
-
-	return data;
-}
-
-static inline unsigned long cheetah_get_ldtlb_tag(int entry)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" ((0 << 16) | (entry << 3)),
-			     "i" (ASI_DTLB_TAG_READ));
-
-	return tag;
-}
-
-static inline unsigned long cheetah_get_litlb_tag(int entry)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" ((0 << 16) | (entry << 3)),
-			     "i" (ASI_ITLB_TAG_READ));
-
-	return tag;
-}
-
-static inline void cheetah_put_ldtlb_data(int entry, unsigned long data)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data),
-			       "r" ((0 << 16) | (entry << 3)),
-			       "i" (ASI_DTLB_DATA_ACCESS));
-}
-
-static inline void cheetah_put_litlb_data(int entry, unsigned long data)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data),
-			       "r" ((0 << 16) | (entry << 3)),
-			       "i" (ASI_ITLB_DATA_ACCESS));
-}
-
-static inline unsigned long cheetah_get_dtlb_data(int entry, int tlb)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
-			     "ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" ((tlb << 16) | (entry << 3)), "i" (ASI_DTLB_DATA_ACCESS));
-
-	return data;
-}
-
-static inline unsigned long cheetah_get_dtlb_tag(int entry, int tlb)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" ((tlb << 16) | (entry << 3)), "i" (ASI_DTLB_TAG_READ));
-	return tag;
-}
-
-static inline void cheetah_put_dtlb_data(int entry, unsigned long data, int tlb)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data),
-			       "r" ((tlb << 16) | (entry << 3)),
-			       "i" (ASI_DTLB_DATA_ACCESS));
-}
-
-static inline unsigned long cheetah_get_itlb_data(int entry)
-{
-	unsigned long data;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %%g0\n\t"
-			     "ldxa	[%1] %2, %0"
-			     : "=r" (data)
-			     : "r" ((2 << 16) | (entry << 3)),
-                               "i" (ASI_ITLB_DATA_ACCESS));
-
-	return data;
-}
-
-static inline unsigned long cheetah_get_itlb_tag(int entry)
-{
-	unsigned long tag;
-
-	__asm__ __volatile__("ldxa	[%1] %2, %0"
-			     : "=r" (tag)
-			     : "r" ((2 << 16) | (entry << 3)), "i" (ASI_ITLB_TAG_READ));
-	return tag;
-}
-
-static inline void cheetah_put_itlb_data(int entry, unsigned long data)
-{
-	__asm__ __volatile__("stxa	%0, [%1] %2\n\t"
-			     "membar	#Sync"
-			     : /* No outputs */
-			     : "r" (data), "r" ((2 << 16) | (entry << 3)),
-			       "i" (ASI_ITLB_DATA_ACCESS));
-}
-
-#endif /* !(__ASSEMBLY__) */
-
-#endif /* !(_SPARC64_SPITFIRE_H) */
+#include <asm-sparc/spitfire.h>
diff --git a/include/asm-sparc64/sstate.h b/include/asm-sparc64/sstate.h
index a7c35dbcb28..97720ce2fd4 100644
--- a/include/asm-sparc64/sstate.h
+++ b/include/asm-sparc64/sstate.h
@@ -1,13 +1 @@
-#ifndef _SPARC64_SSTATE_H
-#define _SPARC64_SSTATE_H
-
-extern void sstate_booting(void);
-extern void sstate_running(void);
-extern void sstate_halt(void);
-extern void sstate_poweroff(void);
-extern void sstate_panic(void);
-extern void sstate_reboot(void);
-
-extern void sun4v_sstate_init(void);
-
-#endif /* _SPARC64_SSTATE_H */
+#include <asm-sparc/sstate.h>
diff --git a/include/asm-sparc64/stacktrace.h b/include/asm-sparc64/stacktrace.h
index 6cee39adf6d..adc9b92c0ef 100644
--- a/include/asm-sparc64/stacktrace.h
+++ b/include/asm-sparc64/stacktrace.h
@@ -1,6 +1 @@
-#ifndef _SPARC64_STACKTRACE_H
-#define _SPARC64_STACKTRACE_H
-
-extern void stack_trace_flush(void);
-
-#endif /* _SPARC64_STACKTRACE_H */
+#include <asm-sparc/stacktrace.h>
diff --git a/include/asm-sparc64/starfire.h b/include/asm-sparc64/starfire.h
index 07bafd31e33..db97daa3bed 100644
--- a/include/asm-sparc64/starfire.h
+++ b/include/asm-sparc64/starfire.h
@@ -1,21 +1 @@
-/*
- * starfire.h: Group all starfire specific code together.
- *
- * Copyright (C) 2000 Anton Blanchard (anton@samba.org)
- */
-
-#ifndef _SPARC64_STARFIRE_H
-#define _SPARC64_STARFIRE_H
-
-#ifndef __ASSEMBLY__
-
-extern int this_is_starfire;
-
-extern void check_if_starfire(void);
-extern void starfire_cpu_setup(void);
-extern int starfire_hard_smp_processor_id(void);
-extern void starfire_hookup(int);
-extern unsigned int starfire_translate(unsigned long imap, unsigned int upaid);
-
-#endif
-#endif
+#include <asm-sparc/starfire.h>
diff --git a/include/asm-sparc64/stat.h b/include/asm-sparc64/stat.h
index 9650fdea847..b108a866256 100644
--- a/include/asm-sparc64/stat.h
+++ b/include/asm-sparc64/stat.h
@@ -1,47 +1 @@
-#ifndef _SPARC64_STAT_H
-#define _SPARC64_STAT_H
-
-#include <linux/types.h>
-
-struct stat {
-	unsigned   st_dev;
-	ino_t   st_ino;
-	mode_t  st_mode;
-	short   st_nlink;
-	uid_t   st_uid;
-	gid_t   st_gid;
-	unsigned   st_rdev;
-	off_t   st_size;
-	time_t  st_atime;
-	time_t  st_mtime;
-	time_t  st_ctime;
-	off_t   st_blksize;
-	off_t   st_blocks;
-	unsigned long  __unused4[2];
-};
-
-struct stat64 {
-	unsigned long	st_dev;
-	unsigned long	st_ino;
-	unsigned long	st_nlink;
-
-	unsigned int	st_mode;
-	unsigned int	st_uid;
-	unsigned int	st_gid;
-	unsigned int	__pad0;
-
-	unsigned long	st_rdev;
-	long		st_size;
-	long		st_blksize;
-	long		st_blocks;
-
-	unsigned long	st_atime;
-	unsigned long	st_atime_nsec;
-	unsigned long	st_mtime;
-	unsigned long	st_mtime_nsec;
-	unsigned long	st_ctime;
-	unsigned long	st_ctime_nsec;
-	long		__unused[3];
-};
-
-#endif
+#include <asm-sparc/stat.h>
diff --git a/include/asm-sparc64/statfs.h b/include/asm-sparc64/statfs.h
index 79b3c890a5f..5503d6a4c67 100644
--- a/include/asm-sparc64/statfs.h
+++ b/include/asm-sparc64/statfs.h
@@ -1,54 +1 @@
-#ifndef _SPARC64_STATFS_H
-#define _SPARC64_STATFS_H
-
-#ifndef __KERNEL_STRICT_NAMES
-
-#include <linux/types.h>
-
-typedef __kernel_fsid_t	fsid_t;
-
-#endif
-
-struct statfs {
-	long f_type;
-	long f_bsize;
-	long f_blocks;
-	long f_bfree;
-	long f_bavail;
-	long f_files;
-	long f_ffree;
-	__kernel_fsid_t f_fsid;
-	long f_namelen;
-	long f_frsize;
-	long f_spare[5];
-};
-
-struct statfs64 {
-	long f_type;
-	long f_bsize;
-	long f_blocks;
-	long f_bfree;
-	long f_bavail;
-	long f_files;
-	long f_ffree;
-	__kernel_fsid_t f_fsid;
-	long f_namelen;
-	long f_frsize;
-	long f_spare[5];
-};
-
-struct compat_statfs64 {
-	__u32 f_type;
-	__u32 f_bsize;
-	__u64 f_blocks;
-	__u64 f_bfree;
-	__u64 f_bavail;
-	__u64 f_files;
-	__u64 f_ffree;
-	__kernel_fsid_t f_fsid;
-	__u32 f_namelen;
-	__u32 f_frsize;
-	__u32 f_spare[5];
-};
-
-#endif
+#include <asm-sparc/statfs.h>
diff --git a/include/asm-sparc64/string.h b/include/asm-sparc64/string.h
index 43161f2d17e..5018cd8b6ad 100644
--- a/include/asm-sparc64/string.h
+++ b/include/asm-sparc64/string.h
@@ -1,83 +1 @@
-/*
- * string.h: External definitions for optimized assembly string
- *           routines for the Linux Kernel.
- *
- * Copyright (C) 1995,1996 David S. Miller (davem@caip.rutgers.edu)
- * Copyright (C) 1996,1997,1999 Jakub Jelinek (jakub@redhat.com)
- */
-
-#ifndef __SPARC64_STRING_H__
-#define __SPARC64_STRING_H__
-
-/* Really, userland/ksyms should not see any of this stuff. */
-
-#ifdef __KERNEL__
-
-#include <asm/asi.h>
-
-extern void *__memset(void *,int,__kernel_size_t);
-
-#ifndef EXPORT_SYMTAB_STROPS
-
-/* First the mem*() things. */
-#define __HAVE_ARCH_MEMMOVE
-extern void *memmove(void *, const void *, __kernel_size_t);
-
-#define __HAVE_ARCH_MEMCPY
-extern void *memcpy(void *, const void *, __kernel_size_t);
-
-#define __HAVE_ARCH_MEMSET
-extern void *__builtin_memset(void *,int,__kernel_size_t);
-
-static inline void *__constant_memset(void *s, int c, __kernel_size_t count)
-{
-	extern __kernel_size_t __bzero(void *, __kernel_size_t);
-
-	if (!c) {
-		__bzero(s, count);
-		return s;
-	} else
-		return __memset(s, c, count);
-}
-
-#undef memset
-#define memset(s, c, count) \
-((__builtin_constant_p(count) && (count) <= 32) ? \
- __builtin_memset((s), (c), (count)) : \
- (__builtin_constant_p(c) ? \
-  __constant_memset((s), (c), (count)) : \
-  __memset((s), (c), (count))))
-
-#define __HAVE_ARCH_MEMSCAN
-
-#undef memscan
-#define memscan(__arg0, __char, __arg2)					\
-({									\
-	extern void *__memscan_zero(void *, size_t);			\
-	extern void *__memscan_generic(void *, int, size_t);		\
-	void *__retval, *__addr = (__arg0);				\
-	size_t __size = (__arg2);					\
-									\
-	if(__builtin_constant_p(__char) && !(__char))			\
-		__retval = __memscan_zero(__addr, __size);		\
-	else								\
-		__retval = __memscan_generic(__addr, (__char), __size);	\
-									\
-	__retval;							\
-})
-
-#define __HAVE_ARCH_MEMCMP
-extern int memcmp(const void *,const void *,__kernel_size_t);
-
-/* Now the str*() stuff... */
-#define __HAVE_ARCH_STRLEN
-extern __kernel_size_t strlen(const char *);
-
-#define __HAVE_ARCH_STRNCMP
-extern int strncmp(const char *, const char *, __kernel_size_t);
-
-#endif /* !EXPORT_SYMTAB_STROPS */
-
-#endif /* __KERNEL__ */
-
-#endif /* !(__SPARC64_STRING_H__) */
+#include <asm-sparc/string.h>
diff --git a/include/asm-sparc64/sunbpp.h b/include/asm-sparc64/sunbpp.h
index 8e171b7a9f4..9632be290eb 100644
--- a/include/asm-sparc64/sunbpp.h
+++ b/include/asm-sparc64/sunbpp.h
@@ -1,80 +1 @@
-/*
- * include/asm-sparc64/sunbpp.h
- */
-
-#ifndef _ASM_SPARC64_SUNBPP_H
-#define _ASM_SPARC64_SUNBPP_H
-
-struct bpp_regs {
-  /* DMA registers */
-  __volatile__ __u32 p_csr;		/* DMA Control/Status Register */
-  __volatile__ __u32 p_addr;		/* Address Register */
-  __volatile__ __u32 p_bcnt;		/* Byte Count Register */
-  __volatile__ __u32 p_tst_csr;		/* Test Control/Status (DMA2 only) */
-  /* Parallel Port registers */
-  __volatile__ __u16 p_hcr;		/* Hardware Configuration Register */
-  __volatile__ __u16 p_ocr;		/* Operation Configuration Register */
-  __volatile__ __u8 p_dr;		/* Parallel Data Register */
-  __volatile__ __u8 p_tcr;		/* Transfer Control Register */
-  __volatile__ __u8 p_or;		/* Output Register */
-  __volatile__ __u8 p_ir;		/* Input Register */
-  __volatile__ __u16 p_icr;		/* Interrupt Control Register */
-};
-
-/* P_HCR. Time is in increments of SBus clock. */
-#define P_HCR_TEST      0x8000      /* Allows buried counters to be read */
-#define P_HCR_DSW       0x7f00      /* Data strobe width (in ticks) */
-#define P_HCR_DDS       0x007f      /* Data setup before strobe (in ticks) */
-
-/* P_OCR. */
-#define P_OCR_MEM_CLR   0x8000
-#define P_OCR_DATA_SRC  0x4000      /* )                  */
-#define P_OCR_DS_DSEL   0x2000      /* )  Bidirectional      */
-#define P_OCR_BUSY_DSEL 0x1000      /* )    selects            */
-#define P_OCR_ACK_DSEL  0x0800      /* )                  */
-#define P_OCR_EN_DIAG   0x0400
-#define P_OCR_BUSY_OP   0x0200      /* Busy operation */
-#define P_OCR_ACK_OP    0x0100      /* Ack operation */
-#define P_OCR_SRST      0x0080      /* Reset state machines. Not selfcleaning. */
-#define P_OCR_IDLE      0x0008      /* PP data transfer state machine is idle */
-#define P_OCR_V_ILCK    0x0002      /* Versatec faded. Zebra only. */
-#define P_OCR_EN_VER    0x0001      /* Enable Versatec (0 - enable). Zebra only. */
-
-/* P_TCR */
-#define P_TCR_DIR       0x08
-#define P_TCR_BUSY      0x04
-#define P_TCR_ACK       0x02
-#define P_TCR_DS        0x01        /* Strobe */
-
-/* P_OR */
-#define P_OR_V3         0x20        /* )                 */
-#define P_OR_V2         0x10        /* ) on Zebra only   */
-#define P_OR_V1         0x08        /* )                 */
-#define P_OR_INIT       0x04
-#define P_OR_AFXN       0x02        /* Auto Feed */
-#define P_OR_SLCT_IN    0x01
-
-/* P_IR */
-#define P_IR_PE         0x04
-#define P_IR_SLCT       0x02
-#define P_IR_ERR        0x01
-
-/* P_ICR */
-#define P_DS_IRQ        0x8000      /* RW1  */
-#define P_ACK_IRQ       0x4000      /* RW1  */
-#define P_BUSY_IRQ      0x2000      /* RW1  */
-#define P_PE_IRQ        0x1000      /* RW1  */
-#define P_SLCT_IRQ      0x0800      /* RW1  */
-#define P_ERR_IRQ       0x0400      /* RW1  */
-#define P_DS_IRQ_EN     0x0200      /* RW   Always on rising edge */
-#define P_ACK_IRQ_EN    0x0100      /* RW   Always on rising edge */
-#define P_BUSY_IRP      0x0080      /* RW   1= rising edge */
-#define P_BUSY_IRQ_EN   0x0040      /* RW   */
-#define P_PE_IRP        0x0020      /* RW   1= rising edge */
-#define P_PE_IRQ_EN     0x0010      /* RW   */
-#define P_SLCT_IRP      0x0008      /* RW   1= rising edge */
-#define P_SLCT_IRQ_EN   0x0004      /* RW   */
-#define P_ERR_IRP       0x0002      /* RW1  1= rising edge */
-#define P_ERR_IRQ_EN    0x0001      /* RW   */
-
-#endif /* !(_ASM_SPARC64_SUNBPP_H) */
+#include <asm-sparc/sunbpp.h>
diff --git a/include/asm-sparc64/syscalls.h b/include/asm-sparc64/syscalls.h
index 45a43f637a1..3477b16e30c 100644
--- a/include/asm-sparc64/syscalls.h
+++ b/include/asm-sparc64/syscalls.h
@@ -1,13 +1 @@
-#ifndef _SPARC64_SYSCALLS_H
-#define _SPARC64_SYSCALLS_H
-
-struct pt_regs;
-
-extern asmlinkage long sparc_do_fork(unsigned long clone_flags,
-				     unsigned long stack_start,
-				     struct pt_regs *regs,
-				     unsigned long stack_size);
-
-extern asmlinkage int sparc_execve(struct pt_regs *regs);
-
-#endif /* _SPARC64_SYSCALLS_H */
+#include <asm-sparc/syscalls.h>
diff --git a/include/asm-sparc64/system.h b/include/asm-sparc64/system.h
index 6897ac31be4..be2603c2e52 100644
--- a/include/asm-sparc64/system.h
+++ b/include/asm-sparc64/system.h
@@ -1,355 +1 @@
-#ifndef __SPARC64_SYSTEM_H
-#define __SPARC64_SYSTEM_H
-
-#include <asm/ptrace.h>
-#include <asm/processor.h>
-#include <asm/visasm.h>
-
-#ifndef __ASSEMBLY__
-
-#include <linux/irqflags.h>
-#include <asm-generic/cmpxchg-local.h>
-
-/*
- * Sparc (general) CPU types
- */
-enum sparc_cpu {
-  sun4        = 0x00,
-  sun4c       = 0x01,
-  sun4m       = 0x02,
-  sun4d       = 0x03,
-  sun4e       = 0x04,
-  sun4u       = 0x05, /* V8 ploos ploos */
-  sun_unknown = 0x06,
-  ap1000      = 0x07, /* almost a sun4m */
-};
-                  
-#define sparc_cpu_model sun4u
-
-/* This cannot ever be a sun4c nor sun4 :) That's just history. */
-#define ARCH_SUN4C_SUN4 0
-#define ARCH_SUN4 0
-
-extern char reboot_command[];
-
-/* These are here in an effort to more fully work around Spitfire Errata
- * #51.  Essentially, if a memory barrier occurs soon after a mispredicted
- * branch, the chip can stop executing instructions until a trap occurs.
- * Therefore, if interrupts are disabled, the chip can hang forever.
- *
- * It used to be believed that the memory barrier had to be right in the
- * delay slot, but a case has been traced recently wherein the memory barrier
- * was one instruction after the branch delay slot and the chip still hung.
- * The offending sequence was the following in sym_wakeup_done() of the
- * sym53c8xx_2 driver:
- *
- *	call	sym_ccb_from_dsa, 0
- *	 movge	%icc, 0, %l0
- *	brz,pn	%o0, .LL1303
- *	 mov	%o0, %l2
- *	membar	#LoadLoad
- *
- * The branch has to be mispredicted for the bug to occur.  Therefore, we put
- * the memory barrier explicitly into a "branch always, predicted taken"
- * delay slot to avoid the problem case.
- */
-#define membar_safe(type) \
-do {	__asm__ __volatile__("ba,pt	%%xcc, 1f\n\t" \
-			     " membar	" type "\n" \
-			     "1:\n" \
-			     : : : "memory"); \
-} while (0)
-
-#define mb()	\
-	membar_safe("#LoadLoad | #LoadStore | #StoreStore | #StoreLoad")
-#define rmb()	\
-	membar_safe("#LoadLoad")
-#define wmb()	\
-	membar_safe("#StoreStore")
-#define membar_storeload() \
-	membar_safe("#StoreLoad")
-#define membar_storeload_storestore() \
-	membar_safe("#StoreLoad | #StoreStore")
-#define membar_storeload_loadload() \
-	membar_safe("#StoreLoad | #LoadLoad")
-#define membar_storestore_loadstore() \
-	membar_safe("#StoreStore | #LoadStore")
-
-#endif
-
-#define nop() 		__asm__ __volatile__ ("nop")
-
-#define read_barrier_depends()		do { } while(0)
-#define set_mb(__var, __value) \
-	do { __var = __value; membar_storeload_storestore(); } while(0)
-
-#ifdef CONFIG_SMP
-#define smp_mb()	mb()
-#define smp_rmb()	rmb()
-#define smp_wmb()	wmb()
-#define smp_read_barrier_depends()	read_barrier_depends()
-#else
-#define smp_mb()	__asm__ __volatile__("":::"memory")
-#define smp_rmb()	__asm__ __volatile__("":::"memory")
-#define smp_wmb()	__asm__ __volatile__("":::"memory")
-#define smp_read_barrier_depends()	do { } while(0)
-#endif
-
-#define flushi(addr)	__asm__ __volatile__ ("flush %0" : : "r" (addr) : "memory")
-
-#define flushw_all()	__asm__ __volatile__("flushw")
-
-/* Performance counter register access. */
-#define read_pcr(__p)  __asm__ __volatile__("rd	%%pcr, %0" : "=r" (__p))
-#define write_pcr(__p) __asm__ __volatile__("wr	%0, 0x0, %%pcr" : : "r" (__p))
-#define read_pic(__p)  __asm__ __volatile__("rd %%pic, %0" : "=r" (__p))
-
-/* Blackbird errata workaround.  See commentary in
- * arch/sparc64/kernel/smp.c:smp_percpu_timer_interrupt()
- * for more information.
- */
-#define reset_pic()    						\
-	__asm__ __volatile__("ba,pt	%xcc, 99f\n\t"		\
-			     ".align	64\n"			\
-			  "99:wr	%g0, 0x0, %pic\n\t"	\
-			     "rd	%pic, %g0")
-
-#ifndef __ASSEMBLY__
-
-extern void sun_do_break(void);
-extern int stop_a_enabled;
-
-extern void fault_in_user_windows(void);
-extern void synchronize_user_stack(void);
-
-extern void __flushw_user(void);
-#define flushw_user() __flushw_user()
-
-#define flush_user_windows flushw_user
-#define flush_register_windows flushw_all
-
-/* Don't hold the runqueue lock over context switch */
-#define __ARCH_WANT_UNLOCKED_CTXSW
-#define prepare_arch_switch(next)		\
-do {						\
-	flushw_all();				\
-} while (0)
-
-	/* See what happens when you design the chip correctly?
-	 *
-	 * We tell gcc we clobber all non-fixed-usage registers except
-	 * for l0/l1.  It will use one for 'next' and the other to hold
-	 * the output value of 'last'.  'next' is not referenced again
-	 * past the invocation of switch_to in the scheduler, so we need
-	 * not preserve it's value.  Hairy, but it lets us remove 2 loads
-	 * and 2 stores in this critical code path.  -DaveM
-	 */
-#define switch_to(prev, next, last)					\
-do {	if (test_thread_flag(TIF_PERFCTR)) {				\
-		unsigned long __tmp;					\
-		read_pcr(__tmp);					\
-		current_thread_info()->pcr_reg = __tmp;			\
-		read_pic(__tmp);					\
-		current_thread_info()->kernel_cntd0 += (unsigned int)(__tmp);\
-		current_thread_info()->kernel_cntd1 += ((__tmp) >> 32);	\
-	}								\
-	flush_tlb_pending();						\
-	save_and_clear_fpu();						\
-	/* If you are tempted to conditionalize the following */	\
-	/* so that ASI is only written if it changes, think again. */	\
-	__asm__ __volatile__("wr %%g0, %0, %%asi"			\
-	: : "r" (__thread_flag_byte_ptr(task_thread_info(next))[TI_FLAG_BYTE_CURRENT_DS]));\
-	trap_block[current_thread_info()->cpu].thread =			\
-		task_thread_info(next);					\
-	__asm__ __volatile__(						\
-	"mov	%%g4, %%g7\n\t"						\
-	"stx	%%i6, [%%sp + 2047 + 0x70]\n\t"				\
-	"stx	%%i7, [%%sp + 2047 + 0x78]\n\t"				\
-	"rdpr	%%wstate, %%o5\n\t"					\
-	"stx	%%o6, [%%g6 + %6]\n\t"					\
-	"stb	%%o5, [%%g6 + %5]\n\t"					\
-	"rdpr	%%cwp, %%o5\n\t"					\
-	"stb	%%o5, [%%g6 + %8]\n\t"					\
-	"mov	%4, %%g6\n\t"						\
-	"ldub	[%4 + %8], %%g1\n\t"					\
-	"wrpr	%%g1, %%cwp\n\t"					\
-	"ldx	[%%g6 + %6], %%o6\n\t"					\
-	"ldub	[%%g6 + %5], %%o5\n\t"					\
-	"ldub	[%%g6 + %7], %%o7\n\t"					\
-	"wrpr	%%o5, 0x0, %%wstate\n\t"				\
-	"ldx	[%%sp + 2047 + 0x70], %%i6\n\t"				\
-	"ldx	[%%sp + 2047 + 0x78], %%i7\n\t"				\
-	"ldx	[%%g6 + %9], %%g4\n\t"					\
-	"brz,pt %%o7, switch_to_pc\n\t"					\
-	" mov	%%g7, %0\n\t"						\
-	"sethi	%%hi(ret_from_syscall), %%g1\n\t"			\
-	"jmpl	%%g1 + %%lo(ret_from_syscall), %%g0\n\t"		\
-	" nop\n\t"							\
-	".globl switch_to_pc\n\t"					\
-	"switch_to_pc:\n\t"						\
-	: "=&r" (last), "=r" (current), "=r" (current_thread_info_reg),	\
-	  "=r" (__local_per_cpu_offset)					\
-	: "0" (task_thread_info(next)),					\
-	  "i" (TI_WSTATE), "i" (TI_KSP), "i" (TI_NEW_CHILD),            \
-	  "i" (TI_CWP), "i" (TI_TASK)					\
-	: "cc",								\
-	        "g1", "g2", "g3",                   "g7",		\
-	        "l1", "l2", "l3", "l4", "l5", "l6", "l7",		\
-	  "i0", "i1", "i2", "i3", "i4", "i5",				\
-	  "o0", "o1", "o2", "o3", "o4", "o5",       "o7");		\
-	/* If you fuck with this, update ret_from_syscall code too. */	\
-	if (test_thread_flag(TIF_PERFCTR)) {				\
-		write_pcr(current_thread_info()->pcr_reg);		\
-		reset_pic();						\
-	}								\
-} while(0)
-
-static inline unsigned long xchg32(__volatile__ unsigned int *m, unsigned int val)
-{
-	unsigned long tmp1, tmp2;
-
-	__asm__ __volatile__(
-"	membar		#StoreLoad | #LoadLoad\n"
-"	mov		%0, %1\n"
-"1:	lduw		[%4], %2\n"
-"	cas		[%4], %2, %0\n"
-"	cmp		%2, %0\n"
-"	bne,a,pn	%%icc, 1b\n"
-"	 mov		%1, %0\n"
-"	membar		#StoreLoad | #StoreStore\n"
-	: "=&r" (val), "=&r" (tmp1), "=&r" (tmp2)
-	: "0" (val), "r" (m)
-	: "cc", "memory");
-	return val;
-}
-
-static inline unsigned long xchg64(__volatile__ unsigned long *m, unsigned long val)
-{
-	unsigned long tmp1, tmp2;
-
-	__asm__ __volatile__(
-"	membar		#StoreLoad | #LoadLoad\n"
-"	mov		%0, %1\n"
-"1:	ldx		[%4], %2\n"
-"	casx		[%4], %2, %0\n"
-"	cmp		%2, %0\n"
-"	bne,a,pn	%%xcc, 1b\n"
-"	 mov		%1, %0\n"
-"	membar		#StoreLoad | #StoreStore\n"
-	: "=&r" (val), "=&r" (tmp1), "=&r" (tmp2)
-	: "0" (val), "r" (m)
-	: "cc", "memory");
-	return val;
-}
-
-#define xchg(ptr,x) ((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
-
-extern void __xchg_called_with_bad_pointer(void);
-
-static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr,
-				       int size)
-{
-	switch (size) {
-	case 4:
-		return xchg32(ptr, x);
-	case 8:
-		return xchg64(ptr, x);
-	};
-	__xchg_called_with_bad_pointer();
-	return x;
-}
-
-extern void die_if_kernel(char *str, struct pt_regs *regs) __attribute__ ((noreturn));
-
-/* 
- * Atomic compare and exchange.  Compare OLD with MEM, if identical,
- * store NEW in MEM.  Return the initial value in MEM.  Success is
- * indicated by comparing RETURN with OLD.
- */
-
-#define __HAVE_ARCH_CMPXCHG 1
-
-static inline unsigned long
-__cmpxchg_u32(volatile int *m, int old, int new)
-{
-	__asm__ __volatile__("membar #StoreLoad | #LoadLoad\n"
-			     "cas [%2], %3, %0\n\t"
-			     "membar #StoreLoad | #StoreStore"
-			     : "=&r" (new)
-			     : "0" (new), "r" (m), "r" (old)
-			     : "memory");
-
-	return new;
-}
-
-static inline unsigned long
-__cmpxchg_u64(volatile long *m, unsigned long old, unsigned long new)
-{
-	__asm__ __volatile__("membar #StoreLoad | #LoadLoad\n"
-			     "casx [%2], %3, %0\n\t"
-			     "membar #StoreLoad | #StoreStore"
-			     : "=&r" (new)
-			     : "0" (new), "r" (m), "r" (old)
-			     : "memory");
-
-	return new;
-}
-
-/* This function doesn't exist, so you'll get a linker error
-   if something tries to do an invalid cmpxchg().  */
-extern void __cmpxchg_called_with_bad_pointer(void);
-
-static inline unsigned long
-__cmpxchg(volatile void *ptr, unsigned long old, unsigned long new, int size)
-{
-	switch (size) {
-		case 4:
-			return __cmpxchg_u32(ptr, old, new);
-		case 8:
-			return __cmpxchg_u64(ptr, old, new);
-	}
-	__cmpxchg_called_with_bad_pointer();
-	return old;
-}
-
-#define cmpxchg(ptr,o,n)						 \
-  ({									 \
-     __typeof__(*(ptr)) _o_ = (o);					 \
-     __typeof__(*(ptr)) _n_ = (n);					 \
-     (__typeof__(*(ptr))) __cmpxchg((ptr), (unsigned long)_o_,		 \
-				    (unsigned long)_n_, sizeof(*(ptr))); \
-  })
-
-/*
- * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
- * them available.
- */
-
-static inline unsigned long __cmpxchg_local(volatile void *ptr,
-				      unsigned long old,
-				      unsigned long new, int size)
-{
-	switch (size) {
-	case 4:
-	case 8:	return __cmpxchg(ptr, old, new, size);
-	default:
-		return __cmpxchg_local_generic(ptr, old, new, size);
-	}
-
-	return old;
-}
-
-#define cmpxchg_local(ptr, o, n)				  	\
-	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
-			(unsigned long)(n), sizeof(*(ptr))))
-#define cmpxchg64_local(ptr, o, n)					\
-  ({									\
-	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
-	cmpxchg_local((ptr), (o), (n));					\
-  })
-
-#endif /* !(__ASSEMBLY__) */
-
-#define arch_align_stack(x) (x)
-
-#endif /* !(__SPARC64_SYSTEM_H) */
+#include <asm-sparc/system.h>
diff --git a/include/asm-sparc64/termbits.h b/include/asm-sparc64/termbits.h
index ebe31c152f1..e03f97592c7 100644
--- a/include/asm-sparc64/termbits.h
+++ b/include/asm-sparc64/termbits.h
@@ -1,260 +1 @@
-#ifndef _SPARC64_TERMBITS_H
-#define _SPARC64_TERMBITS_H
-
-#include <linux/posix_types.h>
-
-typedef unsigned char   cc_t;
-typedef unsigned int    speed_t;
-typedef unsigned int    tcflag_t;
-
-#define NCC 8
-struct termio {
-	unsigned short c_iflag;		/* input mode flags */
-	unsigned short c_oflag;		/* output mode flags */
-	unsigned short c_cflag;		/* control mode flags */
-	unsigned short c_lflag;		/* local mode flags */
-	unsigned char c_line;		/* line discipline */
-	unsigned char c_cc[NCC];	/* control characters */
-};
-
-#define NCCS 17
-struct termios {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-#ifdef __KERNEL__
-#define SIZEOF_USER_TERMIOS sizeof (struct termios) - (2*sizeof (cc_t))
-	cc_t _x_cc[2];                  /* We need them to hold vmin/vtime */
-#endif
-};
-
-struct termios2 {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-	cc_t _x_cc[2];                  /* padding to match ktermios */
-	speed_t c_ispeed;		/* input speed */
-	speed_t c_ospeed;		/* output speed */
-};
-
-struct ktermios {
-	tcflag_t c_iflag;		/* input mode flags */
-	tcflag_t c_oflag;		/* output mode flags */
-	tcflag_t c_cflag;		/* control mode flags */
-	tcflag_t c_lflag;		/* local mode flags */
-	cc_t c_line;			/* line discipline */
-	cc_t c_cc[NCCS];		/* control characters */
-	cc_t _x_cc[2];                  /* We need them to hold vmin/vtime */
-	speed_t c_ispeed;		/* input speed */
-	speed_t c_ospeed;		/* output speed */
-};
-
-/* c_cc characters */
-#define VINTR    0
-#define VQUIT    1
-#define VERASE   2
-#define VKILL    3
-#define VEOF     4
-#define VEOL     5
-#define VEOL2    6
-#define VSWTC    7
-#define VSTART   8
-#define VSTOP    9
-
-
-
-#define VSUSP    10
-#define VDSUSP   11  /* SunOS POSIX nicety I do believe... */
-#define VREPRINT 12
-#define VDISCARD 13
-#define VWERASE  14
-#define VLNEXT   15
-
-/* Kernel keeps vmin/vtime separated, user apps assume vmin/vtime is
- * shared with eof/eol
- */
-#ifdef __KERNEL__
-#define VMIN     16
-#define VTIME    17
-#else
-#define VMIN     VEOF
-#define VTIME    VEOL
-#endif
-
-/* c_iflag bits */
-#define IGNBRK	0x00000001
-#define BRKINT	0x00000002
-#define IGNPAR	0x00000004
-#define PARMRK	0x00000008
-#define INPCK	0x00000010
-#define ISTRIP	0x00000020
-#define INLCR	0x00000040
-#define IGNCR	0x00000080
-#define ICRNL	0x00000100
-#define IUCLC	0x00000200
-#define IXON	0x00000400
-#define IXANY	0x00000800
-#define IXOFF	0x00001000
-#define IMAXBEL	0x00002000
-#define IUTF8	0x00004000
-
-/* c_oflag bits */
-#define OPOST	0x00000001
-#define OLCUC	0x00000002
-#define ONLCR	0x00000004
-#define OCRNL	0x00000008
-#define ONOCR	0x00000010
-#define ONLRET	0x00000020
-#define OFILL	0x00000040
-#define OFDEL	0x00000080
-#define NLDLY	0x00000100
-#define   NL0	0x00000000
-#define   NL1	0x00000100
-#define CRDLY	0x00000600
-#define   CR0	0x00000000
-#define   CR1	0x00000200
-#define   CR2	0x00000400
-#define   CR3	0x00000600
-#define TABDLY	0x00001800
-#define   TAB0	0x00000000
-#define   TAB1	0x00000800
-#define   TAB2	0x00001000
-#define   TAB3	0x00001800
-#define   XTABS	0x00001800
-#define BSDLY	0x00002000
-#define   BS0	0x00000000
-#define   BS1	0x00002000
-#define VTDLY	0x00004000
-#define   VT0	0x00000000
-#define   VT1	0x00004000
-#define FFDLY	0x00008000
-#define   FF0	0x00000000
-#define   FF1	0x00008000
-#define PAGEOUT 0x00010000  /* SUNOS specific */
-#define WRAP    0x00020000  /* SUNOS specific */
-
-/* c_cflag bit meaning */
-#define CBAUD	  0x0000100f
-#define  B0	  0x00000000   /* hang up */
-#define  B50	  0x00000001
-#define  B75	  0x00000002
-#define  B110	  0x00000003
-#define  B134	  0x00000004
-#define  B150	  0x00000005
-#define  B200	  0x00000006
-#define  B300	  0x00000007
-#define  B600	  0x00000008
-#define  B1200	  0x00000009
-#define  B1800	  0x0000000a
-#define  B2400	  0x0000000b
-#define  B4800	  0x0000000c
-#define  B9600	  0x0000000d
-#define  B19200	  0x0000000e
-#define  B38400	  0x0000000f
-#define EXTA      B19200
-#define EXTB      B38400
-#define  CSIZE    0x00000030
-#define   CS5	  0x00000000
-#define   CS6	  0x00000010
-#define   CS7	  0x00000020
-#define   CS8	  0x00000030
-#define CSTOPB	  0x00000040
-#define CREAD	  0x00000080
-#define PARENB	  0x00000100
-#define PARODD	  0x00000200
-#define HUPCL	  0x00000400
-#define CLOCAL	  0x00000800
-#define CBAUDEX   0x00001000
-#define  BOTHER   0x00001000
-#define  B57600   0x00001001
-#define  B115200  0x00001002
-#define  B230400  0x00001003
-#define  B460800  0x00001004
-/* This is what we can do with the Zilogs. */
-#define  B76800   0x00001005
-/* This is what we can do with the SAB82532. */
-#define  B153600  0x00001006
-#define  B307200  0x00001007
-#define  B614400  0x00001008
-#define  B921600  0x00001009
-/* And these are the rest... */
-#define  B500000  0x0000100a
-#define  B576000  0x0000100b
-#define B1000000  0x0000100c
-#define B1152000  0x0000100d
-#define B1500000  0x0000100e
-#define B2000000  0x0000100f
-/* These have totally bogus values and nobody uses them
-   so far. Later on we'd have to use say 0x10000x and
-   adjust CBAUD constant and drivers accordingly.
-#define B2500000  0x00001010
-#define B3000000  0x00001011
-#define B3500000  0x00001012
-#define B4000000  0x00001013  */
-#define CIBAUD	  0x100f0000  /* input baud rate (not used) */
-#define CMSPAR    0x40000000  /* mark or space (stick) parity */
-#define CRTSCTS	  0x80000000  /* flow control */
-
-#define IBSHIFT	  16		/* Shift from CBAUD to CIBAUD */
-
-/* c_lflag bits */
-#define ISIG	0x00000001
-#define ICANON	0x00000002
-#define XCASE	0x00000004
-#define ECHO	0x00000008
-#define ECHOE	0x00000010
-#define ECHOK	0x00000020
-#define ECHONL	0x00000040
-#define NOFLSH	0x00000080
-#define TOSTOP	0x00000100
-#define ECHOCTL	0x00000200
-#define ECHOPRT	0x00000400
-#define ECHOKE	0x00000800
-#define DEFECHO 0x00001000  /* SUNOS thing, what is it? */
-#define FLUSHO	0x00002000
-#define PENDIN	0x00004000
-#define IEXTEN	0x00008000
-
-/* modem lines */
-#define TIOCM_LE	0x001
-#define TIOCM_DTR	0x002
-#define TIOCM_RTS	0x004
-#define TIOCM_ST	0x008
-#define TIOCM_SR	0x010
-#define TIOCM_CTS	0x020
-#define TIOCM_CAR	0x040
-#define TIOCM_RNG	0x080
-#define TIOCM_DSR	0x100
-#define TIOCM_CD	TIOCM_CAR
-#define TIOCM_RI	TIOCM_RNG
-#define TIOCM_OUT1	0x2000
-#define TIOCM_OUT2	0x4000
-#define TIOCM_LOOP	0x8000
-
-/* ioctl (fd, TIOCSERGETLSR, &result) where result may be as below */
-#define TIOCSER_TEMT    0x01	/* Transmitter physically empty */
-
-
-/* tcflow() and TCXONC use these */
-#define	TCOOFF		0
-#define	TCOON		1
-#define	TCIOFF		2
-#define	TCION		3
-
-/* tcflush() and TCFLSH use these */
-#define	TCIFLUSH	0
-#define	TCOFLUSH	1
-#define	TCIOFLUSH	2
-
-/* tcsetattr uses these */
-#define	TCSANOW		0
-#define	TCSADRAIN	1
-#define	TCSAFLUSH	2
-
-#endif /* !(_SPARC64_TERMBITS_H) */
+#include <asm-sparc/termbits.h>
diff --git a/include/asm-sparc64/termios.h b/include/asm-sparc64/termios.h
index 1f5dab25dda..940495eb05c 100644
--- a/include/asm-sparc64/termios.h
+++ b/include/asm-sparc64/termios.h
@@ -1,186 +1 @@
-#ifndef _SPARC64_TERMIOS_H
-#define _SPARC64_TERMIOS_H
-
-#include <asm/ioctls.h>
-#include <asm/termbits.h>
-
-#if defined(__KERNEL__) || defined(__DEFINE_BSD_TERMIOS)
-struct sgttyb {
-	char	sg_ispeed;
-	char	sg_ospeed;
-	char	sg_erase;
-	char	sg_kill;
-	short	sg_flags;
-};
-
-struct tchars {
-	char	t_intrc;
-	char	t_quitc;
-	char	t_startc;
-	char	t_stopc;
-	char	t_eofc;
-	char	t_brkc;
-};
-
-struct ltchars {
-	char	t_suspc;
-	char	t_dsuspc;
-	char	t_rprntc;
-	char	t_flushc;
-	char	t_werasc;
-	char	t_lnextc;
-};
-#endif /* __KERNEL__ */
-
-struct winsize {
-	unsigned short ws_row;
-	unsigned short ws_col;
-	unsigned short ws_xpixel;
-	unsigned short ws_ypixel;
-};
-
-#ifdef __KERNEL__
-#include <linux/module.h>
-
-/*
- * c_cc characters in the termio structure.  Oh, how I love being
- * backwardly compatible.  Notice that character 4 and 5 are
- * interpreted differently depending on whether ICANON is set in
- * c_lflag.  If it's set, they are used as _VEOF and _VEOL, otherwise
- * as _VMIN and V_TIME.  This is for compatibility with OSF/1 (which
- * is compatible with sysV)...
- */
-#define _VMIN	4
-#define _VTIME	5
-
-/*	intr=^C		quit=^\		erase=del	kill=^U
-	eof=^D		eol=\0		eol2=\0		sxtc=\0
-	start=^Q	stop=^S		susp=^Z		dsusp=^Y
-	reprint=^R	discard=^U	werase=^W	lnext=^V
-	vmin=\1         vtime=\0
-*/
-#define INIT_C_CC "\003\034\177\025\004\000\000\000\021\023\032\031\022\025\027\026\001"
-
-/*
- * Translate a "termio" structure into a "termios". Ugh.
- */
-#define user_termio_to_kernel_termios(termios, termio) \
-({ \
-	unsigned short tmp; \
-	int err; \
-	err = get_user(tmp, &(termio)->c_iflag); \
-	(termios)->c_iflag = (0xffff0000 & ((termios)->c_iflag)) | tmp; \
-	err |= get_user(tmp, &(termio)->c_oflag); \
-	(termios)->c_oflag = (0xffff0000 & ((termios)->c_oflag)) | tmp; \
-	err |= get_user(tmp, &(termio)->c_cflag); \
-	(termios)->c_cflag = (0xffff0000 & ((termios)->c_cflag)) | tmp; \
-	err |= get_user(tmp, &(termio)->c_lflag); \
-	(termios)->c_lflag = (0xffff0000 & ((termios)->c_lflag)) | tmp; \
-	err |= copy_from_user((termios)->c_cc, (termio)->c_cc, NCC); \
-	err; \
-})
-
-/*
- * Translate a "termios" structure into a "termio". Ugh.
- *
- * Note the "fun" _VMIN overloading.
- */
-#define kernel_termios_to_user_termio(termio, termios) \
-({ \
-	int err; \
-	err  = put_user((termios)->c_iflag, &(termio)->c_iflag); \
-	err |= put_user((termios)->c_oflag, &(termio)->c_oflag); \
-	err |= put_user((termios)->c_cflag, &(termio)->c_cflag); \
-	err |= put_user((termios)->c_lflag, &(termio)->c_lflag); \
-	err |= put_user((termios)->c_line,  &(termio)->c_line); \
-	err |= copy_to_user((termio)->c_cc, (termios)->c_cc, NCC); \
-	if (!((termios)->c_lflag & ICANON)) { \
-		err |= put_user((termios)->c_cc[VMIN], &(termio)->c_cc[_VMIN]); \
-		err |= put_user((termios)->c_cc[VTIME], &(termio)->c_cc[_VTIME]); \
-	} \
-	err; \
-})
-
-#define user_termios_to_kernel_termios(k, u) \
-({ \
-	int err; \
-	err  = get_user((k)->c_iflag, &(u)->c_iflag); \
-	err |= get_user((k)->c_oflag, &(u)->c_oflag); \
-	err |= get_user((k)->c_cflag, &(u)->c_cflag); \
-	err |= get_user((k)->c_lflag, &(u)->c_lflag); \
-	err |= get_user((k)->c_line,  &(u)->c_line); \
-	err |= copy_from_user((k)->c_cc, (u)->c_cc, NCCS); \
-	if((k)->c_lflag & ICANON) { \
-		err |= get_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		err |= get_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
-	} else { \
-		err |= get_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		err |= get_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
-	} \
-	err |= get_user((k)->c_ispeed,  &(u)->c_ispeed); \
-	err |= get_user((k)->c_ospeed,  &(u)->c_ospeed); \
-	err; \
-})
-
-#define kernel_termios_to_user_termios(u, k) \
-({ \
-	int err; \
-	err  = put_user((k)->c_iflag, &(u)->c_iflag); \
-	err |= put_user((k)->c_oflag, &(u)->c_oflag); \
-	err |= put_user((k)->c_cflag, &(u)->c_cflag); \
-	err |= put_user((k)->c_lflag, &(u)->c_lflag); \
-	err |= put_user((k)->c_line, &(u)->c_line); \
-	err |= copy_to_user((u)->c_cc, (k)->c_cc, NCCS); \
-	if(!((k)->c_lflag & ICANON)) { \
-		err |= put_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		err |= put_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
-	} else { \
-		err |= put_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		err |= put_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
-	} \
-	err |= put_user((k)->c_ispeed, &(u)->c_ispeed); \
-	err |= put_user((k)->c_ospeed, &(u)->c_ospeed); \
-	err; \
-})
-
-#define user_termios_to_kernel_termios_1(k, u) \
-({ \
-	int err; \
-	err  = get_user((k)->c_iflag, &(u)->c_iflag); \
-	err |= get_user((k)->c_oflag, &(u)->c_oflag); \
-	err |= get_user((k)->c_cflag, &(u)->c_cflag); \
-	err |= get_user((k)->c_lflag, &(u)->c_lflag); \
-	err |= get_user((k)->c_line,  &(u)->c_line); \
-	err |= copy_from_user((k)->c_cc, (u)->c_cc, NCCS); \
-	if((k)->c_lflag & ICANON) { \
-		err |= get_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		err |= get_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
-	} else { \
-		err |= get_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		err |= get_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
-	} \
-	err; \
-})
-
-#define kernel_termios_to_user_termios_1(u, k) \
-({ \
-	int err; \
-	err  = put_user((k)->c_iflag, &(u)->c_iflag); \
-	err |= put_user((k)->c_oflag, &(u)->c_oflag); \
-	err |= put_user((k)->c_cflag, &(u)->c_cflag); \
-	err |= put_user((k)->c_lflag, &(u)->c_lflag); \
-	err |= put_user((k)->c_line, &(u)->c_line); \
-	err |= copy_to_user((u)->c_cc, (k)->c_cc, NCCS); \
-	if(!((k)->c_lflag & ICANON)) { \
-		err |= put_user((k)->c_cc[VMIN],  &(u)->c_cc[_VMIN]); \
-		err |= put_user((k)->c_cc[VTIME], &(u)->c_cc[_VTIME]); \
-	} else { \
-		err |= put_user((k)->c_cc[VEOF], &(u)->c_cc[VEOF]); \
-		err |= put_user((k)->c_cc[VEOL], &(u)->c_cc[VEOL]); \
-	} \
-	err; \
-})
-
-#endif	/* __KERNEL__ */
-
-#endif /* _SPARC64_TERMIOS_H */
+#include <asm-sparc/termios.h>
diff --git a/include/asm-sparc64/thread_info.h b/include/asm-sparc64/thread_info.h
index e5873e38530..92bed791339 100644
--- a/include/asm-sparc64/thread_info.h
+++ b/include/asm-sparc64/thread_info.h
@@ -1,277 +1 @@
-/* thread_info.h: sparc64 low-level thread information
- *
- * Copyright (C) 2002  David S. Miller (davem@redhat.com)
- */
-
-#ifndef _ASM_THREAD_INFO_H
-#define _ASM_THREAD_INFO_H
-
-#ifdef __KERNEL__
-
-#define NSWINS		7
-
-#define TI_FLAG_BYTE_FAULT_CODE		0
-#define TI_FLAG_FAULT_CODE_SHIFT	56
-#define TI_FLAG_BYTE_WSTATE		1
-#define TI_FLAG_WSTATE_SHIFT		48
-#define TI_FLAG_BYTE_CWP		2
-#define TI_FLAG_CWP_SHIFT		40
-#define TI_FLAG_BYTE_CURRENT_DS		3
-#define TI_FLAG_CURRENT_DS_SHIFT	32
-#define TI_FLAG_BYTE_FPDEPTH		4
-#define TI_FLAG_FPDEPTH_SHIFT		24
-#define TI_FLAG_BYTE_WSAVED		5
-#define TI_FLAG_WSAVED_SHIFT		16
-
-#include <asm/page.h>
-
-#ifndef __ASSEMBLY__
-
-#include <asm/ptrace.h>
-#include <asm/types.h>
-
-struct task_struct;
-struct exec_domain;
-
-struct thread_info {
-	/* D$ line 1 */
-	struct task_struct	*task;
-	unsigned long		flags;
-	__u8			fpsaved[7];
-	__u8			status;
-	unsigned long		ksp;
-
-	/* D$ line 2 */
-	unsigned long		fault_address;
-	struct pt_regs		*kregs;
-	struct exec_domain	*exec_domain;
-	int			preempt_count;	/* 0 => preemptable, <0 => BUG */
-	__u8			new_child;
-	__u8			syscall_noerror;
-	__u16			cpu;
-
-	unsigned long		*utraps;
-
-	struct reg_window 	reg_window[NSWINS];
-	unsigned long 		rwbuf_stkptrs[NSWINS];
-
-	unsigned long		gsr[7];
-	unsigned long		xfsr[7];
-
-	__u64			__user *user_cntd0;
-	__u64			__user *user_cntd1;
-	__u64			kernel_cntd0, kernel_cntd1;
-	__u64			pcr_reg;
-
-	struct restart_block	restart_block;
-
-	struct pt_regs		*kern_una_regs;
-	unsigned int		kern_una_insn;
-
-	unsigned long		fpregs[0] __attribute__ ((aligned(64)));
-};
-
-#endif /* !(__ASSEMBLY__) */
-
-/* offsets into the thread_info struct for assembly code access */
-#define TI_TASK		0x00000000
-#define TI_FLAGS	0x00000008
-#define TI_FAULT_CODE	(TI_FLAGS + TI_FLAG_BYTE_FAULT_CODE)
-#define TI_WSTATE	(TI_FLAGS + TI_FLAG_BYTE_WSTATE)
-#define TI_CWP		(TI_FLAGS + TI_FLAG_BYTE_CWP)
-#define TI_CURRENT_DS	(TI_FLAGS + TI_FLAG_BYTE_CURRENT_DS)
-#define TI_FPDEPTH	(TI_FLAGS + TI_FLAG_BYTE_FPDEPTH)
-#define TI_WSAVED	(TI_FLAGS + TI_FLAG_BYTE_WSAVED)
-#define TI_FPSAVED	0x00000010
-#define TI_KSP		0x00000018
-#define TI_FAULT_ADDR	0x00000020
-#define TI_KREGS	0x00000028
-#define TI_EXEC_DOMAIN	0x00000030
-#define TI_PRE_COUNT	0x00000038
-#define TI_NEW_CHILD	0x0000003c
-#define TI_SYS_NOERROR	0x0000003d
-#define TI_CPU		0x0000003e
-#define TI_UTRAPS	0x00000040
-#define TI_REG_WINDOW	0x00000048
-#define TI_RWIN_SPTRS	0x000003c8	
-#define TI_GSR		0x00000400
-#define TI_XFSR		0x00000438
-#define TI_USER_CNTD0	0x00000470
-#define TI_USER_CNTD1	0x00000478
-#define TI_KERN_CNTD0	0x00000480
-#define TI_KERN_CNTD1	0x00000488
-#define TI_PCR		0x00000490
-#define TI_RESTART_BLOCK 0x00000498
-#define TI_KUNA_REGS	0x000004c0
-#define TI_KUNA_INSN	0x000004c8
-#define TI_FPREGS	0x00000500
-
-/* We embed this in the uppermost byte of thread_info->flags */
-#define FAULT_CODE_WRITE	0x01	/* Write access, implies D-TLB	   */
-#define FAULT_CODE_DTLB		0x02	/* Miss happened in D-TLB	   */
-#define FAULT_CODE_ITLB		0x04	/* Miss happened in I-TLB	   */
-#define FAULT_CODE_WINFIXUP	0x08	/* Miss happened during spill/fill */
-#define FAULT_CODE_BLKCOMMIT	0x10	/* Use blk-commit ASI in copy_page */
-
-#if PAGE_SHIFT == 13
-#define THREAD_SIZE (2*PAGE_SIZE)
-#define THREAD_SHIFT (PAGE_SHIFT + 1)
-#else /* PAGE_SHIFT == 13 */
-#define THREAD_SIZE PAGE_SIZE
-#define THREAD_SHIFT PAGE_SHIFT
-#endif /* PAGE_SHIFT == 13 */
-
-#define PREEMPT_ACTIVE		0x4000000
-
-/*
- * macros/functions for gaining access to the thread information structure
- *
- * preempt_count needs to be 1 initially, until the scheduler is functional.
- */
-#ifndef __ASSEMBLY__
-
-#define INIT_THREAD_INFO(tsk)				\
-{							\
-	.task		=	&tsk,			\
-	.flags		= ((unsigned long)ASI_P) << TI_FLAG_CURRENT_DS_SHIFT,	\
-	.exec_domain	=	&default_exec_domain,	\
-	.preempt_count	=	1,			\
-	.restart_block	= {				\
-		.fn	=	do_no_restart_syscall,	\
-	},						\
-}
-
-#define init_thread_info	(init_thread_union.thread_info)
-#define init_stack		(init_thread_union.stack)
-
-/* how to get the thread information struct from C */
-register struct thread_info *current_thread_info_reg asm("g6");
-#define current_thread_info()	(current_thread_info_reg)
-
-/* thread information allocation */
-#if PAGE_SHIFT == 13
-#define __THREAD_INFO_ORDER	1
-#else /* PAGE_SHIFT == 13 */
-#define __THREAD_INFO_ORDER	0
-#endif /* PAGE_SHIFT == 13 */
-
-#ifdef CONFIG_DEBUG_STACK_USAGE
-#define alloc_thread_info(tsk)					\
-({								\
-	struct thread_info *ret;				\
-								\
-	ret = (struct thread_info *)				\
-	  __get_free_pages(GFP_KERNEL, __THREAD_INFO_ORDER);	\
-	if (ret)						\
-		memset(ret, 0, PAGE_SIZE<<__THREAD_INFO_ORDER);	\
-	ret;							\
-})
-#else
-#define alloc_thread_info(tsk) \
-	((struct thread_info *)__get_free_pages(GFP_KERNEL, __THREAD_INFO_ORDER))
-#endif
-
-#define free_thread_info(ti) \
-	free_pages((unsigned long)(ti),__THREAD_INFO_ORDER)
-
-#define __thread_flag_byte_ptr(ti)	\
-	((unsigned char *)(&((ti)->flags)))
-#define __cur_thread_flag_byte_ptr	__thread_flag_byte_ptr(current_thread_info())
-
-#define get_thread_fault_code()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FAULT_CODE])
-#define set_thread_fault_code(val)	(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FAULT_CODE] = (val))
-#define get_thread_wstate()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSTATE])
-#define set_thread_wstate(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSTATE] = (val))
-#define get_thread_cwp()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CWP])
-#define set_thread_cwp(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CWP] = (val))
-#define get_thread_current_ds()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CURRENT_DS])
-#define set_thread_current_ds(val)	(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_CURRENT_DS] = (val))
-#define get_thread_fpdepth()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FPDEPTH])
-#define set_thread_fpdepth(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_FPDEPTH] = (val))
-#define get_thread_wsaved()		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSAVED])
-#define set_thread_wsaved(val)		(__cur_thread_flag_byte_ptr[TI_FLAG_BYTE_WSAVED] = (val))
-
-#endif /* !(__ASSEMBLY__) */
-
-/*
- * Thread information flags, only 16 bits are available as we encode
- * other values into the upper 6 bytes.
- *
- * On trap return we need to test several values:
- *
- * user:	need_resched, notify_resume, sigpending, wsaved, perfctr
- * kernel:	fpdepth
- *
- * So to check for work in the kernel case we simply load the fpdepth
- * byte out of the flags and test it.  For the user case we encode the
- * lower 3 bytes of flags as follows:
- *	----------------------------------------
- *	| wsaved | flags byte 1 | flags byte 2 |
- *	----------------------------------------
- * This optimizes the user test into:
- *	ldx		[%g6 + TI_FLAGS], REG1
- *	sethi		%hi(_TIF_USER_WORK_MASK), REG2
- *	or		REG2, %lo(_TIF_USER_WORK_MASK), REG2
- *	andcc		REG1, REG2, %g0
- *	be,pt		no_work_to_do
- *	 nop
- */
-#define TIF_SYSCALL_TRACE	0	/* syscall trace active */
-/* flags bit 1 is available */
-#define TIF_SIGPENDING		2	/* signal pending */
-#define TIF_NEED_RESCHED	3	/* rescheduling necessary */
-#define TIF_PERFCTR		4	/* performance counters active */
-#define TIF_UNALIGNED		5	/* allowed to do unaligned accesses */
-/* flag bit 6 is available */
-#define TIF_32BIT		7	/* 32-bit binary */
-/* flag bit 8 is available */
-#define TIF_SECCOMP		9	/* secure computing */
-#define TIF_SYSCALL_AUDIT	10	/* syscall auditing active */
-/* flag bit 11 is available */
-/* NOTE: Thread flags >= 12 should be ones we have no interest
- *       in using in assembly, else we can't use the mask as
- *       an immediate value in instructions such as andcc.
- */
-#define TIF_ABI_PENDING		12
-#define TIF_MEMDIE		13
-#define TIF_POLLING_NRFLAG	14
-
-#define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
-#define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
-#define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
-#define _TIF_PERFCTR		(1<<TIF_PERFCTR)
-#define _TIF_UNALIGNED		(1<<TIF_UNALIGNED)
-#define _TIF_32BIT		(1<<TIF_32BIT)
-#define _TIF_SECCOMP		(1<<TIF_SECCOMP)
-#define _TIF_SYSCALL_AUDIT	(1<<TIF_SYSCALL_AUDIT)
-#define _TIF_ABI_PENDING	(1<<TIF_ABI_PENDING)
-#define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
-
-#define _TIF_USER_WORK_MASK	((0xff << TI_FLAG_WSAVED_SHIFT) | \
-				 (_TIF_SIGPENDING | \
-				  _TIF_NEED_RESCHED | _TIF_PERFCTR))
-
-/*
- * Thread-synchronous status.
- *
- * This is different from the flags in that nobody else
- * ever touches our thread-synchronous status, so we don't
- * have to worry about atomic accesses.
- *
- * Note that there are only 8 bits available.
- */
-#define TS_RESTORE_SIGMASK	0x0001	/* restore signal mask in do_signal() */
-
-#ifndef __ASSEMBLY__
-#define HAVE_SET_RESTORE_SIGMASK	1
-static inline void set_restore_sigmask(void)
-{
-	struct thread_info *ti = current_thread_info();
-	ti->status |= TS_RESTORE_SIGMASK;
-	set_bit(TIF_SIGPENDING, &ti->flags);
-}
-#endif	/* !__ASSEMBLY__ */
-
-#endif /* __KERNEL__ */
-
-#endif /* _ASM_THREAD_INFO_H */
+#include <asm-sparc/thread_info.h>
diff --git a/include/asm-sparc64/timer.h b/include/asm-sparc64/timer.h
index 5b779fd1f78..88026d83cc9 100644
--- a/include/asm-sparc64/timer.h
+++ b/include/asm-sparc64/timer.h
@@ -1,30 +1 @@
-/* timer.h: System timer definitions for sun5.
- *
- * Copyright (C) 1997, 2008 David S. Miller (davem@davemloft.net)
- */
-
-#ifndef _SPARC64_TIMER_H
-#define _SPARC64_TIMER_H
-
-#include <linux/types.h>
-#include <linux/init.h>
-
-struct sparc64_tick_ops {
-	unsigned long (*get_tick)(void);
-	int (*add_compare)(unsigned long);
-	unsigned long softint_mask;
-	void (*disable_irq)(void);
-
-	void (*init_tick)(void);
-	unsigned long (*add_tick)(unsigned long);
-
-	char *name;
-};
-
-extern struct sparc64_tick_ops *tick_ops;
-
-extern unsigned long sparc64_get_clock_tick(unsigned int cpu);
-extern void __devinit setup_sparc64_timer(void);
-extern void __init time_init(void);
-
-#endif /* _SPARC64_TIMER_H */
+#include <asm-sparc/timer.h>
diff --git a/include/asm-sparc64/timex.h b/include/asm-sparc64/timex.h
index c622535c456..8dd59ee24b4 100644
--- a/include/asm-sparc64/timex.h
+++ b/include/asm-sparc64/timex.h
@@ -1,19 +1 @@
-/*
- * linux/include/asm-sparc64/timex.h
- *
- * sparc64 architecture timex specifications
- */
-#ifndef _ASMsparc64_TIMEX_H
-#define _ASMsparc64_TIMEX_H
-
-#include <asm/timer.h>
-
-#define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
-
-/* Getting on the cycle counter on sparc64. */
-typedef unsigned long cycles_t;
-#define get_cycles()	tick_ops->get_tick()
-
-#define ARCH_HAS_READ_CURRENT_TIMER
-
-#endif
+#include <asm-sparc/timex.h>
diff --git a/include/asm-sparc64/tlb.h b/include/asm-sparc64/tlb.h
index ec81cdedef2..ae92fce1093 100644
--- a/include/asm-sparc64/tlb.h
+++ b/include/asm-sparc64/tlb.h
@@ -1,111 +1 @@
-#ifndef _SPARC64_TLB_H
-#define _SPARC64_TLB_H
-
-#include <linux/swap.h>
-#include <linux/pagemap.h>
-#include <asm/pgalloc.h>
-#include <asm/tlbflush.h>
-#include <asm/mmu_context.h>
-
-#define TLB_BATCH_NR	192
-
-/*
- * For UP we don't need to worry about TLB flush
- * and page free order so much..
- */
-#ifdef CONFIG_SMP
-  #define FREE_PTE_NR	506
-  #define tlb_fast_mode(bp) ((bp)->pages_nr == ~0U)
-#else
-  #define FREE_PTE_NR	1
-  #define tlb_fast_mode(bp) 1
-#endif
-
-struct mmu_gather {
-	struct mm_struct *mm;
-	unsigned int pages_nr;
-	unsigned int need_flush;
-	unsigned int fullmm;
-	unsigned int tlb_nr;
-	unsigned long vaddrs[TLB_BATCH_NR];
-	struct page *pages[FREE_PTE_NR];
-};
-
-DECLARE_PER_CPU(struct mmu_gather, mmu_gathers);
-
-#ifdef CONFIG_SMP
-extern void smp_flush_tlb_pending(struct mm_struct *,
-				  unsigned long, unsigned long *);
-#endif
-
-extern void __flush_tlb_pending(unsigned long, unsigned long, unsigned long *);
-extern void flush_tlb_pending(void);
-
-static inline struct mmu_gather *tlb_gather_mmu(struct mm_struct *mm, unsigned int full_mm_flush)
-{
-	struct mmu_gather *mp = &get_cpu_var(mmu_gathers);
-
-	BUG_ON(mp->tlb_nr);
-
-	mp->mm = mm;
-	mp->pages_nr = num_online_cpus() > 1 ? 0U : ~0U;
-	mp->fullmm = full_mm_flush;
-
-	return mp;
-}
-
-
-static inline void tlb_flush_mmu(struct mmu_gather *mp)
-{
-	if (mp->need_flush) {
-		free_pages_and_swap_cache(mp->pages, mp->pages_nr);
-		mp->pages_nr = 0;
-		mp->need_flush = 0;
-	}
-
-}
-
-#ifdef CONFIG_SMP
-extern void smp_flush_tlb_mm(struct mm_struct *mm);
-#define do_flush_tlb_mm(mm) smp_flush_tlb_mm(mm)
-#else
-#define do_flush_tlb_mm(mm) __flush_tlb_mm(CTX_HWBITS(mm->context), SECONDARY_CONTEXT)
-#endif
-
-static inline void tlb_finish_mmu(struct mmu_gather *mp, unsigned long start, unsigned long end)
-{
-	tlb_flush_mmu(mp);
-
-	if (mp->fullmm)
-		mp->fullmm = 0;
-	else
-		flush_tlb_pending();
-
-	/* keep the page table cache within bounds */
-	check_pgt_cache();
-
-	put_cpu_var(mmu_gathers);
-}
-
-static inline void tlb_remove_page(struct mmu_gather *mp, struct page *page)
-{
-	if (tlb_fast_mode(mp)) {
-		free_page_and_swap_cache(page);
-		return;
-	}
-	mp->need_flush = 1;
-	mp->pages[mp->pages_nr++] = page;
-	if (mp->pages_nr >= FREE_PTE_NR)
-		tlb_flush_mmu(mp);
-}
-
-#define tlb_remove_tlb_entry(mp,ptep,addr) do { } while (0)
-#define pte_free_tlb(mp, ptepage) pte_free((mp)->mm, ptepage)
-#define pmd_free_tlb(mp, pmdp) pmd_free((mp)->mm, pmdp)
-#define pud_free_tlb(tlb,pudp) __pud_free_tlb(tlb,pudp)
-
-#define tlb_migrate_finish(mm)	do { } while (0)
-#define tlb_start_vma(tlb, vma) do { } while (0)
-#define tlb_end_vma(tlb, vma)	do { } while (0)
-
-#endif /* _SPARC64_TLB_H */
+#include <asm-sparc/tlb.h>
diff --git a/include/asm-sparc64/tlbflush.h b/include/asm-sparc64/tlbflush.h
index fbb675dbe0c..a43979a06cd 100644
--- a/include/asm-sparc64/tlbflush.h
+++ b/include/asm-sparc64/tlbflush.h
@@ -1,44 +1 @@
-#ifndef _SPARC64_TLBFLUSH_H
-#define _SPARC64_TLBFLUSH_H
-
-#include <linux/mm.h>
-#include <asm/mmu_context.h>
-
-/* TSB flush operations. */
-struct mmu_gather;
-extern void flush_tsb_kernel_range(unsigned long start, unsigned long end);
-extern void flush_tsb_user(struct mmu_gather *mp);
-
-/* TLB flush operations. */
-
-extern void flush_tlb_pending(void);
-
-#define flush_tlb_range(vma,start,end)	\
-	do { (void)(start); flush_tlb_pending(); } while (0)
-#define flush_tlb_page(vma,addr)	flush_tlb_pending()
-#define flush_tlb_mm(mm)		flush_tlb_pending()
-
-/* Local cpu only.  */
-extern void __flush_tlb_all(void);
-
-extern void __flush_tlb_kernel_range(unsigned long start, unsigned long end);
-
-#ifndef CONFIG_SMP
-
-#define flush_tlb_kernel_range(start,end) \
-do {	flush_tsb_kernel_range(start,end); \
-	__flush_tlb_kernel_range(start,end); \
-} while (0)
-
-#else /* CONFIG_SMP */
-
-extern void smp_flush_tlb_kernel_range(unsigned long start, unsigned long end);
-
-#define flush_tlb_kernel_range(start, end) \
-do {	flush_tsb_kernel_range(start,end); \
-	smp_flush_tlb_kernel_range(start, end); \
-} while (0)
-
-#endif /* ! CONFIG_SMP */
-
-#endif /* _SPARC64_TLBFLUSH_H */
+#include <asm-sparc/tlbflush.h>
diff --git a/include/asm-sparc64/topology.h b/include/asm-sparc64/topology.h
index 001c04027c8..46999b60fbb 100644
--- a/include/asm-sparc64/topology.h
+++ b/include/asm-sparc64/topology.h
@@ -1,86 +1 @@
-#ifndef _ASM_SPARC64_TOPOLOGY_H
-#define _ASM_SPARC64_TOPOLOGY_H
-
-#ifdef CONFIG_NUMA
-
-#include <asm/mmzone.h>
-
-static inline int cpu_to_node(int cpu)
-{
-	return numa_cpu_lookup_table[cpu];
-}
-
-#define parent_node(node)	(node)
-
-static inline cpumask_t node_to_cpumask(int node)
-{
-	return numa_cpumask_lookup_table[node];
-}
-
-/* Returns a pointer to the cpumask of CPUs on Node 'node'. */
-#define node_to_cpumask_ptr(v, node)		\
-		cpumask_t *v = &(numa_cpumask_lookup_table[node])
-
-#define node_to_cpumask_ptr_next(v, node)	\
-			   v = &(numa_cpumask_lookup_table[node])
-
-static inline int node_to_first_cpu(int node)
-{
-	cpumask_t tmp;
-	tmp = node_to_cpumask(node);
-	return first_cpu(tmp);
-}
-
-struct pci_bus;
-#ifdef CONFIG_PCI
-extern int pcibus_to_node(struct pci_bus *pbus);
-#else
-static inline int pcibus_to_node(struct pci_bus *pbus)
-{
-	return -1;
-}
-#endif
-
-#define pcibus_to_cpumask(bus)	\
-	(pcibus_to_node(bus) == -1 ? \
-	 CPU_MASK_ALL : \
-	 node_to_cpumask(pcibus_to_node(bus)))
-
-#define SD_NODE_INIT (struct sched_domain) {		\
-	.min_interval		= 8,			\
-	.max_interval		= 32,			\
-	.busy_factor		= 32,			\
-	.imbalance_pct		= 125,			\
-	.cache_nice_tries	= 2,			\
-	.busy_idx		= 3,			\
-	.idle_idx		= 2,			\
-	.newidle_idx		= 0, 			\
-	.wake_idx		= 1,			\
-	.forkexec_idx		= 1,			\
-	.flags			= SD_LOAD_BALANCE	\
-				| SD_BALANCE_FORK	\
-				| SD_BALANCE_EXEC	\
-				| SD_SERIALIZE		\
-				| SD_WAKE_BALANCE,	\
-	.last_balance		= jiffies,		\
-	.balance_interval	= 1,			\
-}
-
-#else /* CONFIG_NUMA */
-
-#include <asm-generic/topology.h>
-
-#endif /* !(CONFIG_NUMA) */
-
-#ifdef CONFIG_SMP
-#define topology_physical_package_id(cpu)	(cpu_data(cpu).proc_id)
-#define topology_core_id(cpu)			(cpu_data(cpu).core_id)
-#define topology_core_siblings(cpu)		(cpu_core_map[cpu])
-#define topology_thread_siblings(cpu)		(per_cpu(cpu_sibling_map, cpu))
-#define mc_capable()				(sparc64_multi_core)
-#define smt_capable()				(sparc64_multi_core)
-#endif /* CONFIG_SMP */
-
-#define cpu_coregroup_map(cpu)			(cpu_core_map[cpu])
-
-#endif /* _ASM_SPARC64_TOPOLOGY_H */
+#include <asm-sparc/topology.h>
diff --git a/include/asm-sparc64/tsb.h b/include/asm-sparc64/tsb.h
index 76e4299dd9b..3677a302ea3 100644
--- a/include/asm-sparc64/tsb.h
+++ b/include/asm-sparc64/tsb.h
@@ -1,283 +1 @@
-#ifndef _SPARC64_TSB_H
-#define _SPARC64_TSB_H
-
-/* The sparc64 TSB is similar to the powerpc hashtables.  It's a
- * power-of-2 sized table of TAG/PTE pairs.  The cpu precomputes
- * pointers into this table for 8K and 64K page sizes, and also a
- * comparison TAG based upon the virtual address and context which
- * faults.
- *
- * TLB miss trap handler software does the actual lookup via something
- * of the form:
- *
- * 	ldxa		[%g0] ASI_{D,I}MMU_TSB_8KB_PTR, %g1
- * 	ldxa		[%g0] ASI_{D,I}MMU, %g6
- *	sllx		%g6, 22, %g6
- *	srlx		%g6, 22, %g6
- * 	ldda		[%g1] ASI_NUCLEUS_QUAD_LDD, %g4
- * 	cmp		%g4, %g6
- * 	bne,pn	%xcc, tsb_miss_{d,i}tlb
- * 	 mov		FAULT_CODE_{D,I}TLB, %g3
- * 	stxa		%g5, [%g0] ASI_{D,I}TLB_DATA_IN
- * 	retry
- *
- *
- * Each 16-byte slot of the TSB is the 8-byte tag and then the 8-byte
- * PTE.  The TAG is of the same layout as the TLB TAG TARGET mmu
- * register which is:
- *
- * -------------------------------------------------
- * |  -  |  CONTEXT |  -  |    VADDR bits 63:22    |
- * -------------------------------------------------
- *  63 61 60      48 47 42 41                     0
- *
- * But actually, since we use per-mm TSB's, we zero out the CONTEXT
- * field.
- *
- * Like the powerpc hashtables we need to use locking in order to
- * synchronize while we update the entries.  PTE updates need locking
- * as well.
- *
- * We need to carefully choose a lock bits for the TSB entry.  We
- * choose to use bit 47 in the tag.  Also, since we never map anything
- * at page zero in context zero, we use zero as an invalid tag entry.
- * When the lock bit is set, this forces a tag comparison failure.
- */
-
-#define TSB_TAG_LOCK_BIT	47
-#define TSB_TAG_LOCK_HIGH	(1 << (TSB_TAG_LOCK_BIT - 32))
-
-#define TSB_TAG_INVALID_BIT	46
-#define TSB_TAG_INVALID_HIGH	(1 << (TSB_TAG_INVALID_BIT - 32))
-
-#define TSB_MEMBAR	membar	#StoreStore
-
-/* Some cpus support physical address quad loads.  We want to use
- * those if possible so we don't need to hard-lock the TSB mapping
- * into the TLB.  We encode some instruction patching in order to
- * support this.
- *
- * The kernel TSB is locked into the TLB by virtue of being in the
- * kernel image, so we don't play these games for swapper_tsb access.
- */
-#ifndef __ASSEMBLY__
-struct tsb_ldquad_phys_patch_entry {
-	unsigned int	addr;
-	unsigned int	sun4u_insn;
-	unsigned int	sun4v_insn;
-};
-extern struct tsb_ldquad_phys_patch_entry __tsb_ldquad_phys_patch,
-	__tsb_ldquad_phys_patch_end;
-
-struct tsb_phys_patch_entry {
-	unsigned int	addr;
-	unsigned int	insn;
-};
-extern struct tsb_phys_patch_entry __tsb_phys_patch, __tsb_phys_patch_end;
-#endif
-#define TSB_LOAD_QUAD(TSB, REG)	\
-661:	ldda		[TSB] ASI_NUCLEUS_QUAD_LDD, REG; \
-	.section	.tsb_ldquad_phys_patch, "ax"; \
-	.word		661b; \
-	ldda		[TSB] ASI_QUAD_LDD_PHYS, REG; \
-	ldda		[TSB] ASI_QUAD_LDD_PHYS_4V, REG; \
-	.previous
-
-#define TSB_LOAD_TAG_HIGH(TSB, REG) \
-661:	lduwa		[TSB] ASI_N, REG; \
-	.section	.tsb_phys_patch, "ax"; \
-	.word		661b; \
-	lduwa		[TSB] ASI_PHYS_USE_EC, REG; \
-	.previous
-
-#define TSB_LOAD_TAG(TSB, REG) \
-661:	ldxa		[TSB] ASI_N, REG; \
-	.section	.tsb_phys_patch, "ax"; \
-	.word		661b; \
-	ldxa		[TSB] ASI_PHYS_USE_EC, REG; \
-	.previous
-
-#define TSB_CAS_TAG_HIGH(TSB, REG1, REG2) \
-661:	casa		[TSB] ASI_N, REG1, REG2; \
-	.section	.tsb_phys_patch, "ax"; \
-	.word		661b; \
-	casa		[TSB] ASI_PHYS_USE_EC, REG1, REG2; \
-	.previous
-
-#define TSB_CAS_TAG(TSB, REG1, REG2) \
-661:	casxa		[TSB] ASI_N, REG1, REG2; \
-	.section	.tsb_phys_patch, "ax"; \
-	.word		661b; \
-	casxa		[TSB] ASI_PHYS_USE_EC, REG1, REG2; \
-	.previous
-
-#define TSB_STORE(ADDR, VAL) \
-661:	stxa		VAL, [ADDR] ASI_N; \
-	.section	.tsb_phys_patch, "ax"; \
-	.word		661b; \
-	stxa		VAL, [ADDR] ASI_PHYS_USE_EC; \
-	.previous
-
-#define TSB_LOCK_TAG(TSB, REG1, REG2)	\
-99:	TSB_LOAD_TAG_HIGH(TSB, REG1);	\
-	sethi	%hi(TSB_TAG_LOCK_HIGH), REG2;\
-	andcc	REG1, REG2, %g0;	\
-	bne,pn	%icc, 99b;		\
-	 nop;				\
-	TSB_CAS_TAG_HIGH(TSB, REG1, REG2);	\
-	cmp	REG1, REG2;		\
-	bne,pn	%icc, 99b;		\
-	 nop;				\
-	TSB_MEMBAR
-
-#define TSB_WRITE(TSB, TTE, TAG) \
-	add	TSB, 0x8, TSB;   \
-	TSB_STORE(TSB, TTE);     \
-	sub	TSB, 0x8, TSB;   \
-	TSB_MEMBAR;              \
-	TSB_STORE(TSB, TAG);
-
-#define KTSB_LOAD_QUAD(TSB, REG) \
-	ldda		[TSB] ASI_NUCLEUS_QUAD_LDD, REG;
-
-#define KTSB_STORE(ADDR, VAL) \
-	stxa		VAL, [ADDR] ASI_N;
-
-#define KTSB_LOCK_TAG(TSB, REG1, REG2)	\
-99:	lduwa	[TSB] ASI_N, REG1;	\
-	sethi	%hi(TSB_TAG_LOCK_HIGH), REG2;\
-	andcc	REG1, REG2, %g0;	\
-	bne,pn	%icc, 99b;		\
-	 nop;				\
-	casa	[TSB] ASI_N, REG1, REG2;\
-	cmp	REG1, REG2;		\
-	bne,pn	%icc, 99b;		\
-	 nop;				\
-	TSB_MEMBAR
-
-#define KTSB_WRITE(TSB, TTE, TAG) \
-	add	TSB, 0x8, TSB;   \
-	stxa	TTE, [TSB] ASI_N;     \
-	sub	TSB, 0x8, TSB;   \
-	TSB_MEMBAR;              \
-	stxa	TAG, [TSB] ASI_N;
-
-	/* Do a kernel page table walk.  Leaves physical PTE pointer in
-	 * REG1.  Jumps to FAIL_LABEL on early page table walk termination.
-	 * VADDR will not be clobbered, but REG2 will.
-	 */
-#define KERN_PGTABLE_WALK(VADDR, REG1, REG2, FAIL_LABEL)	\
-	sethi		%hi(swapper_pg_dir), REG1; \
-	or		REG1, %lo(swapper_pg_dir), REG1; \
-	sllx		VADDR, 64 - (PGDIR_SHIFT + PGDIR_BITS), REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	andn		REG2, 0x3, REG2; \
-	lduw		[REG1 + REG2], REG1; \
-	brz,pn		REG1, FAIL_LABEL; \
-	 sllx		VADDR, 64 - (PMD_SHIFT + PMD_BITS), REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	sllx		REG1, 11, REG1; \
-	andn		REG2, 0x3, REG2; \
-	lduwa		[REG1 + REG2] ASI_PHYS_USE_EC, REG1; \
-	brz,pn		REG1, FAIL_LABEL; \
-	 sllx		VADDR, 64 - PMD_SHIFT, REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	sllx		REG1, 11, REG1; \
-	andn		REG2, 0x7, REG2; \
-	add		REG1, REG2, REG1;
-
-	/* Do a user page table walk in MMU globals.  Leaves physical PTE
-	 * pointer in REG1.  Jumps to FAIL_LABEL on early page table walk
-	 * termination.  Physical base of page tables is in PHYS_PGD which
-	 * will not be modified.
-	 *
-	 * VADDR will not be clobbered, but REG1 and REG2 will.
-	 */
-#define USER_PGTABLE_WALK_TL1(VADDR, PHYS_PGD, REG1, REG2, FAIL_LABEL)	\
-	sllx		VADDR, 64 - (PGDIR_SHIFT + PGDIR_BITS), REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	andn		REG2, 0x3, REG2; \
-	lduwa		[PHYS_PGD + REG2] ASI_PHYS_USE_EC, REG1; \
-	brz,pn		REG1, FAIL_LABEL; \
-	 sllx		VADDR, 64 - (PMD_SHIFT + PMD_BITS), REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	sllx		REG1, 11, REG1; \
-	andn		REG2, 0x3, REG2; \
-	lduwa		[REG1 + REG2] ASI_PHYS_USE_EC, REG1; \
-	brz,pn		REG1, FAIL_LABEL; \
-	 sllx		VADDR, 64 - PMD_SHIFT, REG2; \
-	srlx		REG2, 64 - PAGE_SHIFT, REG2; \
-	sllx		REG1, 11, REG1; \
-	andn		REG2, 0x7, REG2; \
-	add		REG1, REG2, REG1;
-
-/* Lookup a OBP mapping on VADDR in the prom_trans[] table at TL>0.
- * If no entry is found, FAIL_LABEL will be branched to.  On success
- * the resulting PTE value will be left in REG1.  VADDR is preserved
- * by this routine.
- */
-#define OBP_TRANS_LOOKUP(VADDR, REG1, REG2, REG3, FAIL_LABEL) \
-	sethi		%hi(prom_trans), REG1; \
-	or		REG1, %lo(prom_trans), REG1; \
-97:	ldx		[REG1 + 0x00], REG2; \
-	brz,pn		REG2, FAIL_LABEL; \
-	 nop; \
-	ldx		[REG1 + 0x08], REG3; \
-	add		REG2, REG3, REG3; \
-	cmp		REG2, VADDR; \
-	bgu,pt		%xcc, 98f; \
-	 cmp		VADDR, REG3; \
-	bgeu,pt		%xcc, 98f; \
-	 ldx		[REG1 + 0x10], REG3; \
-	sub		VADDR, REG2, REG2; \
-	ba,pt		%xcc, 99f; \
-	 add		REG3, REG2, REG1; \
-98:	ba,pt		%xcc, 97b; \
-	 add		REG1, (3 * 8), REG1; \
-99:
-
-	/* We use a 32K TSB for the whole kernel, this allows to
-	 * handle about 16MB of modules and vmalloc mappings without
-	 * incurring many hash conflicts.
-	 */
-#define KERNEL_TSB_SIZE_BYTES	(32 * 1024)
-#define KERNEL_TSB_NENTRIES	\
-	(KERNEL_TSB_SIZE_BYTES / 16)
-#define KERNEL_TSB4M_NENTRIES	4096
-
-	/* Do a kernel TSB lookup at tl>0 on VADDR+TAG, branch to OK_LABEL
-	 * on TSB hit.  REG1, REG2, REG3, and REG4 are used as temporaries
-	 * and the found TTE will be left in REG1.  REG3 and REG4 must
-	 * be an even/odd pair of registers.
-	 *
-	 * VADDR and TAG will be preserved and not clobbered by this macro.
-	 */
-#define KERN_TSB_LOOKUP_TL1(VADDR, TAG, REG1, REG2, REG3, REG4, OK_LABEL) \
-	sethi		%hi(swapper_tsb), REG1; \
-	or		REG1, %lo(swapper_tsb), REG1; \
-	srlx		VADDR, PAGE_SHIFT, REG2; \
-	and		REG2, (KERNEL_TSB_NENTRIES - 1), REG2; \
-	sllx		REG2, 4, REG2; \
-	add		REG1, REG2, REG2; \
-	KTSB_LOAD_QUAD(REG2, REG3); \
-	cmp		REG3, TAG; \
-	be,a,pt		%xcc, OK_LABEL; \
-	 mov		REG4, REG1;
-
-#ifndef CONFIG_DEBUG_PAGEALLOC
-	/* This version uses a trick, the TAG is already (VADDR >> 22) so
-	 * we can make use of that for the index computation.
-	 */
-#define KERN_TSB4M_LOOKUP_TL1(TAG, REG1, REG2, REG3, REG4, OK_LABEL) \
-	sethi		%hi(swapper_4m_tsb), REG1; \
-	or		REG1, %lo(swapper_4m_tsb), REG1; \
-	and		TAG, (KERNEL_TSB4M_NENTRIES - 1), REG2; \
-	sllx		REG2, 4, REG2; \
-	add		REG1, REG2, REG2; \
-	KTSB_LOAD_QUAD(REG2, REG3); \
-	cmp		REG3, TAG; \
-	be,a,pt		%xcc, OK_LABEL; \
-	 mov		REG4, REG1;
-#endif
-
-#endif /* !(_SPARC64_TSB_H) */
+#include <asm-sparc/tsb.h>
diff --git a/include/asm-sparc64/ttable.h b/include/asm-sparc64/ttable.h
index 0ba199587e0..a550f1bf6f9 100644
--- a/include/asm-sparc64/ttable.h
+++ b/include/asm-sparc64/ttable.h
@@ -1,658 +1 @@
-#ifndef _SPARC64_TTABLE_H
-#define _SPARC64_TTABLE_H
-
-#include <asm/utrap.h>
-
-#ifdef __ASSEMBLY__
-#include <asm/thread_info.h>
-#endif
-
-#define BOOT_KERNEL b sparc64_boot; nop; nop; nop; nop; nop; nop; nop;
-
-/* We need a "cleaned" instruction... */
-#define CLEAN_WINDOW							\
-	rdpr	%cleanwin, %l0;		add	%l0, 1, %l0;		\
-	wrpr	%l0, 0x0, %cleanwin;					\
-	clr	%o0;	clr	%o1;	clr	%o2;	clr	%o3;	\
-	clr	%o4;	clr	%o5;	clr	%o6;	clr	%o7;	\
-	clr	%l0;	clr	%l1;	clr	%l2;	clr	%l3;	\
-	clr	%l4;	clr	%l5;	clr	%l6;	clr	%l7;	\
-	retry;								\
-	nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;
-
-#define TRAP(routine)					\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etrap;				\
-109:	 or	%g7, %lo(109b), %g7;			\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o0;			\
-	ba,pt	%xcc, rtrap;				\
-	 nop;						\
-	nop;
-
-#define TRAP_7INSNS(routine)				\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etrap;				\
-109:	 or	%g7, %lo(109b), %g7;			\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o0;			\
-	ba,pt	%xcc, rtrap;				\
-	 nop;
-
-#define TRAP_SAVEFPU(routine)				\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, do_fptrap;			\
-109:	 or	%g7, %lo(109b), %g7;			\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o0;			\
-	ba,pt	%xcc, rtrap;				\
-	 nop;						\
-	nop;
-
-#define TRAP_NOSAVE(routine)				\
-	ba,pt	%xcc, routine;				\
-	 nop;						\
-	nop; nop; nop; nop; nop; nop;
-	
-#define TRAP_NOSAVE_7INSNS(routine)			\
-	ba,pt	%xcc, routine;				\
-	 nop;						\
-	nop; nop; nop; nop; nop;
-	
-#define TRAPTL1(routine)				\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etraptl1;				\
-109:	 or	%g7, %lo(109b), %g7;			\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o0;			\
-	ba,pt	%xcc, rtrap;				\
-	 nop;						\
-	nop;
-	
-#define TRAP_ARG(routine, arg)				\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etrap;				\
-109:	 or	%g7, %lo(109b), %g7;			\
-	add	%sp, PTREGS_OFF, %o0;			\
-	call	routine;				\
-	 mov	arg, %o1;				\
-	ba,pt	%xcc, rtrap;				\
-	 nop;
-	
-#define TRAPTL1_ARG(routine, arg)			\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etraptl1;				\
-109:	 or	%g7, %lo(109b), %g7;			\
-	add	%sp, PTREGS_OFF, %o0;			\
-	call	routine;				\
-	 mov	arg, %o1;				\
-	ba,pt	%xcc, rtrap;				\
-	 nop;
-	
-#define SYSCALL_TRAP(routine, systbl)			\
-	rdpr	%pil, %g2;				\
-	mov	TSTATE_SYSCALL, %g3;			\
-	sethi	%hi(109f), %g7;				\
-	ba,pt	%xcc, etrap_syscall;			\
-109:	 or	%g7, %lo(109b), %g7;			\
-	sethi	%hi(systbl), %l7;			\
-	ba,pt	%xcc, routine;				\
-	 or	%l7, %lo(systbl), %l7;
-	
-#define TRAP_UTRAP(handler,lvl)				\
-	mov	handler, %g3;				\
-	ba,pt	%xcc, utrap_trap;			\
-	 mov	lvl, %g4;				\
-	nop;						\
-	nop;						\
-	nop;						\
-	nop;						\
-	nop;
-
-#ifdef CONFIG_COMPAT
-#define	LINUX_32BIT_SYSCALL_TRAP SYSCALL_TRAP(linux_sparc_syscall32, sys_call_table32)
-#else
-#define	LINUX_32BIT_SYSCALL_TRAP BTRAP(0x110)
-#endif
-#define LINUX_64BIT_SYSCALL_TRAP SYSCALL_TRAP(linux_sparc_syscall, sys_call_table64)
-#define GETCC_TRAP TRAP(getcc)
-#define SETCC_TRAP TRAP(setcc)
-#define BREAKPOINT_TRAP TRAP(breakpoint_trap)
-
-#ifdef CONFIG_TRACE_IRQFLAGS
-
-#define TRAP_IRQ(routine, level)			\
-	rdpr	%pil, %g2;				\
-	wrpr	%g0, 15, %pil;				\
-	sethi	%hi(1f-4), %g7;				\
-	ba,pt	%xcc, etrap_irq;			\
-	 or	%g7, %lo(1f-4), %g7;			\
-	nop;						\
-	nop;						\
-	nop;						\
-	.subsection	2;				\
-1:	call	trace_hardirqs_off;			\
-	 nop;						\
-	mov	level, %o0;				\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o1;			\
-	ba,a,pt	%xcc, rtrap_irq;			\
-	.previous;
-
-#else
-
-#define TRAP_IRQ(routine, level)			\
-	rdpr	%pil, %g2;				\
-	wrpr	%g0, 15, %pil;				\
-	ba,pt	%xcc, etrap_irq;			\
-	 rd	%pc, %g7;				\
-	mov	level, %o0;				\
-	call	routine;				\
-	 add	%sp, PTREGS_OFF, %o1;			\
-	ba,a,pt	%xcc, rtrap_irq;
-	
-#endif
-
-#define TRAP_IVEC TRAP_NOSAVE(do_ivec)
-
-#define BTRAP(lvl) TRAP_ARG(bad_trap, lvl)
-
-#define BTRAPTL1(lvl) TRAPTL1_ARG(bad_trap_tl1, lvl)
-
-#define FLUSH_WINDOW_TRAP						\
-	ba,pt	%xcc, etrap;						\
-	 rd	%pc, %g7;						\
-	flushw;								\
-	ldx	[%sp + PTREGS_OFF + PT_V9_TNPC], %l1;			\
-	add	%l1, 4, %l2;						\
-	stx	%l1, [%sp + PTREGS_OFF + PT_V9_TPC];			\
-	ba,pt	%xcc, rtrap;						\
-	 stx	%l2, [%sp + PTREGS_OFF + PT_V9_TNPC];
-	        
-#ifdef CONFIG_KPROBES
-#define KPROBES_TRAP(lvl) TRAP_IRQ(kprobe_trap, lvl)
-#else
-#define KPROBES_TRAP(lvl) TRAP_ARG(bad_trap, lvl)
-#endif
-
-#ifdef CONFIG_KGDB
-#define KGDB_TRAP(lvl) TRAP_IRQ(kgdb_trap, lvl)
-#else
-#define KGDB_TRAP(lvl) TRAP_ARG(bad_trap, lvl)
-#endif
-
-#define SUN4V_ITSB_MISS					\
-	ldxa	[%g0] ASI_SCRATCHPAD, %g2;		\
-	ldx	[%g2 + HV_FAULT_I_ADDR_OFFSET], %g4;	\
-	ldx	[%g2 + HV_FAULT_I_CTX_OFFSET], %g5;	\
-	srlx	%g4, 22, %g6;				\
-	ba,pt	%xcc, sun4v_itsb_miss;			\
-	 nop;						\
-	nop;						\
-	nop;
-
-#define SUN4V_DTSB_MISS					\
-	ldxa	[%g0] ASI_SCRATCHPAD, %g2;		\
-	ldx	[%g2 + HV_FAULT_D_ADDR_OFFSET], %g4;	\
-	ldx	[%g2 + HV_FAULT_D_CTX_OFFSET], %g5;	\
-	srlx	%g4, 22, %g6;				\
-	ba,pt	%xcc, sun4v_dtsb_miss;			\
-	 nop;						\
-	nop;						\
-	nop;
-
-/* Before touching these macros, you owe it to yourself to go and
- * see how arch/sparc64/kernel/winfixup.S works... -DaveM
- *
- * For the user cases we used to use the %asi register, but
- * it turns out that the "wr xxx, %asi" costs ~5 cycles, so
- * now we use immediate ASI loads and stores instead.  Kudos
- * to Greg Onufer for pointing out this performance anomaly.
- *
- * Further note that we cannot use the g2, g4, g5, and g7 alternate
- * globals in the spill routines, check out the save instruction in
- * arch/sparc64/kernel/etrap.S to see what I mean about g2, and
- * g4/g5 are the globals which are preserved by etrap processing
- * for the caller of it.  The g7 register is the return pc for
- * etrap.  Finally, g6 is the current thread register so we cannot
- * us it in the spill handlers either.  Most of these rules do not
- * apply to fill processing, only g6 is not usable.
- */
-
-/* Normal kernel spill */
-#define SPILL_0_NORMAL					\
-	stx	%l0, [%sp + STACK_BIAS + 0x00];		\
-	stx	%l1, [%sp + STACK_BIAS + 0x08];		\
-	stx	%l2, [%sp + STACK_BIAS + 0x10];		\
-	stx	%l3, [%sp + STACK_BIAS + 0x18];		\
-	stx	%l4, [%sp + STACK_BIAS + 0x20];		\
-	stx	%l5, [%sp + STACK_BIAS + 0x28];		\
-	stx	%l6, [%sp + STACK_BIAS + 0x30];		\
-	stx	%l7, [%sp + STACK_BIAS + 0x38];		\
-	stx	%i0, [%sp + STACK_BIAS + 0x40];		\
-	stx	%i1, [%sp + STACK_BIAS + 0x48];		\
-	stx	%i2, [%sp + STACK_BIAS + 0x50];		\
-	stx	%i3, [%sp + STACK_BIAS + 0x58];		\
-	stx	%i4, [%sp + STACK_BIAS + 0x60];		\
-	stx	%i5, [%sp + STACK_BIAS + 0x68];		\
-	stx	%i6, [%sp + STACK_BIAS + 0x70];		\
-	stx	%i7, [%sp + STACK_BIAS + 0x78];		\
-	saved; retry; nop; nop; nop; nop; nop; nop;	\
-	nop; nop; nop; nop; nop; nop; nop; nop;
-
-#define SPILL_0_NORMAL_ETRAP				\
-etrap_kernel_spill:					\
-	stx	%l0, [%sp + STACK_BIAS + 0x00];		\
-	stx	%l1, [%sp + STACK_BIAS + 0x08];		\
-	stx	%l2, [%sp + STACK_BIAS + 0x10];		\
-	stx	%l3, [%sp + STACK_BIAS + 0x18];		\
-	stx	%l4, [%sp + STACK_BIAS + 0x20];		\
-	stx	%l5, [%sp + STACK_BIAS + 0x28];		\
-	stx	%l6, [%sp + STACK_BIAS + 0x30];		\
-	stx	%l7, [%sp + STACK_BIAS + 0x38];		\
-	stx	%i0, [%sp + STACK_BIAS + 0x40];		\
-	stx	%i1, [%sp + STACK_BIAS + 0x48];		\
-	stx	%i2, [%sp + STACK_BIAS + 0x50];		\
-	stx	%i3, [%sp + STACK_BIAS + 0x58];		\
-	stx	%i4, [%sp + STACK_BIAS + 0x60];		\
-	stx	%i5, [%sp + STACK_BIAS + 0x68];		\
-	stx	%i6, [%sp + STACK_BIAS + 0x70];		\
-	stx	%i7, [%sp + STACK_BIAS + 0x78];		\
-	saved;						\
-	sub	%g1, 2, %g1;				\
-	ba,pt	%xcc, etrap_save;			\
-	wrpr	%g1, %cwp;				\
-	nop; nop; nop; nop; nop; nop; nop; nop;		\
-	nop; nop; nop; nop;
-
-/* Normal 64bit spill */
-#define SPILL_1_GENERIC(ASI)				\
-	add	%sp, STACK_BIAS + 0x00, %g1;		\
-	stxa	%l0, [%g1 + %g0] ASI;			\
-	mov	0x08, %g3;				\
-	stxa	%l1, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%l2, [%g1 + %g0] ASI;			\
-	stxa	%l3, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%l4, [%g1 + %g0] ASI;			\
-	stxa	%l5, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%l6, [%g1 + %g0] ASI;			\
-	stxa	%l7, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%i0, [%g1 + %g0] ASI;			\
-	stxa	%i1, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%i2, [%g1 + %g0] ASI;			\
-	stxa	%i3, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%i4, [%g1 + %g0] ASI;			\
-	stxa	%i5, [%g1 + %g3] ASI;			\
-	add	%g1, 0x10, %g1;				\
-	stxa	%i6, [%g1 + %g0] ASI;			\
-	stxa	%i7, [%g1 + %g3] ASI;			\
-	saved;						\
-	retry; nop; nop;				\
-	b,a,pt	%xcc, spill_fixup_dax;			\
-	b,a,pt	%xcc, spill_fixup_mna;			\
-	b,a,pt	%xcc, spill_fixup;
-
-#define SPILL_1_GENERIC_ETRAP				\
-etrap_user_spill_64bit:					\
-	stxa	%l0, [%sp + STACK_BIAS + 0x00] %asi;	\
-	stxa	%l1, [%sp + STACK_BIAS + 0x08] %asi;	\
-	stxa	%l2, [%sp + STACK_BIAS + 0x10] %asi;	\
-	stxa	%l3, [%sp + STACK_BIAS + 0x18] %asi;	\
-	stxa	%l4, [%sp + STACK_BIAS + 0x20] %asi;	\
-	stxa	%l5, [%sp + STACK_BIAS + 0x28] %asi;	\
-	stxa	%l6, [%sp + STACK_BIAS + 0x30] %asi;	\
-	stxa	%l7, [%sp + STACK_BIAS + 0x38] %asi;	\
-	stxa	%i0, [%sp + STACK_BIAS + 0x40] %asi;	\
-	stxa	%i1, [%sp + STACK_BIAS + 0x48] %asi;	\
-	stxa	%i2, [%sp + STACK_BIAS + 0x50] %asi;	\
-	stxa	%i3, [%sp + STACK_BIAS + 0x58] %asi;	\
-	stxa	%i4, [%sp + STACK_BIAS + 0x60] %asi;	\
-	stxa	%i5, [%sp + STACK_BIAS + 0x68] %asi;	\
-	stxa	%i6, [%sp + STACK_BIAS + 0x70] %asi;	\
-	stxa	%i7, [%sp + STACK_BIAS + 0x78] %asi;	\
-	saved;						\
-	sub	%g1, 2, %g1;				\
-	ba,pt	%xcc, etrap_save;			\
-	 wrpr	%g1, %cwp;				\
-	nop; nop; nop; nop; nop;			\
-	nop; nop; nop; nop;				\
-	ba,a,pt	%xcc, etrap_spill_fixup_64bit;		\
-	ba,a,pt	%xcc, etrap_spill_fixup_64bit;		\
-	ba,a,pt	%xcc, etrap_spill_fixup_64bit;
-
-#define SPILL_1_GENERIC_ETRAP_FIXUP			\
-etrap_spill_fixup_64bit:				\
-	ldub	[%g6 + TI_WSAVED], %g1;			\
-	sll	%g1, 3, %g3;				\
-	add	%g6, %g3, %g3;				\
-	stx	%sp, [%g3 + TI_RWIN_SPTRS];		\
-	sll	%g1, 7, %g3;				\
-	add	%g6, %g3, %g3;				\
-	stx	%l0, [%g3 + TI_REG_WINDOW + 0x00];	\
-	stx	%l1, [%g3 + TI_REG_WINDOW + 0x08];	\
-	stx	%l2, [%g3 + TI_REG_WINDOW + 0x10];	\
-	stx	%l3, [%g3 + TI_REG_WINDOW + 0x18];	\
-	stx	%l4, [%g3 + TI_REG_WINDOW + 0x20];	\
-	stx	%l5, [%g3 + TI_REG_WINDOW + 0x28];	\
-	stx	%l6, [%g3 + TI_REG_WINDOW + 0x30];	\
-	stx	%l7, [%g3 + TI_REG_WINDOW + 0x38];	\
-	stx	%i0, [%g3 + TI_REG_WINDOW + 0x40];	\
-	stx	%i1, [%g3 + TI_REG_WINDOW + 0x48];	\
-	stx	%i2, [%g3 + TI_REG_WINDOW + 0x50];	\
-	stx	%i3, [%g3 + TI_REG_WINDOW + 0x58];	\
-	stx	%i4, [%g3 + TI_REG_WINDOW + 0x60];	\
-	stx	%i5, [%g3 + TI_REG_WINDOW + 0x68];	\
-	stx	%i6, [%g3 + TI_REG_WINDOW + 0x70];	\
-	stx	%i7, [%g3 + TI_REG_WINDOW + 0x78];	\
-	add	%g1, 1, %g1;				\
-	stb	%g1, [%g6 + TI_WSAVED];			\
-	saved;						\
-	rdpr	%cwp, %g1;				\
-	sub	%g1, 2, %g1;				\
-	ba,pt	%xcc, etrap_save;			\
-	 wrpr	%g1, %cwp;				\
-	nop; nop; nop
-
-/* Normal 32bit spill */
-#define SPILL_2_GENERIC(ASI)				\
-	srl	%sp, 0, %sp;				\
-	stwa	%l0, [%sp + %g0] ASI;			\
-	mov	0x04, %g3;				\
-	stwa	%l1, [%sp + %g3] ASI;			\
-	add	%sp, 0x08, %g1;				\
-	stwa	%l2, [%g1 + %g0] ASI;			\
-	stwa	%l3, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%l4, [%g1 + %g0] ASI;			\
-	stwa	%l5, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%l6, [%g1 + %g0] ASI;			\
-	stwa	%l7, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%i0, [%g1 + %g0] ASI;			\
-	stwa	%i1, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%i2, [%g1 + %g0] ASI;			\
-	stwa	%i3, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%i4, [%g1 + %g0] ASI;			\
-	stwa	%i5, [%g1 + %g3] ASI;			\
-	add	%g1, 0x08, %g1;				\
-	stwa	%i6, [%g1 + %g0] ASI;			\
-	stwa	%i7, [%g1 + %g3] ASI;			\
-	saved;						\
-        retry; nop; nop;				\
-	b,a,pt	%xcc, spill_fixup_dax;			\
-	b,a,pt	%xcc, spill_fixup_mna;			\
-	b,a,pt	%xcc, spill_fixup;
-
-#define SPILL_2_GENERIC_ETRAP		\
-etrap_user_spill_32bit:			\
-	srl	%sp, 0, %sp;		\
-	stwa	%l0, [%sp + 0x00] %asi;	\
-	stwa	%l1, [%sp + 0x04] %asi;	\
-	stwa	%l2, [%sp + 0x08] %asi;	\
-	stwa	%l3, [%sp + 0x0c] %asi;	\
-	stwa	%l4, [%sp + 0x10] %asi;	\
-	stwa	%l5, [%sp + 0x14] %asi;	\
-	stwa	%l6, [%sp + 0x18] %asi;	\
-	stwa	%l7, [%sp + 0x1c] %asi;	\
-	stwa	%i0, [%sp + 0x20] %asi;	\
-	stwa	%i1, [%sp + 0x24] %asi;	\
-	stwa	%i2, [%sp + 0x28] %asi;	\
-	stwa	%i3, [%sp + 0x2c] %asi;	\
-	stwa	%i4, [%sp + 0x30] %asi;	\
-	stwa	%i5, [%sp + 0x34] %asi;	\
-	stwa	%i6, [%sp + 0x38] %asi;	\
-	stwa	%i7, [%sp + 0x3c] %asi;	\
-	saved;				\
-	sub	%g1, 2, %g1;		\
-	ba,pt	%xcc, etrap_save;	\
-	 wrpr	%g1, %cwp;		\
-	nop; nop; nop; nop;		\
-	nop; nop; nop; nop;		\
-	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
-	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
-	ba,a,pt	%xcc, etrap_spill_fixup_32bit;
-
-#define SPILL_2_GENERIC_ETRAP_FIXUP			\
-etrap_spill_fixup_32bit:				\
-	ldub	[%g6 + TI_WSAVED], %g1;			\
-	sll	%g1, 3, %g3;				\
-	add	%g6, %g3, %g3;				\
-	stx	%sp, [%g3 + TI_RWIN_SPTRS];		\
-	sll	%g1, 7, %g3;				\
-	add	%g6, %g3, %g3;				\
-	stw	%l0, [%g3 + TI_REG_WINDOW + 0x00];	\
-	stw	%l1, [%g3 + TI_REG_WINDOW + 0x04];	\
-	stw	%l2, [%g3 + TI_REG_WINDOW + 0x08];	\
-	stw	%l3, [%g3 + TI_REG_WINDOW + 0x0c];	\
-	stw	%l4, [%g3 + TI_REG_WINDOW + 0x10];	\
-	stw	%l5, [%g3 + TI_REG_WINDOW + 0x14];	\
-	stw	%l6, [%g3 + TI_REG_WINDOW + 0x18];	\
-	stw	%l7, [%g3 + TI_REG_WINDOW + 0x1c];	\
-	stw	%i0, [%g3 + TI_REG_WINDOW + 0x20];	\
-	stw	%i1, [%g3 + TI_REG_WINDOW + 0x24];	\
-	stw	%i2, [%g3 + TI_REG_WINDOW + 0x28];	\
-	stw	%i3, [%g3 + TI_REG_WINDOW + 0x2c];	\
-	stw	%i4, [%g3 + TI_REG_WINDOW + 0x30];	\
-	stw	%i5, [%g3 + TI_REG_WINDOW + 0x34];	\
-	stw	%i6, [%g3 + TI_REG_WINDOW + 0x38];	\
-	stw	%i7, [%g3 + TI_REG_WINDOW + 0x3c];	\
-	add	%g1, 1, %g1;				\
-	stb	%g1, [%g6 + TI_WSAVED];			\
-	saved;						\
-	rdpr	%cwp, %g1;				\
-	sub	%g1, 2, %g1;				\
-	ba,pt	%xcc, etrap_save;			\
-	 wrpr	%g1, %cwp;				\
-	nop; nop; nop
-
-#define SPILL_1_NORMAL SPILL_1_GENERIC(ASI_AIUP)
-#define SPILL_2_NORMAL SPILL_2_GENERIC(ASI_AIUP)
-#define SPILL_3_NORMAL SPILL_0_NORMAL
-#define SPILL_4_NORMAL SPILL_0_NORMAL
-#define SPILL_5_NORMAL SPILL_0_NORMAL
-#define SPILL_6_NORMAL SPILL_0_NORMAL
-#define SPILL_7_NORMAL SPILL_0_NORMAL
-
-#define SPILL_0_OTHER SPILL_0_NORMAL
-#define SPILL_1_OTHER SPILL_1_GENERIC(ASI_AIUS)
-#define SPILL_2_OTHER SPILL_2_GENERIC(ASI_AIUS)
-#define SPILL_3_OTHER SPILL_3_NORMAL
-#define SPILL_4_OTHER SPILL_4_NORMAL
-#define SPILL_5_OTHER SPILL_5_NORMAL
-#define SPILL_6_OTHER SPILL_6_NORMAL
-#define SPILL_7_OTHER SPILL_7_NORMAL
-
-/* Normal kernel fill */
-#define FILL_0_NORMAL					\
-	ldx	[%sp + STACK_BIAS + 0x00], %l0;		\
-	ldx	[%sp + STACK_BIAS + 0x08], %l1;		\
-	ldx	[%sp + STACK_BIAS + 0x10], %l2;		\
-	ldx	[%sp + STACK_BIAS + 0x18], %l3;		\
-	ldx	[%sp + STACK_BIAS + 0x20], %l4;		\
-	ldx	[%sp + STACK_BIAS + 0x28], %l5;		\
-	ldx	[%sp + STACK_BIAS + 0x30], %l6;		\
-	ldx	[%sp + STACK_BIAS + 0x38], %l7;		\
-	ldx	[%sp + STACK_BIAS + 0x40], %i0;		\
-	ldx	[%sp + STACK_BIAS + 0x48], %i1;		\
-	ldx	[%sp + STACK_BIAS + 0x50], %i2;		\
-	ldx	[%sp + STACK_BIAS + 0x58], %i3;		\
-	ldx	[%sp + STACK_BIAS + 0x60], %i4;		\
-	ldx	[%sp + STACK_BIAS + 0x68], %i5;		\
-	ldx	[%sp + STACK_BIAS + 0x70], %i6;		\
-	ldx	[%sp + STACK_BIAS + 0x78], %i7;		\
-	restored; retry; nop; nop; nop; nop; nop; nop;	\
-	nop; nop; nop; nop; nop; nop; nop; nop;
-
-#define FILL_0_NORMAL_RTRAP				\
-kern_rtt_fill:						\
-	rdpr	%cwp, %g1;				\
-	sub	%g1, 1, %g1;				\
-	wrpr	%g1, %cwp;				\
-	ldx	[%sp + STACK_BIAS + 0x00], %l0;		\
-	ldx	[%sp + STACK_BIAS + 0x08], %l1;		\
-	ldx	[%sp + STACK_BIAS + 0x10], %l2;		\
-	ldx	[%sp + STACK_BIAS + 0x18], %l3;		\
-	ldx	[%sp + STACK_BIAS + 0x20], %l4;		\
-	ldx	[%sp + STACK_BIAS + 0x28], %l5;		\
-	ldx	[%sp + STACK_BIAS + 0x30], %l6;		\
-	ldx	[%sp + STACK_BIAS + 0x38], %l7;		\
-	ldx	[%sp + STACK_BIAS + 0x40], %i0;		\
-	ldx	[%sp + STACK_BIAS + 0x48], %i1;		\
-	ldx	[%sp + STACK_BIAS + 0x50], %i2;		\
-	ldx	[%sp + STACK_BIAS + 0x58], %i3;		\
-	ldx	[%sp + STACK_BIAS + 0x60], %i4;		\
-	ldx	[%sp + STACK_BIAS + 0x68], %i5;		\
-	ldx	[%sp + STACK_BIAS + 0x70], %i6;		\
-	ldx	[%sp + STACK_BIAS + 0x78], %i7;		\
-	restored;					\
-	add	%g1, 1, %g1;				\
-	ba,pt	%xcc, kern_rtt_restore;			\
-	 wrpr	%g1, %cwp;				\
-	nop; nop; nop; nop; nop;			\
-	nop; nop; nop; nop;
-
-
-/* Normal 64bit fill */
-#define FILL_1_GENERIC(ASI)				\
-	add	%sp, STACK_BIAS + 0x00, %g1;		\
-	ldxa	[%g1 + %g0] ASI, %l0;			\
-	mov	0x08, %g2;				\
-	mov	0x10, %g3;				\
-	ldxa	[%g1 + %g2] ASI, %l1;			\
-	mov	0x18, %g5;				\
-	ldxa	[%g1 + %g3] ASI, %l2;			\
-	ldxa	[%g1 + %g5] ASI, %l3;			\
-	add	%g1, 0x20, %g1;				\
-	ldxa	[%g1 + %g0] ASI, %l4;			\
-	ldxa	[%g1 + %g2] ASI, %l5;			\
-	ldxa	[%g1 + %g3] ASI, %l6;			\
-	ldxa	[%g1 + %g5] ASI, %l7;			\
-	add	%g1, 0x20, %g1;				\
-	ldxa	[%g1 + %g0] ASI, %i0;			\
-	ldxa	[%g1 + %g2] ASI, %i1;			\
-	ldxa	[%g1 + %g3] ASI, %i2;			\
-	ldxa	[%g1 + %g5] ASI, %i3;			\
-	add	%g1, 0x20, %g1;				\
-	ldxa	[%g1 + %g0] ASI, %i4;			\
-	ldxa	[%g1 + %g2] ASI, %i5;			\
-	ldxa	[%g1 + %g3] ASI, %i6;			\
-	ldxa	[%g1 + %g5] ASI, %i7;			\
-	restored;					\
-	retry; nop; nop; nop; nop;			\
-	b,a,pt	%xcc, fill_fixup_dax;			\
-	b,a,pt	%xcc, fill_fixup_mna;			\
-	b,a,pt	%xcc, fill_fixup;
-
-#define FILL_1_GENERIC_RTRAP				\
-user_rtt_fill_64bit:					\
-	ldxa	[%sp + STACK_BIAS + 0x00] %asi, %l0;	\
-	ldxa	[%sp + STACK_BIAS + 0x08] %asi, %l1;	\
-	ldxa	[%sp + STACK_BIAS + 0x10] %asi, %l2;	\
-	ldxa	[%sp + STACK_BIAS + 0x18] %asi, %l3;	\
-	ldxa	[%sp + STACK_BIAS + 0x20] %asi, %l4;	\
-	ldxa	[%sp + STACK_BIAS + 0x28] %asi, %l5;	\
-	ldxa	[%sp + STACK_BIAS + 0x30] %asi, %l6;	\
-	ldxa	[%sp + STACK_BIAS + 0x38] %asi, %l7;	\
-	ldxa	[%sp + STACK_BIAS + 0x40] %asi, %i0;	\
-	ldxa	[%sp + STACK_BIAS + 0x48] %asi, %i1;	\
-	ldxa	[%sp + STACK_BIAS + 0x50] %asi, %i2;	\
-	ldxa	[%sp + STACK_BIAS + 0x58] %asi, %i3;	\
-	ldxa	[%sp + STACK_BIAS + 0x60] %asi, %i4;	\
-	ldxa	[%sp + STACK_BIAS + 0x68] %asi, %i5;	\
-	ldxa	[%sp + STACK_BIAS + 0x70] %asi, %i6;	\
-	ldxa	[%sp + STACK_BIAS + 0x78] %asi, %i7;	\
-	ba,pt	%xcc, user_rtt_pre_restore;		\
-	 restored;					\
-	nop; nop; nop; nop; nop; nop;			\
-	nop; nop; nop; nop; nop;			\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;
-
-
-/* Normal 32bit fill */
-#define FILL_2_GENERIC(ASI)				\
-	srl	%sp, 0, %sp;				\
-	lduwa	[%sp + %g0] ASI, %l0;			\
-	mov	0x04, %g2;				\
-	mov	0x08, %g3;				\
-	lduwa	[%sp + %g2] ASI, %l1;			\
-	mov	0x0c, %g5;				\
-	lduwa	[%sp + %g3] ASI, %l2;			\
-	lduwa	[%sp + %g5] ASI, %l3;			\
-	add	%sp, 0x10, %g1;				\
-	lduwa	[%g1 + %g0] ASI, %l4;			\
-	lduwa	[%g1 + %g2] ASI, %l5;			\
-	lduwa	[%g1 + %g3] ASI, %l6;			\
-	lduwa	[%g1 + %g5] ASI, %l7;			\
-	add	%g1, 0x10, %g1;				\
-	lduwa	[%g1 + %g0] ASI, %i0;			\
-	lduwa	[%g1 + %g2] ASI, %i1;			\
-	lduwa	[%g1 + %g3] ASI, %i2;			\
-	lduwa	[%g1 + %g5] ASI, %i3;			\
-	add	%g1, 0x10, %g1;				\
-	lduwa	[%g1 + %g0] ASI, %i4;			\
-	lduwa	[%g1 + %g2] ASI, %i5;			\
-	lduwa	[%g1 + %g3] ASI, %i6;			\
-	lduwa	[%g1 + %g5] ASI, %i7;			\
-	restored;					\
-	retry; nop; nop; nop; nop;			\
-	b,a,pt	%xcc, fill_fixup_dax;			\
-	b,a,pt	%xcc, fill_fixup_mna;			\
-	b,a,pt	%xcc, fill_fixup;
-
-#define FILL_2_GENERIC_RTRAP				\
-user_rtt_fill_32bit:					\
-	srl	%sp, 0, %sp;				\
-	lduwa	[%sp + 0x00] %asi, %l0;			\
-	lduwa	[%sp + 0x04] %asi, %l1;			\
-	lduwa	[%sp + 0x08] %asi, %l2;			\
-	lduwa	[%sp + 0x0c] %asi, %l3;			\
-	lduwa	[%sp + 0x10] %asi, %l4;			\
-	lduwa	[%sp + 0x14] %asi, %l5;			\
-	lduwa	[%sp + 0x18] %asi, %l6;			\
-	lduwa	[%sp + 0x1c] %asi, %l7;			\
-	lduwa	[%sp + 0x20] %asi, %i0;			\
-	lduwa	[%sp + 0x24] %asi, %i1;			\
-	lduwa	[%sp + 0x28] %asi, %i2;			\
-	lduwa	[%sp + 0x2c] %asi, %i3;			\
-	lduwa	[%sp + 0x30] %asi, %i4;			\
-	lduwa	[%sp + 0x34] %asi, %i5;			\
-	lduwa	[%sp + 0x38] %asi, %i6;			\
-	lduwa	[%sp + 0x3c] %asi, %i7;			\
-	ba,pt	%xcc, user_rtt_pre_restore;		\
-	 restored;					\
-	nop; nop; nop; nop; nop;			\
-	nop; nop; nop; nop; nop;			\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
-	ba,a,pt	%xcc, user_rtt_fill_fixup;
-		
-
-#define FILL_1_NORMAL FILL_1_GENERIC(ASI_AIUP)
-#define FILL_2_NORMAL FILL_2_GENERIC(ASI_AIUP)
-#define FILL_3_NORMAL FILL_0_NORMAL
-#define FILL_4_NORMAL FILL_0_NORMAL
-#define FILL_5_NORMAL FILL_0_NORMAL
-#define FILL_6_NORMAL FILL_0_NORMAL
-#define FILL_7_NORMAL FILL_0_NORMAL
-
-#define FILL_0_OTHER FILL_0_NORMAL
-#define FILL_1_OTHER FILL_1_GENERIC(ASI_AIUS)
-#define FILL_2_OTHER FILL_2_GENERIC(ASI_AIUS)
-#define FILL_3_OTHER FILL_3_NORMAL
-#define FILL_4_OTHER FILL_4_NORMAL
-#define FILL_5_OTHER FILL_5_NORMAL
-#define FILL_6_OTHER FILL_6_NORMAL
-#define FILL_7_OTHER FILL_7_NORMAL
-
-#endif /* !(_SPARC64_TTABLE_H) */
+#include <asm-sparc/ttable.h>
diff --git a/include/asm-sparc64/types.h b/include/asm-sparc64/types.h
index b27ccc85202..cfbfad5043e 100644
--- a/include/asm-sparc64/types.h
+++ b/include/asm-sparc64/types.h
@@ -1,34 +1 @@
-#ifndef _SPARC64_TYPES_H
-#define _SPARC64_TYPES_H
-
-/*
- * This file is never included by application software unless
- * explicitly requested (e.g., via linux/types.h) in which case the
- * application is Linux specific so (user-) name space pollution is
- * not a major issue.  However, for interoperability, libraries still
- * need to be careful to avoid a name clashes.
- */
-#include <asm-generic/int-l64.h>
-
-#ifndef __ASSEMBLY__
-
-typedef unsigned short umode_t;
-
-#endif /* __ASSEMBLY__ */
-
-#ifdef __KERNEL__
-
-#define BITS_PER_LONG 64
-
-#ifndef __ASSEMBLY__
-
-/* Dma addresses come in generic and 64-bit flavours.  */
-
-typedef u32 dma_addr_t;
-typedef u64 dma64_addr_t;
-
-#endif /* __ASSEMBLY__ */
-
-#endif /* __KERNEL__ */
-
-#endif /* defined(_SPARC64_TYPES_H) */
+#include <asm-sparc/types.h>
diff --git a/include/asm-sparc64/uaccess.h b/include/asm-sparc64/uaccess.h
index 5fcbaf68c3f..2872d22844f 100644
--- a/include/asm-sparc64/uaccess.h
+++ b/include/asm-sparc64/uaccess.h
@@ -1,273 +1 @@
-#ifndef _ASM_UACCESS_H
-#define _ASM_UACCESS_H
-
-/*
- * User space memory access functions
- */
-
-#ifdef __KERNEL__
-#include <linux/compiler.h>
-#include <linux/sched.h>
-#include <linux/string.h>
-#include <asm/asi.h>
-#include <asm/system.h>
-#include <asm/spitfire.h>
-#include <asm-generic/uaccess.h>
-#endif
-
-#ifndef __ASSEMBLY__
-
-/*
- * Sparc64 is segmented, though more like the M68K than the I386. 
- * We use the secondary ASI to address user memory, which references a
- * completely different VM map, thus there is zero chance of the user
- * doing something queer and tricking us into poking kernel memory.
- *
- * What is left here is basically what is needed for the other parts of
- * the kernel that expect to be able to manipulate, erum, "segments".
- * Or perhaps more properly, permissions.
- *
- * "For historical reasons, these macros are grossly misnamed." -Linus
- */
-
-#define KERNEL_DS   ((mm_segment_t) { ASI_P })
-#define USER_DS     ((mm_segment_t) { ASI_AIUS })	/* har har har */
-
-#define VERIFY_READ	0
-#define VERIFY_WRITE	1
-
-#define get_fs() ((mm_segment_t) { get_thread_current_ds() })
-#define get_ds() (KERNEL_DS)
-
-#define segment_eq(a,b)  ((a).seg == (b).seg)
-
-#define set_fs(val)								\
-do {										\
-	set_thread_current_ds((val).seg);					\
-	__asm__ __volatile__ ("wr %%g0, %0, %%asi" : : "r" ((val).seg));	\
-} while(0)
-
-static inline int __access_ok(const void __user * addr, unsigned long size)
-{
-	return 1;
-}
-
-static inline int access_ok(int type, const void __user * addr, unsigned long size)
-{
-	return 1;
-}
-
-/*
- * The exception table consists of pairs of addresses: the first is the
- * address of an instruction that is allowed to fault, and the second is
- * the address at which the program should continue.  No registers are
- * modified, so it is entirely up to the continuation code to figure out
- * what to do.
- *
- * All the routines below use bits of fixup code that are out of line
- * with the main instruction path.  This means when everything is well,
- * we don't even have to jump over them.  Further, they do not intrude
- * on our cache or tlb entries.
- */
-
-struct exception_table_entry {
-        unsigned int insn, fixup;
-};
-
-extern void __ret_efault(void);
-extern void __retl_efault(void);
-
-/* Uh, these should become the main single-value transfer routines..
- * They automatically use the right size if we just have the right
- * pointer type..
- *
- * This gets kind of ugly. We want to return _two_ values in "get_user()"
- * and yet we don't want to do any pointers, because that is too much
- * of a performance impact. Thus we have a few rather ugly macros here,
- * and hide all the ugliness from the user.
- */
-#define put_user(x,ptr) ({ \
-unsigned long __pu_addr = (unsigned long)(ptr); \
-__chk_user_ptr(ptr); \
-__put_user_nocheck((__typeof__(*(ptr)))(x),__pu_addr,sizeof(*(ptr))); })
-
-#define get_user(x,ptr) ({ \
-unsigned long __gu_addr = (unsigned long)(ptr); \
-__chk_user_ptr(ptr); \
-__get_user_nocheck((x),__gu_addr,sizeof(*(ptr)),__typeof__(*(ptr))); })
-
-#define __put_user(x,ptr) put_user(x,ptr)
-#define __get_user(x,ptr) get_user(x,ptr)
-
-struct __large_struct { unsigned long buf[100]; };
-#define __m(x) ((struct __large_struct *)(x))
-
-#define __put_user_nocheck(data,addr,size) ({ \
-register int __pu_ret; \
-switch (size) { \
-case 1: __put_user_asm(data,b,addr,__pu_ret); break; \
-case 2: __put_user_asm(data,h,addr,__pu_ret); break; \
-case 4: __put_user_asm(data,w,addr,__pu_ret); break; \
-case 8: __put_user_asm(data,x,addr,__pu_ret); break; \
-default: __pu_ret = __put_user_bad(); break; \
-} __pu_ret; })
-
-#define __put_user_asm(x,size,addr,ret)					\
-__asm__ __volatile__(							\
-	"/* Put user asm, inline. */\n"					\
-"1:\t"	"st"#size "a %1, [%2] %%asi\n\t"				\
-	"clr	%0\n"							\
-"2:\n\n\t"								\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"sethi	%%hi(2b), %0\n\t"					\
-	"jmpl	%0 + %%lo(2b), %%g0\n\t"				\
-	" mov	%3, %0\n\n\t"						\
-	".previous\n\t"							\
-	".section __ex_table,\"a\"\n\t"					\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\t"						\
-	".previous\n\n\t"						\
-       : "=r" (ret) : "r" (x), "r" (__m(addr)),				\
-	 "i" (-EFAULT))
-
-extern int __put_user_bad(void);
-
-#define __get_user_nocheck(data,addr,size,type) ({ \
-register int __gu_ret; \
-register unsigned long __gu_val; \
-switch (size) { \
-case 1: __get_user_asm(__gu_val,ub,addr,__gu_ret); break; \
-case 2: __get_user_asm(__gu_val,uh,addr,__gu_ret); break; \
-case 4: __get_user_asm(__gu_val,uw,addr,__gu_ret); break; \
-case 8: __get_user_asm(__gu_val,x,addr,__gu_ret); break; \
-default: __gu_val = 0; __gu_ret = __get_user_bad(); break; \
-} data = (type) __gu_val; __gu_ret; })
-
-#define __get_user_nocheck_ret(data,addr,size,type,retval) ({ \
-register unsigned long __gu_val __asm__ ("l1"); \
-switch (size) { \
-case 1: __get_user_asm_ret(__gu_val,ub,addr,retval); break; \
-case 2: __get_user_asm_ret(__gu_val,uh,addr,retval); break; \
-case 4: __get_user_asm_ret(__gu_val,uw,addr,retval); break; \
-case 8: __get_user_asm_ret(__gu_val,x,addr,retval); break; \
-default: if (__get_user_bad()) return retval; \
-} data = (type) __gu_val; })
-
-#define __get_user_asm(x,size,addr,ret)					\
-__asm__ __volatile__(							\
-	"/* Get user asm, inline. */\n"					\
-"1:\t"	"ld"#size "a [%2] %%asi, %1\n\t"				\
-	"clr	%0\n"							\
-"2:\n\n\t"								\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"sethi	%%hi(2b), %0\n\t"					\
-	"clr	%1\n\t"							\
-	"jmpl	%0 + %%lo(2b), %%g0\n\t"				\
-	" mov	%3, %0\n\n\t"						\
-	".previous\n\t"							\
-	".section __ex_table,\"a\"\n\t"					\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\n\t"						\
-	".previous\n\t"							\
-       : "=r" (ret), "=r" (x) : "r" (__m(addr)),			\
-	 "i" (-EFAULT))
-
-#define __get_user_asm_ret(x,size,addr,retval)				\
-if (__builtin_constant_p(retval) && retval == -EFAULT)			\
-__asm__ __volatile__(							\
-	"/* Get user asm ret, inline. */\n"				\
-"1:\t"	"ld"#size "a [%1] %%asi, %0\n\n\t"				\
-	".section __ex_table,\"a\"\n\t"					\
-	".align	4\n\t"							\
-	".word	1b,__ret_efault\n\n\t"					\
-	".previous\n\t"							\
-       : "=r" (x) : "r" (__m(addr)));					\
-else									\
-__asm__ __volatile__(							\
-	"/* Get user asm ret, inline. */\n"				\
-"1:\t"	"ld"#size "a [%1] %%asi, %0\n\n\t"				\
-	".section .fixup,#alloc,#execinstr\n\t"				\
-	".align	4\n"							\
-"3:\n\t"								\
-	"ret\n\t"							\
-	" restore %%g0, %2, %%o0\n\n\t"					\
-	".previous\n\t"							\
-	".section __ex_table,\"a\"\n\t"					\
-	".align	4\n\t"							\
-	".word	1b, 3b\n\n\t"						\
-	".previous\n\t"							\
-       : "=r" (x) : "r" (__m(addr)), "i" (retval))
-
-extern int __get_user_bad(void);
-
-extern unsigned long __must_check ___copy_from_user(void *to,
-						    const void __user *from,
-						    unsigned long size);
-extern unsigned long copy_from_user_fixup(void *to, const void __user *from,
-					  unsigned long size);
-static inline unsigned long __must_check
-copy_from_user(void *to, const void __user *from, unsigned long size)
-{
-	unsigned long ret = ___copy_from_user(to, from, size);
-
-	if (unlikely(ret))
-		ret = copy_from_user_fixup(to, from, size);
-	return ret;
-}
-#define __copy_from_user copy_from_user
-
-extern unsigned long __must_check ___copy_to_user(void __user *to,
-						  const void *from,
-						  unsigned long size);
-extern unsigned long copy_to_user_fixup(void __user *to, const void *from,
-					unsigned long size);
-static inline unsigned long __must_check
-copy_to_user(void __user *to, const void *from, unsigned long size)
-{
-	unsigned long ret = ___copy_to_user(to, from, size);
-
-	if (unlikely(ret))
-		ret = copy_to_user_fixup(to, from, size);
-	return ret;
-}
-#define __copy_to_user copy_to_user
-
-extern unsigned long __must_check ___copy_in_user(void __user *to,
-						  const void __user *from,
-						  unsigned long size);
-extern unsigned long copy_in_user_fixup(void __user *to, void __user *from,
-					unsigned long size);
-static inline unsigned long __must_check
-copy_in_user(void __user *to, void __user *from, unsigned long size)
-{
-	unsigned long ret = ___copy_in_user(to, from, size);
-
-	if (unlikely(ret))
-		ret = copy_in_user_fixup(to, from, size);
-	return ret;
-}
-#define __copy_in_user copy_in_user
-
-extern unsigned long __must_check __clear_user(void __user *, unsigned long);
-
-#define clear_user __clear_user
-
-extern long __must_check __strncpy_from_user(char *dest, const char __user *src, long count);
-
-#define strncpy_from_user __strncpy_from_user
-
-extern long __strlen_user(const char __user *);
-extern long __strnlen_user(const char __user *, long len);
-
-#define strlen_user __strlen_user
-#define strnlen_user __strnlen_user
-#define __copy_to_user_inatomic __copy_to_user
-#define __copy_from_user_inatomic __copy_from_user
-
-#endif  /* __ASSEMBLY__ */
-
-#endif /* _ASM_UACCESS_H */
+#include <asm-sparc/uaccess.h>
diff --git a/include/asm-sparc64/uctx.h b/include/asm-sparc64/uctx.h
index dc937c75ffd..9e1b5794b07 100644
--- a/include/asm-sparc64/uctx.h
+++ b/include/asm-sparc64/uctx.h
@@ -1,71 +1 @@
-/*
- * uctx.h: Sparc64 {set,get}context() register state layouts.
- *
- * Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#ifndef __SPARC64_UCTX_H
-#define __SPARC64_UCTX_H
-
-#define MC_TSTATE	0
-#define MC_PC		1
-#define MC_NPC		2
-#define MC_Y		3
-#define MC_G1		4
-#define MC_G2		5
-#define MC_G3		6
-#define MC_G4		7
-#define MC_G5		8
-#define MC_G6		9
-#define MC_G7		10
-#define MC_O0		11
-#define MC_O1		12
-#define MC_O2		13
-#define MC_O3		14
-#define MC_O4		15
-#define MC_O5		16
-#define MC_O6		17
-#define MC_O7		18
-#define MC_NGREG	19
-
-typedef unsigned long mc_greg_t;
-typedef mc_greg_t mc_gregset_t[MC_NGREG];
-
-#define MC_MAXFPQ	16
-struct mc_fq {
-	unsigned long	*mcfq_addr;
-	unsigned int	mcfq_insn;
-};
-
-struct mc_fpu {
-	union {
-		unsigned int	sregs[32];
-		unsigned long	dregs[32];
-		long double	qregs[16];
-	} mcfpu_fregs;
-	unsigned long	mcfpu_fsr;
-	unsigned long	mcfpu_fprs;
-	unsigned long	mcfpu_gsr;
-	struct mc_fq	*mcfpu_fq;
-	unsigned char	mcfpu_qcnt;
-	unsigned char	mcfpu_qentsz;
-	unsigned char	mcfpu_enab;
-};
-typedef struct mc_fpu mc_fpu_t;
-
-typedef struct {
-	mc_gregset_t	mc_gregs;
-	mc_greg_t	mc_fp;
-	mc_greg_t	mc_i7;
-	mc_fpu_t	mc_fpregs;
-} mcontext_t;
-
-struct ucontext {
-	struct ucontext		*uc_link;
-	unsigned long		uc_flags;
-	sigset_t		uc_sigmask;
-	mcontext_t		uc_mcontext;
-};
-typedef struct ucontext ucontext_t;
-
-#endif /* __SPARC64_UCTX_H */
+#include <asm-sparc/uctx.h>
diff --git a/include/asm-sparc64/unaligned.h b/include/asm-sparc64/unaligned.h
index edcebb09441..19fbf9508ac 100644
--- a/include/asm-sparc64/unaligned.h
+++ b/include/asm-sparc64/unaligned.h
@@ -1,10 +1 @@
-#ifndef _ASM_SPARC64_UNALIGNED_H
-#define _ASM_SPARC64_UNALIGNED_H
-
-#include <linux/unaligned/be_struct.h>
-#include <linux/unaligned/le_byteshift.h>
-#include <linux/unaligned/generic.h>
-#define get_unaligned	__get_unaligned_be
-#define put_unaligned	__put_unaligned_be
-
-#endif /* _ASM_SPARC64_UNALIGNED_H */
+#include <asm-sparc/unaligned.h>
diff --git a/include/asm-sparc64/unistd.h b/include/asm-sparc64/unistd.h
index 13be4453a1f..ad86e0b7a45 100644
--- a/include/asm-sparc64/unistd.h
+++ b/include/asm-sparc64/unistd.h
@@ -1,373 +1 @@
-#ifndef _SPARC64_UNISTD_H
-#define _SPARC64_UNISTD_H
-
-/*
- * System calls under the Sparc.
- *
- * Don't be scared by the ugly clobbers, it is the only way I can
- * think of right now to force the arguments into fixed registers
- * before the trap into the system call with gcc 'asm' statements.
- *
- * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
- *
- * SunOS compatibility based upon preliminary work which is:
- *
- * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu)
- */
-
-#define __NR_restart_syscall      0 /* Linux Specific				   */
-#define __NR_exit                 1 /* Common                                      */
-#define __NR_fork                 2 /* Common                                      */
-#define __NR_read                 3 /* Common                                      */
-#define __NR_write                4 /* Common                                      */
-#define __NR_open                 5 /* Common                                      */
-#define __NR_close                6 /* Common                                      */
-#define __NR_wait4                7 /* Common                                      */
-#define __NR_creat                8 /* Common                                      */
-#define __NR_link                 9 /* Common                                      */
-#define __NR_unlink              10 /* Common                                      */
-#define __NR_execv               11 /* SunOS Specific                              */
-#define __NR_chdir               12 /* Common                                      */
-#define __NR_chown		 13 /* Common					   */
-#define __NR_mknod               14 /* Common                                      */
-#define __NR_chmod               15 /* Common                                      */
-#define __NR_lchown              16 /* Common                                      */
-#define __NR_brk                 17 /* Common                                      */
-#define __NR_perfctr             18 /* Performance counter operations              */
-#define __NR_lseek               19 /* Common                                      */
-#define __NR_getpid              20 /* Common                                      */
-#define __NR_capget		 21 /* Linux Specific				   */
-#define __NR_capset		 22 /* Linux Specific				   */
-#define __NR_setuid              23 /* Implemented via setreuid in SunOS           */
-#define __NR_getuid              24 /* Common                                      */
-#define __NR_vmsplice	         25 /* ENOSYS under SunOS			   */
-#define __NR_ptrace              26 /* Common                                      */
-#define __NR_alarm               27 /* Implemented via setitimer in SunOS          */
-#define __NR_sigaltstack	 28 /* Common					   */
-#define __NR_pause               29 /* Is sigblock(0)->sigpause() in SunOS         */
-#define __NR_utime               30 /* Implemented via utimes() under SunOS        */
-/* #define __NR_lchown32         31    Linux sparc32 specific                      */
-/* #define __NR_fchown32         32    Linux sparc32 specific                      */
-#define __NR_access              33 /* Common                                      */
-#define __NR_nice                34 /* Implemented via get/setpriority() in SunOS  */
-/* #define __NR_chown32          35    Linux sparc32 specific                      */
-#define __NR_sync                36 /* Common                                      */
-#define __NR_kill                37 /* Common                                      */
-#define __NR_stat                38 /* Common                                      */
-#define __NR_sendfile		 39 /* Linux Specific				   */
-#define __NR_lstat               40 /* Common                                      */
-#define __NR_dup                 41 /* Common                                      */
-#define __NR_pipe                42 /* Common                                      */
-#define __NR_times               43 /* Implemented via getrusage() in SunOS        */
-/* #define __NR_getuid32         44    Linux sparc32 specific                      */
-#define __NR_umount2             45 /* Linux Specific                              */
-#define __NR_setgid              46 /* Implemented via setregid() in SunOS         */
-#define __NR_getgid              47 /* Common                                      */
-#define __NR_signal              48 /* Implemented via sigvec() in SunOS           */
-#define __NR_geteuid             49 /* SunOS calls getuid()                        */
-#define __NR_getegid             50 /* SunOS calls getgid()                        */
-#define __NR_acct                51 /* Common                                      */
-#define __NR_memory_ordering	 52 /* Linux Specific				   */
-/* #define __NR_getgid32         53    Linux sparc32 specific                      */
-#define __NR_ioctl               54 /* Common                                      */
-#define __NR_reboot              55 /* Common                                      */
-/* #define __NR_mmap2		 56    Linux sparc32 Specific                      */
-#define __NR_symlink             57 /* Common                                      */
-#define __NR_readlink            58 /* Common                                      */
-#define __NR_execve              59 /* Common                                      */
-#define __NR_umask               60 /* Common                                      */
-#define __NR_chroot              61 /* Common                                      */
-#define __NR_fstat               62 /* Common                                      */
-#define __NR_fstat64             63 /* Linux Specific                              */
-#define __NR_getpagesize         64 /* Common                                      */
-#define __NR_msync               65 /* Common in newer 1.3.x revs...               */
-#define __NR_vfork               66 /* Common                                      */
-#define __NR_pread64             67 /* Linux Specific                              */
-#define __NR_pwrite64            68 /* Linux Specific                              */
-/* #define __NR_geteuid32        69    Linux sparc32, sbrk under SunOS             */
-/* #define __NR_getegid32        70    Linux sparc32, sstk under SunOS             */
-#define __NR_mmap                71 /* Common                                      */
-/* #define __NR_setreuid32       72    Linux sparc32, vadvise under SunOS          */
-#define __NR_munmap              73 /* Common                                      */
-#define __NR_mprotect            74 /* Common                                      */
-#define __NR_madvise             75 /* Common                                      */
-#define __NR_vhangup             76 /* Common                                      */
-/* #define __NR_truncate64       77    Linux sparc32 Specific			   */
-#define __NR_mincore             78 /* Common                                      */
-#define __NR_getgroups           79 /* Common                                      */
-#define __NR_setgroups           80 /* Common                                      */
-#define __NR_getpgrp             81 /* Common                                      */
-/* #define __NR_setgroups32      82    Linux sparc32, setpgrp under SunOS          */
-#define __NR_setitimer           83 /* Common                                      */
-/* #define __NR_ftruncate64      84    Linux sparc32 Specific			   */
-#define __NR_swapon              85 /* Common                                      */
-#define __NR_getitimer           86 /* Common                                      */
-/* #define __NR_setuid32         87    Linux sparc32, gethostname under SunOS      */
-#define __NR_sethostname         88 /* Common                                      */
-/* #define __NR_setgid32         89    Linux sparc32, getdtablesize under SunOS    */
-#define __NR_dup2                90 /* Common                                      */
-/* #define __NR_setfsuid32       91    Linux sparc32, getdopt under SunOS          */
-#define __NR_fcntl               92 /* Common                                      */
-#define __NR_select              93 /* Common                                      */
-/* #define __NR_setfsgid32       94    Linux sparc32, setdopt under SunOS          */
-#define __NR_fsync               95 /* Common                                      */
-#define __NR_setpriority         96 /* Common                                      */
-#define __NR_socket              97 /* Common                                      */
-#define __NR_connect             98 /* Common                                      */
-#define __NR_accept              99 /* Common                                      */
-#define __NR_getpriority        100 /* Common                                      */
-#define __NR_rt_sigreturn       101 /* Linux Specific                              */
-#define __NR_rt_sigaction       102 /* Linux Specific                              */
-#define __NR_rt_sigprocmask     103 /* Linux Specific                              */
-#define __NR_rt_sigpending      104 /* Linux Specific                              */
-#define __NR_rt_sigtimedwait    105 /* Linux Specific                              */
-#define __NR_rt_sigqueueinfo    106 /* Linux Specific                              */
-#define __NR_rt_sigsuspend      107 /* Linux Specific                              */
-#define __NR_setresuid          108 /* Linux Specific, sigvec under SunOS	   */
-#define __NR_getresuid          109 /* Linux Specific, sigblock under SunOS	   */
-#define __NR_setresgid          110 /* Linux Specific, sigsetmask under SunOS	   */
-#define __NR_getresgid          111 /* Linux Specific, sigpause under SunOS	   */
-/* #define __NR_setregid32       75    Linux sparc32, sigstack under SunOS         */
-#define __NR_recvmsg            113 /* Common                                      */
-#define __NR_sendmsg            114 /* Common                                      */
-/* #define __NR_getgroups32     115    Linux sparc32, vtrace under SunOS           */
-#define __NR_gettimeofday       116 /* Common                                      */
-#define __NR_getrusage          117 /* Common                                      */
-#define __NR_getsockopt         118 /* Common                                      */
-#define __NR_getcwd		119 /* Linux Specific				   */
-#define __NR_readv              120 /* Common                                      */
-#define __NR_writev             121 /* Common                                      */
-#define __NR_settimeofday       122 /* Common                                      */
-#define __NR_fchown             123 /* Common                                      */
-#define __NR_fchmod             124 /* Common                                      */
-#define __NR_recvfrom           125 /* Common                                      */
-#define __NR_setreuid           126 /* Common                                      */
-#define __NR_setregid           127 /* Common                                      */
-#define __NR_rename             128 /* Common                                      */
-#define __NR_truncate           129 /* Common                                      */
-#define __NR_ftruncate          130 /* Common                                      */
-#define __NR_flock              131 /* Common                                      */
-#define __NR_lstat64		132 /* Linux Specific                              */
-#define __NR_sendto             133 /* Common                                      */
-#define __NR_shutdown           134 /* Common                                      */
-#define __NR_socketpair         135 /* Common                                      */
-#define __NR_mkdir              136 /* Common                                      */
-#define __NR_rmdir              137 /* Common                                      */
-#define __NR_utimes             138 /* SunOS Specific                              */
-#define __NR_stat64		139 /* Linux Specific			           */
-#define __NR_sendfile64         140 /* adjtime under SunOS                         */
-#define __NR_getpeername        141 /* Common                                      */
-#define __NR_futex              142 /* gethostid under SunOS                       */
-#define __NR_gettid             143 /* ENOSYS under SunOS                          */
-#define __NR_getrlimit		144 /* Common                                      */
-#define __NR_setrlimit          145 /* Common                                      */
-#define __NR_pivot_root		146 /* Linux Specific, killpg under SunOS          */
-#define __NR_prctl		147 /* ENOSYS under SunOS                          */
-#define __NR_pciconfig_read	148 /* ENOSYS under SunOS                          */
-#define __NR_pciconfig_write	149 /* ENOSYS under SunOS                          */
-#define __NR_getsockname        150 /* Common                                      */
-#define __NR_inotify_init       151 /* Linux specific                              */
-#define __NR_inotify_add_watch  152 /* Linux specific                              */
-#define __NR_poll               153 /* Common                                      */
-#define __NR_getdents64		154 /* Linux specific				   */
-/* #define __NR_fcntl64         155    Linux sparc32 Specific                      */
-#define __NR_inotify_rm_watch   156 /* Linux specific				   */
-#define __NR_statfs             157 /* Common                                      */
-#define __NR_fstatfs            158 /* Common                                      */
-#define __NR_umount             159 /* Common                                      */
-#define __NR_sched_set_affinity 160 /* Linux specific, async_daemon under SunOS    */
-#define __NR_sched_get_affinity 161 /* Linux specific, getfh under SunOS           */
-#define __NR_getdomainname      162 /* SunOS Specific                              */
-#define __NR_setdomainname      163 /* Common                                      */
-#define __NR_utrap_install	164 /* SYSV ABI/v9 required			   */
-#define __NR_quotactl           165 /* Common                                      */
-#define __NR_set_tid_address    166 /* Linux specific, exportfs under SunOS        */
-#define __NR_mount              167 /* Common                                      */
-#define __NR_ustat              168 /* Common                                      */
-#define __NR_setxattr           169 /* SunOS: semsys                               */
-#define __NR_lsetxattr          170 /* SunOS: msgsys                               */
-#define __NR_fsetxattr          171 /* SunOS: shmsys                               */
-#define __NR_getxattr           172 /* SunOS: auditsys                             */
-#define __NR_lgetxattr          173 /* SunOS: rfssys                               */
-#define __NR_getdents           174 /* Common                                      */
-#define __NR_setsid             175 /* Common                                      */
-#define __NR_fchdir             176 /* Common                                      */
-#define __NR_fgetxattr          177 /* SunOS: fchroot                              */
-#define __NR_listxattr          178 /* SunOS: vpixsys                              */
-#define __NR_llistxattr         179 /* SunOS: aioread                              */
-#define __NR_flistxattr         180 /* SunOS: aiowrite                             */
-#define __NR_removexattr        181 /* SunOS: aiowait                              */
-#define __NR_lremovexattr       182 /* SunOS: aiocancel                            */
-#define __NR_sigpending         183 /* Common                                      */
-#define __NR_query_module	184 /* Linux Specific				   */
-#define __NR_setpgid            185 /* Common                                      */
-#define __NR_fremovexattr       186 /* SunOS: pathconf                             */
-#define __NR_tkill              187 /* SunOS: fpathconf                            */
-#define __NR_exit_group		188 /* Linux specific, sysconf undef SunOS         */
-#define __NR_uname              189 /* Linux Specific                              */
-#define __NR_init_module        190 /* Linux Specific                              */
-#define __NR_personality        191 /* Linux Specific                              */
-#define __NR_remap_file_pages   192 /* Linux Specific                              */
-#define __NR_epoll_create       193 /* Linux Specific                              */
-#define __NR_epoll_ctl          194 /* Linux Specific                              */
-#define __NR_epoll_wait         195 /* Linux Specific                              */
-#define __NR_ioprio_set         196 /* Linux Specific                              */
-#define __NR_getppid            197 /* Linux Specific                              */
-#define __NR_sigaction          198 /* Linux Specific                              */
-#define __NR_sgetmask           199 /* Linux Specific                              */
-#define __NR_ssetmask           200 /* Linux Specific                              */
-#define __NR_sigsuspend         201 /* Linux Specific                              */
-#define __NR_oldlstat           202 /* Linux Specific                              */
-#define __NR_uselib             203 /* Linux Specific                              */
-#define __NR_readdir            204 /* Linux Specific                              */
-#define __NR_readahead          205 /* Linux Specific                              */
-#define __NR_socketcall         206 /* Linux Specific                              */
-#define __NR_syslog             207 /* Linux Specific                              */
-#define __NR_lookup_dcookie     208 /* Linux Specific                              */
-#define __NR_fadvise64          209 /* Linux Specific                              */
-#define __NR_fadvise64_64       210 /* Linux Specific                              */
-#define __NR_tgkill             211 /* Linux Specific                              */
-#define __NR_waitpid            212 /* Linux Specific                              */
-#define __NR_swapoff            213 /* Linux Specific                              */
-#define __NR_sysinfo            214 /* Linux Specific                              */
-#define __NR_ipc                215 /* Linux Specific                              */
-#define __NR_sigreturn          216 /* Linux Specific                              */
-#define __NR_clone              217 /* Linux Specific                              */
-#define __NR_ioprio_get         218 /* Linux Specific                              */
-#define __NR_adjtimex           219 /* Linux Specific                              */
-#define __NR_sigprocmask        220 /* Linux Specific                              */
-#define __NR_create_module      221 /* Linux Specific                              */
-#define __NR_delete_module      222 /* Linux Specific                              */
-#define __NR_get_kernel_syms    223 /* Linux Specific                              */
-#define __NR_getpgid            224 /* Linux Specific                              */
-#define __NR_bdflush            225 /* Linux Specific                              */
-#define __NR_sysfs              226 /* Linux Specific                              */
-#define __NR_afs_syscall        227 /* Linux Specific                              */
-#define __NR_setfsuid           228 /* Linux Specific                              */
-#define __NR_setfsgid           229 /* Linux Specific                              */
-#define __NR__newselect         230 /* Linux Specific                              */
-#ifdef __KERNEL__
-#define __NR_time		231 /* Linux sparc32                               */
-#endif
-#define __NR_splice             232 /* Linux Specific                              */
-#define __NR_stime              233 /* Linux Specific                              */
-#define __NR_statfs64           234 /* Linux Specific                              */
-#define __NR_fstatfs64          235 /* Linux Specific                              */
-#define __NR__llseek            236 /* Linux Specific                              */
-#define __NR_mlock              237
-#define __NR_munlock            238
-#define __NR_mlockall           239
-#define __NR_munlockall         240
-#define __NR_sched_setparam     241
-#define __NR_sched_getparam     242
-#define __NR_sched_setscheduler 243
-#define __NR_sched_getscheduler 244
-#define __NR_sched_yield        245
-#define __NR_sched_get_priority_max 246
-#define __NR_sched_get_priority_min 247
-#define __NR_sched_rr_get_interval  248
-#define __NR_nanosleep          249
-#define __NR_mremap             250
-#define __NR__sysctl            251
-#define __NR_getsid             252
-#define __NR_fdatasync          253
-#define __NR_nfsservctl         254
-#define __NR_sync_file_range	255
-#define __NR_clock_settime	256
-#define __NR_clock_gettime	257
-#define __NR_clock_getres	258
-#define __NR_clock_nanosleep	259
-#define __NR_sched_getaffinity	260
-#define __NR_sched_setaffinity	261
-#define __NR_timer_settime	262
-#define __NR_timer_gettime	263
-#define __NR_timer_getoverrun	264
-#define __NR_timer_delete	265
-#define __NR_timer_create	266
-/* #define __NR_vserver		267 Reserved for VSERVER */
-#define __NR_io_setup		268
-#define __NR_io_destroy		269
-#define __NR_io_submit		270
-#define __NR_io_cancel		271
-#define __NR_io_getevents	272
-#define __NR_mq_open		273
-#define __NR_mq_unlink		274
-#define __NR_mq_timedsend	275
-#define __NR_mq_timedreceive	276
-#define __NR_mq_notify		277
-#define __NR_mq_getsetattr	278
-#define __NR_waitid		279
-#define __NR_tee		280
-#define __NR_add_key		281
-#define __NR_request_key	282
-#define __NR_keyctl		283
-#define __NR_openat		284
-#define __NR_mkdirat		285
-#define __NR_mknodat		286
-#define __NR_fchownat		287
-#define __NR_futimesat		288
-#define __NR_fstatat64		289
-#define __NR_unlinkat		290
-#define __NR_renameat		291
-#define __NR_linkat		292
-#define __NR_symlinkat		293
-#define __NR_readlinkat		294
-#define __NR_fchmodat		295
-#define __NR_faccessat		296
-#define __NR_pselect6		297
-#define __NR_ppoll		298
-#define __NR_unshare		299
-#define __NR_set_robust_list	300
-#define __NR_get_robust_list	301
-#define __NR_migrate_pages	302
-#define __NR_mbind		303
-#define __NR_get_mempolicy	304
-#define __NR_set_mempolicy	305
-#define __NR_kexec_load		306
-#define __NR_move_pages		307
-#define __NR_getcpu		308
-#define __NR_epoll_pwait	309
-#define __NR_utimensat		310
-#define __NR_signalfd		311
-#define __NR_timerfd_create	312
-#define __NR_eventfd		313
-#define __NR_fallocate		314
-#define __NR_timerfd_settime	315
-#define __NR_timerfd_gettime	316
-
-#define NR_SYSCALLS		317
-
-#ifdef __KERNEL__
-#define __ARCH_WANT_IPC_PARSE_VERSION
-#define __ARCH_WANT_OLD_READDIR
-#define __ARCH_WANT_STAT64
-#define __ARCH_WANT_SYS_ALARM
-#define __ARCH_WANT_SYS_GETHOSTNAME
-#define __ARCH_WANT_SYS_PAUSE
-#define __ARCH_WANT_SYS_SGETMASK
-#define __ARCH_WANT_SYS_SIGNAL
-#define __ARCH_WANT_SYS_TIME
-#define __ARCH_WANT_COMPAT_SYS_TIME
-#define __ARCH_WANT_SYS_UTIME
-#define __ARCH_WANT_SYS_WAITPID
-#define __ARCH_WANT_SYS_SOCKETCALL
-#define __ARCH_WANT_SYS_FADVISE64
-#define __ARCH_WANT_SYS_GETPGRP
-#define __ARCH_WANT_SYS_LLSEEK
-#define __ARCH_WANT_SYS_NICE
-#define __ARCH_WANT_SYS_OLDUMOUNT
-#define __ARCH_WANT_SYS_SIGPENDING
-#define __ARCH_WANT_SYS_SIGPROCMASK
-#define __ARCH_WANT_SYS_RT_SIGSUSPEND
-#define __ARCH_WANT_COMPAT_SYS_RT_SIGSUSPEND
-
-/*
- * "Conditional" syscalls
- *
- * What we want is __attribute__((weak,alias("sys_ni_syscall"))),
- * but it doesn't work on all toolchains, so we just do it by hand
- */
-#define cond_syscall(x) asm(".weak\t" #x "\n\t.set\t" #x ",sys_ni_syscall")
-
-#endif /* __KERNEL__ */
-#endif /* _SPARC64_UNISTD_H */
+#include <asm-sparc/unistd.h>
diff --git a/include/asm-sparc64/upa.h b/include/asm-sparc64/upa.h
index 5b1633223f9..aab72930815 100644
--- a/include/asm-sparc64/upa.h
+++ b/include/asm-sparc64/upa.h
@@ -1,109 +1 @@
-#ifndef _SPARC64_UPA_H
-#define _SPARC64_UPA_H
-
-#include <asm/asi.h>
-
-/* UPA level registers and defines. */
-
-/* UPA Config Register */
-#define UPA_CONFIG_RESV		0xffffffffc0000000 /* Reserved.                    */
-#define UPA_CONFIG_PCON		0x000000003fc00000 /* Depth of various sys queues. */
-#define UPA_CONFIG_MID		0x00000000003e0000 /* Module ID.                   */
-#define UPA_CONFIG_PCAP		0x000000000001ffff /* Port Capabilities.           */
-
-/* UPA Port ID Register */
-#define UPA_PORTID_FNP		0xff00000000000000 /* Hardcoded to 0xfc on ultra.  */
-#define UPA_PORTID_RESV		0x00fffff800000000 /* Reserved.                    */
-#define UPA_PORTID_ECCVALID     0x0000000400000000 /* Zero if mod can generate ECC */
-#define UPA_PORTID_ONEREAD      0x0000000200000000 /* Set if mod generates P_RASB  */
-#define UPA_PORTID_PINTRDQ      0x0000000180000000 /* # outstanding P_INT_REQ's    */
-#define UPA_PORTID_PREQDQ       0x000000007e000000 /* slave-wr's to mod supported  */
-#define UPA_PORTID_PREQRD       0x0000000001e00000 /* # incoming P_REQ's supported */
-#define UPA_PORTID_UPACAP       0x00000000001f0000 /* UPA capabilities of mod      */
-#define UPA_PORTID_ID           0x000000000000ffff /* Module Identification bits  */
-
-/* UPA I/O space accessors */
-#if defined(__KERNEL__) && !defined(__ASSEMBLY__)
-static inline unsigned char _upa_readb(unsigned long addr)
-{
-	unsigned char ret;
-
-	__asm__ __volatile__("lduba\t[%1] %2, %0\t/* upa_readb */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline unsigned short _upa_readw(unsigned long addr)
-{
-	unsigned short ret;
-
-	__asm__ __volatile__("lduha\t[%1] %2, %0\t/* upa_readw */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline unsigned int _upa_readl(unsigned long addr)
-{
-	unsigned int ret;
-
-	__asm__ __volatile__("lduwa\t[%1] %2, %0\t/* upa_readl */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline unsigned long _upa_readq(unsigned long addr)
-{
-	unsigned long ret;
-
-	__asm__ __volatile__("ldxa\t[%1] %2, %0\t/* upa_readq */"
-			     : "=r" (ret)
-			     : "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-
-	return ret;
-}
-
-static inline void _upa_writeb(unsigned char b, unsigned long addr)
-{
-	__asm__ __volatile__("stba\t%0, [%1] %2\t/* upa_writeb */"
-			     : /* no outputs */
-			     : "r" (b), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _upa_writew(unsigned short w, unsigned long addr)
-{
-	__asm__ __volatile__("stha\t%0, [%1] %2\t/* upa_writew */"
-			     : /* no outputs */
-			     : "r" (w), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _upa_writel(unsigned int l, unsigned long addr)
-{
-	__asm__ __volatile__("stwa\t%0, [%1] %2\t/* upa_writel */"
-			     : /* no outputs */
-			     : "r" (l), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-static inline void _upa_writeq(unsigned long q, unsigned long addr)
-{
-	__asm__ __volatile__("stxa\t%0, [%1] %2\t/* upa_writeq */"
-			     : /* no outputs */
-			     : "r" (q), "r" (addr), "i" (ASI_PHYS_BYPASS_EC_E));
-}
-
-#define upa_readb(__addr)		(_upa_readb((unsigned long)(__addr)))
-#define upa_readw(__addr)		(_upa_readw((unsigned long)(__addr)))
-#define upa_readl(__addr)		(_upa_readl((unsigned long)(__addr)))
-#define upa_readq(__addr)		(_upa_readq((unsigned long)(__addr)))
-#define upa_writeb(__b, __addr)		(_upa_writeb((__b), (unsigned long)(__addr)))
-#define upa_writew(__w, __addr)		(_upa_writew((__w), (unsigned long)(__addr)))
-#define upa_writel(__l, __addr)		(_upa_writel((__l), (unsigned long)(__addr)))
-#define upa_writeq(__q, __addr)		(_upa_writeq((__q), (unsigned long)(__addr)))
-#endif /* __KERNEL__ && !__ASSEMBLY__ */
-
-#endif /* !(_SPARC64_UPA_H) */
+#include <asm-sparc/upa.h>
diff --git a/include/asm-sparc64/utrap.h b/include/asm-sparc64/utrap.h
index e49e5c46ad6..b030a41f189 100644
--- a/include/asm-sparc64/utrap.h
+++ b/include/asm-sparc64/utrap.h
@@ -1,51 +1 @@
-/*
- * include/asm-sparc64/utrap.h
- *
- * Copyright (C) 1997 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
- */
-
-#ifndef __ASM_SPARC64_UTRAP_H
-#define __ASM_SPARC64_UTRAP_H
-
-#define UT_INSTRUCTION_EXCEPTION		1
-#define UT_INSTRUCTION_ERROR			2
-#define UT_INSTRUCTION_PROTECTION		3
-#define UT_ILLTRAP_INSTRUCTION			4
-#define UT_ILLEGAL_INSTRUCTION			5
-#define UT_PRIVILEGED_OPCODE			6
-#define UT_FP_DISABLED				7
-#define UT_FP_EXCEPTION_IEEE_754		8
-#define UT_FP_EXCEPTION_OTHER			9
-#define UT_TAG_OVERVIEW				10
-#define UT_DIVISION_BY_ZERO			11
-#define UT_DATA_EXCEPTION			12
-#define UT_DATA_ERROR				13
-#define UT_DATA_PROTECTION			14
-#define UT_MEM_ADDRESS_NOT_ALIGNED		15
-#define UT_PRIVILEGED_ACTION			16
-#define UT_ASYNC_DATA_ERROR			17
-#define UT_TRAP_INSTRUCTION_16			18
-#define UT_TRAP_INSTRUCTION_17			19
-#define UT_TRAP_INSTRUCTION_18			20
-#define UT_TRAP_INSTRUCTION_19			21
-#define UT_TRAP_INSTRUCTION_20			22
-#define UT_TRAP_INSTRUCTION_21			23
-#define UT_TRAP_INSTRUCTION_22			24
-#define UT_TRAP_INSTRUCTION_23			25
-#define UT_TRAP_INSTRUCTION_24			26
-#define UT_TRAP_INSTRUCTION_25			27
-#define UT_TRAP_INSTRUCTION_26			28
-#define UT_TRAP_INSTRUCTION_27			29
-#define UT_TRAP_INSTRUCTION_28			30
-#define UT_TRAP_INSTRUCTION_29			31
-#define UT_TRAP_INSTRUCTION_30			32
-#define UT_TRAP_INSTRUCTION_31			33
-
-#define	UTH_NOCHANGE				(-1)
-
-#ifndef __ASSEMBLY__ 
-typedef int utrap_entry_t;
-typedef void *utrap_handler_t;
-#endif /* __ASSEMBLY__ */
-
-#endif /* !(__ASM_SPARC64_PROCESSOR_H) */
+#include <asm-sparc/utrap.h>
diff --git a/include/asm-sparc64/vga.h b/include/asm-sparc64/vga.h
index c69d5b2ba19..fbf4d58a56f 100644
--- a/include/asm-sparc64/vga.h
+++ b/include/asm-sparc64/vga.h
@@ -1,33 +1 @@
-/*
- *	Access to VGA videoram
- *
- *	(c) 1998 Martin Mares <mj@ucw.cz>
- */
-
-#ifndef _LINUX_ASM_VGA_H_
-#define _LINUX_ASM_VGA_H_
-
-#include <asm/types.h>
-
-#define VT_BUF_HAVE_RW
-
-#undef scr_writew
-#undef scr_readw
-
-static inline void scr_writew(u16 val, u16 *addr)
-{
-	BUG_ON((long) addr >= 0);
-
-	*addr = val;
-}
-
-static inline u16 scr_readw(const u16 *addr)
-{
-	BUG_ON((long) addr >= 0);
-
-	return *addr;
-}
-
-#define VGA_MAP_MEM(x,s) (x)
-
-#endif
+#include <asm-sparc/vga.h>
diff --git a/include/asm-sparc64/vio.h b/include/asm-sparc64/vio.h
index d4de32f0f8a..299b26ab81a 100644
--- a/include/asm-sparc64/vio.h
+++ b/include/asm-sparc64/vio.h
@@ -1,406 +1 @@
-#ifndef _SPARC64_VIO_H
-#define _SPARC64_VIO_H
-
-#include <linux/kernel.h>
-#include <linux/device.h>
-#include <linux/mod_devicetable.h>
-#include <linux/timer.h>
-#include <linux/spinlock.h>
-#include <linux/completion.h>
-#include <linux/list.h>
-#include <linux/log2.h>
-
-#include <asm/ldc.h>
-#include <asm/mdesc.h>
-
-struct vio_msg_tag {
-	u8			type;
-#define VIO_TYPE_CTRL		0x01
-#define VIO_TYPE_DATA		0x02
-#define VIO_TYPE_ERR		0x04
-
-	u8			stype;
-#define VIO_SUBTYPE_INFO	0x01
-#define VIO_SUBTYPE_ACK		0x02
-#define VIO_SUBTYPE_NACK	0x04
-
-	u16			stype_env;
-#define VIO_VER_INFO		0x0001
-#define VIO_ATTR_INFO		0x0002
-#define VIO_DRING_REG		0x0003
-#define VIO_DRING_UNREG		0x0004
-#define VIO_RDX			0x0005
-#define VIO_PKT_DATA		0x0040
-#define VIO_DESC_DATA		0x0041
-#define VIO_DRING_DATA		0x0042
-#define VNET_MCAST_INFO		0x0101
-
-	u32		sid;
-};
-
-struct vio_rdx {
-	struct vio_msg_tag	tag;
-	u64			resv[6];
-};
-
-struct vio_ver_info {
-	struct vio_msg_tag	tag;
-	u16			major;
-	u16			minor;
-	u8			dev_class;
-#define VDEV_NETWORK		0x01
-#define VDEV_NETWORK_SWITCH	0x02
-#define VDEV_DISK		0x03
-#define VDEV_DISK_SERVER	0x04
-
-	u8			resv1[3];
-	u64			resv2[5];
-};
-
-struct vio_dring_register {
-	struct vio_msg_tag	tag;
-	u64			dring_ident;
-	u32			num_descr;
-	u32			descr_size;
-	u16			options;
-#define VIO_TX_DRING		0x0001
-#define VIO_RX_DRING		0x0002
-	u16			resv;
-	u32			num_cookies;
-	struct ldc_trans_cookie	cookies[0];
-};
-
-struct vio_dring_unregister {
-	struct vio_msg_tag	tag;
-	u64			dring_ident;
-	u64			resv[5];
-};
-
-/* Data transfer modes */
-#define VIO_PKT_MODE		0x01 /* Packet based transfer	*/
-#define VIO_DESC_MODE		0x02 /* In-band descriptors	*/
-#define VIO_DRING_MODE		0x03 /* Descriptor rings	*/
-
-struct vio_dring_data {
-	struct vio_msg_tag	tag;
-	u64			seq;
-	u64			dring_ident;
-	u32			start_idx;
-	u32			end_idx;
-	u8			state;
-#define VIO_DRING_ACTIVE	0x01
-#define VIO_DRING_STOPPED	0x02
-
-	u8			__pad1;
-	u16			__pad2;
-	u32			__pad3;
-	u64			__par4[2];
-};
-
-struct vio_dring_hdr {
-	u8			state;
-#define VIO_DESC_FREE		0x01
-#define VIO_DESC_READY		0x02
-#define VIO_DESC_ACCEPTED	0x03
-#define VIO_DESC_DONE		0x04
-	u8			ack;
-#define VIO_ACK_ENABLE		0x01
-#define VIO_ACK_DISABLE		0x00
-
-	u16			__pad1;
-	u32			__pad2;
-};
-
-/* VIO disk specific structures and defines */
-struct vio_disk_attr_info {
-	struct vio_msg_tag	tag;
-	u8			xfer_mode;
-	u8			vdisk_type;
-#define VD_DISK_TYPE_SLICE	0x01 /* Slice in block device	*/
-#define VD_DISK_TYPE_DISK	0x02 /* Entire block device	*/
-	u16			resv1;
-	u32			vdisk_block_size;
-	u64			operations;
-	u64			vdisk_size;
-	u64			max_xfer_size;
-	u64			resv2[2];
-};
-
-struct vio_disk_desc {
-	struct vio_dring_hdr	hdr;
-	u64			req_id;
-	u8			operation;
-#define VD_OP_BREAD		0x01 /* Block read			*/
-#define VD_OP_BWRITE		0x02 /* Block write			*/
-#define VD_OP_FLUSH		0x03 /* Flush disk contents		*/
-#define VD_OP_GET_WCE		0x04 /* Get write-cache status		*/
-#define VD_OP_SET_WCE		0x05 /* Enable/disable write-cache	*/
-#define VD_OP_GET_VTOC		0x06 /* Get VTOC			*/
-#define VD_OP_SET_VTOC		0x07 /* Set VTOC			*/
-#define VD_OP_GET_DISKGEOM	0x08 /* Get disk geometry		*/
-#define VD_OP_SET_DISKGEOM	0x09 /* Set disk geometry		*/
-#define VD_OP_SCSICMD		0x0a /* SCSI control command		*/
-#define VD_OP_GET_DEVID		0x0b /* Get device ID			*/
-#define VD_OP_GET_EFI		0x0c /* Get EFI				*/
-#define VD_OP_SET_EFI		0x0d /* Set EFI				*/
-	u8			slice;
-	u16			resv1;
-	u32			status;
-	u64			offset;
-	u64			size;
-	u32			ncookies;
-	u32			resv2;
-	struct ldc_trans_cookie	cookies[0];
-};
-
-#define VIO_DISK_VNAME_LEN	8
-#define VIO_DISK_ALABEL_LEN	128
-#define VIO_DISK_NUM_PART	8
-
-struct vio_disk_vtoc {
-	u8			volume_name[VIO_DISK_VNAME_LEN];
-	u16			sector_size;
-	u16			num_partitions;
-	u8			ascii_label[VIO_DISK_ALABEL_LEN];
-	struct {
-		u16		id;
-		u16		perm_flags;
-		u32		resv;
-		u64		start_block;
-		u64		num_blocks;
-	} partitions[VIO_DISK_NUM_PART];
-};
-
-struct vio_disk_geom {
-	u16			num_cyl; /* Num data cylinders		*/
-	u16			alt_cyl; /* Num alternate cylinders	*/
-	u16			beg_cyl; /* Cyl off of fixed head area	*/
-	u16			num_hd;  /* Num heads			*/
-	u16			num_sec; /* Num sectors			*/
-	u16			ifact;   /* Interleave factor		*/
-	u16			apc;     /* Alts per cylinder (SCSI)	*/
-	u16			rpm;	 /* Revolutions per minute	*/
-	u16			phy_cyl; /* Num physical cylinders	*/
-	u16			wr_skip; /* Num sects to skip, writes	*/
-	u16			rd_skip; /* Num sects to skip, writes	*/
-};
-
-struct vio_disk_devid {
-	u16			resv;
-	u16			type;
-	u32			len;
-	char			id[0];
-};
-
-struct vio_disk_efi {
-	u64			lba;
-	u64			len;
-	char			data[0];
-};
-
-/* VIO net specific structures and defines */
-struct vio_net_attr_info {
-	struct vio_msg_tag	tag;
-	u8			xfer_mode;
-	u8			addr_type;
-#define VNET_ADDR_ETHERMAC	0x01
-	u16			ack_freq;
-	u32			resv1;
-	u64			addr;
-	u64			mtu;
-	u64			resv2[3];
-};
-
-#define VNET_NUM_MCAST		7
-
-struct vio_net_mcast_info {
-	struct vio_msg_tag	tag;
-	u8			set;
-	u8			count;
-	u8			mcast_addr[VNET_NUM_MCAST * 6];
-	u32			resv;
-};
-
-struct vio_net_desc {
-	struct vio_dring_hdr	hdr;
-	u32			size;
-	u32			ncookies;
-	struct ldc_trans_cookie	cookies[0];
-};
-
-#define VIO_MAX_RING_COOKIES	24
-
-struct vio_dring_state {
-	u64			ident;
-	void			*base;
-	u64			snd_nxt;
-	u64			rcv_nxt;
-	u32			entry_size;
-	u32			num_entries;
-	u32			prod;
-	u32			cons;
-	u32			pending;
-	int			ncookies;
-	struct ldc_trans_cookie	cookies[VIO_MAX_RING_COOKIES];
-};
-
-static inline void *vio_dring_cur(struct vio_dring_state *dr)
-{
-	return dr->base + (dr->entry_size * dr->prod);
-}
-
-static inline void *vio_dring_entry(struct vio_dring_state *dr,
-				    unsigned int index)
-{
-	return dr->base + (dr->entry_size * index);
-}
-
-static inline u32 vio_dring_avail(struct vio_dring_state *dr,
-				  unsigned int ring_size)
-{
-	BUILD_BUG_ON(!is_power_of_2(ring_size));
-
-	return (dr->pending -
-		((dr->prod - dr->cons) & (ring_size - 1)));
-}
-
-#define VIO_MAX_TYPE_LEN	32
-#define VIO_MAX_COMPAT_LEN	64
-
-struct vio_dev {
-	u64			mp;
-	struct device_node	*dp;
-
-	char			type[VIO_MAX_TYPE_LEN];
-	char			compat[VIO_MAX_COMPAT_LEN];
-	int			compat_len;
-
-	u64			dev_no;
-
-	unsigned long		channel_id;
-
-	unsigned int		tx_irq;
-	unsigned int		rx_irq;
-
-	struct device		dev;
-};
-
-struct vio_driver {
-	struct list_head		node;
-	const struct vio_device_id	*id_table;
-	int (*probe)(struct vio_dev *dev, const struct vio_device_id *id);
-	int (*remove)(struct vio_dev *dev);
-	void (*shutdown)(struct vio_dev *dev);
-	unsigned long			driver_data;
-	struct device_driver		driver;
-};
-
-struct vio_version {
-	u16		major;
-	u16		minor;
-};
-
-struct vio_driver_state;
-struct vio_driver_ops {
-	int	(*send_attr)(struct vio_driver_state *vio);
-	int	(*handle_attr)(struct vio_driver_state *vio, void *pkt);
-	void	(*handshake_complete)(struct vio_driver_state *vio);
-};
-
-struct vio_completion {
-	struct completion	com;
-	int			err;
-	int			waiting_for;
-};
-
-struct vio_driver_state {
-	/* Protects VIO handshake and, optionally, driver private state.  */
-	spinlock_t		lock;
-
-	struct ldc_channel	*lp;
-
-	u32			_peer_sid;
-	u32			_local_sid;
-	struct vio_dring_state	drings[2];
-#define VIO_DRIVER_TX_RING	0
-#define VIO_DRIVER_RX_RING	1
-
-	u8			hs_state;
-#define VIO_HS_INVALID		0x00
-#define VIO_HS_GOTVERS		0x01
-#define VIO_HS_GOT_ATTR		0x04
-#define VIO_HS_SENT_DREG	0x08
-#define VIO_HS_SENT_RDX		0x10
-#define VIO_HS_GOT_RDX_ACK	0x20
-#define VIO_HS_GOT_RDX		0x40
-#define VIO_HS_SENT_RDX_ACK	0x80
-#define VIO_HS_COMPLETE		(VIO_HS_GOT_RDX_ACK | VIO_HS_SENT_RDX_ACK)
-
-	u8			dev_class;
-
-	u8			dr_state;
-#define VIO_DR_STATE_TXREG	0x01
-#define VIO_DR_STATE_RXREG	0x02
-#define VIO_DR_STATE_TXREQ	0x10
-#define VIO_DR_STATE_RXREQ	0x20
-
-	u8			debug;
-#define VIO_DEBUG_HS		0x01
-#define VIO_DEBUG_DATA		0x02
-
-	void			*desc_buf;
-	unsigned int		desc_buf_len;
-
-	struct vio_completion	*cmp;
-
-	struct vio_dev		*vdev;
-
-	struct timer_list	timer;
-
-	struct vio_version	ver;
-
-	struct vio_version	*ver_table;
-	int			ver_table_entries;
-
-	char			*name;
-
-	struct vio_driver_ops	*ops;
-};
-
-#define viodbg(TYPE, f, a...) \
-do {	if (vio->debug & VIO_DEBUG_##TYPE) \
-		printk(KERN_INFO "vio: ID[%lu] " f, \
-		       vio->vdev->channel_id, ## a); \
-} while (0)
-
-extern int vio_register_driver(struct vio_driver *drv);
-extern void vio_unregister_driver(struct vio_driver *drv);
-
-static inline struct vio_driver *to_vio_driver(struct device_driver *drv)
-{
-	return container_of(drv, struct vio_driver, driver);
-}
-
-static inline struct vio_dev *to_vio_dev(struct device *dev)
-{
-	return container_of(dev, struct vio_dev, dev);
-}
-
-extern int vio_ldc_send(struct vio_driver_state *vio, void *data, int len);
-extern void vio_link_state_change(struct vio_driver_state *vio, int event);
-extern void vio_conn_reset(struct vio_driver_state *vio);
-extern int vio_control_pkt_engine(struct vio_driver_state *vio, void *pkt);
-extern int vio_validate_sid(struct vio_driver_state *vio,
-			    struct vio_msg_tag *tp);
-extern u32 vio_send_sid(struct vio_driver_state *vio);
-extern int vio_ldc_alloc(struct vio_driver_state *vio,
-			 struct ldc_channel_config *base_cfg, void *event_arg);
-extern void vio_ldc_free(struct vio_driver_state *vio);
-extern int vio_driver_init(struct vio_driver_state *vio, struct vio_dev *vdev,
-			   u8 dev_class, struct vio_version *ver_table,
-			   int ver_table_size, struct vio_driver_ops *ops,
-			   char *name);
-
-extern void vio_port_up(struct vio_driver_state *vio);
-
-#endif /* _SPARC64_VIO_H */
+#include <asm-sparc/vio.h>
diff --git a/include/asm-sparc64/visasm.h b/include/asm-sparc64/visasm.h
index 34f2ec64933..837a12278f4 100644
--- a/include/asm-sparc64/visasm.h
+++ b/include/asm-sparc64/visasm.h
@@ -1,62 +1 @@
-#ifndef _SPARC64_VISASM_H
-#define _SPARC64_VISASM_H
-
-/* visasm.h:  FPU saving macros for VIS routines
- *
- * Copyright (C) 1998 Jakub Jelinek (jj@ultra.linux.cz)
- */
- 
-#include <asm/pstate.h>
-#include <asm/ptrace.h>
-
-/* Clobbers %o5, %g1, %g2, %g3, %g7, %icc, %xcc */
-
-#define VISEntry					\
-	rd		%fprs, %o5;			\
-	andcc		%o5, (FPRS_FEF|FPRS_DU), %g0;	\
-	be,pt		%icc, 297f;			\
-	 sethi		%hi(297f), %g7;			\
-	sethi		%hi(VISenter), %g1;		\
-	jmpl		%g1 + %lo(VISenter), %g0;	\
-	 or		%g7, %lo(297f), %g7;		\
-297:	wr		%g0, FPRS_FEF, %fprs;		\
-
-#define VISExit						\
-	wr		%g0, 0, %fprs;
-
-/* Clobbers %o5, %g1, %g2, %g3, %g7, %icc, %xcc.
- * Must preserve %o5 between VISEntryHalf and VISExitHalf */
-
-#define VISEntryHalf					\
-	rd		%fprs, %o5;			\
-	andcc		%o5, FPRS_FEF, %g0;		\
-	be,pt		%icc, 297f;			\
-	 sethi		%hi(298f), %g7;			\
-	sethi		%hi(VISenterhalf), %g1;		\
-	jmpl		%g1 + %lo(VISenterhalf), %g0;	\
-	 or		%g7, %lo(298f), %g7;		\
-	clr		%o5;				\
-297:	wr		%o5, FPRS_FEF, %fprs;		\
-298:
-
-#define VISExitHalf					\
-	wr		%o5, 0, %fprs;
-
-#ifndef __ASSEMBLY__	
-static inline void save_and_clear_fpu(void) {
-	__asm__ __volatile__ (
-"		rd %%fprs, %%o5\n"
-"		andcc %%o5, %0, %%g0\n"
-"		be,pt %%icc, 299f\n"
-"		 sethi %%hi(298f), %%g7\n"
-"		sethi %%hi(VISenter), %%g1\n"
-"		jmpl %%g1 + %%lo(VISenter), %%g0\n"
-"		 or %%g7, %%lo(298f), %%g7\n"
-"	298:	wr %%g0, 0, %%fprs\n"
-"	299:\n"
-"		" : : "i" (FPRS_FEF|FPRS_DU) :
-		"o5", "g1", "g2", "g3", "g7", "cc");
-}
-#endif
-
-#endif /* _SPARC64_ASI_H */
+#include <asm-sparc/visasm.h>
diff --git a/include/asm-sparc64/watchdog.h b/include/asm-sparc64/watchdog.h
index 5baf2d3919c..b0f2857145f 100644
--- a/include/asm-sparc64/watchdog.h
+++ b/include/asm-sparc64/watchdog.h
@@ -1,31 +1 @@
-/*
- *
- * watchdog - Driver interface for the hardware watchdog timers
- * present on Sun Microsystems boardsets
- *
- * Copyright (c) 2000 Eric Brower <ebrower@usa.net>
- *
- */
-
-#ifndef _SPARC64_WATCHDOG_H
-#define _SPARC64_WATCHDOG_H
-
-#include <linux/watchdog.h>
-
-/* Solaris compatibility ioctls--
- * Ref. <linux/watchdog.h> for standard linux watchdog ioctls
- */
-#define WIOCSTART _IO (WATCHDOG_IOCTL_BASE, 10)		/* Start Timer		*/
-#define WIOCSTOP  _IO (WATCHDOG_IOCTL_BASE, 11)		/* Stop Timer		*/
-#define WIOCGSTAT _IOR(WATCHDOG_IOCTL_BASE, 12, int)/* Get Timer Status	*/
-
-/* Status flags from WIOCGSTAT ioctl
- */
-#define WD_FREERUN	0x01	/* timer is running, interrupts disabled	*/
-#define WD_EXPIRED	0x02	/* timer has expired						*/
-#define WD_RUNNING	0x04	/* timer is running, interrupts enabled		*/
-#define WD_STOPPED	0x08	/* timer has not been started				*/
-#define WD_SERVICED 0x10	/* timer interrupt was serviced				*/
-
-#endif /* ifndef _SPARC64_WATCHDOG_H */
-
+#include <asm-sparc/watchdog.h>
diff --git a/include/asm-sparc64/xor.h b/include/asm-sparc64/xor.h
index a0233884fc9..ef187cc07ed 100644
--- a/include/asm-sparc64/xor.h
+++ b/include/asm-sparc64/xor.h
@@ -1,70 +1 @@
-/*
- * include/asm-sparc64/xor.h
- *
- * High speed xor_block operation for RAID4/5 utilizing the
- * UltraSparc Visual Instruction Set and Niagara block-init
- * twin-load instructions.
- *
- * Copyright (C) 1997, 1999 Jakub Jelinek (jj@ultra.linux.cz)
- * Copyright (C) 2006 David S. Miller <davem@davemloft.net>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2, or (at your option)
- * any later version.
- *
- * You should have received a copy of the GNU General Public License
- * (for example /usr/src/linux/COPYING); if not, write to the Free
- * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- */
-
-#include <asm/spitfire.h>
-
-extern void xor_vis_2(unsigned long, unsigned long *, unsigned long *);
-extern void xor_vis_3(unsigned long, unsigned long *, unsigned long *,
-		      unsigned long *);
-extern void xor_vis_4(unsigned long, unsigned long *, unsigned long *,
-		      unsigned long *, unsigned long *);
-extern void xor_vis_5(unsigned long, unsigned long *, unsigned long *,
-		      unsigned long *, unsigned long *, unsigned long *);
-
-/* XXX Ugh, write cheetah versions... -DaveM */
-
-static struct xor_block_template xor_block_VIS = {
-        .name	= "VIS",
-        .do_2	= xor_vis_2,
-        .do_3	= xor_vis_3,
-        .do_4	= xor_vis_4,
-        .do_5	= xor_vis_5,
-};
-
-extern void xor_niagara_2(unsigned long, unsigned long *, unsigned long *);
-extern void xor_niagara_3(unsigned long, unsigned long *, unsigned long *,
-			  unsigned long *);
-extern void xor_niagara_4(unsigned long, unsigned long *, unsigned long *,
-			  unsigned long *, unsigned long *);
-extern void xor_niagara_5(unsigned long, unsigned long *, unsigned long *,
-			  unsigned long *, unsigned long *, unsigned long *);
-
-static struct xor_block_template xor_block_niagara = {
-        .name	= "Niagara",
-        .do_2	= xor_niagara_2,
-        .do_3	= xor_niagara_3,
-        .do_4	= xor_niagara_4,
-        .do_5	= xor_niagara_5,
-};
-
-#undef XOR_TRY_TEMPLATES
-#define XOR_TRY_TEMPLATES				\
-	do {						\
-		xor_speed(&xor_block_VIS);		\
-		xor_speed(&xor_block_niagara);		\
-	} while (0)
-
-/* For VIS for everything except Niagara.  */
-#define XOR_SELECT_TEMPLATE(FASTEST) \
-	((tlb_type == hypervisor && \
-	  (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 || \
-	   sun4v_chip_type == SUN4V_CHIP_NIAGARA2)) ? \
-	 &xor_block_niagara : \
-	 &xor_block_VIS)
+#include <asm-sparc/xor.h>
author	Linus Torvalds <torvalds@linux-foundation.org>	2008-07-21 09:40:26 -0700
committer	Linus Torvalds <torvalds@linux-foundation.org>	2008-07-21 09:40:26 -0700
commit	f8b71a3a92d69cb1c9c2162f9235d3dd550e4aa0 (patch)
tree	d0853186f33314a6e45903876d0e94d1619fcca9
parent	42a77a1b8fe4fd78cf8d8fccbaac918bbe888192 (diff)
parent	e4f25060b87a627f5cda84b8134911d43c919458 (diff)