Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

18 files changed, 338 insertions, 392 deletions

diff --git a/arch/parisc/include/asm/atomic.h b/arch/parisc/include/asm/atomic.h
index 4054b31e0fa..6c6defc2461 100644
--- a/arch/parisc/include/asm/atomic.h
+++ b/arch/parisc/include/asm/atomic.h
@@ -6,7 +6,7 @@
 #define _ASM_PARISC_ATOMIC_H_
 
 #include <linux/types.h>
-#include <asm/system.h>
+#include <asm/cmpxchg.h>
 
 /*
  * Atomic operations that C can't guarantee us.  Useful for
@@ -49,112 +49,6 @@ extern arch_spinlock_t __atomic_hash[ATOMIC_HASH_SIZE] __lock_aligned;
 #  define _atomic_spin_unlock_irqrestore(l,f) do { local_irq_restore(f); } while (0)
 #endif
 
-/* This should get optimized out since it's never called.
-** Or get a link error if xchg is used "wrong".
-*/
-extern void __xchg_called_with_bad_pointer(void);
-
-
-/* __xchg32/64 defined in arch/parisc/lib/bitops.c */
-extern unsigned long __xchg8(char, char *);
-extern unsigned long __xchg32(int, int *);
-#ifdef CONFIG_64BIT
-extern unsigned long __xchg64(unsigned long, unsigned long *);
-#endif
-
-/* optimizer better get rid of switch since size is a constant */
-static __inline__ unsigned long
-__xchg(unsigned long x, __volatile__ void * ptr, int size)
-{
-	switch(size) {
-#ifdef CONFIG_64BIT
-	case 8: return __xchg64(x,(unsigned long *) ptr);
-#endif
-	case 4: return __xchg32((int) x, (int *) ptr);
-	case 1: return __xchg8((char) x, (char *) ptr);
-	}
-	__xchg_called_with_bad_pointer();
-	return x;
-}
-
-
-/*
-** REVISIT - Abandoned use of LDCW in xchg() for now:
-** o need to test sizeof(*ptr) to avoid clearing adjacent bytes
-** o and while we are at it, could CONFIG_64BIT code use LDCD too?
-**
-**	if (__builtin_constant_p(x) && (x == NULL))
-**		if (((unsigned long)p & 0xf) == 0)
-**			return __ldcw(p);
-*/
-#define xchg(ptr,x) \
-	((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
-
-
-#define __HAVE_ARCH_CMPXCHG	1
-
-/* bug catcher for when unsupported size is used - won't link */
-extern void __cmpxchg_called_with_bad_pointer(void);
-
-/* __cmpxchg_u32/u64 defined in arch/parisc/lib/bitops.c */
-extern unsigned long __cmpxchg_u32(volatile unsigned int *m, unsigned int old, unsigned int new_);
-extern unsigned long __cmpxchg_u64(volatile unsigned long *ptr, unsigned long old, unsigned long 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) {
-#ifdef CONFIG_64BIT
-	case 8: return __cmpxchg_u64((unsigned long *)ptr, old, new_);
-#endif
-	case 4: return __cmpxchg_u32((unsigned int *)ptr, (unsigned int) old, (unsigned int) 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))); \
-  })
-
-#include <asm-generic/cmpxchg-local.h>
-
-static inline unsigned long __cmpxchg_local(volatile void *ptr,
-				      unsigned long old,
-				      unsigned long new_, int size)
-{
-	switch (size) {
-#ifdef CONFIG_64BIT
-	case 8:	return __cmpxchg_u64((unsigned long *)ptr, old, new_);
-#endif
-	case 4:	return __cmpxchg_u32(ptr, old, new_);
-	default:
-		return __cmpxchg_local_generic(ptr, old, new_, size);
-	}
-}
-
-/*
- * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
- * them available.
- */
-#define cmpxchg_local(ptr, o, n)				  	\
-	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
-			(unsigned long)(n), sizeof(*(ptr))))
-#ifdef CONFIG_64BIT
-#define cmpxchg64_local(ptr, o, n)					\
-  ({									\
-	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
-	cmpxchg_local((ptr), (o), (n));					\
-  })
-#else
-#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
-#endif
-
 /*
  * Note that we need not lock read accesses - aligned word writes/reads
  * are atomic, so a reader never sees inconsistent values.
diff --git a/arch/parisc/include/asm/barrier.h b/arch/parisc/include/asm/barrier.h
new file mode 100644
index 00000000000..e77d834aa80
--- /dev/null
+++ b/arch/parisc/include/asm/barrier.h
@@ -0,0 +1,35 @@
+#ifndef __PARISC_BARRIER_H
+#define __PARISC_BARRIER_H
+
+/*
+** This is simply the barrier() macro from linux/kernel.h but when serial.c
+** uses tqueue.h uses smp_mb() defined using barrier(), linux/kernel.h
+** hasn't yet been included yet so it fails, thus repeating the macro here.
+**
+** PA-RISC architecture allows for weakly ordered memory accesses although
+** none of the processors use it. There is a strong ordered bit that is
+** set in the O-bit of the page directory entry. Operating systems that
+** can not tolerate out of order accesses should set this bit when mapping
+** pages. The O-bit of the PSW should also be set to 1 (I don't believe any
+** of the processor implemented the PSW O-bit). The PCX-W ERS states that
+** the TLB O-bit is not implemented so the page directory does not need to
+** have the O-bit set when mapping pages (section 3.1). This section also
+** states that the PSW Y, Z, G, and O bits are not implemented.
+** So it looks like nothing needs to be done for parisc-linux (yet).
+** (thanks to chada for the above comment -ggg)
+**
+** The __asm__ op below simple prevents gcc/ld from reordering
+** instructions across the mb() "call".
+*/
+#define mb()		__asm__ __volatile__("":::"memory")	/* barrier() */
+#define rmb()		mb()
+#define wmb()		mb()
+#define smp_mb()	mb()
+#define smp_rmb()	mb()
+#define smp_wmb()	mb()
+#define smp_read_barrier_depends()	do { } while(0)
+#define read_barrier_depends()		do { } while(0)
+
+#define set_mb(var, value)		do { var = value; mb(); } while (0)
+
+#endif /* __PARISC_BARRIER_H */
diff --git a/arch/parisc/include/asm/cmpxchg.h b/arch/parisc/include/asm/cmpxchg.h
new file mode 100644
index 00000000000..dbd13354ec4
--- /dev/null
+++ b/arch/parisc/include/asm/cmpxchg.h
@@ -0,0 +1,116 @@
+/*
+ * forked from parisc asm/atomic.h which was:
+ *	Copyright (C) 2000 Philipp Rumpf <prumpf@tux.org>
+ *	Copyright (C) 2006 Kyle McMartin <kyle@parisc-linux.org>
+ */
+
+#ifndef _ASM_PARISC_CMPXCHG_H_
+#define _ASM_PARISC_CMPXCHG_H_
+
+/* This should get optimized out since it's never called.
+** Or get a link error if xchg is used "wrong".
+*/
+extern void __xchg_called_with_bad_pointer(void);
+
+/* __xchg32/64 defined in arch/parisc/lib/bitops.c */
+extern unsigned long __xchg8(char, char *);
+extern unsigned long __xchg32(int, int *);
+#ifdef CONFIG_64BIT
+extern unsigned long __xchg64(unsigned long, unsigned long *);
+#endif
+
+/* optimizer better get rid of switch since size is a constant */
+static inline unsigned long
+__xchg(unsigned long x, __volatile__ void *ptr, int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8: return __xchg64(x, (unsigned long *) ptr);
+#endif
+	case 4: return __xchg32((int) x, (int *) ptr);
+	case 1: return __xchg8((char) x, (char *) ptr);
+	}
+	__xchg_called_with_bad_pointer();
+	return x;
+}
+
+/*
+** REVISIT - Abandoned use of LDCW in xchg() for now:
+** o need to test sizeof(*ptr) to avoid clearing adjacent bytes
+** o and while we are at it, could CONFIG_64BIT code use LDCD too?
+**
+**	if (__builtin_constant_p(x) && (x == NULL))
+**		if (((unsigned long)p & 0xf) == 0)
+**			return __ldcw(p);
+*/
+#define xchg(ptr, x) \
+	((__typeof__(*(ptr)))__xchg((unsigned long)(x), (ptr), sizeof(*(ptr))))
+
+#define __HAVE_ARCH_CMPXCHG	1
+
+/* bug catcher for when unsupported size is used - won't link */
+extern void __cmpxchg_called_with_bad_pointer(void);
+
+/* __cmpxchg_u32/u64 defined in arch/parisc/lib/bitops.c */
+extern unsigned long __cmpxchg_u32(volatile unsigned int *m, unsigned int old,
+				   unsigned int new_);
+extern unsigned long __cmpxchg_u64(volatile unsigned long *ptr,
+				   unsigned long old, unsigned long 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) {
+#ifdef CONFIG_64BIT
+	case 8: return __cmpxchg_u64((unsigned long *)ptr, old, new_);
+#endif
+	case 4: return __cmpxchg_u32((unsigned int *)ptr,
+				     (unsigned int)old, (unsigned int)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))); \
+})
+
+#include <asm-generic/cmpxchg-local.h>
+
+static inline unsigned long __cmpxchg_local(volatile void *ptr,
+				      unsigned long old,
+				      unsigned long new_, int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8:	return __cmpxchg_u64((unsigned long *)ptr, old, new_);
+#endif
+	case 4:	return __cmpxchg_u32(ptr, old, new_);
+	default:
+		return __cmpxchg_local_generic(ptr, old, new_, size);
+	}
+}
+
+/*
+ * cmpxchg_local and cmpxchg64_local are atomic wrt current CPU. Always make
+ * them available.
+ */
+#define cmpxchg_local(ptr, o, n)					\
+	((__typeof__(*(ptr)))__cmpxchg_local((ptr), (unsigned long)(o),	\
+			(unsigned long)(n), sizeof(*(ptr))))
+#ifdef CONFIG_64BIT
+#define cmpxchg64_local(ptr, o, n)					\
+({									\
+	BUILD_BUG_ON(sizeof(*(ptr)) != 8);				\
+	cmpxchg_local((ptr), (o), (n));					\
+})
+#else
+#define cmpxchg64_local(ptr, o, n) __cmpxchg64_local_generic((ptr), (o), (n))
+#endif
+
+#endif /* _ASM_PARISC_CMPXCHG_H_ */
diff --git a/arch/parisc/include/asm/delay.h b/arch/parisc/include/asm/delay.h
index 7a75e984674..912ee7e6a57 100644
--- a/arch/parisc/include/asm/delay.h
+++ b/arch/parisc/include/asm/delay.h
@@ -1,7 +1,7 @@
 #ifndef _PARISC_DELAY_H
 #define _PARISC_DELAY_H
 
-#include <asm/system.h>    /* for mfctl() */
+#include <asm/special_insns.h>    /* for mfctl() */
 #include <asm/processor.h> /* for boot_cpu_data */
 
 
diff --git a/arch/parisc/include/asm/dma.h b/arch/parisc/include/asm/dma.h
index f7a18f96870..fd48ae2de95 100644
--- a/arch/parisc/include/asm/dma.h
+++ b/arch/parisc/include/asm/dma.h
@@ -9,7 +9,6 @@
 #define _ASM_DMA_H
 
 #include <asm/io.h>		/* need byte IO */
-#include <asm/system.h>	
 
 #define dma_outb	outb
 #define dma_inb		inb
diff --git a/arch/parisc/include/asm/exec.h b/arch/parisc/include/asm/exec.h
new file mode 100644
index 00000000000..6bb5af75b17
--- /dev/null
+++ b/arch/parisc/include/asm/exec.h
@@ -0,0 +1,6 @@
+#ifndef __PARISC_EXEC_H
+#define __PARISC_EXEC_H
+
+#define arch_align_stack(x) (x)
+
+#endif /* __PARISC_EXEC_H */
diff --git a/arch/parisc/include/asm/futex.h b/arch/parisc/include/asm/futex.h
index 2388bdb3283..49df14805a9 100644
--- a/arch/parisc/include/asm/futex.h
+++ b/arch/parisc/include/asm/futex.h
@@ -8,6 +8,29 @@
 #include <asm/atomic.h>
 #include <asm/errno.h>
 
+/* The following has to match the LWS code in syscall.S.  We have
+   sixteen four-word locks. */
+
+static inline void
+_futex_spin_lock_irqsave(u32 __user *uaddr, unsigned long int *flags)
+{
+	extern u32 lws_lock_start[];
+	long index = ((long)uaddr & 0xf0) >> 2;
+	arch_spinlock_t *s = (arch_spinlock_t *)&lws_lock_start[index];
+	local_irq_save(*flags);
+	arch_spin_lock(s);
+}
+
+static inline void
+_futex_spin_unlock_irqrestore(u32 __user *uaddr, unsigned long int *flags)
+{
+	extern u32 lws_lock_start[];
+	long index = ((long)uaddr & 0xf0) >> 2;
+	arch_spinlock_t *s = (arch_spinlock_t *)&lws_lock_start[index];
+	arch_spin_unlock(s);
+	local_irq_restore(*flags);
+}
+
 static inline int
 futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
 {
@@ -26,7 +49,7 @@ futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
 
 	pagefault_disable();
 
-	_atomic_spin_lock_irqsave(uaddr, flags);
+	_futex_spin_lock_irqsave(uaddr, &flags);
 
 	switch (op) {
 	case FUTEX_OP_SET:
@@ -71,7 +94,7 @@ futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
 		ret = -ENOSYS;
 	}
 
-	_atomic_spin_unlock_irqrestore(uaddr, flags);
+	_futex_spin_unlock_irqrestore(uaddr, &flags);
 
 	pagefault_enable();
 
@@ -113,7 +136,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
 	 * address. This should scale to a couple of CPUs.
 	 */
 
-	_atomic_spin_lock_irqsave(uaddr, flags);
+	_futex_spin_lock_irqsave(uaddr, &flags);
 
 	ret = get_user(val, uaddr);
 
@@ -122,7 +145,7 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
 
 	*uval = val;
 
-	_atomic_spin_unlock_irqrestore(uaddr, flags);
+	_futex_spin_unlock_irqrestore(uaddr, &flags);
 
 	return ret;
 }
diff --git a/arch/parisc/include/asm/ldcw.h b/arch/parisc/include/asm/ldcw.h
new file mode 100644
index 00000000000..d2d11b7055b
--- /dev/null
+++ b/arch/parisc/include/asm/ldcw.h
@@ -0,0 +1,48 @@
+#ifndef __PARISC_LDCW_H
+#define __PARISC_LDCW_H
+
+#ifndef CONFIG_PA20
+/* Because kmalloc only guarantees 8-byte alignment for kmalloc'd data,
+   and GCC only guarantees 8-byte alignment for stack locals, we can't
+   be assured of 16-byte alignment for atomic lock data even if we
+   specify "__attribute ((aligned(16)))" in the type declaration.  So,
+   we use a struct containing an array of four ints for the atomic lock
+   type and dynamically select the 16-byte aligned int from the array
+   for the semaphore.  */
+
+#define __PA_LDCW_ALIGNMENT	16
+#define __ldcw_align(a) ({					\
+	unsigned long __ret = (unsigned long) &(a)->lock[0];	\
+	__ret = (__ret + __PA_LDCW_ALIGNMENT - 1)		\
+		& ~(__PA_LDCW_ALIGNMENT - 1);			\
+	(volatile unsigned int *) __ret;			\
+})
+#define __LDCW	"ldcw"
+
+#else /*CONFIG_PA20*/
+/* From: "Jim Hull" <jim.hull of hp.com>
+   I've attached a summary of the change, but basically, for PA 2.0, as
+   long as the ",CO" (coherent operation) completer is specified, then the
+   16-byte alignment requirement for ldcw and ldcd is relaxed, and instead
+   they only require "natural" alignment (4-byte for ldcw, 8-byte for
+   ldcd). */
+
+#define __PA_LDCW_ALIGNMENT	4
+#define __ldcw_align(a) (&(a)->slock)
+#define __LDCW	"ldcw,co"
+
+#endif /*!CONFIG_PA20*/
+
+/* LDCW, the only atomic read-write operation PA-RISC has. *sigh*.  */
+#define __ldcw(a) ({						\
+	unsigned __ret;						\
+	__asm__ __volatile__(__LDCW " 0(%2),%0"			\
+		: "=r" (__ret), "+m" (*(a)) : "r" (a));		\
+	__ret;							\
+})
+
+#ifdef CONFIG_SMP
+# define __lock_aligned __attribute__((__section__(".data..lock_aligned")))
+#endif
+
+#endif /* __PARISC_LDCW_H */
diff --git a/arch/parisc/include/asm/posix_types.h b/arch/parisc/include/asm/posix_types.h
index 00da29a340b..5212b0357da 100644
--- a/arch/parisc/include/asm/posix_types.h
+++ b/arch/parisc/include/asm/posix_types.h
@@ -6,123 +6,22 @@
  * be a little careful about namespace pollution etc.  Also, we cannot
  * assume GCC is being used.
  */
-typedef unsigned long		__kernel_ino_t;
+
 typedef unsigned short		__kernel_mode_t;
+#define __kernel_mode_t __kernel_mode_t
+
 typedef unsigned short		__kernel_nlink_t;
-typedef long			__kernel_off_t;
-typedef int			__kernel_pid_t;
+#define __kernel_nlink_t __kernel_nlink_t
+
 typedef unsigned short		__kernel_ipc_pid_t;
-typedef unsigned int		__kernel_uid_t;
-typedef unsigned int		__kernel_gid_t;
-typedef int			__kernel_suseconds_t;
-typedef long			__kernel_clock_t;
-typedef int			__kernel_timer_t;
-typedef int			__kernel_clockid_t;
-typedef int			__kernel_daddr_t;
-/* Note these change from narrow to wide kernels */
-#ifdef CONFIG_64BIT
-typedef unsigned long		__kernel_size_t;
-typedef long			__kernel_ssize_t;
-typedef long			__kernel_ptrdiff_t;
-#else
-typedef unsigned int		__kernel_size_t;
-typedef int			__kernel_ssize_t;
-typedef int			__kernel_ptrdiff_t;
-#endif
-typedef long			__kernel_time_t;
-typedef char *			__kernel_caddr_t;
+#define __kernel_ipc_pid_t __kernel_ipc_pid_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 int			__kernel_suseconds_t;
+#define __kernel_suseconds_t __kernel_suseconds_t
 
-#ifdef __GNUC__
-typedef long long		__kernel_loff_t;
 typedef long long		__kernel_off64_t;
 typedef unsigned long long	__kernel_ino64_t;
-#endif
-
-typedef unsigned int		__kernel_old_dev_t;
-
-typedef struct {
-	int	val[2];
-} __kernel_fsid_t;
-
-/* compatibility stuff */
-typedef __kernel_uid_t __kernel_old_uid_t;
-typedef __kernel_gid_t __kernel_old_gid_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 case (8 ints,
- * for a 256-bit fd_set)
- */
-#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 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__) */
+#include <asm-generic/posix_types.h>
 
 #endif
diff --git a/arch/parisc/include/asm/processor.h b/arch/parisc/include/asm/processor.h
index 7213ec9e594..acdf4cad612 100644
--- a/arch/parisc/include/asm/processor.h
+++ b/arch/parisc/include/asm/processor.h
@@ -16,7 +16,6 @@
 #include <asm/pdc.h>
 #include <asm/ptrace.h>
 #include <asm/types.h>
-#include <asm/system.h>
 #include <asm/percpu.h>
 
 #endif /* __ASSEMBLY__ */
@@ -169,6 +168,7 @@ struct thread_struct {
  * Return saved PC of a blocked thread.  This is used by ps mostly.
  */
 
+struct task_struct;
 unsigned long thread_saved_pc(struct task_struct *t);
 void show_trace(struct task_struct *task, unsigned long *stack);
 
diff --git a/arch/parisc/include/asm/psw.h b/arch/parisc/include/asm/psw.h
index 5a3e23c9ce6..ad69a35e9c0 100644
--- a/arch/parisc/include/asm/psw.h
+++ b/arch/parisc/include/asm/psw.h
@@ -59,4 +59,45 @@
 #define USER_PSW_MASK (WIDE_PSW | PSW_T | PSW_N | PSW_X | PSW_B | PSW_V | PSW_CB)
 #define USER_PSW      (PSW_C | PSW_Q | PSW_P | PSW_D | PSW_I)
 
+#ifndef __ASSEMBLY__
+
+/* The program status word as bitfields.  */
+struct pa_psw {
+	unsigned int y:1;
+	unsigned int z:1;
+	unsigned int rv:2;
+	unsigned int w:1;
+	unsigned int e:1;
+	unsigned int s:1;
+	unsigned int t:1;
+
+	unsigned int h:1;
+	unsigned int l:1;
+	unsigned int n:1;
+	unsigned int x:1;
+	unsigned int b:1;
+	unsigned int c:1;
+	unsigned int v:1;
+	unsigned int m:1;
+
+	unsigned int cb:8;
+
+	unsigned int o:1;
+	unsigned int g:1;
+	unsigned int f:1;
+	unsigned int r:1;
+	unsigned int q:1;
+	unsigned int p:1;
+	unsigned int d:1;
+	unsigned int i:1;
+};
+
+#ifdef CONFIG_64BIT
+#define pa_psw(task) ((struct pa_psw *) ((char *) (task) + TASK_PT_PSW + 4))
+#else
+#define pa_psw(task) ((struct pa_psw *) ((char *) (task) + TASK_PT_PSW))
+#endif
+
+#endif /* !__ASSEMBLY__ */
+
 #endif
diff --git a/arch/parisc/include/asm/special_insns.h b/arch/parisc/include/asm/special_insns.h
new file mode 100644
index 00000000000..d306b75bc77
--- /dev/null
+++ b/arch/parisc/include/asm/special_insns.h
@@ -0,0 +1,40 @@
+#ifndef __PARISC_SPECIAL_INSNS_H
+#define __PARISC_SPECIAL_INSNS_H
+
+#define mfctl(reg)	({		\
+	unsigned long cr;		\
+	__asm__ __volatile__(		\
+		"mfctl " #reg ",%0" :	\
+		 "=r" (cr)		\
+	);				\
+	cr;				\
+})
+
+#define mtctl(gr, cr) \
+	__asm__ __volatile__("mtctl %0,%1" \
+		: /* no outputs */ \
+		: "r" (gr), "i" (cr) : "memory")
+
+/* these are here to de-mystefy the calling code, and to provide hooks */
+/* which I needed for debugging EIEM problems -PB */
+#define get_eiem() mfctl(15)
+static inline void set_eiem(unsigned long val)
+{
+	mtctl(val, 15);
+}
+
+#define mfsp(reg)	({		\
+	unsigned long cr;		\
+	__asm__ __volatile__(		\
+		"mfsp " #reg ",%0" :	\
+		 "=r" (cr)		\
+	);				\
+	cr;				\
+})
+
+#define mtsp(gr, cr) \
+	__asm__ __volatile__("mtsp %0,%1" \
+		: /* no outputs */ \
+		: "r" (gr), "i" (cr) : "memory")
+
+#endif /* __PARISC_SPECIAL_INSNS_H */
diff --git a/arch/parisc/include/asm/spinlock.h b/arch/parisc/include/asm/spinlock.h
index 74036f436a3..804aa28ab1d 100644
--- a/arch/parisc/include/asm/spinlock.h
+++ b/arch/parisc/include/asm/spinlock.h
@@ -1,7 +1,6 @@
 #ifndef __ASM_SPINLOCK_H
 #define __ASM_SPINLOCK_H
 
-#include <asm/system.h>
 #include <asm/processor.h>
 #include <asm/spinlock_types.h>
 
diff --git a/arch/parisc/include/asm/switch_to.h b/arch/parisc/include/asm/switch_to.h
new file mode 100644
index 00000000000..8ed8fea1e78
--- /dev/null
+++ b/arch/parisc/include/asm/switch_to.h
@@ -0,0 +1,12 @@
+#ifndef __PARISC_SWITCH_TO_H
+#define __PARISC_SWITCH_TO_H
+
+struct task_struct;
+
+extern struct task_struct *_switch_to(struct task_struct *, struct task_struct *);
+
+#define switch_to(prev, next, last) do {			\
+	(last) = _switch_to(prev, next);			\
+} while(0)
+
+#endif /* __PARISC_SWITCH_TO_H */
diff --git a/arch/parisc/include/asm/system.h b/arch/parisc/include/asm/system.h
deleted file mode 100644
index b19e63a8e84..00000000000
--- a/arch/parisc/include/asm/system.h
+++ /dev/null
@@ -1,165 +0,0 @@
-#ifndef __PARISC_SYSTEM_H
-#define __PARISC_SYSTEM_H
-
-#include <linux/irqflags.h>
-
-/* The program status word as bitfields.  */
-struct pa_psw {
-	unsigned int y:1;
-	unsigned int z:1;
-	unsigned int rv:2;
-	unsigned int w:1;
-	unsigned int e:1;
-	unsigned int s:1;
-	unsigned int t:1;
-
-	unsigned int h:1;
-	unsigned int l:1;
-	unsigned int n:1;
-	unsigned int x:1;
-	unsigned int b:1;
-	unsigned int c:1;
-	unsigned int v:1;
-	unsigned int m:1;
-
-	unsigned int cb:8;
-
-	unsigned int o:1;
-	unsigned int g:1;
-	unsigned int f:1;
-	unsigned int r:1;
-	unsigned int q:1;
-	unsigned int p:1;
-	unsigned int d:1;
-	unsigned int i:1;
-};
-
-#ifdef CONFIG_64BIT
-#define pa_psw(task) ((struct pa_psw *) ((char *) (task) + TASK_PT_PSW + 4))
-#else
-#define pa_psw(task) ((struct pa_psw *) ((char *) (task) + TASK_PT_PSW))
-#endif
-
-struct task_struct;
-
-extern struct task_struct *_switch_to(struct task_struct *, struct task_struct *);
-
-#define switch_to(prev, next, last) do {			\
-	(last) = _switch_to(prev, next);			\
-} while(0)
-
-#define mfctl(reg)	({		\
-	unsigned long cr;		\
-	__asm__ __volatile__(		\
-		"mfctl " #reg ",%0" :	\
-		 "=r" (cr)		\
-	);				\
-	cr;				\
-})
-
-#define mtctl(gr, cr) \
-	__asm__ __volatile__("mtctl %0,%1" \
-		: /* no outputs */ \
-		: "r" (gr), "i" (cr) : "memory")
-
-/* these are here to de-mystefy the calling code, and to provide hooks */
-/* which I needed for debugging EIEM problems -PB */
-#define get_eiem() mfctl(15)
-static inline void set_eiem(unsigned long val)
-{
-	mtctl(val, 15);
-}
-
-#define mfsp(reg)	({		\
-	unsigned long cr;		\
-	__asm__ __volatile__(		\
-		"mfsp " #reg ",%0" :	\
-		 "=r" (cr)		\
-	);				\
-	cr;				\
-})
-
-#define mtsp(gr, cr) \
-	__asm__ __volatile__("mtsp %0,%1" \
-		: /* no outputs */ \
-		: "r" (gr), "i" (cr) : "memory")
-
-
-/*
-** This is simply the barrier() macro from linux/kernel.h but when serial.c
-** uses tqueue.h uses smp_mb() defined using barrier(), linux/kernel.h
-** hasn't yet been included yet so it fails, thus repeating the macro here.
-**
-** PA-RISC architecture allows for weakly ordered memory accesses although
-** none of the processors use it. There is a strong ordered bit that is
-** set in the O-bit of the page directory entry. Operating systems that
-** can not tolerate out of order accesses should set this bit when mapping
-** pages. The O-bit of the PSW should also be set to 1 (I don't believe any
-** of the processor implemented the PSW O-bit). The PCX-W ERS states that
-** the TLB O-bit is not implemented so the page directory does not need to
-** have the O-bit set when mapping pages (section 3.1). This section also
-** states that the PSW Y, Z, G, and O bits are not implemented.
-** So it looks like nothing needs to be done for parisc-linux (yet).
-** (thanks to chada for the above comment -ggg)
-**
-** The __asm__ op below simple prevents gcc/ld from reordering
-** instructions across the mb() "call".
-*/
-#define mb()		__asm__ __volatile__("":::"memory")	/* barrier() */
-#define rmb()		mb()
-#define wmb()		mb()
-#define smp_mb()	mb()
-#define smp_rmb()	mb()
-#define smp_wmb()	mb()
-#define smp_read_barrier_depends()	do { } while(0)
-#define read_barrier_depends()		do { } while(0)
-
-#define set_mb(var, value)		do { var = value; mb(); } while (0)
-
-#ifndef CONFIG_PA20
-/* Because kmalloc only guarantees 8-byte alignment for kmalloc'd data,
-   and GCC only guarantees 8-byte alignment for stack locals, we can't
-   be assured of 16-byte alignment for atomic lock data even if we
-   specify "__attribute ((aligned(16)))" in the type declaration.  So,
-   we use a struct containing an array of four ints for the atomic lock
-   type and dynamically select the 16-byte aligned int from the array
-   for the semaphore.  */
-
-#define __PA_LDCW_ALIGNMENT	16
-#define __ldcw_align(a) ({					\
-	unsigned long __ret = (unsigned long) &(a)->lock[0];	\
-	__ret = (__ret + __PA_LDCW_ALIGNMENT - 1)		\
-		& ~(__PA_LDCW_ALIGNMENT - 1);			\
-	(volatile unsigned int *) __ret;			\
-})
-#define __LDCW	"ldcw"
-
-#else /*CONFIG_PA20*/
-/* From: "Jim Hull" <jim.hull of hp.com>
-   I've attached a summary of the change, but basically, for PA 2.0, as
-   long as the ",CO" (coherent operation) completer is specified, then the
-   16-byte alignment requirement for ldcw and ldcd is relaxed, and instead
-   they only require "natural" alignment (4-byte for ldcw, 8-byte for
-   ldcd). */
-
-#define __PA_LDCW_ALIGNMENT	4
-#define __ldcw_align(a) (&(a)->slock)
-#define __LDCW	"ldcw,co"
-
-#endif /*!CONFIG_PA20*/
-
-/* LDCW, the only atomic read-write operation PA-RISC has. *sigh*.  */
-#define __ldcw(a) ({						\
-	unsigned __ret;						\
-	__asm__ __volatile__(__LDCW " 0(%2),%0"			\
-		: "=r" (__ret), "+m" (*(a)) : "r" (a));		\
-	__ret;							\
-})
-
-#ifdef CONFIG_SMP
-# define __lock_aligned __attribute__((__section__(".data..lock_aligned")))
-#endif
-
-#define arch_align_stack(x) (x)
-
-#endif
diff --git a/arch/parisc/include/asm/thread_info.h b/arch/parisc/include/asm/thread_info.h
index 6d9c7c7973d..83ae7dd4d99 100644
--- a/arch/parisc/include/asm/thread_info.h
+++ b/arch/parisc/include/asm/thread_info.h
@@ -5,6 +5,7 @@
 
 #ifndef __ASSEMBLY__
 #include <asm/processor.h>
+#include <asm/special_insns.h>
 
 struct thread_info {
 	struct task_struct *task;	/* main task structure */
diff --git a/arch/parisc/include/asm/timex.h b/arch/parisc/include/asm/timex.h
index 3b68d77273d..2bd51f6d832 100644
--- a/arch/parisc/include/asm/timex.h
+++ b/arch/parisc/include/asm/timex.h
@@ -6,7 +6,6 @@
 #ifndef _ASMPARISC_TIMEX_H
 #define _ASMPARISC_TIMEX_H
 
-#include <asm/system.h>
 
 #define CLOCK_TICK_RATE	1193180 /* Underlying HZ */
 
diff --git a/arch/parisc/include/asm/uaccess.h b/arch/parisc/include/asm/uaccess.h
index ff4cf9dab8d..9ac066086f0 100644
--- a/arch/parisc/include/asm/uaccess.h
+++ b/arch/parisc/include/asm/uaccess.h
@@ -5,7 +5,6 @@
  * User space memory access functions
  */
 #include <asm/page.h>
-#include <asm/system.h>
 #include <asm/cache.h>
 #include <asm/errno.h>
 #include <asm-generic/uaccess-unaligned.h>