summaryrefslogtreecommitdiffstats
path: root/arch
diff options
context:
space:
mode:
Diffstat (limited to 'arch')
-rw-r--r--arch/alpha/kernel/osf_sys.c7
-rw-r--r--arch/arm/mach-s3c2410/clock.c2
-rw-r--r--arch/arm/mach-s3c2410/s3c2440.c6
-rw-r--r--arch/arm/mm/Kconfig21
-rw-r--r--arch/arm/mm/copypage-v4mc.S80
-rw-r--r--arch/arm/mm/copypage-v4mc.c111
-rw-r--r--arch/arm/mm/copypage-v6.c28
-rw-r--r--arch/arm/mm/flush.c37
-rw-r--r--arch/arm/mm/mm-armv.c27
-rw-r--r--arch/i386/kernel/cpu/amd.c7
-rw-r--r--arch/i386/kernel/cpu/common.c7
-rw-r--r--arch/i386/pci/fixup.c2
-rw-r--r--arch/ia64/ia32/ia32_ioctl.c1
-rw-r--r--arch/mips/vr41xx/common/pmu.c55
-rw-r--r--arch/ppc/kernel/setup.c2
-rw-r--r--arch/um/kernel/irq_user.c10
-rw-r--r--arch/x86_64/Kconfig14
-rw-r--r--arch/x86_64/defconfig58
-rw-r--r--arch/x86_64/kernel/Makefile1
-rw-r--r--arch/x86_64/kernel/apic.c5
-rw-r--r--arch/x86_64/kernel/entry.S11
-rw-r--r--arch/x86_64/kernel/io_apic.c70
-rw-r--r--arch/x86_64/kernel/mpparse.c22
-rw-r--r--arch/x86_64/kernel/nmi.c248
-rw-r--r--arch/x86_64/kernel/pmtimer.c101
-rw-r--r--arch/x86_64/kernel/ptrace.c13
-rw-r--r--arch/x86_64/kernel/setup.c17
-rw-r--r--arch/x86_64/kernel/smpboot.c262
-rw-r--r--arch/x86_64/kernel/time.c62
-rw-r--r--arch/x86_64/kernel/vsyscall.c5
-rw-r--r--arch/x86_64/mm/fault.c11
-rw-r--r--arch/x86_64/mm/ioremap.c2
32 files changed, 869 insertions, 436 deletions
diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c
index 64e450dddb4..167fd89f870 100644
--- a/arch/alpha/kernel/osf_sys.c
+++ b/arch/alpha/kernel/osf_sys.c
@@ -1150,16 +1150,13 @@ osf_usleep_thread(struct timeval32 __user *sleep, struct timeval32 __user *remai
if (get_tv32(&tmp, sleep))
goto fault;
- ticks = tmp.tv_usec;
- ticks = (ticks + (1000000 / HZ) - 1) / (1000000 / HZ);
- ticks += tmp.tv_sec * HZ;
+ ticks = timeval_to_jiffies(&tmp);
current->state = TASK_INTERRUPTIBLE;
ticks = schedule_timeout(ticks);
if (remain) {
- tmp.tv_sec = ticks / HZ;
- tmp.tv_usec = ticks % HZ;
+ jiffies_to_timeval(ticks, &tmp);
if (put_tv32(remain, &tmp))
goto fault;
}
diff --git a/arch/arm/mach-s3c2410/clock.c b/arch/arm/mach-s3c2410/clock.c
index e23f534d4e1..8d986b8401c 100644
--- a/arch/arm/mach-s3c2410/clock.c
+++ b/arch/arm/mach-s3c2410/clock.c
@@ -478,7 +478,7 @@ static int s3c2440_clk_add(struct sys_device *sysdev)
{
unsigned long upllcon = __raw_readl(S3C2410_UPLLCON);
- s3c2440_clk_upll.rate = s3c2410_get_pll(upllcon, clk_xtal.rate) * 2;
+ s3c2440_clk_upll.rate = s3c2410_get_pll(upllcon, clk_xtal.rate);
printk("S3C2440: Clock Support, UPLL %ld.%03ld MHz\n",
print_mhz(s3c2440_clk_upll.rate));
diff --git a/arch/arm/mach-s3c2410/s3c2440.c b/arch/arm/mach-s3c2410/s3c2440.c
index 9a8cc5ae225..d4c8281b55f 100644
--- a/arch/arm/mach-s3c2410/s3c2440.c
+++ b/arch/arm/mach-s3c2410/s3c2440.c
@@ -192,9 +192,11 @@ void __init s3c2440_map_io(struct map_desc *mach_desc, int size)
iotable_init(s3c2440_iodesc, ARRAY_SIZE(s3c2440_iodesc));
iotable_init(mach_desc, size);
+
/* rename any peripherals used differing from the s3c2410 */
- s3c_device_i2c.name = "s3c2440-i2c";
+ s3c_device_i2c.name = "s3c2440-i2c";
+ s3c_device_nand.name = "s3c2440-nand";
/* change irq for watchdog */
@@ -225,7 +227,7 @@ void __init s3c2440_init_clocks(int xtal)
break;
case S3C2440_CLKDIVN_HDIVN_2:
- hdiv = 1;
+ hdiv = 2;
break;
case S3C2440_CLKDIVN_HDIVN_4_8:
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig
index c4fc6be629d..48bac7da8c7 100644
--- a/arch/arm/mm/Kconfig
+++ b/arch/arm/mm/Kconfig
@@ -412,21 +412,20 @@ config CPU_BPREDICT_DISABLE
config TLS_REG_EMUL
bool
- default y if (SMP || CPU_32v6) && (CPU_32v5 || CPU_32v4 || CPU_32v3)
+ default y if SMP && (CPU_32v5 || CPU_32v4 || CPU_32v3)
help
- We might be running on an ARMv6+ processor which should have the TLS
- register but for some reason we can't use it, or maybe an SMP system
- using a pre-ARMv6 processor (there are apparently a few prototypes
- like that in existence) and therefore access to that register must
- be emulated.
+ An SMP system using a pre-ARMv6 processor (there are apparently
+ a few prototypes like that in existence) and therefore access to
+ that required register must be emulated.
config HAS_TLS_REG
bool
- depends on CPU_32v6
- default y if !TLS_REG_EMUL
+ depends on !TLS_REG_EMUL
+ default y if SMP || CPU_32v7
help
This selects support for the CP15 thread register.
- It is defined to be available on ARMv6 or later. If a particular
- ARMv6 or later CPU doesn't support it then it must omc;ide "select
- TLS_REG_EMUL" along with its other caracteristics.
+ It is defined to be available on some ARMv6 processors (including
+ all SMP capable ARMv6's) or later processors. User space may
+ assume directly accessing that register and always obtain the
+ expected value only on ARMv7 and above.
diff --git a/arch/arm/mm/copypage-v4mc.S b/arch/arm/mm/copypage-v4mc.S
deleted file mode 100644
index 305af3dab3d..00000000000
--- a/arch/arm/mm/copypage-v4mc.S
+++ /dev/null
@@ -1,80 +0,0 @@
-/*
- * linux/arch/arm/lib/copy_page-armv4mc.S
- *
- * Copyright (C) 1995-2001 Russell King
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- *
- * ASM optimised string functions
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/constants.h>
-
- .text
- .align 5
-/*
- * ARMv4 mini-dcache optimised copy_user_page
- *
- * We flush the destination cache lines just before we write the data into the
- * corresponding address. Since the Dcache is read-allocate, this removes the
- * Dcache aliasing issue. The writes will be forwarded to the write buffer,
- * and merged as appropriate.
- *
- * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
- * instruction. If your processor does not supply this, you have to write your
- * own copy_user_page that does the right thing.
- */
-ENTRY(v4_mc_copy_user_page)
- stmfd sp!, {r4, lr} @ 2
- mov r4, r0
- mov r0, r1
- bl map_page_minicache
- mov r1, #PAGE_SZ/64 @ 1
- ldmia r0!, {r2, r3, ip, lr} @ 4
-1: mcr p15, 0, r4, c7, c6, 1 @ 1 invalidate D line
- stmia r4!, {r2, r3, ip, lr} @ 4
- ldmia r0!, {r2, r3, ip, lr} @ 4+1
- stmia r4!, {r2, r3, ip, lr} @ 4
- ldmia r0!, {r2, r3, ip, lr} @ 4
- mcr p15, 0, r4, c7, c6, 1 @ 1 invalidate D line
- stmia r4!, {r2, r3, ip, lr} @ 4
- ldmia r0!, {r2, r3, ip, lr} @ 4
- subs r1, r1, #1 @ 1
- stmia r4!, {r2, r3, ip, lr} @ 4
- ldmneia r0!, {r2, r3, ip, lr} @ 4
- bne 1b @ 1
- ldmfd sp!, {r4, pc} @ 3
-
- .align 5
-/*
- * ARMv4 optimised clear_user_page
- *
- * Same story as above.
- */
-ENTRY(v4_mc_clear_user_page)
- str lr, [sp, #-4]!
- mov r1, #PAGE_SZ/64 @ 1
- mov r2, #0 @ 1
- mov r3, #0 @ 1
- mov ip, #0 @ 1
- mov lr, #0 @ 1
-1: mcr p15, 0, r0, c7, c6, 1 @ 1 invalidate D line
- stmia r0!, {r2, r3, ip, lr} @ 4
- stmia r0!, {r2, r3, ip, lr} @ 4
- mcr p15, 0, r0, c7, c6, 1 @ 1 invalidate D line
- stmia r0!, {r2, r3, ip, lr} @ 4
- stmia r0!, {r2, r3, ip, lr} @ 4
- subs r1, r1, #1 @ 1
- bne 1b @ 1
- ldr pc, [sp], #4
-
- __INITDATA
-
- .type v4_mc_user_fns, #object
-ENTRY(v4_mc_user_fns)
- .long v4_mc_clear_user_page
- .long v4_mc_copy_user_page
- .size v4_mc_user_fns, . - v4_mc_user_fns
diff --git a/arch/arm/mm/copypage-v4mc.c b/arch/arm/mm/copypage-v4mc.c
new file mode 100644
index 00000000000..fc69dccdace
--- /dev/null
+++ b/arch/arm/mm/copypage-v4mc.c
@@ -0,0 +1,111 @@
+/*
+ * linux/arch/arm/lib/copypage-armv4mc.S
+ *
+ * Copyright (C) 1995-2005 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This handles the mini data cache, as found on SA11x0 and XScale
+ * processors. When we copy a user page page, we map it in such a way
+ * that accesses to this page will not touch the main data cache, but
+ * will be cached in the mini data cache. This prevents us thrashing
+ * the main data cache on page faults.
+ */
+#include <linux/init.h>
+#include <linux/mm.h>
+
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/tlbflush.h>
+
+/*
+ * 0xffff8000 to 0xffffffff is reserved for any ARM architecture
+ * specific hacks for copying pages efficiently.
+ */
+#define minicache_pgprot __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | \
+ L_PTE_CACHEABLE)
+
+#define TOP_PTE(x) pte_offset_kernel(top_pmd, x)
+
+static DEFINE_SPINLOCK(minicache_lock);
+
+/*
+ * ARMv4 mini-dcache optimised copy_user_page
+ *
+ * We flush the destination cache lines just before we write the data into the
+ * corresponding address. Since the Dcache is read-allocate, this removes the
+ * Dcache aliasing issue. The writes will be forwarded to the write buffer,
+ * and merged as appropriate.
+ *
+ * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
+ * instruction. If your processor does not supply this, you have to write your
+ * own copy_user_page that does the right thing.
+ */
+static void __attribute__((naked))
+mc_copy_user_page(void *from, void *to)
+{
+ asm volatile(
+ "stmfd sp!, {r4, lr} @ 2\n\
+ mov r4, %2 @ 1\n\
+ ldmia %0!, {r2, r3, ip, lr} @ 4\n\
+1: mcr p15, 0, %1, c7, c6, 1 @ 1 invalidate D line\n\
+ stmia %1!, {r2, r3, ip, lr} @ 4\n\
+ ldmia %0!, {r2, r3, ip, lr} @ 4+1\n\
+ stmia %1!, {r2, r3, ip, lr} @ 4\n\
+ ldmia %0!, {r2, r3, ip, lr} @ 4\n\
+ mcr p15, 0, %1, c7, c6, 1 @ 1 invalidate D line\n\
+ stmia %1!, {r2, r3, ip, lr} @ 4\n\
+ ldmia %0!, {r2, r3, ip, lr} @ 4\n\
+ subs r4, r4, #1 @ 1\n\
+ stmia %1!, {r2, r3, ip, lr} @ 4\n\
+ ldmneia %0!, {r2, r3, ip, lr} @ 4\n\
+ bne 1b @ 1\n\
+ ldmfd sp!, {r4, pc} @ 3"
+ :
+ : "r" (from), "r" (to), "I" (PAGE_SIZE / 64));
+}
+
+void v4_mc_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+ spin_lock(&minicache_lock);
+
+ set_pte(TOP_PTE(0xffff8000), pfn_pte(__pa(kfrom) >> PAGE_SHIFT, minicache_pgprot));
+ flush_tlb_kernel_page(0xffff8000);
+
+ mc_copy_user_page((void *)0xffff8000, kto);
+
+ spin_unlock(&minicache_lock);
+}
+
+/*
+ * ARMv4 optimised clear_user_page
+ */
+void __attribute__((naked))
+v4_mc_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+ asm volatile(
+ "str lr, [sp, #-4]!\n\
+ mov r1, %0 @ 1\n\
+ mov r2, #0 @ 1\n\
+ mov r3, #0 @ 1\n\
+ mov ip, #0 @ 1\n\
+ mov lr, #0 @ 1\n\
+1: mcr p15, 0, r0, c7, c6, 1 @ 1 invalidate D line\n\
+ stmia r0!, {r2, r3, ip, lr} @ 4\n\
+ stmia r0!, {r2, r3, ip, lr} @ 4\n\
+ mcr p15, 0, r0, c7, c6, 1 @ 1 invalidate D line\n\
+ stmia r0!, {r2, r3, ip, lr} @ 4\n\
+ stmia r0!, {r2, r3, ip, lr} @ 4\n\
+ subs r1, r1, #1 @ 1\n\
+ bne 1b @ 1\n\
+ ldr pc, [sp], #4"
+ :
+ : "I" (PAGE_SIZE / 64));
+}
+
+struct cpu_user_fns v4_mc_user_fns __initdata = {
+ .cpu_clear_user_page = v4_mc_clear_user_page,
+ .cpu_copy_user_page = v4_mc_copy_user_page,
+};
diff --git a/arch/arm/mm/copypage-v6.c b/arch/arm/mm/copypage-v6.c
index 694ac820885..a8c00236bd3 100644
--- a/arch/arm/mm/copypage-v6.c
+++ b/arch/arm/mm/copypage-v6.c
@@ -26,8 +26,8 @@
#define to_address (0xffffc000)
#define to_pgprot PAGE_KERNEL
-static pte_t *from_pte;
-static pte_t *to_pte;
+#define TOP_PTE(x) pte_offset_kernel(top_pmd, x)
+
static DEFINE_SPINLOCK(v6_lock);
#define DCACHE_COLOUR(vaddr) ((vaddr & (SHMLBA - 1)) >> PAGE_SHIFT)
@@ -74,8 +74,8 @@ void v6_copy_user_page_aliasing(void *kto, const void *kfrom, unsigned long vadd
*/
spin_lock(&v6_lock);
- set_pte(from_pte + offset, pfn_pte(__pa(kfrom) >> PAGE_SHIFT, from_pgprot));
- set_pte(to_pte + offset, pfn_pte(__pa(kto) >> PAGE_SHIFT, to_pgprot));
+ set_pte(TOP_PTE(from_address) + offset, pfn_pte(__pa(kfrom) >> PAGE_SHIFT, from_pgprot));
+ set_pte(TOP_PTE(to_address) + offset, pfn_pte(__pa(kto) >> PAGE_SHIFT, to_pgprot));
from = from_address + (offset << PAGE_SHIFT);
to = to_address + (offset << PAGE_SHIFT);
@@ -114,7 +114,7 @@ void v6_clear_user_page_aliasing(void *kaddr, unsigned long vaddr)
*/
spin_lock(&v6_lock);
- set_pte(to_pte + offset, pfn_pte(__pa(kaddr) >> PAGE_SHIFT, to_pgprot));
+ set_pte(TOP_PTE(to_address) + offset, pfn_pte(__pa(kaddr) >> PAGE_SHIFT, to_pgprot));
flush_tlb_kernel_page(to);
clear_page((void *)to);
@@ -129,21 +129,6 @@ struct cpu_user_fns v6_user_fns __initdata = {
static int __init v6_userpage_init(void)
{
if (cache_is_vipt_aliasing()) {
- pgd_t *pgd;
- pmd_t *pmd;
-
- pgd = pgd_offset_k(from_address);
- pmd = pmd_alloc(&init_mm, pgd, from_address);
- if (!pmd)
- BUG();
- from_pte = pte_alloc_kernel(&init_mm, pmd, from_address);
- if (!from_pte)
- BUG();
-
- to_pte = pte_alloc_kernel(&init_mm, pmd, to_address);
- if (!to_pte)
- BUG();
-
cpu_user.cpu_clear_user_page = v6_clear_user_page_aliasing;
cpu_user.cpu_copy_user_page = v6_copy_user_page_aliasing;
}
@@ -151,5 +136,4 @@ static int __init v6_userpage_init(void)
return 0;
}
-__initcall(v6_userpage_init);
-
+core_initcall(v6_userpage_init);
diff --git a/arch/arm/mm/flush.c b/arch/arm/mm/flush.c
index c6de48d8950..4085ed983e4 100644
--- a/arch/arm/mm/flush.c
+++ b/arch/arm/mm/flush.c
@@ -13,6 +13,29 @@
#include <asm/cacheflush.h>
#include <asm/system.h>
+#include <asm/tlbflush.h>
+
+#ifdef CONFIG_CPU_CACHE_VIPT
+#define ALIAS_FLUSH_START 0xffff4000
+
+#define TOP_PTE(x) pte_offset_kernel(top_pmd, x)
+
+static void flush_pfn_alias(unsigned long pfn, unsigned long vaddr)
+{
+ unsigned long to = ALIAS_FLUSH_START + (CACHE_COLOUR(vaddr) << PAGE_SHIFT);
+
+ set_pte(TOP_PTE(to), pfn_pte(pfn, PAGE_KERNEL));
+ flush_tlb_kernel_page(to);
+
+ asm( "mcrr p15, 0, %1, %0, c14\n"
+ " mcrr p15, 0, %1, %0, c5\n"
+ :
+ : "r" (to), "r" (to + PAGE_SIZE - L1_CACHE_BYTES)
+ : "cc");
+}
+#else
+#define flush_pfn_alias(pfn,vaddr) do { } while (0)
+#endif
static void __flush_dcache_page(struct address_space *mapping, struct page *page)
{
@@ -37,6 +60,18 @@ static void __flush_dcache_page(struct address_space *mapping, struct page *page
return;
/*
+ * This is a page cache page. If we have a VIPT cache, we
+ * only need to do one flush - which would be at the relevant
+ * userspace colour, which is congruent with page->index.
+ */
+ if (cache_is_vipt()) {
+ if (cache_is_vipt_aliasing())
+ flush_pfn_alias(page_to_pfn(page),
+ page->index << PAGE_CACHE_SHIFT);
+ return;
+ }
+
+ /*
* There are possible user space mappings of this page:
* - VIVT cache: we need to also write back and invalidate all user
* data in the current VM view associated with this page.
@@ -57,8 +92,6 @@ static void __flush_dcache_page(struct address_space *mapping, struct page *page
continue;
offset = (pgoff - mpnt->vm_pgoff) << PAGE_SHIFT;
flush_cache_page(mpnt, mpnt->vm_start + offset, page_to_pfn(page));
- if (cache_is_vipt())
- break;
}
flush_dcache_mmap_unlock(mapping);
}
diff --git a/arch/arm/mm/mm-armv.c b/arch/arm/mm/mm-armv.c
index 585dfb8e20b..2c2b93d77d4 100644
--- a/arch/arm/mm/mm-armv.c
+++ b/arch/arm/mm/mm-armv.c
@@ -37,6 +37,8 @@ pgprot_t pgprot_kernel;
EXPORT_SYMBOL(pgprot_kernel);
+pmd_t *top_pmd;
+
struct cachepolicy {
const char policy[16];
unsigned int cr_mask;
@@ -142,6 +144,16 @@ __setup("noalign", noalign_setup);
#define FIRST_KERNEL_PGD_NR (FIRST_USER_PGD_NR + USER_PTRS_PER_PGD)
+static inline pmd_t *pmd_off(pgd_t *pgd, unsigned long virt)
+{
+ return pmd_offset(pgd, virt);
+}
+
+static inline pmd_t *pmd_off_k(unsigned long virt)
+{
+ return pmd_off(pgd_offset_k(virt), virt);
+}
+
/*
* need to get a 16k page for level 1
*/
@@ -220,7 +232,7 @@ void free_pgd_slow(pgd_t *pgd)
return;
/* pgd is always present and good */
- pmd = (pmd_t *)pgd;
+ pmd = pmd_off(pgd, 0);
if (pmd_none(*pmd))
goto free;
if (pmd_bad(*pmd)) {
@@ -246,9 +258,8 @@ free:
static inline void
alloc_init_section(unsigned long virt, unsigned long phys, int prot)
{
- pmd_t *pmdp;
+ pmd_t *pmdp = pmd_off_k(virt);
- pmdp = pmd_offset(pgd_offset_k(virt), virt);
if (virt & (1 << 20))
pmdp++;
@@ -283,11 +294,9 @@ alloc_init_supersection(unsigned long virt, unsigned long phys, int prot)
static inline void
alloc_init_page(unsigned long virt, unsigned long phys, unsigned int prot_l1, pgprot_t prot)
{
- pmd_t *pmdp;
+ pmd_t *pmdp = pmd_off_k(virt);
pte_t *ptep;
- pmdp = pmd_offset(pgd_offset_k(virt), virt);
-
if (pmd_none(*pmdp)) {
unsigned long pmdval;
ptep = alloc_bootmem_low_pages(2 * PTRS_PER_PTE *
@@ -310,7 +319,7 @@ alloc_init_page(unsigned long virt, unsigned long phys, unsigned int prot_l1, pg
*/
static inline void clear_mapping(unsigned long virt)
{
- pmd_clear(pmd_offset(pgd_offset_k(virt), virt));
+ pmd_clear(pmd_off_k(virt));
}
struct mem_types {
@@ -578,7 +587,7 @@ void setup_mm_for_reboot(char mode)
PMD_TYPE_SECT;
if (cpu_arch <= CPU_ARCH_ARMv5)
pmdval |= PMD_BIT4;
- pmd = pmd_offset(pgd + i, i << PGDIR_SHIFT);
+ pmd = pmd_off(pgd, i << PGDIR_SHIFT);
pmd[0] = __pmd(pmdval);
pmd[1] = __pmd(pmdval + (1 << (PGDIR_SHIFT - 1)));
flush_pmd_entry(pmd);
@@ -675,6 +684,8 @@ void __init memtable_init(struct meminfo *mi)
flush_cache_all();
flush_tlb_all();
+
+ top_pmd = pmd_off_k(0xffff0000);
}
/*
diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c
index 16dbc4151be..fa34a06c0d7 100644
--- a/arch/i386/kernel/cpu/amd.c
+++ b/arch/i386/kernel/cpu/amd.c
@@ -24,9 +24,6 @@ __asm__(".align 4\nvide: ret");
static void __init init_amd(struct cpuinfo_x86 *c)
{
-#ifdef CONFIG_X86_SMP
- int cpu = c == &boot_cpu_data ? 0 : c - cpu_data;
-#endif
u32 l, h;
int mbytes = num_physpages >> (20-PAGE_SHIFT);
int r;
@@ -205,7 +202,9 @@ static void __init init_amd(struct cpuinfo_x86 *c)
* of two.
*/
if (c->x86_num_cores > 1) {
- cpu_core_id[cpu] = cpu >> hweight32(c->x86_num_cores - 1);
+ int cpu = smp_processor_id();
+ /* Fix up the APIC ID following AMD specifications. */
+ cpu_core_id[cpu] >>= hweight32(c->x86_num_cores - 1);
printk(KERN_INFO "CPU %d(%d) -> Core %d\n",
cpu, c->x86_num_cores, cpu_core_id[cpu]);
}
diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c
index 6be0310e3cd..11e6e6f23fa 100644
--- a/arch/i386/kernel/cpu/common.c
+++ b/arch/i386/kernel/cpu/common.c
@@ -243,6 +243,13 @@ static void __init early_cpu_detect(void)
}
early_intel_workaround(c);
+
+#ifdef CONFIG_SMP
+#ifdef CONFIG_X86_HT
+ phys_proc_id[smp_processor_id()] =
+#endif
+ cpu_core_id[smp_processor_id()] = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
}
void __init generic_identify(struct cpuinfo_x86 * c)
diff --git a/arch/i386/pci/fixup.c b/arch/i386/pci/fixup.c
index be52c5ac4e0..8e8e895e1b5 100644
--- a/arch/i386/pci/fixup.c
+++ b/arch/i386/pci/fixup.c
@@ -253,7 +253,7 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE2, pci
#define MAX_PCIEROOT 6
static int quirk_aspm_offset[MAX_PCIEROOT << 3];
-#define GET_INDEX(a, b) (((a - PCI_DEVICE_ID_INTEL_MCH_PA) << 3) + b)
+#define GET_INDEX(a, b) ((((a) - PCI_DEVICE_ID_INTEL_MCH_PA) << 3) + ((b) & 7))
static int quirk_pcie_aspm_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
{
diff --git a/arch/ia64/ia32/ia32_ioctl.c b/arch/ia64/ia32/ia32_ioctl.c
index 9845dabe261..164b211f417 100644
--- a/arch/ia64/ia32/ia32_ioctl.c
+++ b/arch/ia64/ia32/ia32_ioctl.c
@@ -13,7 +13,6 @@
#define INCLUDES
#include "compat_ioctl.c"
-#include <asm/ioctl32.h>
#define IOCTL_NR(a) ((a) & ~(_IOC_SIZEMASK << _IOC_SIZESHIFT))
diff --git a/arch/mips/vr41xx/common/pmu.c b/arch/mips/vr41xx/common/pmu.c
index c5f1043de93..53166f3598b 100644
--- a/arch/mips/vr41xx/common/pmu.c
+++ b/arch/mips/vr41xx/common/pmu.c
@@ -1,7 +1,7 @@
/*
* pmu.c, Power Management Unit routines for NEC VR4100 series.
*
- * Copyright (C) 2003-2004 Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
+ * Copyright (C) 2003-2005 Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
*
* 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
@@ -17,7 +17,9 @@
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
+#include <linux/errno.h>
#include <linux/init.h>
+#include <linux/ioport.h>
#include <linux/kernel.h>
#include <linux/smp.h>
#include <linux/types.h>
@@ -27,20 +29,31 @@
#include <asm/reboot.h>
#include <asm/system.h>
-#define PMUCNT2REG KSEG1ADDR(0x0f0000c6)
+#define PMU_TYPE1_BASE 0x0b0000a0UL
+#define PMU_TYPE1_SIZE 0x0eUL
+
+#define PMU_TYPE2_BASE 0x0f0000c0UL
+#define PMU_TYPE2_SIZE 0x10UL
+
+#define PMUCNT2REG 0x06
#define SOFTRST 0x0010
+static void __iomem *pmu_base;
+
+#define pmu_read(offset) readw(pmu_base + (offset))
+#define pmu_write(offset, value) writew((value), pmu_base + (offset))
+
static inline void software_reset(void)
{
- uint16_t val;
+ uint16_t pmucnt2;
switch (current_cpu_data.cputype) {
case CPU_VR4122:
case CPU_VR4131:
case CPU_VR4133:
- val = readw(PMUCNT2REG);
- val |= SOFTRST;
- writew(val, PMUCNT2REG);
+ pmucnt2 = pmu_read(PMUCNT2REG);
+ pmucnt2 |= SOFTRST;
+ pmu_write(PMUCNT2REG, pmucnt2);
break;
default:
break;
@@ -71,6 +84,34 @@ static void vr41xx_power_off(void)
static int __init vr41xx_pmu_init(void)
{
+ unsigned long start, size;
+
+ switch (current_cpu_data.cputype) {
+ case CPU_VR4111:
+ case CPU_VR4121:
+ start = PMU_TYPE1_BASE;
+ size = PMU_TYPE1_SIZE;
+ break;
+ case CPU_VR4122:
+ case CPU_VR4131:
+ case CPU_VR4133:
+ start = PMU_TYPE2_BASE;
+ size = PMU_TYPE2_SIZE;
+ break;
+ default:
+ printk("Unexpected CPU of NEC VR4100 series\n");
+ return -ENODEV;
+ }
+
+ if (request_mem_region(start, size, "PMU") == NULL)
+ return -EBUSY;
+
+ pmu_base = ioremap(start, size);
+ if (pmu_base == NULL) {
+ release_mem_region(start, size);
+ return -EBUSY;
+ }
+
_machine_restart = vr41xx_restart;
_machine_halt = vr41xx_halt;
_machine_power_off = vr41xx_power_off;
@@ -78,4 +119,4 @@ static int __init vr41xx_pmu_init(void)
return 0;
}
-early_initcall(vr41xx_pmu_init);
+core_initcall(vr41xx_pmu_init);
diff --git a/arch/ppc/kernel/setup.c b/arch/ppc/kernel/setup.c
index 5dfb42f1a15..309797d7f96 100644
--- a/arch/ppc/kernel/setup.c
+++ b/arch/ppc/kernel/setup.c
@@ -753,6 +753,8 @@ void __init setup_arch(char **cmdline_p)
strlcpy(saved_command_line, cmd_line, COMMAND_LINE_SIZE);
*cmdline_p = cmd_line;
+ parse_early_param();
+
/* set up the bootmem stuff with available memory */
do_init_bootmem();
if ( ppc_md.progress ) ppc_md.progress("setup_arch: bootmem", 0x3eab);
diff --git a/arch/um/kernel/irq_user.c b/arch/um/kernel/irq_user.c
index 6d6f9484b88..b3074cbaa47 100644
--- a/arch/um/kernel/irq_user.c
+++ b/arch/um/kernel/irq_user.c
@@ -236,9 +236,15 @@ static void free_irq_by_cb(int (*test)(struct irq_fd *, void *), void *arg)
(*prev)->fd, pollfds[i].fd);
goto out;
}
- memcpy(&pollfds[i], &pollfds[i + 1],
- (pollfds_num - i - 1) * sizeof(pollfds[0]));
+
pollfds_num--;
+
+ /* This moves the *whole* array after pollfds[i] (though
+ * it doesn't spot as such)! */
+
+ memmove(&pollfds[i], &pollfds[i + 1],
+ (pollfds_num - i) * sizeof(pollfds[0]));
+
if(last_irq_ptr == &old_fd->next)
last_irq_ptr = prev;
*prev = (*prev)->next;
diff --git a/arch/x86_64/Kconfig b/arch/x86_64/Kconfig
index 44ee7f6acf7..82cb2a3f127 100644
--- a/arch/x86_64/Kconfig
+++ b/arch/x86_64/Kconfig
@@ -303,6 +303,20 @@ config HPET_TIMER
as it is off-chip. You can find the HPET spec at
<http://www.intel.com/labs/platcomp/hpet/hpetspec.htm>.
+config X86_PM_TIMER
+ bool "PM timer"
+ default y
+ help
+ Support the ACPI PM timer for time keeping. This is slow,
+ but is useful on some chipsets without HPET on systems with more
+ than one CPU. On a single processor or single socket multi core
+ system it is normally not required.
+ When the PM timer is active 64bit vsyscalls are disabled
+ and should not be enabled (/proc/sys/kernel/vsyscall64 should
+ not be changed).
+ The kernel selects the PM timer only as a last resort, so it is
+ useful to enable just in case.
+
config HPET_EMULATE_RTC
bool "Provide RTC interrupt"
depends on HPET_TIMER && RTC=y
diff --git a/arch/x86_64/defconfig b/arch/x86_64/defconfig
index 9ce51dee30b..569595b74c7 100644
--- a/arch/x86_64/defconfig
+++ b/arch/x86_64/defconfig
@@ -1,7 +1,7 @@
#
# Automatically generated make config: don't edit
-# Linux kernel version: 2.6.11-bk7
-# Sat Mar 12 23:43:44 2005
+# Linux kernel version: 2.6.12-rc4
+# Fri May 13 06:39:11 2005
#
CONFIG_X86_64=y
CONFIG_64BIT=y
@@ -11,8 +11,6 @@ CONFIG_RWSEM_GENERIC_SPINLOCK=y
CONFIG_GENERIC_CALIBRATE_DELAY=y
CONFIG_X86_CMPXCHG=y
CONFIG_EARLY_PRINTK=y
-CONFIG_HPET_TIMER=y
-CONFIG_HPET_EMULATE_RTC=y
CONFIG_GENERIC_ISA_DMA=y
CONFIG_GENERIC_IOMAP=y
@@ -22,6 +20,7 @@ CONFIG_GENERIC_IOMAP=y
CONFIG_EXPERIMENTAL=y
CONFIG_CLEAN_COMPILE=y
CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
#
# General setup
@@ -33,7 +32,6 @@ CONFIG_POSIX_MQUEUE=y
# CONFIG_BSD_PROCESS_ACCT is not set
CONFIG_SYSCTL=y
# CONFIG_AUDIT is not set
-CONFIG_LOG_BUF_SHIFT=18
# CONFIG_HOTPLUG is not set
CONFIG_KOBJECT_UEVENT=y
CONFIG_IKCONFIG=y
@@ -43,10 +41,11 @@ CONFIG_IKCONFIG_PROC=y
CONFIG_KALLSYMS=y
CONFIG_KALLSYMS_ALL=y
# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_PRINTK=y
+CONFIG_BUG=y
CONFIG_BASE_FULL=y
CONFIG_FUTEX=y
CONFIG_EPOLL=y
-# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
CONFIG_SHMEM=y
CONFIG_CC_ALIGN_FUNCTIONS=0
CONFIG_CC_ALIGN_LABELS=0
@@ -93,6 +92,9 @@ CONFIG_DISCONTIGMEM=y
CONFIG_NUMA=y
CONFIG_HAVE_DEC_LOCK=y
CONFIG_NR_CPUS=8
+CONFIG_HPET_TIMER=y
+CONFIG_X86_PM_TIMER=y
+CONFIG_HPET_EMULATE_RTC=y
CONFIG_GART_IOMMU=y
CONFIG_SWIOTLB=y
CONFIG_X86_MCE=y
@@ -100,6 +102,7 @@ CONFIG_X86_MCE_INTEL=y
CONFIG_SECCOMP=y
CONFIG_GENERIC_HARDIRQS=y
CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_ISA_DMA_API=y
#
# Power management options
@@ -129,7 +132,7 @@ CONFIG_ACPI_NUMA=y
# CONFIG_ACPI_IBM is not set
CONFIG_ACPI_TOSHIBA=y
CONFIG_ACPI_BLACKLIST_YEAR=2001
-CONFIG_ACPI_DEBUG=y
+# CONFIG_ACPI_DEBUG is not set
CONFIG_ACPI_BUS=y
CONFIG_ACPI_EC=y
CONFIG_ACPI_POWER=y
@@ -141,6 +144,7 @@ CONFIG_ACPI_SYSTEM=y
# CPU Frequency scaling
#
CONFIG_CPU_FREQ=y
+CONFIG_CPU_FREQ_TABLE=y
# CONFIG_CPU_FREQ_DEBUG is not set
CONFIG_CPU_FREQ_STAT=y
# CONFIG_CPU_FREQ_STAT_DETAILS is not set
@@ -150,7 +154,6 @@ CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
# CONFIG_CPU_FREQ_GOV_POWERSAVE is not set
CONFIG_CPU_FREQ_GOV_USERSPACE=y
CONFIG_CPU_FREQ_GOV_ONDEMAND=y
-CONFIG_CPU_FREQ_TABLE=y
#
# CPUFreq processor drivers
@@ -164,6 +167,7 @@ CONFIG_X86_ACPI_CPUFREQ=y
# shared options
#
CONFIG_X86_ACPI_CPUFREQ_PROC_INTF=y
+# CONFIG_X86_SPEEDSTEP_LIB is not set
#
# Bus options (PCI etc.)
@@ -172,9 +176,11 @@ CONFIG_PCI=y
CONFIG_PCI_DIRECT=y
CONFIG_PCI_MMCONFIG=y
CONFIG_UNORDERED_IO=y
+# CONFIG_PCIEPORTBUS is not set
CONFIG_PCI_MSI=y
# CONFIG_PCI_LEGACY_PROC is not set
# CONFIG_PCI_NAMES is not set
+# CONFIG_PCI_DEBUG is not set
#
# PCCARD (PCMCIA/CardBus) support
@@ -182,10 +188,6 @@ CONFIG_PCI_MSI=y
# CONFIG_PCCARD is not set
#
-# PC-card bridges
-#
-
-#
# PCI Hotplug Support
#
# CONFIG_HOTPLUG_PCI is not set
@@ -254,7 +256,7 @@ CONFIG_LBD=y
# IO Schedulers
#
CONFIG_IOSCHED_NOOP=y
-CONFIG_IOSCHED_AS=y
+# CONFIG_IOSCHED_AS is not set
CONFIG_IOSCHED_DEADLINE=y
CONFIG_IOSCHED_CFQ=y
# CONFIG_ATA_OVER_ETH is not set
@@ -308,7 +310,8 @@ CONFIG_BLK_DEV_AMD74XX=y
CONFIG_BLK_DEV_PIIX=y
# CONFIG_BLK_DEV_NS87415 is not set
# CONFIG_BLK_DEV_PDC202XX_OLD is not set
-# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+CONFIG_BLK_DEV_PDC202XX_NEW=y
+# CONFIG_PDC202XX_FORCE is not set
# CONFIG_BLK_DEV_SVWKS is not set
# CONFIG_BLK_DEV_SIIMAGE is not set
# CONFIG_BLK_DEV_SIS5513 is not set
@@ -353,7 +356,7 @@ CONFIG_BLK_DEV_SD=y
#
# SCSI low-level drivers
#
-CONFIG_BLK_DEV_3W_XXXX_RAID=y
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
# CONFIG_SCSI_3W_9XXX is not set
# CONFIG_SCSI_ACARD is not set
# CONFIG_SCSI_AACRAID is not set
@@ -384,7 +387,6 @@ CONFIG_SCSI_SATA_VIA=y
# CONFIG_SCSI_BUSLOGIC is not set
# CONFIG_SCSI_DMX3191D is not set
# CONFIG_SCSI_EATA is not set
-# CONFIG_SCSI_EATA_PIO is not set
# CONFIG_SCSI_FUTURE_DOMAIN is not set
# CONFIG_SCSI_GDTH is not set
# CONFIG_SCSI_IPS is not set
@@ -392,7 +394,6 @@ CONFIG_SCSI_SATA_VIA=y
# CONFIG_SCSI_INIA100 is not set
# CONFIG_SCSI_SYM53C8XX_2 is not set
# CONFIG_SCSI_IPR is not set
-# CONFIG_SCSI_QLOGIC_ISP is not set
# CONFIG_SCSI_QLOGIC_FC is not set
# CONFIG_SCSI_QLOGIC_1280 is not set
CONFIG_SCSI_QLA2XXX=y
@@ -401,6 +402,7 @@ CONFIG_SCSI_QLA2XXX=y
# CONFIG_SCSI_QLA2300 is not set
# CONFIG_SCSI_QLA2322 is not set
# CONFIG_SCSI_QLA6312 is not set
+# CONFIG_SCSI_LPFC is not set
# CONFIG_SCSI_DC395x is not set
# CONFIG_SCSI_DC390T is not set
# CONFIG_SCSI_DEBUG is not set
@@ -437,7 +439,6 @@ CONFIG_NET=y
#
CONFIG_PACKET=y
# CONFIG_PACKET_MMAP is not set
-# CONFIG_NETLINK_DEV is not set
CONFIG_UNIX=y
# CONFIG_NET_KEY is not set
CONFIG_INET=y
@@ -502,7 +503,7 @@ CONFIG_NETDEVICES=y
# CONFIG_DUMMY is not set
# CONFIG_BONDING is not set
# CONFIG_EQUALIZER is not set
-# CONFIG_TUN is not set
+CONFIG_TUN=y
#
# ARCnet devices
@@ -525,8 +526,7 @@ CONFIG_MII=y
# CONFIG_HP100 is not set
CONFIG_NET_PCI=y
# CONFIG_PCNET32 is not set
-CONFIG_AMD8111_ETH=y
-# CONFIG_AMD8111E_NAPI is not set
+# CONFIG_AMD8111_ETH is not set
# CONFIG_ADAPTEC_STARFIRE is not set
# CONFIG_B44 is not set
CONFIG_FORCEDETH=y
@@ -536,7 +536,7 @@ CONFIG_FORCEDETH=y
# CONFIG_FEALNX is not set
# CONFIG_NATSEMI is not set
# CONFIG_NE2K_PCI is not set
-CONFIG_8139CP=m
+CONFIG_8139CP=y
CONFIG_8139TOO=y
# CONFIG_8139TOO_PIO is not set
# CONFIG_8139TOO_TUNE_TWISTER is not set
@@ -671,6 +671,7 @@ CONFIG_SERIAL_8250_NR_UARTS=4
#
CONFIG_SERIAL_CORE=y
CONFIG_SERIAL_CORE_CONSOLE=y
+# CONFIG_SERIAL_JSM is not set
CONFIG_UNIX98_PTYS=y
CONFIG_LEGACY_PTYS=y
CONFIG_LEGACY_PTY_COUNT=256
@@ -696,6 +697,7 @@ CONFIG_RTC=y
#
CONFIG_AGP=y
CONFIG_AGP_AMD64=y
+CONFIG_AGP_INTEL=y
# CONFIG_DRM is not set
# CONFIG_MWAVE is not set
CONFIG_RAW_DRIVER=y
@@ -703,7 +705,7 @@ CONFIG_HPET=y
# CONFIG_HPET_RTC_IRQ is not set
CONFIG_HPET_MMAP=y
CONFIG_MAX_RAW_DEVS=256
-CONFIG_HANGCHECK_TIMER=y
+# CONFIG_HANGCHECK_TIMER is not set
#
# TPM devices
@@ -786,6 +788,8 @@ CONFIG_SOUND_ICH=y
#
# USB support
#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
CONFIG_USB=y
# CONFIG_USB_DEBUG is not set
@@ -797,8 +801,6 @@ CONFIG_USB_DEVICEFS=y
# CONFIG_USB_DYNAMIC_MINORS is not set
# CONFIG_USB_SUSPEND is not set
# CONFIG_USB_OTG is not set
-CONFIG_USB_ARCH_HAS_HCD=y
-CONFIG_USB_ARCH_HAS_OHCI=y
#
# USB Host Controller Drivers
@@ -826,7 +828,6 @@ CONFIG_USB_PRINTER=y
#
CONFIG_USB_STORAGE=y
# CONFIG_USB_STORAGE_DEBUG is not set
-# CONFIG_USB_STORAGE_RW_DETECT is not set
# CONFIG_USB_STORAGE_DATAFAB is not set
# CONFIG_USB_STORAGE_FREECOM is not set
# CONFIG_USB_STORAGE_ISD200 is not set
@@ -965,7 +966,7 @@ CONFIG_AUTOFS_FS=y
# CD-ROM/DVD Filesystems
#
CONFIG_ISO9660_FS=y
-# CONFIG_JOLIET is not set
+CONFIG_JOLIET=y
# CONFIG_ZISOFS is not set
# CONFIG_UDF_FS is not set
@@ -1092,9 +1093,10 @@ CONFIG_OPROFILE=y
#
# Kernel hacking
#
+# CONFIG_PRINTK_TIME is not set
CONFIG_DEBUG_KERNEL=y
CONFIG_MAGIC_SYSRQ=y
-# CONFIG_PRINTK_TIME is not set
+CONFIG_LOG_BUF_SHIFT=18
# CONFIG_SCHEDSTATS is not set
# CONFIG_DEBUG_SLAB is not set
# CONFIG_DEBUG_SPINLOCK is not set
diff --git a/arch/x86_64/kernel/Makefile b/arch/x86_64/kernel/Makefile
index 0a3318e08ab..5ca4a4598fd 100644
--- a/arch/x86_64/kernel/Makefile
+++ b/arch/x86_64/kernel/Makefile
@@ -28,6 +28,7 @@ obj-$(CONFIG_GART_IOMMU) += pci-gart.o aperture.o
obj-$(CONFIG_DUMMY_IOMMU) += pci-nommu.o pci-dma.o
obj-$(CONFIG_SWIOTLB) += swiotlb.o
obj-$(CONFIG_KPROBES) += kprobes.o
+obj-$(CONFIG_X86_PM_TIMER) += pmtimer.o
obj-$(CONFIG_MODULES) += module.o
diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c
index 7e13545748e..f8e6cc4fecd 100644
--- a/arch/x86_64/kernel/apic.c
+++ b/arch/x86_64/kernel/apic.c
@@ -33,6 +33,7 @@
#include <asm/mpspec.h>
#include <asm/pgalloc.h>
#include <asm/mach_apic.h>
+#include <asm/nmi.h>
int apic_verbosity;
@@ -925,7 +926,7 @@ __init int oem_force_hpet_timer(void)
unsigned id;
DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS);
- bitmap_empty(clustermap, NUM_APIC_CLUSTERS);
+ bitmap_zero(clustermap, NUM_APIC_CLUSTERS);
for (i = 0; i < NR_CPUS; i++) {
id = bios_cpu_apicid[i];
@@ -1056,7 +1057,7 @@ int __init APIC_init_uniprocessor (void)
nr_ioapics = 0;
#endif
setup_boot_APIC_clock();
-
+ check_nmi_watchdog();
return 0;
}
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 1086b5fcac2..28817490fdc 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -220,13 +220,18 @@ sysret_careful:
jmp sysret_check
/* Handle a signal */
- /* edx: work flags (arg3) */
sysret_signal:
sti
+ testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+ jz 1f
+
+ /* Really a signal */
+ /* edx: work flags (arg3) */
leaq do_notify_resume(%rip),%rax
leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1
xorl %esi,%esi # oldset -> arg2
call ptregscall_common
+1: movl $_TIF_NEED_RESCHED,%edi
jmp sysret_check
/* Do syscall tracing */
@@ -484,6 +489,8 @@ retint_careful:
jmp retint_check
retint_signal:
+ testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+ jz retint_swapgs
sti
SAVE_REST
movq $-1,ORIG_RAX(%rsp)
@@ -492,8 +499,8 @@ retint_signal:
call do_notify_resume
RESTORE_REST
cli
+ movl $_TIF_NEED_RESCHED,%edi
GET_THREAD_INFO(%rcx)
- movl $_TIF_WORK_MASK,%edi
jmp retint_check
#ifdef CONFIG_PREEMPT
diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c
index 60be58617eb..ac768432495 100644
--- a/arch/x86_64/kernel/io_apic.c
+++ b/arch/x86_64/kernel/io_apic.c
@@ -1804,76 +1804,6 @@ device_initcall(ioapic_init_sysfs);
#define IO_APIC_MAX_ID 0xFE
-int __init io_apic_get_unique_id (int ioapic, int apic_id)
-{
- union IO_APIC_reg_00 reg_00;
- static physid_mask_t apic_id_map;
- unsigned long flags;
- int i = 0;
-
- /*
- * The P4 platform supports up to 256 APIC IDs on two separate APIC
- * buses (one for LAPICs, one for IOAPICs), where predecessors only
- * supports up to 16 on one shared APIC bus.
- *
- * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full
- * advantage of new APIC bus architecture.
- */
-
- if (physids_empty(apic_id_map))
- apic_id_map = phys_cpu_present_map;
-
- spin_lock_irqsave(&ioapic_lock, flags);
- reg_00.raw = io_apic_read(ioapic, 0);
- spin_unlock_irqrestore(&ioapic_lock, flags);
-
- if (apic_id >= IO_APIC_MAX_ID) {
- apic_printk(APIC_QUIET, KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying "
- "%d\n", ioapic, apic_id, reg_00.bits.ID);
- apic_id = reg_00.bits.ID;
- }
-
- /*
- * Every APIC in a system must have a unique ID or we get lots of nice
- * 'stuck on smp_invalidate_needed IPI wait' messages.
- */
- if (physid_isset(apic_id, apic_id_map)) {
-
- for (i = 0; i < IO_APIC_MAX_ID; i++) {
- if (!physid_isset(i, apic_id_map))
- break;
- }
-
- if (i == IO_APIC_MAX_ID)
- panic("Max apic_id exceeded!\n");
-
- apic_printk(APIC_VERBOSE, KERN_WARNING "IOAPIC[%d]: apic_id %d already used, "
- "trying %d\n", ioapic, apic_id, i);
-
- apic_id = i;
- }
-
- physid_set(apic_id, apic_id_map);
-
- if (reg_00.bits.ID != apic_id) {
- reg_00.bits.ID = apic_id;
-
- spin_lock_irqsave(&ioapic_lock, flags);
- io_apic_write(ioapic, 0, reg_00.raw);
- reg_00.raw = io_apic_read(ioapic, 0);
- spin_unlock_irqrestore(&ioapic_lock, flags);
-
- /* Sanity check */
- if (reg_00.bits.ID != apic_id)
- panic("IOAPIC[%d]: Unable change apic_id!\n", ioapic);
- }
-
- apic_printk(APIC_VERBOSE,KERN_INFO "IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id);
-
- return apic_id;
-}
-
-
int __init io_apic_get_version (int ioapic)
{
union IO_APIC_reg_01 reg_01;
diff --git a/arch/x86_64/kernel/mpparse.c b/arch/x86_64/kernel/mpparse.c
index 7ec031c6ca1..f86d9db94bf 100644
--- a/arch/x86_64/kernel/mpparse.c
+++ b/arch/x86_64/kernel/mpparse.c
@@ -107,6 +107,7 @@ static int __init mpf_checksum(unsigned char *mp, int len)
static void __init MP_processor_info (struct mpc_config_processor *m)
{
int ver;
+ static int found_bsp=0;
if (!(m->mpc_cpuflag & CPU_ENABLED))
return;
@@ -126,11 +127,6 @@ static void __init MP_processor_info (struct mpc_config_processor *m)
" Processor ignored.\n", NR_CPUS);
return;
}
- if (num_processors >= maxcpus) {
- printk(KERN_WARNING "WARNING: maxcpus limit of %i reached."
- " Processor ignored.\n", maxcpus);
- return;
- }
num_processors++;
@@ -150,7 +146,19 @@ static void __init MP_processor_info (struct mpc_config_processor *m)
ver = 0x10;
}
apic_version[m->mpc_apicid] = ver;
- bios_cpu_apicid[num_processors - 1] = m->mpc_apicid;
+ if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) {
+ /*
+ * bios_cpu_apicid is required to have processors listed
+ * in same order as logical cpu numbers. Hence the first
+ * entry is BSP, and so on.
+ */
+ bios_cpu_apicid[0] = m->mpc_apicid;
+ x86_cpu_to_apicid[0] = m->mpc_apicid;
+ found_bsp = 1;
+ } else {
+ bios_cpu_apicid[num_processors - found_bsp] = m->mpc_apicid;
+ x86_cpu_to_apicid[num_processors - found_bsp] = m->mpc_apicid;
+ }
}
static void __init MP_bus_info (struct mpc_config_bus *m)
@@ -759,7 +767,7 @@ void __init mp_register_ioapic (
mp_ioapics[idx].mpc_apicaddr = address;
set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
- mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id);
+ mp_ioapics[idx].mpc_apicid = id;
mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx);
/*
diff --git a/arch/x86_64/kernel/nmi.c b/arch/x86_64/kernel/nmi.c
index 61de0b34a01..31c0f2e6ac9 100644
--- a/arch/x86_64/kernel/nmi.c
+++ b/arch/x86_64/kernel/nmi.c
@@ -33,6 +33,7 @@
#include <asm/msr.h>
#include <asm/proto.h>
#include <asm/kdebug.h>
+#include <asm/local.h>
/*
* lapic_nmi_owner tracks the ownership of the lapic NMI hardware:
@@ -59,7 +60,8 @@ int panic_on_timeout;
unsigned int nmi_watchdog = NMI_DEFAULT;
static unsigned int nmi_hz = HZ;
-unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */
+static unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */
+static unsigned int nmi_p4_cccr_val;
/* Note that these events don't tick when the CPU idles. This means
the frequency varies with CPU load. */
@@ -71,61 +73,87 @@ unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */
#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76
#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING
-#define P6_EVNTSEL0_ENABLE (1 << 22)
-#define P6_EVNTSEL_INT (1 << 20)
-#define P6_EVNTSEL_OS (1 << 17)
-#define P6_EVNTSEL_USR (1 << 16)
-#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79
-#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED
+#define MSR_P4_MISC_ENABLE 0x1A0
+#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7)
+#define MSR_P4_MISC_ENABLE_PEBS_UNAVAIL (1<<12)
+#define MSR_P4_PERFCTR0 0x300
+#define MSR_P4_CCCR0 0x360
+#define P4_ESCR_EVENT_SELECT(N) ((N)<<25)
+#define P4_ESCR_OS (1<<3)
+#define P4_ESCR_USR (1<<2)
+#define P4_CCCR_OVF_PMI0 (1<<26)
+#define P4_CCCR_OVF_PMI1 (1<<27)
+#define P4_CCCR_THRESHOLD(N) ((N)<<20)
+#define P4_CCCR_COMPLEMENT (1<<19)
+#define P4_CCCR_COMPARE (1<<18)
+#define P4_CCCR_REQUIRED (3<<16)
+#define P4_CCCR_ESCR_SELECT(N) ((N)<<13)
+#define P4_CCCR_ENABLE (1<<12)
+/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter
+ CRU_ESCR0 (with any non-null event selector) through a complemented
+ max threshold. [IA32-Vol3, Section 14.9.9] */
+#define MSR_P4_IQ_COUNTER0 0x30C
+#define P4_NMI_CRU_ESCR0 (P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR)
+#define P4_NMI_IQ_CCCR0 \
+ (P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT| \
+ P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE)
+
+static __init inline int nmi_known_cpu(void)
+{
+ switch (boot_cpu_data.x86_vendor) {
+ case X86_VENDOR_AMD:
+ return boot_cpu_data.x86 == 15;
+ case X86_VENDOR_INTEL:
+ return boot_cpu_data.x86 == 15;
+ }
+ return 0;
+}
/* Run after command line and cpu_init init, but before all other checks */
void __init nmi_watchdog_default(void)
{
if (nmi_watchdog != NMI_DEFAULT)
return;
-
- /* For some reason the IO APIC watchdog doesn't work on the AMD
- 8111 chipset. For now switch to local APIC mode using
- perfctr0 there. On Intel CPUs we don't have code to handle
- the perfctr and the IO-APIC seems to work, so use that. */
-
- if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
- nmi_watchdog = NMI_LOCAL_APIC;
- printk(KERN_INFO
- "Using local APIC NMI watchdog using perfctr0\n");
- } else {
- printk(KERN_INFO "Using IO APIC NMI watchdog\n");
+ if (nmi_known_cpu())
+ nmi_watchdog = NMI_LOCAL_APIC;
+ else
nmi_watchdog = NMI_IO_APIC;
- }
}
-/* Why is there no CPUID flag for this? */
-static __init int cpu_has_lapic(void)
+#ifdef CONFIG_SMP
+/* The performance counters used by NMI_LOCAL_APIC don't trigger when
+ * the CPU is idle. To make sure the NMI watchdog really ticks on all
+ * CPUs during the test make them busy.
+ */
+static __init void nmi_cpu_busy(void *data)
{
- switch (boot_cpu_data.x86_vendor) {
- case X86_VENDOR_INTEL:
- case X86_VENDOR_AMD:
- return boot_cpu_data.x86 >= 6;
- /* .... add more cpus here or find a different way to figure this out. */
- default:
- return 0;
- }
+ volatile int *endflag = data;
+ local_irq_enable();
+ /* Intentionally don't use cpu_relax here. This is
+ to make sure that the performance counter really ticks,
+ even if there is a simulator or similar that catches the
+ pause instruction. On a real HT machine this is fine because
+ all other CPUs are busy with "useless" delay loops and don't
+ care if they get somewhat less cycles. */
+ while (*endflag == 0)
+ barrier();
}
+#endif
-static int __init check_nmi_watchdog (void)
+int __init check_nmi_watchdog (void)
{
- int counts[NR_CPUS];
+ volatile int endflag = 0;
+ int *counts;
int cpu;
- if (nmi_watchdog == NMI_NONE)
- return 0;
+ counts = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL);
+ if (!counts)
+ return -1;
- if (nmi_watchdog == NMI_LOCAL_APIC && !cpu_has_lapic()) {
- nmi_watchdog = NMI_NONE;
- return -1;
- }
+ printk(KERN_INFO "testing NMI watchdog ... ");
- printk(KERN_INFO "Testing NMI watchdog ... ");
+ if (nmi_watchdog == NMI_LOCAL_APIC)
+ smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0);
for (cpu = 0; cpu < NR_CPUS; cpu++)
counts[cpu] = cpu_pda[cpu].__nmi_count;
@@ -133,15 +161,22 @@ static int __init check_nmi_watchdog (void)
mdelay((10*1000)/nmi_hz); // wait 10 ticks
for (cpu = 0; cpu < NR_CPUS; cpu++) {
+ if (!cpu_online(cpu))
+ continue;
if (cpu_pda[cpu].__nmi_count - counts[cpu] <= 5) {
- printk("CPU#%d: NMI appears to be stuck (%d)!\n",
+ endflag = 1;
+ printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n",
cpu,
+ counts[cpu],
cpu_pda[cpu].__nmi_count);
nmi_active = 0;
lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG;
+ nmi_perfctr_msr = 0;
+ kfree(counts);
return -1;
}
}
+ endflag = 1;
printk("OK.\n");
/* now that we know it works we can reduce NMI frequency to
@@ -149,10 +184,9 @@ static int __init check_nmi_watchdog (void)
if (nmi_watchdog == NMI_LOCAL_APIC)
nmi_hz = 1;
+ kfree(counts);
return 0;
}
-/* Have this called later during boot so counters are updating */
-late_initcall(check_nmi_watchdog);
int __init setup_nmi_watchdog(char *str)
{
@@ -170,7 +204,7 @@ int __init setup_nmi_watchdog(char *str)
if (nmi >= NMI_INVALID)
return 0;
- nmi_watchdog = nmi;
+ nmi_watchdog = nmi;
return 1;
}
@@ -185,7 +219,10 @@ static void disable_lapic_nmi_watchdog(void)
wrmsr(MSR_K7_EVNTSEL0, 0, 0);
break;
case X86_VENDOR_INTEL:
- wrmsr(MSR_IA32_EVNTSEL0, 0, 0);
+ if (boot_cpu_data.x86 == 15) {
+ wrmsr(MSR_P4_IQ_CCCR0, 0, 0);
+ wrmsr(MSR_P4_CRU_ESCR0, 0, 0);
+ }
break;
}
nmi_active = -1;
@@ -253,7 +290,7 @@ void enable_timer_nmi_watchdog(void)
static int nmi_pm_active; /* nmi_active before suspend */
-static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state)
+static int lapic_nmi_suspend(struct sys_device *dev, u32 state)
{
nmi_pm_active = nmi_active;
disable_lapic_nmi_watchdog();
@@ -300,22 +337,27 @@ late_initcall(init_lapic_nmi_sysfs);
* Original code written by Keith Owens.
*/
+static void clear_msr_range(unsigned int base, unsigned int n)
+{
+ unsigned int i;
+
+ for(i = 0; i < n; ++i)
+ wrmsr(base+i, 0, 0);
+}
+
static void setup_k7_watchdog(void)
{
int i;
unsigned int evntsel;
- /* No check, so can start with slow frequency */
- nmi_hz = 1;
-
- /* XXX should check these in EFER */
-
nmi_perfctr_msr = MSR_K7_PERFCTR0;
for(i = 0; i < 4; ++i) {
/* Simulator may not support it */
- if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL))
+ if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL)) {
+ nmi_perfctr_msr = 0;
return;
+ }
wrmsrl(MSR_K7_PERFCTR0+i, 0UL);
}
@@ -325,12 +367,54 @@ static void setup_k7_watchdog(void)
| K7_NMI_EVENT;
wrmsr(MSR_K7_EVNTSEL0, evntsel, 0);
- wrmsrl(MSR_K7_PERFCTR0, -((u64)cpu_khz*1000) / nmi_hz);
+ wrmsr(MSR_K7_PERFCTR0, -(cpu_khz/nmi_hz*1000), -1);
apic_write(APIC_LVTPC, APIC_DM_NMI);
evntsel |= K7_EVNTSEL_ENABLE;
wrmsr(MSR_K7_EVNTSEL0, evntsel, 0);
}
+
+static int setup_p4_watchdog(void)
+{
+ unsigned int misc_enable, dummy;
+
+ rdmsr(MSR_P4_MISC_ENABLE, misc_enable, dummy);
+ if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL))
+ return 0;
+
+ nmi_perfctr_msr = MSR_P4_IQ_COUNTER0;
+ nmi_p4_cccr_val = P4_NMI_IQ_CCCR0;
+#ifdef CONFIG_SMP
+ if (smp_num_siblings == 2)
+ nmi_p4_cccr_val |= P4_CCCR_OVF_PMI1;
+#endif
+
+ if (!(misc_enable & MSR_P4_MISC_ENABLE_PEBS_UNAVAIL))
+ clear_msr_range(0x3F1, 2);
+ /* MSR 0x3F0 seems to have a default value of 0xFC00, but current
+ docs doesn't fully define it, so leave it alone for now. */
+ if (boot_cpu_data.x86_model >= 0x3) {
+ /* MSR_P4_IQ_ESCR0/1 (0x3ba/0x3bb) removed */
+ clear_msr_range(0x3A0, 26);
+ clear_msr_range(0x3BC, 3);
+ } else {
+ clear_msr_range(0x3A0, 31);
+ }
+ clear_msr_range(0x3C0, 6);
+ clear_msr_range(0x3C8, 6);
+ clear_msr_range(0x3E0, 2);
+ clear_msr_range(MSR_P4_CCCR0, 18);
+ clear_msr_range(MSR_P4_PERFCTR0, 18);
+
+ wrmsr(MSR_P4_CRU_ESCR0, P4_NMI_CRU_ESCR0, 0);
+ wrmsr(MSR_P4_IQ_CCCR0, P4_NMI_IQ_CCCR0 & ~P4_CCCR_ENABLE, 0);
+ Dprintk("setting P4_IQ_COUNTER0 to 0x%08lx\n", -(cpu_khz/nmi_hz*1000));
+ wrmsr(MSR_P4_IQ_COUNTER0, -(cpu_khz/nmi_hz*1000), -1);
+ apic_write(APIC_LVTPC, APIC_DM_NMI);
+ wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0);
+ return 1;
+}
+
void setup_apic_nmi_watchdog(void)
{
switch (boot_cpu_data.x86_vendor) {
@@ -341,6 +425,13 @@ void setup_apic_nmi_watchdog(void)
return;
setup_k7_watchdog();
break;
+ case X86_VENDOR_INTEL:
+ if (boot_cpu_data.x86 != 15)
+ return;
+ if (!setup_p4_watchdog())
+ return;
+ break;
+
default:
return;
}
@@ -355,56 +446,67 @@ void setup_apic_nmi_watchdog(void)
*
* as these watchdog NMI IRQs are generated on every CPU, we only
* have to check the current processor.
- *
- * since NMIs don't listen to _any_ locks, we have to be extremely
- * careful not to rely on unsafe variables. The printk might lock
- * up though, so we have to break up any console locks first ...
- * [when there will be more tty-related locks, break them up
- * here too!]
*/
-static unsigned int
- last_irq_sums [NR_CPUS],
- alert_counter [NR_CPUS];
+static DEFINE_PER_CPU(unsigned, last_irq_sum);
+static DEFINE_PER_CPU(local_t, alert_counter);
+static DEFINE_PER_CPU(int, nmi_touch);
void touch_nmi_watchdog (void)
{
int i;
/*
- * Just reset the alert counters, (other CPUs might be
- * spinning on locks we hold):
+ * Tell other CPUs to reset their alert counters. We cannot
+ * do it ourselves because the alert count increase is not
+ * atomic.
*/
for (i = 0; i < NR_CPUS; i++)
- alert_counter[i] = 0;
+ per_cpu(nmi_touch, i) = 1;
}
void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason)
{
- int sum, cpu;
+ int sum;
+ int touched = 0;
- cpu = safe_smp_processor_id();
sum = read_pda(apic_timer_irqs);
- if (last_irq_sums[cpu] == sum) {
+ if (__get_cpu_var(nmi_touch)) {
+ __get_cpu_var(nmi_touch) = 0;
+ touched = 1;
+ }
+ if (!touched && __get_cpu_var(last_irq_sum) == sum) {
/*
* Ayiee, looks like this CPU is stuck ...
* wait a few IRQs (5 seconds) before doing the oops ...
*/
- alert_counter[cpu]++;
- if (alert_counter[cpu] == 5*nmi_hz) {
+ local_inc(&__get_cpu_var(alert_counter));
+ if (local_read(&__get_cpu_var(alert_counter)) == 5*nmi_hz) {
if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT)
== NOTIFY_STOP) {
- alert_counter[cpu] = 0;
+ local_set(&__get_cpu_var(alert_counter), 0);
return;
}
die_nmi("NMI Watchdog detected LOCKUP on CPU%d", regs);
}
} else {
- last_irq_sums[cpu] = sum;
- alert_counter[cpu] = 0;
+ __get_cpu_var(last_irq_sum) = sum;
+ local_set(&__get_cpu_var(alert_counter), 0);
}
- if (nmi_perfctr_msr)
+ if (nmi_perfctr_msr) {
+ if (nmi_perfctr_msr == MSR_P4_IQ_COUNTER0) {
+ /*
+ * P4 quirks:
+ * - An overflown perfctr will assert its interrupt
+ * until the OVF flag in its CCCR is cleared.
+ * - LVTPC is masked on interrupt and must be
+ * unmasked by the LVTPC handler.
+ */
+ wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0);
+ apic_write(APIC_LVTPC, APIC_DM_NMI);
+ }
wrmsr(nmi_perfctr_msr, -(cpu_khz/nmi_hz*1000), -1);
+ }
}
static int dummy_nmi_callback(struct pt_regs * regs, int cpu)
diff --git a/arch/x86_64/kernel/pmtimer.c b/arch/x86_64/kernel/pmtimer.c
new file mode 100644
index 00000000000..feb5f108dd2
--- /dev/null
+++ b/arch/x86_64/kernel/pmtimer.c
@@ -0,0 +1,101 @@
+/* Ported over from i386 by AK, original copyright was:
+ *
+ * (C) Dominik Brodowski <linux@brodo.de> 2003
+ *
+ * Driver to use the Power Management Timer (PMTMR) available in some
+ * southbridges as primary timing source for the Linux kernel.
+ *
+ * Based on parts of linux/drivers/acpi/hardware/hwtimer.c, timer_pit.c,
+ * timer_hpet.c, and on Arjan van de Ven's implementation for 2.4.
+ *
+ * This file is licensed under the GPL v2.
+ *
+ * Dropped all the hardware bug workarounds for now. Hopefully they
+ * are not needed on 64bit chipsets.
+ */
+
+#include <linux/jiffies.h>
+#include <linux/kernel.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/cpumask.h>
+#include <asm/io.h>
+#include <asm/proto.h>
+#include <asm/msr.h>
+#include <asm/vsyscall.h>
+
+/* The I/O port the PMTMR resides at.
+ * The location is detected during setup_arch(),
+ * in arch/i386/kernel/acpi/boot.c */
+u32 pmtmr_ioport;
+
+/* value of the Power timer at last timer interrupt */
+static u32 offset_delay;
+static u32 last_pmtmr_tick;
+
+#define ACPI_PM_MASK 0xFFFFFF /* limit it to 24 bits */
+
+static inline u32 cyc2us(u32 cycles)
+{
+ /* The Power Management Timer ticks at 3.579545 ticks per microsecond.
+ * 1 / PM_TIMER_FREQUENCY == 0.27936511 =~ 286/1024 [error: 0.024%]
+ *
+ * Even with HZ = 100, delta is at maximum 35796 ticks, so it can
+ * easily be multiplied with 286 (=0x11E) without having to fear
+ * u32 overflows.
+ */
+ cycles *= 286;
+ return (cycles >> 10);
+}
+
+int pmtimer_mark_offset(void)
+{
+ static int first_run = 1;
+ unsigned long tsc;
+ u32 lost;
+
+ u32 tick = inl(pmtmr_ioport);
+ u32 delta;
+
+ delta = cyc2us((tick - last_pmtmr_tick) & ACPI_PM_MASK);
+
+ last_pmtmr_tick = tick;
+ monotonic_base += delta * NSEC_PER_USEC;
+
+ delta += offset_delay;
+
+ lost = delta / (USEC_PER_SEC / HZ);
+ offset_delay = delta % (USEC_PER_SEC / HZ);
+
+ rdtscll(tsc);
+ vxtime.last_tsc = tsc - offset_delay * cpu_khz;
+
+ /* don't calculate delay for first run,
+ or if we've got less then a tick */
+ if (first_run || (lost < 1)) {
+ first_run = 0;
+ offset_delay = 0;
+ }
+
+ return lost - 1;
+}
+
+unsigned int do_gettimeoffset_pm(void)
+{
+ u32 now, offset, delta = 0;
+
+ offset = last_pmtmr_tick;
+ now = inl(pmtmr_ioport);
+ delta = (now - offset) & ACPI_PM_MASK;
+
+ return offset_delay + cyc2us(delta);
+}
+
+
+static int __init nopmtimer_setup(char *s)
+{
+ pmtmr_ioport = 0;
+ return 0;
+}
+
+__setup("nopmtimer", nopmtimer_setup);
diff --git a/arch/x86_64/kernel/ptrace.c b/arch/x86_64/kernel/ptrace.c
index e26e86bb56f..60dc9b98951 100644
--- a/arch/x86_64/kernel/ptrace.c
+++ b/arch/x86_64/kernel/ptrace.c
@@ -257,13 +257,13 @@ static int putreg(struct task_struct *child,
value &= 0xffff;
return 0;
case offsetof(struct user_regs_struct,fs_base):
- if (!((value >> 48) == 0 || (value >> 48) == 0xffff))
- return -EIO;
+ if (value >= TASK_SIZE)
+ return -EIO;
child->thread.fs = value;
return 0;
case offsetof(struct user_regs_struct,gs_base):
- if (!((value >> 48) == 0 || (value >> 48) == 0xffff))
- return -EIO;
+ if (value >= TASK_SIZE)
+ return -EIO;
child->thread.gs = value;
return 0;
case offsetof(struct user_regs_struct, eflags):
@@ -277,6 +277,11 @@ static int putreg(struct task_struct *child,
return -EIO;
value &= 0xffff;
break;
+ case offsetof(struct user_regs_struct, rip):
+ /* Check if the new RIP address is canonical */
+ if (value >= TASK_SIZE)
+ return -EIO;
+ break;
}
put_stack_long(child, regno - sizeof(struct pt_regs), value);
return 0;
diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c
index 2129cf9ba6b..b9fd0252c27 100644
--- a/arch/x86_64/kernel/setup.c
+++ b/arch/x86_64/kernel/setup.c
@@ -727,11 +727,12 @@ static void __init display_cacheinfo(struct cpuinfo_x86 *c)
static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
{
#ifdef CONFIG_SMP
- int cpu = c->x86_apicid;
+ int cpu = smp_processor_id();
int node = 0;
if (c->x86_num_cores == 1)
return;
- cpu_core_id[cpu] = cpu >> hweight32(c->x86_num_cores - 1);
+ /* Fix up the APIC ID following the AMD specification. */
+ cpu_core_id[cpu] >>= hweight32(c->x86_num_cores - 1);
#ifdef CONFIG_NUMA
/* When an ACPI SRAT table is available use the mappings from SRAT
@@ -745,6 +746,9 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
node = cpu_to_node[cpu];
}
#endif
+ /* For now: - better than BAD_APIC_ID at least*/
+ phys_proc_id[cpu] = cpu_core_id[cpu];
+
printk(KERN_INFO "CPU %d(%d) -> Node %d -> Core %d\n",
cpu, c->x86_num_cores, node, cpu_core_id[cpu]);
#endif
@@ -925,7 +929,6 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c)
c->x86_clflush_size = 64;
c->x86_cache_alignment = c->x86_clflush_size;
c->x86_num_cores = 1;
- c->x86_apicid = c == &boot_cpu_data ? 0 : c - cpu_data;
c->extended_cpuid_level = 0;
memset(&c->x86_capability, 0, sizeof c->x86_capability);
@@ -954,11 +957,15 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c)
}
if (c->x86_capability[0] & (1<<19))
c->x86_clflush_size = ((misc >> 8) & 0xff) * 8;
- c->x86_apicid = misc >> 24;
} else {
/* Have CPUID level 0 only - unheard of */
c->x86 = 4;
}
+
+#ifdef CONFIG_SMP
+ phys_proc_id[smp_processor_id()] =
+ cpu_core_id[smp_processor_id()] = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
}
/*
@@ -1088,7 +1095,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
/* Other (Linux-defined) */
- "cxmmx", NULL, "cyrix_arr", "centaur_mcr", "k8c+",
+ "cxmmx", NULL, "cyrix_arr", "centaur_mcr", NULL,
"constant_tsc", NULL, NULL,
NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c
index 73f7e8b9543..5abdee1e16a 100644
--- a/arch/x86_64/kernel/smpboot.c
+++ b/arch/x86_64/kernel/smpboot.c
@@ -56,6 +56,7 @@
#include <asm/kdebug.h>
#include <asm/tlbflush.h>
#include <asm/proto.h>
+#include <asm/nmi.h>
/* Change for real CPU hotplug. Note other files need to be fixed
first too. */
@@ -125,96 +126,210 @@ static void __cpuinit smp_store_cpu_info(int id)
*c = boot_cpu_data;
identify_cpu(c);
+ print_cpu_info(c);
}
/*
- * Synchronize TSCs of CPUs
+ * New Funky TSC sync algorithm borrowed from IA64.
+ * Main advantage is that it doesn't reset the TSCs fully and
+ * in general looks more robust and it works better than my earlier
+ * attempts. I believe it was written by David Mosberger. Some minor
+ * adjustments for x86-64 by me -AK
*
- * This new algorithm is less accurate than the old "zero TSCs"
- * one, but we cannot zero TSCs anymore in the new hotplug CPU
- * model.
+ * Original comment reproduced below.
+ *
+ * Synchronize TSC of the current (slave) CPU with the TSC of the
+ * MASTER CPU (normally the time-keeper CPU). We use a closed loop to
+ * eliminate the possibility of unaccounted-for errors (such as
+ * getting a machine check in the middle of a calibration step). The
+ * basic idea is for the slave to ask the master what itc value it has
+ * and to read its own itc before and after the master responds. Each
+ * iteration gives us three timestamps:
+ *
+ * slave master
+ *
+ * t0 ---\
+ * ---\
+ * --->
+ * tm
+ * /---
+ * /---
+ * t1 <---
+ *
+ *
+ * The goal is to adjust the slave's TSC such that tm falls exactly
+ * half-way between t0 and t1. If we achieve this, the clocks are
+ * synchronized provided the interconnect between the slave and the
+ * master is symmetric. Even if the interconnect were asymmetric, we
+ * would still know that the synchronization error is smaller than the
+ * roundtrip latency (t0 - t1).
+ *
+ * When the interconnect is quiet and symmetric, this lets us
+ * synchronize the TSC to within one or two cycles. However, we can
+ * only *guarantee* that the synchronization is accurate to within a
+ * round-trip time, which is typically in the range of several hundred
+ * cycles (e.g., ~500 cycles). In practice, this means that the TSCs
+ * are usually almost perfectly synchronized, but we shouldn't assume
+ * that the accuracy is much better than half a micro second or so.
+ *
+ * [there are other errors like the latency of RDTSC and of the
+ * WRMSR. These can also account to hundreds of cycles. So it's
+ * probably worse. It claims 153 cycles error on a dual Opteron,
+ * but I suspect the numbers are actually somewhat worse -AK]
*/
-static atomic_t __cpuinitdata tsc_flag;
+#define MASTER 0
+#define SLAVE (SMP_CACHE_BYTES/8)
+
+/* Intentionally don't use cpu_relax() while TSC synchronization
+ because we don't want to go into funky power save modi or cause
+ hypervisors to schedule us away. Going to sleep would likely affect
+ latency and low latency is the primary objective here. -AK */
+#define no_cpu_relax() barrier()
+
static __cpuinitdata DEFINE_SPINLOCK(tsc_sync_lock);
-static unsigned long long __cpuinitdata bp_tsc, ap_tsc;
+static volatile __cpuinitdata unsigned long go[SLAVE + 1];
+static int notscsync __cpuinitdata;
+
+#undef DEBUG_TSC_SYNC
-#define NR_LOOPS 5
+#define NUM_ROUNDS 64 /* magic value */
+#define NUM_ITERS 5 /* likewise */
-static void __cpuinit sync_tsc_bp_init(int init)
+/* Callback on boot CPU */
+static __cpuinit void sync_master(void *arg)
{
- if (init)
- _raw_spin_lock(&tsc_sync_lock);
- else
- _raw_spin_unlock(&tsc_sync_lock);
- atomic_set(&tsc_flag, 0);
+ unsigned long flags, i;
+
+ if (smp_processor_id() != boot_cpu_id)
+ return;
+
+ go[MASTER] = 0;
+
+ local_irq_save(flags);
+ {
+ for (i = 0; i < NUM_ROUNDS*NUM_ITERS; ++i) {
+ while (!go[MASTER])
+ no_cpu_relax();
+ go[MASTER] = 0;
+ rdtscll(go[SLAVE]);
+ }
+ }
+ local_irq_restore(flags);
}
/*
- * Synchronize TSC on AP with BP.
+ * Return the number of cycles by which our tsc differs from the tsc
+ * on the master (time-keeper) CPU. A positive number indicates our
+ * tsc is ahead of the master, negative that it is behind.
*/
-static void __cpuinit __sync_tsc_ap(void)
+static inline long
+get_delta(long *rt, long *master)
{
- if (!cpu_has_tsc)
- return;
- Dprintk("AP %d syncing TSC\n", smp_processor_id());
+ unsigned long best_t0 = 0, best_t1 = ~0UL, best_tm = 0;
+ unsigned long tcenter, t0, t1, tm;
+ int i;
- while (atomic_read(&tsc_flag) != 0)
- cpu_relax();
- atomic_inc(&tsc_flag);
- mb();
- _raw_spin_lock(&tsc_sync_lock);
- wrmsrl(MSR_IA32_TSC, bp_tsc);
- _raw_spin_unlock(&tsc_sync_lock);
- rdtscll(ap_tsc);
- mb();
- atomic_inc(&tsc_flag);
- mb();
+ for (i = 0; i < NUM_ITERS; ++i) {
+ rdtscll(t0);
+ go[MASTER] = 1;
+ while (!(tm = go[SLAVE]))
+ no_cpu_relax();
+ go[SLAVE] = 0;
+ rdtscll(t1);
+
+ if (t1 - t0 < best_t1 - best_t0)
+ best_t0 = t0, best_t1 = t1, best_tm = tm;
+ }
+
+ *rt = best_t1 - best_t0;
+ *master = best_tm - best_t0;
+
+ /* average best_t0 and best_t1 without overflow: */
+ tcenter = (best_t0/2 + best_t1/2);
+ if (best_t0 % 2 + best_t1 % 2 == 2)
+ ++tcenter;
+ return tcenter - best_tm;
}
-static void __cpuinit sync_tsc_ap(void)
+static __cpuinit void sync_tsc(void)
{
- int i;
- for (i = 0; i < NR_LOOPS; i++)
- __sync_tsc_ap();
+ int i, done = 0;
+ long delta, adj, adjust_latency = 0;
+ unsigned long flags, rt, master_time_stamp, bound;
+#if DEBUG_TSC_SYNC
+ static struct syncdebug {
+ long rt; /* roundtrip time */
+ long master; /* master's timestamp */
+ long diff; /* difference between midpoint and master's timestamp */
+ long lat; /* estimate of tsc adjustment latency */
+ } t[NUM_ROUNDS] __cpuinitdata;
+#endif
+
+ go[MASTER] = 1;
+
+ smp_call_function(sync_master, NULL, 1, 0);
+
+ while (go[MASTER]) /* wait for master to be ready */
+ no_cpu_relax();
+
+ spin_lock_irqsave(&tsc_sync_lock, flags);
+ {
+ for (i = 0; i < NUM_ROUNDS; ++i) {
+ delta = get_delta(&rt, &master_time_stamp);
+ if (delta == 0) {
+ done = 1; /* let's lock on to this... */
+ bound = rt;
+ }
+
+ if (!done) {
+ unsigned long t;
+ if (i > 0) {
+ adjust_latency += -delta;
+ adj = -delta + adjust_latency/4;
+ } else
+ adj = -delta;
+
+ rdtscll(t);
+ wrmsrl(MSR_IA32_TSC, t + adj);
+ }
+#if DEBUG_TSC_SYNC
+ t[i].rt = rt;
+ t[i].master = master_time_stamp;
+ t[i].diff = delta;
+ t[i].lat = adjust_latency/4;
+#endif
+ }
+ }
+ spin_unlock_irqrestore(&tsc_sync_lock, flags);
+
+#if DEBUG_TSC_SYNC
+ for (i = 0; i < NUM_ROUNDS; ++i)
+ printk("rt=%5ld master=%5ld diff=%5ld adjlat=%5ld\n",
+ t[i].rt, t[i].master, t[i].diff, t[i].lat);
+#endif
+
+ printk(KERN_INFO
+ "CPU %d: synchronized TSC with CPU %u (last diff %ld cycles, "
+ "maxerr %lu cycles)\n",
+ smp_processor_id(), boot_cpu_id, delta, rt);
}
-/*
- * Synchronize TSC from BP to AP.
- */
-static void __cpuinit __sync_tsc_bp(int cpu)
+static void __cpuinit tsc_sync_wait(void)
{
- if (!cpu_has_tsc)
+ if (notscsync || !cpu_has_tsc)
return;
-
- /* Wait for AP */
- while (atomic_read(&tsc_flag) == 0)
- cpu_relax();
- /* Save BPs TSC */
- sync_core();
- rdtscll(bp_tsc);
- /* Don't do the sync core here to avoid too much latency. */
- mb();
- /* Start the AP */
- _raw_spin_unlock(&tsc_sync_lock);
- /* Wait for AP again */
- while (atomic_read(&tsc_flag) < 2)
- cpu_relax();
- rdtscl(bp_tsc);
- barrier();
+ printk(KERN_INFO "CPU %d: Syncing TSC to CPU %u.\n", smp_processor_id(),
+ boot_cpu_id);
+ sync_tsc();
}
-static void __cpuinit sync_tsc_bp(int cpu)
+static __init int notscsync_setup(char *s)
{
- int i;
- for (i = 0; i < NR_LOOPS - 1; i++) {
- __sync_tsc_bp(cpu);
- sync_tsc_bp_init(1);
- }
- __sync_tsc_bp(cpu);
- printk(KERN_INFO "Synced TSC of CPU %d difference %Ld\n",
- cpu, ap_tsc - bp_tsc);
+ notscsync = 1;
+ return 0;
}
+__setup("notscsync", notscsync_setup);
static atomic_t init_deasserted __cpuinitdata;
@@ -315,11 +430,6 @@ void __cpuinit start_secondary(void)
cpu_init();
smp_callin();
- /*
- * Synchronize the TSC with the BP
- */
- sync_tsc_ap();
-
/* otherwise gcc will move up the smp_processor_id before the cpu_init */
barrier();
@@ -334,7 +444,6 @@ void __cpuinit start_secondary(void)
enable_8259A_irq(0);
}
-
enable_APIC_timer();
/*
@@ -343,6 +452,11 @@ void __cpuinit start_secondary(void)
cpu_set(smp_processor_id(), cpu_online_map);
mb();
+ /* Wait for TSC sync to not schedule things before.
+ We still process interrupts, which could see an inconsistent
+ time in that window unfortunately. */
+ tsc_sync_wait();
+
cpu_idle();
}
@@ -531,7 +645,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid)
printk("failed fork for CPU %d\n", cpu);
return PTR_ERR(idle);
}
- x86_cpu_to_apicid[cpu] = apicid;
cpu_pda[cpu].pcurrent = idle;
@@ -600,8 +713,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid)
if (cpu_isset(cpu, cpu_callin_map)) {
/* number CPUs logically, starting from 1 (BSP is 0) */
- Dprintk("OK.\n");
- print_cpu_info(&cpu_data[cpu]);
Dprintk("CPU has booted.\n");
} else {
boot_error = 1;
@@ -842,7 +953,6 @@ void __cpuinit smp_prepare_cpus(unsigned int max_cpus)
GET_APIC_ID(apic_read(APIC_ID)), boot_cpu_id);
/* Or can we switch back to PIC here? */
}
- x86_cpu_to_apicid[0] = boot_cpu_id;
/*
* Now start the IO-APICs
@@ -889,18 +999,14 @@ int __cpuinit __cpu_up(unsigned int cpu)
printk("__cpu_up: bad cpu %d\n", cpu);
return -EINVAL;
}
- sync_tsc_bp_init(1);
/* Boot it! */
err = do_boot_cpu(cpu, apicid);
if (err < 0) {
- sync_tsc_bp_init(0);
Dprintk("do_boot_cpu failed %d\n", err);
return err;
}
- sync_tsc_bp(cpu);
-
/* Unleash the CPU! */
Dprintk("waiting for cpu %d\n", cpu);
@@ -923,4 +1029,6 @@ void __cpuinit smp_cpus_done(unsigned int max_cpus)
detect_siblings();
time_init_gtod();
+
+ check_nmi_watchdog();
}
diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c
index 0e3b60e1ac5..735b6767c8e 100644
--- a/arch/x86_64/kernel/time.c
+++ b/arch/x86_64/kernel/time.c
@@ -26,6 +26,8 @@
#include <linux/sysdev.h>
#include <linux/bcd.h>
#include <linux/kallsyms.h>
+#include <linux/acpi.h>
+#include <acpi/achware.h> /* for PM timer frequency */
#include <asm/8253pit.h>
#include <asm/pgtable.h>
#include <asm/vsyscall.h>
@@ -396,6 +398,10 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
(offset - vxtime.last)*(NSEC_PER_SEC/HZ) / hpet_tick;
vxtime.last = offset;
+#ifdef CONFIG_X86_PM_TIMER
+ } else if (vxtime.mode == VXTIME_PMTMR) {
+ lost = pmtimer_mark_offset();
+#endif
} else {
offset = (((tsc - vxtime.last_tsc) *
vxtime.tsc_quot) >> 32) - (USEC_PER_SEC / HZ);
@@ -898,6 +904,13 @@ void __init time_init(void)
hpet_period;
cpu_khz = hpet_calibrate_tsc();
timename = "HPET";
+#ifdef CONFIG_X86_PM_TIMER
+ } else if (pmtmr_ioport) {
+ vxtime_hz = PM_TIMER_FREQUENCY;
+ timename = "PM";
+ pit_init();
+ cpu_khz = pit_calibrate_tsc();
+#endif
} else {
pit_init();
cpu_khz = pit_calibrate_tsc();
@@ -923,35 +936,50 @@ void __init time_init(void)
}
/*
+ * Make an educated guess if the TSC is trustworthy and synchronized
+ * over all CPUs.
+ */
+static __init int unsynchronized_tsc(void)
+{
+#ifdef CONFIG_SMP
+ if (oem_force_hpet_timer())
+ return 1;
+ /* Intel systems are normally all synchronized. Exceptions
+ are handled in the OEM check above. */
+ if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL)
+ return 0;
+ /* All in a single socket - should be synchronized */
+ if (cpus_weight(cpu_core_map[0]) == num_online_cpus())
+ return 0;
+#endif
+ /* Assume multi socket systems are not synchronized */
+ return num_online_cpus() > 1;
+}
+
+/*
* Decide after all CPUs are booted what mode gettimeofday should use.
*/
void __init time_init_gtod(void)
{
char *timetype;
- /*
- * AMD systems with more than one CPU don't have fully synchronized
- * TSCs. Always use HPET gettimeofday for these, although it is slower.
- * Intel SMP systems usually have synchronized TSCs, so use always
- * the TSC.
- *
- * Exceptions:
- * IBM Summit2 checked by oem_force_hpet_timer().
- * AMD dual core may also not need HPET. Check me.
- *
- * Can be turned off with "notsc".
- */
- if (num_online_cpus() > 1 &&
- boot_cpu_data.x86_vendor == X86_VENDOR_AMD)
- notsc = 1;
- /* Some systems will want to disable TSC and use HPET. */
- if (oem_force_hpet_timer())
+ if (unsynchronized_tsc())
notsc = 1;
if (vxtime.hpet_address && notsc) {
timetype = "HPET";
vxtime.last = hpet_readl(HPET_T0_CMP) - hpet_tick;
vxtime.mode = VXTIME_HPET;
do_gettimeoffset = do_gettimeoffset_hpet;
+#ifdef CONFIG_X86_PM_TIMER
+ /* Using PM for gettimeofday is quite slow, but we have no other
+ choice because the TSC is too unreliable on some systems. */
+ } else if (pmtmr_ioport && !vxtime.hpet_address && notsc) {
+ timetype = "PM";
+ do_gettimeoffset = do_gettimeoffset_pm;
+ vxtime.mode = VXTIME_PMTMR;
+ sysctl_vsyscall = 0;
+ printk(KERN_INFO "Disabling vsyscall due to use of PM timer\n");
+#endif
} else {
timetype = vxtime.hpet_address ? "HPET/TSC" : "PIT/TSC";
vxtime.mode = VXTIME_TSC;
diff --git a/arch/x86_64/kernel/vsyscall.c b/arch/x86_64/kernel/vsyscall.c
index b4b8dc59663..2e573442594 100644
--- a/arch/x86_64/kernel/vsyscall.c
+++ b/arch/x86_64/kernel/vsyscall.c
@@ -65,7 +65,7 @@ static force_inline void do_vgettimeofday(struct timeval * tv)
usec = (__xtime.tv_nsec / 1000) +
(__jiffies - __wall_jiffies) * (1000000 / HZ);
- if (__vxtime.mode == VXTIME_TSC) {
+ if (__vxtime.mode != VXTIME_HPET) {
sync_core();
rdtscll(t);
if (t < __vxtime.last_tsc)
@@ -217,8 +217,9 @@ static int __init vsyscall_init(void)
BUG_ON((unsigned long) &vtime != VSYSCALL_ADDR(__NR_vtime));
BUG_ON((VSYSCALL_ADDR(0) != __fix_to_virt(VSYSCALL_FIRST_PAGE)));
map_vsyscall();
- sysctl_vsyscall = 1;
+#ifdef CONFIG_SYSCTL
register_sysctl_table(kernel_root_table2, 0);
+#endif
return 0;
}
diff --git a/arch/x86_64/mm/fault.c b/arch/x86_64/mm/fault.c
index e0330921676..5d6b2114cc9 100644
--- a/arch/x86_64/mm/fault.c
+++ b/arch/x86_64/mm/fault.c
@@ -234,6 +234,8 @@ static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs,
/*
* Handle a fault on the vmalloc or module mapping area
+ *
+ * This assumes no large pages in there.
*/
static int vmalloc_fault(unsigned long address)
{
@@ -272,7 +274,10 @@ static int vmalloc_fault(unsigned long address)
if (!pte_present(*pte_ref))
return -1;
pte = pte_offset_kernel(pmd, address);
- if (!pte_present(*pte) || pte_page(*pte) != pte_page(*pte_ref))
+ /* Don't use pte_page here, because the mappings can point
+ outside mem_map, and the NUMA hash lookup cannot handle
+ that. */
+ if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref))
BUG();
__flush_tlb_all();
return 0;
@@ -346,7 +351,9 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long error_code)
* protection error (error_code & 1) == 0.
*/
if (unlikely(address >= TASK_SIZE)) {
- if (!(error_code & 5)) {
+ if (!(error_code & 5) &&
+ ((address >= VMALLOC_START && address < VMALLOC_END) ||
+ (address >= MODULES_VADDR && address < MODULES_END))) {
if (vmalloc_fault(address) < 0)
goto bad_area_nosemaphore;
return;
diff --git a/arch/x86_64/mm/ioremap.c b/arch/x86_64/mm/ioremap.c
index 74ec8554b19..c6fb0cb6999 100644
--- a/arch/x86_64/mm/ioremap.c
+++ b/arch/x86_64/mm/ioremap.c
@@ -272,7 +272,7 @@ void iounmap(volatile void __iomem *addr)
if ((p->flags >> 20) &&
p->phys_addr + p->size - 1 < virt_to_phys(high_memory)) {
/* p->size includes the guard page, but cpa doesn't like that */
- change_page_attr(virt_to_page(__va(p->phys_addr)),
+ change_page_attr_addr((unsigned long)__va(p->phys_addr),
p->size >> PAGE_SHIFT,
PAGE_KERNEL);
global_flush_tlb();