aboutsummaryrefslogtreecommitdiff
path: root/arch/x86
diff options
context:
space:
mode:
Diffstat (limited to 'arch/x86')
-rw-r--r--arch/x86/Kconfig10
-rw-r--r--arch/x86/include/asm/apicdef.h1
-rw-r--r--arch/x86/include/asm/cacheflush.h53
-rw-r--r--arch/x86/include/asm/i387.h8
-rw-r--r--arch/x86/include/asm/io.h3
-rw-r--r--arch/x86/include/asm/mce.h35
-rw-r--r--arch/x86/include/asm/mmzone_32.h43
-rw-r--r--arch/x86/include/asm/msr-index.h5
-rw-r--r--arch/x86/include/asm/percpu.h8
-rw-r--r--arch/x86/include/asm/pgtable.h2
-rw-r--r--arch/x86/include/asm/xen/page.h1
-rw-r--r--arch/x86/kernel/alternative.c17
-rw-r--r--arch/x86/kernel/apic/apic.c15
-rw-r--r--arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c2
-rw-r--r--arch/x86/kernel/cpu/mcheck/Makefile1
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_32.c14
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_64.c530
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd_64.c22
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_intel_64.c207
-rw-r--r--arch/x86/kernel/cpu/mcheck/threshold.c29
-rw-r--r--arch/x86/kernel/i387.c2
-rw-r--r--arch/x86/kernel/irq_32.c29
-rw-r--r--arch/x86/kernel/reboot.c8
-rw-r--r--arch/x86/kernel/setup_percpu.c367
-rw-r--r--arch/x86/math-emu/fpu_aux.c31
-rw-r--r--arch/x86/mm/init_32.c17
-rw-r--r--arch/x86/mm/init_64.c72
-rw-r--r--arch/x86/xen/enlighten.c10
-rw-r--r--arch/x86/xen/mmu.c7
-rw-r--r--arch/x86/xen/smp.c6
30 files changed, 1242 insertions, 313 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 469f3450bf8..31758378bcd 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -138,6 +138,9 @@ config ARCH_HAS_CACHE_LINE_SIZE
config HAVE_SETUP_PER_CPU_AREA
def_bool y
+config HAVE_DYNAMIC_PER_CPU_AREA
+ def_bool y
+
config HAVE_CPUMASK_OF_CPU_MAP
def_bool X86_64_SMP
@@ -780,6 +783,11 @@ config X86_MCE_AMD
Additional support for AMD specific MCE features such as
the DRAM Error Threshold.
+config X86_MCE_THRESHOLD
+ depends on X86_MCE_AMD || X86_MCE_INTEL
+ bool
+ default y
+
config X86_MCE_NONFATAL
tristate "Check for non-fatal errors on AMD Athlon/Duron / Intel Pentium 4"
depends on X86_32 && X86_MCE
@@ -1125,7 +1133,7 @@ config NODES_SHIFT
Specify the maximum number of NUMA Nodes available on the target
system. Increases memory reserved to accomodate various tables.
-config HAVE_ARCH_BOOTMEM_NODE
+config HAVE_ARCH_BOOTMEM
def_bool y
depends on X86_32 && NUMA
diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h
index 63134e31e8b..bc9514fb3b1 100644
--- a/arch/x86/include/asm/apicdef.h
+++ b/arch/x86/include/asm/apicdef.h
@@ -53,6 +53,7 @@
#define APIC_ESR_SENDILL 0x00020
#define APIC_ESR_RECVILL 0x00040
#define APIC_ESR_ILLREGA 0x00080
+#define APIC_LVTCMCI 0x2f0
#define APIC_ICR 0x300
#define APIC_DEST_SELF 0x40000
#define APIC_DEST_ALLINC 0x80000
diff --git a/arch/x86/include/asm/cacheflush.h b/arch/x86/include/asm/cacheflush.h
index 2f8466540fb..5b301b7ff5f 100644
--- a/arch/x86/include/asm/cacheflush.h
+++ b/arch/x86/include/asm/cacheflush.h
@@ -5,24 +5,43 @@
#include <linux/mm.h>
/* Caches aren't brain-dead on the intel. */
-#define flush_cache_all() do { } while (0)
-#define flush_cache_mm(mm) do { } while (0)
-#define flush_cache_dup_mm(mm) do { } while (0)
-#define flush_cache_range(vma, start, end) do { } while (0)
-#define flush_cache_page(vma, vmaddr, pfn) do { } while (0)
-#define flush_dcache_page(page) do { } while (0)
-#define flush_dcache_mmap_lock(mapping) do { } while (0)
-#define flush_dcache_mmap_unlock(mapping) do { } while (0)
-#define flush_icache_range(start, end) do { } while (0)
-#define flush_icache_page(vma, pg) do { } while (0)
-#define flush_icache_user_range(vma, pg, adr, len) do { } while (0)
-#define flush_cache_vmap(start, end) do { } while (0)
-#define flush_cache_vunmap(start, end) do { } while (0)
+static inline void flush_cache_all(void) { }
+static inline void flush_cache_mm(struct mm_struct *mm) { }
+static inline void flush_cache_dup_mm(struct mm_struct *mm) { }
+static inline void flush_cache_range(struct vm_area_struct *vma,
+ unsigned long start, unsigned long end) { }
+static inline void flush_cache_page(struct vm_area_struct *vma,
+ unsigned long vmaddr, unsigned long pfn) { }
+static inline void flush_dcache_page(struct page *page) { }
+static inline void flush_dcache_mmap_lock(struct address_space *mapping) { }
+static inline void flush_dcache_mmap_unlock(struct address_space *mapping) { }
+static inline void flush_icache_range(unsigned long start,
+ unsigned long end) { }
+static inline void flush_icache_page(struct vm_area_struct *vma,
+ struct page *page) { }
+static inline void flush_icache_user_range(struct vm_area_struct *vma,
+ struct page *page,
+ unsigned long addr,
+ unsigned long len) { }
+static inline void flush_cache_vmap(unsigned long start, unsigned long end) { }
+static inline void flush_cache_vunmap(unsigned long start,
+ unsigned long end) { }
-#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
- memcpy((dst), (src), (len))
-#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
- memcpy((dst), (src), (len))
+static inline void copy_to_user_page(struct vm_area_struct *vma,
+ struct page *page, unsigned long vaddr,
+ void *dst, const void *src,
+ unsigned long len)
+{
+ memcpy(dst, src, len);
+}
+
+static inline void copy_from_user_page(struct vm_area_struct *vma,
+ struct page *page, unsigned long vaddr,
+ void *dst, const void *src,
+ unsigned long len)
+{
+ memcpy(dst, src, len);
+}
#define PG_non_WB PG_arch_1
PAGEFLAG(NonWB, non_WB)
diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h
index 48f0004db8c..71c9e518398 100644
--- a/arch/x86/include/asm/i387.h
+++ b/arch/x86/include/asm/i387.h
@@ -172,7 +172,13 @@ static inline void __save_init_fpu(struct task_struct *tsk)
#else /* CONFIG_X86_32 */
-extern void finit(void);
+#ifdef CONFIG_MATH_EMULATION
+extern void finit_task(struct task_struct *tsk);
+#else
+static inline void finit_task(struct task_struct *tsk)
+{
+}
+#endif
static inline void tolerant_fwait(void)
{
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index 683d0b4c00f..e5383e3d2f8 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -172,8 +172,6 @@ static inline void __iomem *ioremap(resource_size_t offset, unsigned long size)
extern void iounmap(volatile void __iomem *addr);
-extern void __iomem *fix_ioremap(unsigned idx, unsigned long phys);
-
#ifdef CONFIG_X86_32
# include "io_32.h"
@@ -198,7 +196,6 @@ extern void early_ioremap_reset(void);
extern void __iomem *early_ioremap(unsigned long offset, unsigned long size);
extern void __iomem *early_memremap(unsigned long offset, unsigned long size);
extern void early_iounmap(void __iomem *addr, unsigned long size);
-extern void __iomem *fix_ioremap(unsigned idx, unsigned long phys);
#define IO_SPACE_LIMIT 0xffff
diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h
index 32c6e17b960..563933e06a3 100644
--- a/arch/x86/include/asm/mce.h
+++ b/arch/x86/include/asm/mce.h
@@ -11,6 +11,8 @@
*/
#define MCG_CTL_P (1UL<<8) /* MCG_CAP register available */
+#define MCG_EXT_P (1ULL<<9) /* Extended registers available */
+#define MCG_CMCI_P (1ULL<<10) /* CMCI supported */
#define MCG_STATUS_RIPV (1UL<<0) /* restart ip valid */
#define MCG_STATUS_EIPV (1UL<<1) /* ip points to correct instruction */
@@ -90,14 +92,29 @@ extern int mce_disabled;
#include <asm/atomic.h>
+void mce_setup(struct mce *m);
void mce_log(struct mce *m);
DECLARE_PER_CPU(struct sys_device, device_mce);
extern void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu);
+/*
+ * To support more than 128 would need to escape the predefined
+ * Linux defined extended banks first.
+ */
+#define MAX_NR_BANKS (MCE_EXTENDED_BANK - 1)
+
#ifdef CONFIG_X86_MCE_INTEL
void mce_intel_feature_init(struct cpuinfo_x86 *c);
+void cmci_clear(void);
+void cmci_reenable(void);
+void cmci_rediscover(int dying);
+void cmci_recheck(void);
#else
static inline void mce_intel_feature_init(struct cpuinfo_x86 *c) { }
+static inline void cmci_clear(void) {}
+static inline void cmci_reenable(void) {}
+static inline void cmci_rediscover(int dying) {}
+static inline void cmci_recheck(void) {}
#endif
#ifdef CONFIG_X86_MCE_AMD
@@ -106,11 +123,23 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c);
static inline void mce_amd_feature_init(struct cpuinfo_x86 *c) { }
#endif
-void mce_log_therm_throt_event(unsigned int cpu, __u64 status);
+extern int mce_available(struct cpuinfo_x86 *c);
+
+void mce_log_therm_throt_event(__u64 status);
extern atomic_t mce_entry;
extern void do_machine_check(struct pt_regs *, long);
+
+typedef DECLARE_BITMAP(mce_banks_t, MAX_NR_BANKS);
+DECLARE_PER_CPU(mce_banks_t, mce_poll_banks);
+
+enum mcp_flags {
+ MCP_TIMESTAMP = (1 << 0), /* log time stamp */
+ MCP_UC = (1 << 1), /* log uncorrected errors */
+};
+extern void machine_check_poll(enum mcp_flags flags, mce_banks_t *b);
+
extern int mce_notify_user(void);
#endif /* !CONFIG_X86_32 */
@@ -120,8 +149,8 @@ extern void mcheck_init(struct cpuinfo_x86 *c);
#else
#define mcheck_init(c) do { } while (0)
#endif
-extern void stop_mce(void);
-extern void restart_mce(void);
+
+extern void (*mce_threshold_vector)(void);
#endif /* __KERNEL__ */
#endif /* _ASM_X86_MCE_H */
diff --git a/arch/x86/include/asm/mmzone_32.h b/arch/x86/include/asm/mmzone_32.h
index 105fb90a063..ede6998bd92 100644
--- a/arch/x86/include/asm/mmzone_32.h
+++ b/arch/x86/include/asm/mmzone_32.h
@@ -91,46 +91,9 @@ static inline int pfn_valid(int pfn)
#endif /* CONFIG_DISCONTIGMEM */
#ifdef CONFIG_NEED_MULTIPLE_NODES
-
-/*
- * Following are macros that are specific to this numa platform.
- */
-#define reserve_bootmem(addr, size, flags) \
- reserve_bootmem_node(NODE_DATA(0), (addr), (size), (flags))
-#define alloc_bootmem(x) \
- __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS))
-#define alloc_bootmem_nopanic(x) \
- __alloc_bootmem_node_nopanic(NODE_DATA(0), (x), SMP_CACHE_BYTES, \
- __pa(MAX_DMA_ADDRESS))
-#define alloc_bootmem_low(x) \
- __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, 0)
-#define alloc_bootmem_pages(x) \
- __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, __pa(MAX_DMA_ADDRESS))
-#define alloc_bootmem_pages_nopanic(x) \
- __alloc_bootmem_node_nopanic(NODE_DATA(0), (x), PAGE_SIZE, \
- __pa(MAX_DMA_ADDRESS))
-#define alloc_bootmem_low_pages(x) \
- __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, 0)
-#define alloc_bootmem_node(pgdat, x) \
-({ \
- struct pglist_data __maybe_unused \
- *__alloc_bootmem_node__pgdat = (pgdat); \
- __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, \
- __pa(MAX_DMA_ADDRESS)); \
-})
-#define alloc_bootmem_pages_node(pgdat, x) \
-({ \
- struct pglist_data __maybe_unused \
- *__alloc_bootmem_node__pgdat = (pgdat); \
- __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, \
- __pa(MAX_DMA_ADDRESS)); \
-})
-#define alloc_bootmem_low_pages_node(pgdat, x) \
-({ \
- struct pglist_data __maybe_unused \
- *__alloc_bootmem_node__pgdat = (pgdat); \
- __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, 0); \
-})
+/* always use node 0 for bootmem on this numa platform */
+#define bootmem_arch_preferred_node(__bdata, size, align, goal, limit) \
+ (NODE_DATA(0)->bdata)
#endif /* CONFIG_NEED_MULTIPLE_NODES */
#endif /* _ASM_X86_MMZONE_32_H */
diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h
index 358acc59ae0..2dbd2314139 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -77,6 +77,11 @@
#define MSR_IA32_MC0_ADDR 0x00000402
#define MSR_IA32_MC0_MISC 0x00000403
+/* These are consecutive and not in the normal 4er MCE bank block */
+#define MSR_IA32_MC0_CTL2 0x00000280
+#define CMCI_EN (1ULL << 30)
+#define CMCI_THRESHOLD_MASK 0xffffULL
+
#define MSR_P6_PERFCTR0 0x000000c1
#define MSR_P6_PERFCTR1 0x000000c2
#define MSR_P6_EVNTSEL0 0x00000186
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index aee103b26d0..8f1d2fbec1d 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -43,6 +43,14 @@
#else /* ...!ASSEMBLY */
#include <linux/stringify.h>
+#include <asm/sections.h>
+
+#define __addr_to_pcpu_ptr(addr) \
+ (void *)((unsigned long)(addr) - (unsigned long)pcpu_base_addr \
+ + (unsigned long)__per_cpu_start)
+#define __pcpu_ptr_to_addr(ptr) \
+ (void *)((unsigned long)(ptr) + (unsigned long)pcpu_base_addr \
+ - (unsigned long)__per_cpu_start)
#ifdef CONFIG_SMP
#define __percpu_arg(x) "%%"__stringify(__percpu_seg)":%P" #x
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 1c097a3a666..d0812e155f1 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -288,6 +288,8 @@ static inline int is_new_memtype_allowed(unsigned long flags,
return 1;
}
+pmd_t *populate_extra_pmd(unsigned long vaddr);
+pte_t *populate_extra_pte(unsigned long vaddr);
#endif /* __ASSEMBLY__ */
#ifdef CONFIG_X86_32
diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h
index 4bd990ee43d..1a918dde46b 100644
--- a/arch/x86/include/asm/xen/page.h
+++ b/arch/x86/include/asm/xen/page.h
@@ -164,6 +164,7 @@ static inline pte_t __pte_ma(pteval_t x)
xmaddr_t arbitrary_virt_to_machine(void *address);
+unsigned long arbitrary_virt_to_mfn(void *vaddr);
void make_lowmem_page_readonly(void *vaddr);
void make_lowmem_page_readwrite(void *vaddr);
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 6907b8e85d5..4c80f155743 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -414,9 +414,17 @@ void __init alternative_instructions(void)
that might execute the to be patched code.
Other CPUs are not running. */
stop_nmi();
-#ifdef CONFIG_X86_MCE
- stop_mce();
-#endif
+
+ /*
+ * Don't stop machine check exceptions while patching.
+ * MCEs only happen when something got corrupted and in this
+ * case we must do something about the corruption.
+ * Ignoring it is worse than a unlikely patching race.
+ * Also machine checks tend to be broadcast and if one CPU
+ * goes into machine check the others follow quickly, so we don't
+ * expect a machine check to cause undue problems during to code
+ * patching.
+ */
apply_alternatives(__alt_instructions, __alt_instructions_end);
@@ -456,9 +464,6 @@ void __init alternative_instructions(void)
(unsigned long)__smp_locks_end);
restart_nmi();
-#ifdef CONFIG_X86_MCE
- restart_mce();
-#endif
}
/**
diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index f9cecdfd05c..30909a258d0 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -46,6 +46,7 @@
#include <asm/idle.h>
#include <asm/mtrr.h>
#include <asm/smp.h>
+#include <asm/mce.h>
unsigned int num_processors;
@@ -842,6 +843,14 @@ void clear_local_APIC(void)
apic_write(APIC_LVTTHMR, v | APIC_LVT_MASKED);
}
#endif
+#ifdef CONFIG_X86_MCE_INTEL
+ if (maxlvt >= 6) {
+ v = apic_read(APIC_LVTCMCI);
+ if (!(v & APIC_LVT_MASKED))
+ apic_write(APIC_LVTCMCI, v | APIC_LVT_MASKED);
+ }
+#endif
+
/*
* Clean APIC state for other OSs:
*/
@@ -1241,6 +1250,12 @@ void __cpuinit setup_local_APIC(void)
apic_write(APIC_LVT1, value);
preempt_enable();
+
+#ifdef CONFIG_X86_MCE_INTEL
+ /* Recheck CMCI information after local APIC is up on CPU #0 */
+ if (smp_processor_id() == 0)
+ cmci_recheck();
+#endif
}
void __cpuinit end_local_APIC_setup(void)
diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
index 4b1c319d30c..22590cf688a 100644
--- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -601,7 +601,7 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
if (!data)
return -ENOMEM;
- data->acpi_data = percpu_ptr(acpi_perf_data, cpu);
+ data->acpi_data = per_cpu_ptr(acpi_perf_data, cpu);
per_cpu(drv_data, cpu) = data;
if (cpu_has(c, X86_FEATURE_CONSTANT_TSC))
diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile
index d7d2323bbb6..b2f89829bbe 100644
--- a/arch/x86/kernel/cpu/mcheck/Makefile
+++ b/arch/x86/kernel/cpu/mcheck/Makefile
@@ -4,3 +4,4 @@ obj-$(CONFIG_X86_32) += k7.o p4.o p5.o p6.o winchip.o
obj-$(CONFIG_X86_MCE_INTEL) += mce_intel_64.o
obj-$(CONFIG_X86_MCE_AMD) += mce_amd_64.o
obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o
+obj-$(CONFIG_X86_MCE_THRESHOLD) += threshold.o
diff --git a/arch/x86/kernel/cpu/mcheck/mce_32.c b/arch/x86/kernel/cpu/mcheck/mce_32.c
index dfaebce3633..3552119b091 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_32.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_32.c
@@ -60,20 +60,6 @@ void mcheck_init(struct cpuinfo_x86 *c)
}
}
-static unsigned long old_cr4 __initdata;
-
-void __init stop_mce(void)
-{
- old_cr4 = read_cr4();
- clear_in_cr4(X86_CR4_MCE);
-}
-
-void __init restart_mce(void)
-{
- if (old_cr4 & X86_CR4_MCE)
- set_in_cr4(X86_CR4_MCE);
-}
-
static int __init mcheck_disable(char *str)
{
mce_disabled = 1;
diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c
index fe79985ce0f..bfbd5323a63 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -3,6 +3,8 @@
* K8 parts Copyright 2002,2003 Andi Kleen, SuSE Labs.
* Rest from unknown author(s).
* 2004 Andi Kleen. Rewrote most of it.
+ * Copyright 2008 Intel Corporation
+ * Author: Andi Kleen
*/
#include <linux/init.h>
@@ -24,6 +26,9 @@
#include <linux/ctype.h>
#include <linux/kmod.h>
#include <linux/kdebug.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/ratelimit.h>
#include <asm/processor.h>
#include <asm/msr.h>
#include <asm/mce.h>
@@ -32,7 +37,6 @@
#include <asm/idle.h>
#define MISC_MCELOG_MINOR 227
-#define NR_SYSFS_BANKS 6
atomic_t mce_entry;
@@ -47,7 +51,7 @@ static int mce_dont_init;
*/
static int tolerant = 1;
static int banks;
-static unsigned long bank[NR_SYSFS_BANKS] = { [0 ... NR_SYSFS_BANKS-1] = ~0UL };
+static u64 *bank;
static unsigned long notify_user;
static int rip_msr;
static int mce_bootlog = -1;
@@ -58,6 +62,19 @@ static char *trigger_argv[2] = { trigger, NULL };
static DECLARE_WAIT_QUEUE_HEAD(mce_wait);
+/* MCA banks polled by the period polling timer for corrected events */
+DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = {
+ [0 ... BITS_TO_LONGS(MAX_NR_BANKS)-1] = ~0UL
+};
+
+/* Do initial initialization of a struct mce */
+void mce_setup(struct mce *m)
+{
+ memset(m, 0, sizeof(struct mce));
+ m->cpu = smp_processor_id();
+ rdtscll(m->tsc);
+}
+
/*
* Lockless MCE logging infrastructure.
* This avoids deadlocks on printk locks without having to break locks. Also
@@ -119,11 +136,11 @@ static void print_mce(struct mce *m)
print_symbol("{%s}", m->ip);
printk("\n");
}
- printk(KERN_EMERG "TSC %Lx ", m->tsc);
+ printk(KERN_EMERG "TSC %llx ", m->tsc);
if (m->addr)
- printk("ADDR %Lx ", m->addr);
+ printk("ADDR %llx ", m->addr);
if (m->misc)
- printk("MISC %Lx ", m->misc);
+ printk("MISC %llx ", m->misc);
printk("\n");
printk(KERN_EMERG "This is not a software problem!\n");
printk(KERN_EMERG "Run through mcelog --ascii to decode "
@@ -149,8 +166,10 @@ static void mce_panic(char *msg, struct mce *backup, unsigned long start)
panic(msg);
}
-static int mce_available(struct cpuinfo_x86 *c)
+int mce_available(struct cpuinfo_x86 *c)
{
+ if (mce_dont_init)
+ return 0;
return cpu_has(c, X86_FEATURE_MCE) && cpu_has(c, X86_FEATURE_MCA);
}
@@ -172,7 +191,77 @@ static inline void mce_get_rip(struct mce *m, struct pt_regs *regs)
}
/*
- * The actual machine check handler
+ * Poll for corrected events or events that happened before reset.
+ * Those are just logged through /dev/mcelog.
+ *
+ * This is executed in standard interrupt context.
+ */
+void machine_check_poll(enum mcp_flags flags, mce_banks_t *b)
+{
+ struct mce m;
+ int i;
+
+ mce_setup(&m);
+
+ rdmsrl(MSR_IA32_MCG_STATUS, m.mcgstatus);
+ for (i = 0; i < banks; i++) {
+ if (!bank[i] || !test_bit(i, *b))
+ continue;
+
+ m.misc = 0;
+ m.addr = 0;
+ m.bank = i;
+ m.tsc = 0;
+
+ barrier();
+ rdmsrl(MSR_IA32_MC0_STATUS + i*4, m.status);
+ if (!(m.status & MCI_STATUS_VAL))
+ continue;
+
+ /*
+ * Uncorrected events are handled by the exception handler
+ * when it is enabled. But when the exception is disabled log
+ * everything.
+ *
+ * TBD do the same check for MCI_STATUS_EN here?
+ */
+ if ((m.status & MCI_STATUS_UC) && !(flags & MCP_UC))
+ continue;
+
+ if (m.status & MCI_STATUS_MISCV)
+ rdmsrl(MSR_IA32_MC0_MISC + i*4, m.misc);
+ if (m.status & MCI_STATUS_ADDRV)
+ rdmsrl(MSR_IA32_MC0_ADDR + i*4, m.addr);
+
+ if (!(flags & MCP_TIMESTAMP))
+ m.tsc = 0;
+ /*
+ * Don't get the IP here because it's unlikely to
+ * have anything to do with the actual error location.
+ */
+
+ mce_log(&m);
+ add_taint(TAINT_MACHINE_CHECK);
+
+ /*
+ * Clear state for this bank.
+ */
+ wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ }
+
+ /*
+ * Don't clear MCG_STATUS here because it's only defined for
+ * exceptions.
+ */
+}
+
+/*
+ * The actual machine check handler. This only handles real
+ * exceptions when something got corrupted coming in through int 18.
+ *
+ * This is executed in NMI context not subject to normal locking rules. This
+ * implies that most kernel services cannot be safely used. Don't even
+ * think about putting a printk in there!
*/
void do_machine_check(struct pt_regs * regs, long error_code)
{
@@ -190,17 +279,18 @@ void do_machine_check(struct pt_regs * regs, long error_code)
* error.
*/
int kill_it = 0;
+ DECLARE_BITMAP(toclear, MAX_NR_BANKS);
atomic_inc(&mce_entry);
- if ((regs
- && notify_die(DIE_NMI, "machine check", regs, error_code,
+ if (notify_die(DIE_NMI, "machine check", regs, error_code,
18, SIGKILL) == NOTIFY_STOP)
- || !banks)
+ goto out2;
+ if (!banks)
goto out2;
- memset(&m, 0, sizeof(struct mce));
- m.cpu = smp_processor_id();
+ mce_setup(&m);
+
rdmsrl(MSR_IA32_MCG_STATUS, m.mcgstatus);
/* if the restart IP is not valid, we're done for */
if (!(m.mcgstatus & MCG_STATUS_RIPV))
@@ -210,18 +300,32 @@ void do_machine_check(struct pt_regs * regs, long error_code)
barrier();
for (i = 0; i < banks; i++) {
- if (i < NR_SYSFS_BANKS && !bank[i])
+ __clear_bit(i, toclear);
+ if (!bank[i])
continue;
m.misc = 0;
m.addr = 0;
m.bank = i;
- m.tsc = 0;
rdmsrl(MSR_IA32_MC0_STATUS + i*4, m.status);
if ((m.status & MCI_STATUS_VAL) == 0)
continue;
+ /*
+ * Non uncorrected errors are handled by machine_check_poll
+ * Leave them alone.
+ */
+ if ((m.status & MCI_STATUS_UC) == 0)
+ continue;
+
+ /*
+ * Set taint even when machine check was not enabled.
+ */
+ add_taint(TAINT_MACHINE_CHECK);
+
+ __set_bit(i, toclear);
+
if (m.status & MCI_STATUS_EN) {
/* if PCC was set, there's no way out */
no_way_out |= !!(m.status & MCI_STATUS_PCC);
@@ -235,6 +339,12 @@ void do_machine_check(struct pt_regs * regs, long error_code)
no_way_out = 1;
kill_it = 1;
}
+ } else {
+ /*
+ * Machine check event was not enabled. Clear, but
+ * ignore.
+ */
+ continue;
}
if (m.status & MCI_STATUS_MISCV)
@@ -243,10 +353,7 @@ void do_machine_check(struct pt_regs * regs, long error_code)
rdmsrl(MSR_IA32_MC0_ADDR + i*4, m.addr);
mce_get_rip(&m, regs);
- if (error_code >= 0)
- rdtscll(m.tsc);
- if (error_code != -2)
- mce_log(&m);
+ mce_log(&m);
/* Did this bank cause the exception? */
/* Assume that the bank with uncorrectable errors did it,
@@ -255,14 +362,8 @@ void do_machine_check(struct pt_regs * regs, long error_code)
panicm = m;
panicm_found = 1;
}
-
- add_taint(TAINT_MACHINE_CHECK);
}
- /* Never do anything final in the polling timer */
- if (!regs)
- goto out;
-
/* If we didn't find an uncorrectable error, pick
the last one (shouldn't happen, just being safe). */
if (!panicm_found)
@@ -309,10 +410,11 @@ void do_machine_check(struct pt_regs * regs, long error_code)
/* notify userspace ASAP */
set_thread_flag(TIF_MCE_NOTIFY);
- out:
/* the last thing we do is clear state */
- for (i = 0; i < banks; i++)
- wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ for (i = 0; i < banks; i++) {
+ if (test_bit(i, toclear))
+ wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
+ }
wrmsrl(MSR_IA32_MCG_STATUS, 0);
out2:
atomic_dec(&mce_entry);
@@ -332,15 +434,13 @@ void do_machine_check(struct pt_regs * regs, long error_code)
* and historically has been the register value of the
* MSR_IA32_THERMAL_STATUS (Intel) msr.
*/
-void mce_log_therm_throt_event(unsigned int cpu, __u64 status)
+void mce_log_therm_throt_event(__u64 status)
{
struct mce m;
- memset(&m, 0, sizeof(m));
- m.cpu = cpu;
+ mce_setup(&m);
m.bank = MCE_THERMAL_BANK;
m.status = status;
- rdtscll(m.tsc);
mce_log(&m);
}
#endif /* CONFIG_X86_MCE_INTEL */
@@ -353,18 +453,18 @@ void mce_log_therm_throt_event(unsigned int cpu, __u64 status)
static int check_interval = 5 * 60; /* 5 minutes */
static int next_interval; /* in jiffies */
-static void mcheck_timer(struct work_struct *work);
-static DECLARE_DELAYED_WORK(mcheck_work, mcheck_timer);
+static void mcheck_timer(unsigned long);
+static DEFINE_PER_CPU(struct timer_list, mce_timer);
-static void mcheck_check_cpu(void *info)
+static void mcheck_timer(unsigned long data)
{
- if (mce_available(&current_cpu_data))
- do_machine_check(NULL, 0);
-}
+ struct timer_list *t = &per_cpu(mce_timer, data);
-static void mcheck_timer(struct work_struct *work)
-{
- on_each_cpu(mcheck_check_cpu, NULL, 1);
+ WARN_ON(smp_processor_id() != data);
+
+ if (mce_available(&current_cpu_data))
+ machine_check_poll(MCP_TIMESTAMP,
+ &__get_cpu_var(mce_poll_banks));
/*
* Alert userspace if needed. If we logged an MCE, reduce the
@@ -377,31 +477,41 @@ static void mcheck_timer(struct work_struct *work)
(int)round_jiffies_relative(check_interval*HZ));
}
- schedule_delayed_work(&mcheck_work, next_interval);
+ t->expires = jiffies + next_interval;
+ add_timer(t);
+}
+
+static void mce_do_trigger(struct work_struct *work)
+{
+ call_usermodehelper(trigger, trigger_argv, NULL, UMH_NO_WAIT);
}
+static DECLARE_WORK(mce_trigger_work, mce_do_trigger);
+
/*
- * This is only called from process context. This is where we do
- * anything we need to alert userspace about new MCEs. This is called
- * directly from the poller and also from entry.S and idle, thanks to
- * TIF_MCE_NOTIFY.
+ * Notify the user(s) about new machine check events.
+ * Can be called from interrupt context, but not from machine check/NMI
+ * context.
*/
int mce_notify_user(void)
{
+ /* Not more than two messages every minute */
+ static DEFINE_RATELIMIT_STATE(ratelimit, 60*HZ, 2);
+
clear_thread_flag(TIF_MCE_NOTIFY);
if (test_and_clear_bit(0, &notify_user)) {
- static unsigned long last_print;
- unsigned long now = jiffies;
-
wake_up_interruptible(&mce_wait);
- if (trigger[0])
- call_usermodehelper(trigger, trigger_argv, NULL,
- UMH_NO_WAIT);
- if (time_after_eq(now, last_print + (check_interval*HZ))) {
- last_print = now;
+ /*
+ * There is no risk of missing notifications because
+ * work_pending is always cleared before the function is
+ * executed.
+ */
+ if (trigger[0] && !work_pending(&mce_trigger_work))
+ schedule_work(&mce_trigger_work);
+
+ if (__ratelimit(&ratelimit))
printk(KERN_INFO "Machine check events logged\n");
- }
return 1;
}
@@ -425,63 +535,78 @@ static struct notifier_block mce_idle_notifier = {
static __init int periodic_mcheck_init(void)
{
- next_interval = check_interval * HZ;
- if (next_interval)
- schedule_delayed_work(&mcheck_work,
- round_jiffies_relative(next_interval));
- idle_notifier_register(&mce_idle_notifier);
- return 0;
+ idle_notifier_register(&mce_idle_notifier);
+ return 0;
}
__initcall(periodic_mcheck_init);
-
/*
* Initialize Machine Checks for a CPU.
*/
-static void mce_init(void *dummy)
+static int mce_cap_init(void)
{
u64 cap;
- int i;
+ unsigned b;
rdmsrl(MSR_IA32_MCG_CAP, cap);
- banks = cap & 0xff;
- if (banks > MCE_EXTENDED_BANK) {
- banks = MCE_EXTENDED_BANK;
- printk(KERN_INFO "MCE: warning: using only %d banks\n",
- MCE_EXTENDED_BANK);
+ b = cap & 0xff;
+ if (b > MAX_NR_BANKS) {
+ printk(KERN_WARNING
+ "MCE: Using only %u machine check banks out of %u\n",
+ MAX_NR_BANKS, b);
+ b = MAX_NR_BANKS;
}
+
+ /* Don't support asymmetric configurations today */
+ WARN_ON(banks != 0 && b != banks);
+ banks = b;
+ if (!bank) {
+ bank = kmalloc(banks * sizeof(u64), GFP_KERNEL);
+ if (!bank)
+ return -ENOMEM;
+ memset(bank, 0xff, banks * sizeof(u64));
+ }
+
/* Use accurate RIP reporting if available. */
if ((cap & (1<<9)) && ((cap >> 16) & 0xff) >= 9)
rip_msr = MSR_IA32_MCG_EIP;
- /* Log the machine checks left over from the previous reset.
- This also clears all registers */
- do_machine_check(NULL, mce_bootlog ? -1 : -2);
+ return 0;
+}
+
+static void mce_init(void *dummy)
+{
+ u64 cap;
+ int i;
+ mce_banks_t all_banks;
+
+ /*
+ * Log the machine checks left over from the previous reset.
+ */
+ bitmap_fill(all_banks, MAX_NR_BANKS);
+ machine_check_poll(MCP_UC, &all_banks);
set_in_cr4(X86_CR4_MCE);
+ rdmsrl(MSR_IA32_MCG_CAP, cap);
if (cap & MCG_CTL_P)
wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff);
for (i = 0; i < banks; i++) {
- if (i < NR_SYSFS_BANKS)
- wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]);
- else
- wrmsrl(MSR_IA32_MC0_CTL+4*i, ~0UL);
-
+ wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]);
wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
}
}
/* Add per CPU specific workarounds here */
-static void __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
+static void mce_cpu_quirks(struct cpuinfo_x86 *c)
{
/* This should be disabled by the BIOS, but isn't always */
if (c->x86_vendor == X86_VENDOR_AMD) {
- if(c->x86 == 15)
+ if (c->x86 == 15 && banks > 4)
/* disable GART TBL walk error reporting, which trips off
incorrectly with the IOMMU & 3ware & Cerberus. */
- clear_bit(10, &bank[4]);
+ clear_bit(10, (unsigned long *)&bank[4]);
if(c->x86 <= 17 && mce_bootlog < 0)
/* Lots of broken BIOS around that don't clear them
by default and leave crap in there. Don't log. */
@@ -504,20 +629,38 @@ static void mce_cpu_features(struct cpuinfo_x86 *c)
}
}
+static void mce_init_timer(void)
+{
+ struct timer_list *t = &__get_cpu_var(mce_timer);
+
+ /* data race harmless because everyone sets to the same value */
+ if (!next_interval)
+ next_interval = check_interval * HZ;
+ if (!next_interval)
+ return;
+ setup_timer(t, mcheck_timer, smp_processor_id());
+ t->expires = round_jiffies_relative(jiffies + next_interval);
+ add_timer(t);
+}
+
/*
* Called for each booted CPU to set up machine checks.
* Must be called with preempt off.
*/
void __cpuinit mcheck_init(struct cpuinfo_x86 *c)
{
- mce_cpu_quirks(c);
+ if (!mce_available(c))
+ return;
- if (mce_dont_init ||
- !mce_available(c))
+ if (mce_cap_init() < 0) {
+ mce_dont_init = 1;
return;
+ }
+ mce_cpu_quirks(c);
mce_init(NULL);
mce_cpu_features(c);
+ mce_init_timer();
}
/*
@@ -573,7 +716,7 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
{
unsigned long *cpu_tsc;
static DEFINE_MUTEX(mce_read_mutex);
- unsigned next;
+ unsigned prev, next;
char __user *buf = ubuf;
int i, err;
@@ -592,25 +735,32 @@ static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
}
err = 0;
- for (i = 0; i < next; i++) {
- unsigned long start = jiffies;
-
- while (!mcelog.entry[i].finished) {
- if (time_after_eq(jiffies, start + 2)) {
- memset(mcelog.entry + i,0, sizeof(struct mce));
- goto timeout;
+ prev = 0;
+ do {
+ for (i = prev; i < next; i++) {
+ unsigned long start = jiffies;
+
+ while (!mcelog.entry[i].finished) {
+ if (time_after_eq(jiffies, start + 2)) {
+ memset(mcelog.entry + i, 0,
+ sizeof(struct mce));
+ goto timeout;
+ }
+ cpu_relax();
}
- cpu_relax();
+ smp_rmb();
+ err |= copy_to_user(buf, mcelog.entry + i,
+ sizeof(struct mce));
+ buf += sizeof(struct mce);
+timeout:
+ ;
}
- smp_rmb();
- err |= copy_to_user(buf, mcelog.entry + i, sizeof(struct mce));
- buf += sizeof(struct mce);
- timeout:
- ;
- }
- memset(mcelog.entry, 0, next * sizeof(struct mce));
- mcelog.next = 0;
+ memset(mcelog.entry + prev, 0,
+ (next - prev) * sizeof(struct mce));
+ prev = next;
+ next = cmpxchg(&mcelog.next, prev, 0);
+ } while (next != prev);
synchronize_sched();
@@ -680,20 +830,6 @@ static struct miscdevice mce_log_device = {
&mce_chrdev_ops,
};
-static unsigned long old_cr4 __initdata;
-
-void __init stop_mce(void)
-{
- old_cr4 = read_cr4();
- clear_in_cr4(X86_CR4_MCE);
-}
-
-void __init restart_mce(void)
-{
- if (old_cr4 & X86_CR4_MCE)
- set_in_cr4(X86_CR4_MCE);
-}
-
/*
* Old style boot options parsing. Only for compatibility.
*/
@@ -703,8 +839,7 @@ static int __init mcheck_disable(char *str)
return 1;
}
-/* mce=off disables machine check. Note you can re-enable it later
- using sysfs.
+/* mce=off disables machine check.
mce=TOLERANCELEVEL (number, see above)
mce=bootlog Log MCEs from before booting. Disabled by default on AMD.
mce=nobootlog Don't log MCEs from before booting. */
@@ -728,6 +863,29 @@ __setup("mce=", mcheck_enable);
* Sysfs support
*/
+/*
+ * Disable machine checks on suspend and shutdown. We can't really handle
+ * them later.
+ */
+static int mce_disable(void)
+{
+ int i;
+
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, 0);
+ return 0;
+}
+
+static int mce_suspend(struct sys_device *dev, pm_message_t state)
+{
+ return mce_disable();
+}
+
+static int mce_shutdown(struct sys_device *dev)
+{
+ return mce_disable();
+}
+
/* On resume clear all MCE state. Don't want to see leftovers from the BIOS.
Only one CPU is active at this time, the others get readded later using
CPU hotplug. */
@@ -738,20 +896,24 @@ static int mce_resume(struct sys_device *dev)
return 0;
}
+static void mce_cpu_restart(void *data)
+{
+ del_timer_sync(&__get_cpu_var(mce_timer));
+ if (mce_available(&current_cpu_data))
+ mce_init(NULL);
+ mce_init_timer();
+}
+
/* Reinit MCEs after user configuration changes */
static void mce_restart(void)
{
- if (next_interval)
- cancel_delayed_work(&mcheck_work);
- /* Timer race is harmless here */
- on_each_cpu(mce_init, NULL, 1);
next_interval = check_interval * HZ;
- if (next_interval)
- schedule_delayed_work(&mcheck_work,
- round_jiffies_relative(next_interval));
+ on_each_cpu(mce_cpu_restart, NULL, 1);
}
static struct sysdev_class mce_sysclass = {
+ .suspend = mce_suspend,
+ .shutdown = mce_shutdown,
.resume = mce_resume,
.name = "machinecheck",
};
@@ -778,16 +940,26 @@ void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu) __cpuinit
} \
static SYSDEV_ATTR(name, 0644, show_ ## name, set_ ## name);
-/*
- * TBD should generate these dynamically based on number of available banks.
- * Have only 6 contol banks in /sysfs until then.
- */
-ACCESSOR(bank0ctl,bank[0],mce_restart())
-ACCESSOR(bank1ctl,bank[1],mce_restart())
-ACCESSOR(bank2ctl,bank[2],mce_restart())
-ACCESSOR(bank3ctl,bank[3],mce_restart())
-ACCESSOR(bank4ctl,bank[4],mce_restart())
-ACCESSOR(bank5ctl,bank[5],mce_restart())
+static struct sysdev_attribute *bank_attrs;
+
+static ssize_t show_bank(struct sys_device *s, struct sysdev_attribute *attr,
+ char *buf)
+{
+ u64 b = bank[attr - bank_attrs];
+ return sprintf(buf, "%llx\n", b);
+}
+
+static ssize_t set_bank(struct sys_device *s, struct sysdev_attribute *attr,
+ const char *buf, size_t siz)
+{
+ char *end;
+ u64 new = simple_strtoull(buf, &end, 0);
+ if (end == buf)
+ return -EINVAL;
+ bank[attr - bank_attrs] = new;
+ mce_restart();
+ return end-buf;
+}
static ssize_t show_trigger(struct sys_device *s, struct sysdev_attribute *attr,
char *buf)
@@ -814,8 +986,6 @@ static SYSDEV_ATTR(trigger, 0644, show_trigger, set_trigger);
static SYSDEV_INT_ATTR(tolerant, 0644, tolerant);
ACCESSOR(check_interval,check_interval,mce_restart())
static struct sysdev_attribute *mce_attributes[] = {
- &attr_bank0ctl, &attr_bank1ctl, &attr_bank2ctl,
- &attr_bank3ctl, &attr_bank4ctl, &attr_bank5ctl,
&attr_tolerant.attr, &attr_check_interval, &attr_trigger,
NULL
};
@@ -845,11 +1015,22 @@ static __cpuinit int mce_create_device(unsigned int cpu)
if (err)
goto error;
}
+ for (i = 0; i < banks; i++) {
+ err = sysdev_create_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
+ if (err)
+ goto error2;
+ }
cpu_set(cpu, mce_device_initialized);
return 0;
+error2:
+ while (--i >= 0) {
+ sysdev_remove_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
+ }
error:
- while (i--) {
+ while (--i >= 0) {
sysdev_remove_file(&per_cpu(device_mce,cpu),
mce_attributes[i]);
}
@@ -868,15 +1049,46 @@ static __cpuinit void mce_remove_device(unsigned int cpu)
for (i = 0; mce_attributes[i]; i++)
sysdev_remove_file(&per_cpu(device_mce,cpu),
mce_attributes[i]);
+ for (i = 0; i < banks; i++)
+ sysdev_remove_file(&per_cpu(device_mce, cpu),
+ &bank_attrs[i]);
sysdev_unregister(&per_cpu(device_mce,cpu));
cpu_clear(cpu, mce_device_initialized);
}
+/* Make sure there are no machine checks on offlined CPUs. */
+static void mce_disable_cpu(void *h)
+{
+ int i;
+ unsigned long action = *(unsigned long *)h;
+
+ if (!mce_available(&current_cpu_data))
+ return;
+ if (!(action & CPU_TASKS_FROZEN))
+ cmci_clear();
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, 0);
+}
+
+static void mce_reenable_cpu(void *h)
+{
+ int i;
+ unsigned long action = *(unsigned long *)h;
+
+ if (!mce_available(&current_cpu_data))
+ return;
+ if (!(action & CPU_TASKS_FROZEN))
+ cmci_reenable();
+ for (i = 0; i < banks; i++)
+ wrmsrl(MSR_IA32_MC0_CTL + i*4, bank[i]);
+}
+
/* Get notified when a cpu comes on/off. Be hotplug friendly. */
static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
unsigned long action, void *hcpu)
{
unsigned int cpu = (unsigned long)hcpu;
+ struct timer_list *t = &per_cpu(mce_timer, cpu);
switch (action) {
case CPU_ONLINE:
@@ -891,6 +1103,21 @@ static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
threshold_cpu_callback(action, cpu);
mce_remove_device(cpu);
break;
+ case CPU_DOWN_PREPARE:
+ case CPU_DOWN_PREPARE_FROZEN:
+ del_timer_sync(t);
+ smp_call_function_single(cpu, mce_disable_cpu, &action, 1);
+ break;
+ case CPU_DOWN_FAILED:
+ case CPU_DOWN_FAILED_FROZEN:
+ t->expires = round_jiffies_relative(jiffies + next_interval);
+ add_timer_on(t, cpu);
+ smp_call_function_single(cpu, mce_reenable_cpu, &action, 1);
+ break;
+ case CPU_POST_DEAD:
+ /* intentionally ignoring frozen here */
+ cmci_rediscover(cpu);
+ break;
}
return NOTIFY_OK;
}
@@ -899,6 +1126,34 @@ static struct notifier_block mce_cpu_notifier __cpuinitdata = {
.notifier_call = mce_cpu_callback,
};
+static __init int mce_init_banks(void)
+{
+ int i;
+
+ bank_attrs = kzalloc(sizeof(struct sysdev_attribute) * banks,
+ GFP_KERNEL);
+ if (!bank_attrs)
+ return -ENOMEM;
+
+ for (i = 0; i < banks; i++) {
+ struct sysdev_attribute *a = &bank_attrs[i];
+ a->attr.name = kasprintf(GFP_KERNEL, "bank%d", i);
+ if (!a->attr.name)
+ goto nomem;
+ a->attr.mode = 0644;
+ a->show = show_bank;
+ a->store = set_bank;
+ }
+ return 0;
+
+nomem:
+ while (--i >= 0)
+ kfree(bank_attrs[i].attr.name);
+ kfree(bank_attrs);
+ bank_attrs = NULL;
+ return -ENOMEM;
+}
+
static __init int mce_init_device(void)
{
int err;
@@ -906,6 +1161,11 @@ static __init int mce_init_device(void)
if (!mce_available(&boot_cpu_data))
return -EIO;
+
+ err = mce_init_banks();
+ if (err)
+ return err;
+
err = sysdev_class_register(&mce_sysclass);
if (err)
return err;
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
index 9817506dd46..c5a32f92d07 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
@@ -79,6 +79,8 @@ static unsigned char shared_bank[NR_BANKS] = {
static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */
+static void amd_threshold_interrupt(void);
+
/*
* CPU Initialization
*/
@@ -174,6 +176,8 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
tr.reset = 0;
tr.old_limit = 0;
threshold_restart_bank(&tr);
+
+ mce_threshold_vector = amd_threshold_interrupt;
}
}
}
@@ -187,19 +191,13 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
* the interrupt goes off when error_count reaches threshold_limit.
* the handler will simply log mcelog w/ software defined bank number.
*/
-asmlinkage void mce_threshold_interrupt(void)
+static void amd_threshold_interrupt(void)
{
unsigned int bank, block;
struct mce m;
u32 low = 0, high = 0, address = 0;
- ack_APIC_irq();
- exit_idle();
- irq_enter();
-
- memset(&m, 0, sizeof(m));
- rdtscll(m.tsc);
- m.cpu = smp_processor_id();
+ mce_setup(&m);
/* assume first bank caused it */
for (bank = 0; bank < NR_BANKS; ++bank) {
@@ -233,7 +231,8 @@ asmlinkage void mce_threshold_interrupt(void)
/* Log the machine check that caused the threshold
event. */
- do_machine_check(NULL, 0);
+ machine_check_poll(MCP_TIMESTAMP,
+ &__get_cpu_var(mce_poll_banks));
if (high & MASK_OVERFLOW_HI) {
rdmsrl(address, m.misc);
@@ -243,13 +242,10 @@ asmlinkage void mce_threshold_interrupt(void)
+ bank * NR_BLOCKS
+ block;
mce_log(&m);
- goto out;
+ return;
}
}
}
-out:
- inc_irq_stat(irq_threshold_count);
- irq_exit();
}
/*
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
index aa5e287c98e..aaa7d973093 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c
@@ -1,6 +1,8 @@
/*
* Intel specific MCE features.
* Copyright 2004 Zwane Mwaikambo <zwane@linuxpower.ca>
+ * Copyright (C) 2008, 2009 Intel Corporation
+ * Author: Andi Kleen
*/
#include <linux/init.h>
@@ -13,6 +15,7 @@
#include <asm/hw_irq.h>
#include <asm/idle.h>
#include <asm/therm_throt.h>
+#include <asm/apic.h>
asmlinkage void smp_thermal_interrupt(void)
{
@@ -25,7 +28,7 @@ asmlinkage void smp_thermal_interrupt(void)
rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
if (therm_throt_process(msr_val & 1))
- mce_log_therm_throt_event(smp_processor_id(), msr_val);
+ mce_log_therm_throt_event(msr_val);
inc_irq_stat(irq_thermal_count);
irq_exit();
@@ -85,7 +88,209 @@ static void intel_init_thermal(struct cpuinfo_x86 *c)
return;
}
+/*
+ * Support for Intel Correct Machine Check Interrupts. This allows
+ * the CPU to raise an interrupt when a corrected machine check happened.
+ * Normally we pick those up using a regular polling timer.
+ * Also supports reliable discovery of shared banks.
+ */
+
+static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
+
+/*
+ * cmci_discover_lock protects against parallel discovery attempts
+ * which could race against each other.
+ */
+static DEFINE_SPINLOCK(cmci_discover_lock);
+
+#define CMCI_THRESHOLD 1
+
+static int cmci_supported(int *banks)
+{
+ u64 cap;
+
+ /*
+ * Vendor check is not strictly needed, but the initial
+ * initialization is vendor keyed and this
+ * makes sure none of the backdoors are entered otherwise.
+ */
+ if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL)
+ return 0;
+ if (!cpu_has_apic || lapic_get_maxlvt() < 6)
+ return 0;
+ rdmsrl(MSR_IA32_MCG_CAP, cap);
+ *banks = min_t(unsigned, MAX_NR_BANKS, cap & 0xff);
+ return !!(cap & MCG_CMCI_P);
+}
+
+/*
+ * The interrupt handler. This is called on every event.
+ * Just call the poller directly to log any events.
+ * This could in theory increase the threshold under high load,
+ * but doesn't for now.
+ */
+static void intel_threshold_interrupt(void)
+{
+ machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned));
+ mce_notify_user();
+}
+
+static void print_update(char *type, int *hdr, int num)
+{
+ if (*hdr == 0)
+ printk(KERN_INFO "CPU %d MCA banks", smp_processor_id());
+ *hdr = 1;
+ printk(KERN_CONT " %s:%d", type, num);
+}
+
+/*
+ * Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks
+ * on this CPU. Use the algorithm recommended in the SDM to discover shared
+ * banks.
+ */
+static void cmci_discover(int banks, int boot)
+{
+ unsigned long *owned = (void *)&__get_cpu_var(mce_banks_owned);
+ int hdr = 0;
+ int i;
+
+ spin_lock(&cmci_discover_lock);
+ for (i = 0; i < banks; i++) {
+ u64 val;
+
+ if (test_bit(i, owned))
+ continue;
+
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+
+ /* Already owned by someone else? */
+ if (val & CMCI_EN) {
+ if (test_and_clear_bit(i, owned) || boot)
+ print_update("SHD", &hdr, i);
+ __clear_bit(i, __get_cpu_var(mce_poll_banks));
+ continue;
+ }
+
+ val |= CMCI_EN | CMCI_THRESHOLD;
+ wrmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+
+ /* Did the enable bit stick? -- the bank supports CMCI */
+ if (val & CMCI_EN) {
+ if (!test_and_set_bit(i, owned) || boot)
+ print_update("CMCI", &hdr, i);
+ __clear_bit(i, __get_cpu_var(mce_poll_banks));
+ } else {
+ WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks)));
+ }
+ }
+ spin_unlock(&cmci_discover_lock);
+ if (hdr)
+ printk(KERN_CONT "\n");
+}
+
+/*
+ * Just in case we missed an event during initialization check
+ * all the CMCI owned banks.
+ */
+void cmci_recheck(void)
+{
+ unsigned long flags;
+ int banks;
+
+ if (!mce_available(&current_cpu_data) || !cmci_supported(&banks))
+ return;
+ local_irq_save(flags);
+ machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned));
+ local_irq_restore(flags);
+}
+
+/*
+ * Disable CMCI on this CPU for all banks it owns when it goes down.
+ * This allows other CPUs to claim the banks on rediscovery.
+ */
+void cmci_clear(void)
+{
+ int i;
+ int banks;
+ u64 val;
+
+ if (!cmci_supported(&banks))
+ return;
+ spin_lock(&cmci_discover_lock);
+ for (i = 0; i < banks; i++) {
+ if (!test_bit(i, __get_cpu_var(mce_banks_owned)))
+ continue;
+ /* Disable CMCI */
+ rdmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ val &= ~(CMCI_EN|CMCI_THRESHOLD_MASK);
+ wrmsrl(MSR_IA32_MC0_CTL2 + i, val);
+ __clear_bit(i, __get_cpu_var(mce_banks_owned));
+ }
+ spin_unlock(&cmci_discover_lock);
+}
+
+/*
+ * After a CPU went down cycle through all the others and rediscover
+ * Must run in process context.
+ */
+void cmci_rediscover(int dying)
+{
+ int banks;
+ int cpu;
+ cpumask_var_t old;
+
+ if (!cmci_supported(&banks))
+ return;
+ if (!alloc_cpumask_var(&old, GFP_KERNEL))
+ return;
+ cpumask_copy(old, &current->cpus_allowed);
+
+ for_each_online_cpu (cpu) {
+ if (cpu == dying)
+ continue;
+ if (set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)))
+ continue;
+ /* Recheck banks in case CPUs don't all have the same */
+ if (cmci_supported(&banks))
+ cmci_discover(banks, 0);
+ }
+
+ set_cpus_allowed_ptr(current, old);
+ free_cpumask_var(old);
+}
+
+/*
+ * Reenable CMCI on this CPU in case a CPU down failed.
+ */
+void cmci_reenable(void)
+{
+ int banks;
+ if (cmci_supported(&banks))
+ cmci_discover(banks, 0);
+}
+
+static __cpuinit void intel_init_cmci(void)
+{
+ int banks;
+
+ if (!cmci_supported(&banks))
+ return;
+
+ mce_threshold_vector = intel_threshold_interrupt;
+ cmci_discover(banks, 1);
+ /*
+ * For CPU #0 this runs with still disabled APIC, but that's
+ * ok because only the vector is set up. We still do another
+ * check for the banks later for CPU #0 just to make sure
+ * to not miss any events.
+ */
+ apic_write(APIC_LVTCMCI, THRESHOLD_APIC_VECTOR|APIC_DM_FIXED);
+ cmci_recheck();
+}
+
void mce_intel_feature_init(struct cpuinfo_x86 *c)
{
intel_init_thermal(c);
+ intel_init_cmci();
}
diff --git a/arch/x86/kernel/cpu/mcheck/threshold.c b/arch/x86/kernel/cpu/mcheck/threshold.c
new file mode 100644
index 00000000000..23ee9e730f7
--- /dev/null
+++ b/arch/x86/kernel/cpu/mcheck/threshold.c
@@ -0,0 +1,29 @@
+/*
+ * Common corrected MCE threshold handler code:
+ */
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+
+#include <asm/irq_vectors.h>
+#include <asm/apic.h>
+#include <asm/idle.h>
+#include <asm/mce.h>
+
+static void default_threshold_interrupt(void)
+{
+ printk(KERN_ERR "Unexpected threshold interrupt at vector %x\n",
+ THRESHOLD_APIC_VECTOR);
+}
+
+void (*mce_threshold_vector)(void) = default_threshold_interrupt;
+
+asmlinkage void mce_threshold_interrupt(void)
+{
+ exit_idle();
+ irq_enter();
+ inc_irq_stat(irq_threshold_count);
+ mce_threshold_vector();
+ irq_exit();
+ /* Ack only at the end to avoid potential reentry */
+ ack_APIC_irq();
+}
diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c
index b0f61f0dcd0..f2f8540a7f3 100644
--- a/arch/x86/kernel/i387.c
+++ b/arch/x86/kernel/i387.c
@@ -136,7 +136,7 @@ int init_fpu(struct task_struct *tsk)
#ifdef CONFIG_X86_32
if (!HAVE_HWFP) {
memset(tsk->thread.xstate, 0, xstate_size);
- finit();
+ finit_task(tsk);
set_stopped_child_used_math(tsk);
return 0;
}
diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c
index 9dc6b2b2427..3b09634a515 100644
--- a/arch/x86/kernel/irq_32.c
+++ b/arch/x86/kernel/irq_32.c
@@ -16,6 +16,7 @@
#include <linux/cpu.h>
#include <linux/delay.h>
#include <linux/uaccess.h>
+#include <linux/percpu.h>
#include <asm/apic.h>
@@ -55,13 +56,13 @@ static inline void print_stack_overflow(void) { }
union irq_ctx {
struct thread_info tinfo;
u32 stack[THREAD_SIZE/sizeof(u32)];
-};
+} __attribute__((aligned(PAGE_SIZE)));
-static union irq_ctx *hardirq_ctx[NR_CPUS] __read_mostly;
-static union irq_ctx *softirq_ctx[NR_CPUS] __read_mostly;
+static DEFINE_PER_CPU(union irq_ctx *, hardirq_ctx);
+static DEFINE_PER_CPU(union irq_ctx *, softirq_ctx);
-static char softirq_stack[NR_CPUS * THREAD_SIZE] __page_aligned_bss;
-static char hardirq_stack[NR_CPUS * THREAD_SIZE] __page_aligned_bss;
+static DEFINE_PER_CPU_PAGE_ALIGNED(union irq_ctx, hardirq_stack);
+static DEFINE_PER_CPU_PAGE_ALIGNED(union irq_ctx, softirq_stack);
static void call_on_stack(void *func, void *stack)
{
@@ -81,7 +82,7 @@ execute_on_irq_stack(int overflow, struct irq_desc *desc, int irq)
u32 *isp, arg1, arg2;
curctx = (union irq_ctx *) current_thread_info();
- irqctx = hardirq_ctx[smp_processor_id()];
+ irqctx = __get_cpu_var(hardirq_ctx);
/*
* this is where we switch to the IRQ stack. However, if we are
@@ -125,34 +126,34 @@ void __cpuinit irq_ctx_init(int cpu)
{
union irq_ctx *irqctx;
- if (hardirq_ctx[cpu])
+ if (per_cpu(hardirq_ctx, cpu))
return;
- irqctx = (union irq_ctx*) &hardirq_stack[cpu*THREAD_SIZE];
+ irqctx = &per_cpu(hardirq_stack, cpu);
irqctx->tinfo.task = NULL;
irqctx->tinfo.exec_domain = NULL;
irqctx->tinfo.cpu = cpu;
irqctx->tinfo.preempt_count = HARDIRQ_OFFSET;
irqctx->tinfo.addr_limit = MAKE_MM_SEG(0);
- hardirq_ctx[cpu] = irqctx;
+ per_cpu(hardirq_ctx, cpu) = irqctx;
- irqctx = (union irq_ctx *) &softirq_stack[cpu*THREAD_SIZE];
+ irqctx = &per_cpu(softirq_stack, cpu);
irqctx->tinfo.task = NULL;
irqctx->tinfo.exec_domain = NULL;
irqctx->tinfo.cpu = cpu;
irqctx->tinfo.preempt_count = 0;
irqctx->tinfo.addr_limit = MAKE_MM_SEG(0);
- softirq_ctx[cpu] = irqctx;
+ per_cpu(softirq_ctx, cpu) = irqctx;
printk(KERN_DEBUG "CPU %u irqstacks, hard=%p soft=%p\n",
- cpu, hardirq_ctx[cpu], softirq_ctx[cpu]);
+ cpu, per_cpu(hardirq_ctx, cpu), per_cpu(softirq_ctx, cpu));
}
void irq_ctx_exit(int cpu)
{
- hardirq_ctx[cpu] = NULL;
+ per_cpu(hardirq_ctx, cpu) = NULL;
}
asmlinkage void do_softirq(void)
@@ -169,7 +170,7 @@ asmlinkage void do_softirq(void)
if (local_softirq_pending()) {
curctx = current_thread_info();
- irqctx = softirq_ctx[smp_processor_id()];
+ irqctx = __get_cpu_var(softirq_ctx);
irqctx->tinfo.task = curctx->task;
irqctx->tinfo.previous_esp = current_stack_pointer;
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 1cc18d439bb..2aef36d8aca 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -216,6 +216,14 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = {
DMI_MATCH(DMI_PRODUCT_NAME, "HP Compaq"),
},
},
+ { /* Handle problems with rebooting on Dell XPS710 */
+ .callback = set_bios_reboot,
+ .ident = "Dell XPS710",
+ .matches = {
+ DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+ DMI_MATCH(DMI_PRODUCT_NAME, "Dell XPS710"),
+ },
+ },
{ }
};
diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c
index d992e6cff73..c29f301d388 100644
--- a/arch/x86/kernel/setup_percpu.c
+++ b/arch/x86/kernel/setup_percpu.c
@@ -7,6 +7,7 @@
#include <linux/crash_dump.h>
#include <linux/smp.h>
#include <linux/topology.h>
+#include <linux/pfn.h>
#include <asm/sections.h>
#include <asm/processor.h>
#include <asm/setup.h>
@@ -41,6 +42,321 @@ unsigned long __per_cpu_offset[NR_CPUS] __read_mostly = {
};
EXPORT_SYMBOL(__per_cpu_offset);
+/**
+ * pcpu_need_numa - determine percpu allocation needs to consider NUMA
+ *
+ * If NUMA is not configured or there is only one NUMA node available,
+ * there is no reason to consider NUMA. This function determines
+ * whether percpu allocation should consider NUMA or not.
+ *
+ * RETURNS:
+ * true if NUMA should be considered; otherwise, false.
+ */
+static bool __init pcpu_need_numa(void)
+{
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+ pg_data_t *last = NULL;
+ unsigned int cpu;
+
+ for_each_possible_cpu(cpu) {
+ int node = early_cpu_to_node(cpu);
+
+ if (node_online(node) && NODE_DATA(node) &&
+ last && last != NODE_DATA(node))
+ return true;
+
+ last = NODE_DATA(node);
+ }
+#endif
+ return false;
+}
+
+/**
+ * pcpu_alloc_bootmem - NUMA friendly alloc_bootmem wrapper for percpu
+ * @cpu: cpu to allocate for
+ * @size: size allocation in bytes
+ * @align: alignment
+ *
+ * Allocate @size bytes aligned at @align for cpu @cpu. This wrapper
+ * does the right thing for NUMA regardless of the current
+ * configuration.
+ *
+ * RETURNS:
+ * Pointer to the allocated area on success, NULL on failure.
+ */
+static void * __init pcpu_alloc_bootmem(unsigned int cpu, unsigned long size,
+ unsigned long align)
+{
+ const unsigned long goal = __pa(MAX_DMA_ADDRESS);
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+ int node = early_cpu_to_node(cpu);
+ void *ptr;
+
+ if (!node_online(node) || !NODE_DATA(node)) {
+ ptr = __alloc_bootmem_nopanic(size, align, goal);
+ pr_info("cpu %d has no node %d or node-local memory\n",
+ cpu, node);
+ pr_debug("per cpu data for cpu%d %lu bytes at %016lx\n",
+ cpu, size, __pa(ptr));
+ } else {
+ ptr = __alloc_bootmem_node_nopanic(NODE_DATA(node),
+ size, align, goal);
+ pr_debug("per cpu data for cpu%d %lu bytes on node%d at "
+ "%016lx\n", cpu, size, node, __pa(ptr));
+ }
+ return ptr;
+#else
+ return __alloc_bootmem_nopanic(size, align, goal);
+#endif
+}
+
+/*
+ * Remap allocator
+ *
+ * This allocator uses PMD page as unit. A PMD page is allocated for
+ * each cpu and each is remapped into vmalloc area using PMD mapping.
+ * As PMD page is quite large, only part of it is used for the first
+ * chunk. Unused part is returned to the bootmem allocator.
+ *
+ * So, the PMD pages are mapped twice - once to the physical mapping
+ * and to the vmalloc area for the first percpu chunk. The double
+ * mapping does add one more PMD TLB entry pressure but still is much
+ * better than only using 4k mappings while still being NUMA friendly.
+ */
+#ifdef CONFIG_NEED_MULTIPLE_NODES
+static size_t pcpur_size __initdata;
+static void **pcpur_ptrs __initdata;
+
+static struct page * __init pcpur_get_page(unsigned int cpu, int pageno)
+{
+ size_t off = (size_t)pageno << PAGE_SHIFT;
+
+ if (off >= pcpur_size)
+ return NULL;
+
+ return virt_to_page(pcpur_ptrs[cpu] + off);
+}
+
+static ssize_t __init setup_pcpu_remap(size_t static_size)
+{
+ static struct vm_struct vm;
+ pg_data_t *last;
+ size_t ptrs_size;
+ unsigned int cpu;
+ ssize_t ret;
+
+ /*
+ * If large page isn't supported, there's no benefit in doing
+ * this. Also, on non-NUMA, embedding is better.
+ */
+ if (!cpu_has_pse || pcpu_need_numa())
+ return -EINVAL;
+
+ last = NULL;
+ for_each_possible_cpu(cpu) {
+ int node = early_cpu_to_node(cpu);
+
+ if (node_online(node) && NODE_DATA(node) &&
+ last && last != NODE_DATA(node))
+ goto proceed;
+
+ last = NODE_DATA(node);
+ }
+ return -EINVAL;
+
+proceed:
+ /*
+ * Currently supports only single page. Supporting multiple
+ * pages won't be too difficult if it ever becomes necessary.
+ */
+ pcpur_size = PFN_ALIGN(static_size + PERCPU_DYNAMIC_RESERVE);
+ if (pcpur_size > PMD_SIZE) {
+ pr_warning("PERCPU: static data is larger than large page, "
+ "can't use large page\n");
+ return -EINVAL;
+ }
+
+ /* allocate pointer array and alloc large pages */
+ ptrs_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpur_ptrs[0]));
+ pcpur_ptrs = alloc_bootmem(ptrs_size);
+
+ for_each_possible_cpu(cpu) {
+ pcpur_ptrs[cpu] = pcpu_alloc_bootmem(cpu, PMD_SIZE, PMD_SIZE);
+ if (!pcpur_ptrs[cpu])
+ goto enomem;
+
+ /*
+ * Only use pcpur_size bytes and give back the rest.
+ *
+ * Ingo: The 2MB up-rounding bootmem is needed to make
+ * sure the partial 2MB page is still fully RAM - it's
+ * not well-specified to have a PAT-incompatible area
+ * (unmapped RAM, device memory, etc.) in that hole.
+ */
+ free_bootmem(__pa(pcpur_ptrs[cpu] + pcpur_size),
+ PMD_SIZE - pcpur_size);
+
+ memcpy(pcpur_ptrs[cpu], __per_cpu_load, static_size);
+ }
+
+ /* allocate address and map */
+ vm.flags = VM_ALLOC;
+ vm.size = num_possible_cpus() * PMD_SIZE;
+ vm_area_register_early(&vm, PMD_SIZE);
+
+ for_each_possible_cpu(cpu) {
+ pmd_t *pmd;
+
+ pmd = populate_extra_pmd((unsigned long)vm.addr
+ + cpu * PMD_SIZE);
+ set_pmd(pmd, pfn_pmd(page_to_pfn(virt_to_page(pcpur_ptrs[cpu])),
+ PAGE_KERNEL_LARGE));
+ }
+
+ /* we're ready, commit */
+ pr_info("PERCPU: Remapped at %p with large pages, static data "
+ "%zu bytes\n", vm.addr, static_size);
+
+ ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, PMD_SIZE,
+ pcpur_size - static_size, vm.addr, NULL);
+ goto out_free_ar;
+
+enomem:
+ for_each_possible_cpu(cpu)
+ if (pcpur_ptrs[cpu])
+ free_bootmem(__pa(pcpur_ptrs[cpu]), PMD_SIZE);
+ ret = -ENOMEM;
+out_free_ar:
+ free_bootmem(__pa(pcpur_ptrs), ptrs_size);
+ return ret;
+}
+#else
+static ssize_t __init setup_pcpu_remap(size_t static_size)
+{
+ return -EINVAL;
+}
+#endif
+
+/*
+ * Embedding allocator
+ *
+ * The first chunk is sized to just contain the static area plus
+ * PERCPU_DYNAMIC_RESERVE and allocated as a contiguous area using
+ * bootmem allocator and used as-is without being mapped into vmalloc
+ * area. This enables the first chunk to piggy back on the linear
+ * physical PMD mapping and doesn't add any additional pressure to
+ * TLB.
+ */
+static void *pcpue_ptr __initdata;
+static size_t pcpue_unit_size __initdata;
+
+static struct page * __init pcpue_get_page(unsigned int cpu, int pageno)
+{
+ return virt_to_page(pcpue_ptr + cpu * pcpue_unit_size
+ + ((size_t)pageno << PAGE_SHIFT));
+}
+
+static ssize_t __init setup_pcpu_embed(size_t static_size)
+{
+ unsigned int cpu;
+
+ /*
+ * If large page isn't supported, there's no benefit in doing
+ * this. Also, embedding allocation doesn't play well with
+ * NUMA.
+ */
+ if (!cpu_has_pse || pcpu_need_numa())
+ return -EINVAL;
+
+ /* allocate and copy */
+ pcpue_unit_size = PFN_ALIGN(static_size + PERCPU_DYNAMIC_RESERVE);
+ pcpue_unit_size = max_t(size_t, pcpue_unit_size, PCPU_MIN_UNIT_SIZE);
+ pcpue_ptr = pcpu_alloc_bootmem(0, num_possible_cpus() * pcpue_unit_size,
+ PAGE_SIZE);
+ if (!pcpue_ptr)
+ return -ENOMEM;
+
+ for_each_possible_cpu(cpu)
+ memcpy(pcpue_ptr + cpu * pcpue_unit_size, __per_cpu_load,
+ static_size);
+
+ /* we're ready, commit */
+ pr_info("PERCPU: Embedded %zu pages at %p, static data %zu bytes\n",
+ pcpue_unit_size >> PAGE_SHIFT, pcpue_ptr, static_size);
+
+ return pcpu_setup_first_chunk(pcpue_get_page, static_size,
+ pcpue_unit_size,
+ pcpue_unit_size - static_size, pcpue_ptr,
+ NULL);
+}
+
+/*
+ * 4k page allocator
+ *
+ * This is the basic allocator. Static percpu area is allocated
+ * page-by-page and most of initialization is done by the generic
+ * setup function.
+ */
+static struct page **pcpu4k_pages __initdata;
+static int pcpu4k_nr_static_pages __initdata;
+
+static struct page * __init pcpu4k_get_page(unsigned int cpu, int pageno)
+{
+ if (pageno < pcpu4k_nr_static_pages)
+ return pcpu4k_pages[cpu * pcpu4k_nr_static_pages + pageno];
+ return NULL;
+}
+
+static void __init pcpu4k_populate_pte(unsigned long addr)
+{
+ populate_extra_pte(addr);
+}
+
+static ssize_t __init setup_pcpu_4k(size_t static_size)
+{
+ size_t pages_size;
+ unsigned int cpu;
+ int i, j;
+ ssize_t ret;
+
+ pcpu4k_nr_static_pages = PFN_UP(static_size);
+
+ /* unaligned allocations can't be freed, round up to page size */
+ pages_size = PFN_ALIGN(pcpu4k_nr_static_pages * num_possible_cpus()
+ * sizeof(pcpu4k_pages[0]));
+ pcpu4k_pages = alloc_bootmem(pages_size);
+
+ /* allocate and copy */
+ j = 0;
+ for_each_possible_cpu(cpu)
+ for (i = 0; i < pcpu4k_nr_static_pages; i++) {
+ void *ptr;
+
+ ptr = pcpu_alloc_bootmem(cpu, PAGE_SIZE, PAGE_SIZE);
+ if (!ptr)
+ goto enomem;
+
+ memcpy(ptr, __per_cpu_load + i * PAGE_SIZE, PAGE_SIZE);
+ pcpu4k_pages[j++] = virt_to_page(ptr);
+ }
+
+ /* we're ready, commit */
+ pr_info("PERCPU: Allocated %d 4k pages, static data %zu bytes\n",
+ pcpu4k_nr_static_pages, static_size);
+
+ ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, 0, 0, NULL,
+ pcpu4k_populate_pte);
+ goto out_free_ar;
+
+enomem:
+ while (--j >= 0)
+ free_bootmem(__pa(page_address(pcpu4k_pages[j])), PAGE_SIZE);
+ ret = -ENOMEM;
+out_free_ar:
+ free_bootmem(__pa(pcpu4k_pages), pages_size);
+ return ret;
+}
+
static inline void setup_percpu_segment(int cpu)
{
#ifdef CONFIG_X86_32
@@ -61,38 +377,35 @@ static inline void setup_percpu_segment(int cpu)
*/
void __init setup_per_cpu_areas(void)
{
- ssize_t size;
- char *ptr;
- int cpu;
-
- /* Copy section for each CPU (we discard the original) */
- size = roundup(PERCPU_ENOUGH_ROOM, PAGE_SIZE);
+ size_t static_size = __per_cpu_end - __per_cpu_start;
+ unsigned int cpu;
+ unsigned long delta;
+ size_t pcpu_unit_size;
+ ssize_t ret;
pr_info("NR_CPUS:%d nr_cpumask_bits:%d nr_cpu_ids:%d nr_node_ids:%d\n",
NR_CPUS, nr_cpumask_bits, nr_cpu_ids, nr_node_ids);
- pr_info("PERCPU: Allocating %zd bytes of per cpu data\n", size);
+ /*
+ * Allocate percpu area. If PSE is supported, try to make use
+ * of large page mappings. Please read comments on top of
+ * each allocator for details.
+ */
+ ret = setup_pcpu_remap(static_size);
+ if (ret < 0)
+ ret = setup_pcpu_embed(static_size);
+ if (ret < 0)
+ ret = setup_pcpu_4k(static_size);
+ if (ret < 0)
+ panic("cannot allocate static percpu area (%zu bytes, err=%zd)",
+ static_size, ret);
- for_each_possible_cpu(cpu) {
-#ifndef CONFIG_NEED_MULTIPLE_NODES
- ptr = alloc_bootmem_pages(size);
-#else
- int node = early_cpu_to_node(cpu);
- if (!node_online(node) || !NODE_DATA(node)) {
- ptr = alloc_bootmem_pages(size);
- pr_info("cpu %d has no node %d or node-local memory\n",
- cpu, node);
- pr_debug("per cpu data for cpu%d at %016lx\n",
- cpu, __pa(ptr));
- } else {
- ptr = alloc_bootmem_pages_node(NODE_DATA(node), size);
- pr_debug("per cpu data for cpu%d on node%d at %016lx\n",
- cpu, node, __pa(ptr));
- }
-#endif
+ pcpu_unit_size = ret;
- memcpy(ptr, __per_cpu_load, __per_cpu_end - __per_cpu_start);
- per_cpu_offset(cpu) = ptr - __per_cpu_start;
+ /* alrighty, percpu areas up and running */
+ delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
+ for_each_possible_cpu(cpu) {
+ per_cpu_offset(cpu) = delta + cpu * pcpu_unit_size;
per_cpu(this_cpu_off, cpu) = per_cpu_offset(cpu);
per_cpu(cpu_number, cpu) = cpu;
setup_percpu_segment(cpu);
@@ -125,8 +438,6 @@ void __init setup_per_cpu_areas(void)
*/
if (cpu == boot_cpu_id)
switch_to_new_gdt(cpu);
-
- DBG("PERCPU: cpu %4d %p\n", cpu, ptr);
}
/* indicate the early static arrays will soon be gone */
diff --git a/arch/x86/math-emu/fpu_aux.c b/arch/x86/math-emu/fpu_aux.c
index 491e737ce54..aa098708877 100644
--- a/arch/x86/math-emu/fpu_aux.c
+++ b/arch/x86/math-emu/fpu_aux.c
@@ -30,20 +30,29 @@ static void fclex(void)
}
/* Needs to be externally visible */
-void finit(void)
+void finit_task(struct task_struct *tsk)
{
- control_word = 0x037f;
- partial_status = 0;
- top = 0; /* We don't keep top in the status word internally. */
- fpu_tag_word = 0xffff;
+ struct i387_soft_struct *soft = &tsk->thread.xstate->soft;
+ struct address *oaddr, *iaddr;
+ soft->cwd = 0x037f;
+ soft->swd = 0;
+ soft->ftop = 0; /* We don't keep top in the status word internally. */
+ soft->twd = 0xffff;
/* The behaviour is different from that detailed in
Section 15.1.6 of the Intel manual */
- operand_address.offset = 0;
- operand_address.selector = 0;
- instruction_address.offset = 0;
- instruction_address.selector = 0;
- instruction_address.opcode = 0;
- no_ip_update = 1;
+ oaddr = (struct address *)&soft->foo;
+ oaddr->offset = 0;
+ oaddr->selector = 0;
+ iaddr = (struct address *)&soft->fip;
+ iaddr->offset = 0;
+ iaddr->selector = 0;
+ iaddr->opcode = 0;
+ soft->no_update = 1;
+}
+
+void finit(void)
+{
+ finit_task(current);
}
/*
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index d57dfffb021..2966c6b8d30 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -131,6 +131,23 @@ static pte_t * __init one_page_table_init(pmd_t *pmd)
return pte_offset_kernel(pmd, 0);
}
+pmd_t * __init populate_extra_pmd(unsigned long vaddr)
+{
+ int pgd_idx = pgd_index(vaddr);
+ int pmd_idx = pmd_index(vaddr);
+
+ return one_md_table_init(swapper_pg_dir + pgd_idx) + pmd_idx;
+}
+
+pte_t * __init populate_extra_pte(unsigned long vaddr)
+{
+ int pte_idx = pte_index(vaddr);
+ pmd_t *pmd;
+
+ pmd = populate_extra_pmd(vaddr);
+ return one_page_table_init(pmd) + pte_idx;
+}
+
static pte_t *__init page_table_kmap_check(pte_t *pte, pmd_t *pmd,
unsigned long vaddr, pte_t *lastpte)
{
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 7dd7ce49d69..8a853bc3b28 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -161,34 +161,51 @@ static __ref void *spp_getpage(void)
return ptr;
}
-void
-set_pte_vaddr_pud(pud_t *pud_page, unsigned long vaddr, pte_t new_pte)
+static pud_t *fill_pud(pgd_t *pgd, unsigned long vaddr)
{
- pud_t *pud;
- pmd_t *pmd;
- pte_t *pte;
+ if (pgd_none(*pgd)) {
+ pud_t *pud = (pud_t *)spp_getpage();
+ pgd_populate(&init_mm, pgd, pud);
+ if (pud != pud_offset(pgd, 0))
+ printk(KERN_ERR "PAGETABLE BUG #00! %p <-> %p\n",
+ pud, pud_offset(pgd, 0));
+ }
+ return pud_offset(pgd, vaddr);
+}
- pud = pud_page + pud_index(vaddr);
+static pmd_t *fill_pmd(pud_t *pud, unsigned long vaddr)
+{
if (pud_none(*pud)) {
- pmd = (pmd_t *) spp_getpage();
+ pmd_t *pmd = (pmd_t *) spp_getpage();
pud_populate(&init_mm, pud, pmd);
- if (pmd != pmd_offset(pud, 0)) {
+ if (pmd != pmd_offset(pud, 0))
printk(KERN_ERR "PAGETABLE BUG #01! %p <-> %p\n",
- pmd, pmd_offset(pud, 0));
- return;
- }
+ pmd, pmd_offset(pud, 0));
}
- pmd = pmd_offset(pud, vaddr);
+ return pmd_offset(pud, vaddr);
+}
+
+static pte_t *fill_pte(pmd_t *pmd, unsigned long vaddr)
+{
if (pmd_none(*pmd)) {
- pte = (pte_t *) spp_getpage();
+ pte_t *pte = (pte_t *) spp_getpage();
pmd_populate_kernel(&init_mm, pmd, pte);
- if (pte != pte_offset_kernel(pmd, 0)) {
+ if (pte != pte_offset_kernel(pmd, 0))
printk(KERN_ERR "PAGETABLE BUG #02!\n");
- return;
- }
}
+ return pte_offset_kernel(pmd, vaddr);
+}
+
+void set_pte_vaddr_pud(pud_t *pud_page, unsigned long vaddr, pte_t new_pte)
+{
+ pud_t *pud;
+ pmd_t *pmd;
+ pte_t *pte;
+
+ pud = pud_page + pud_index(vaddr);
+ pmd = fill_pmd(pud, vaddr);
+ pte = fill_pte(pmd, vaddr);
- pte = pte_offset_kernel(pmd, vaddr);
set_pte(pte, new_pte);
/*
@@ -198,8 +215,7 @@ set_pte_vaddr_pud(pud_t *pud_page, unsigned long vaddr, pte_t new_pte)
__flush_tlb_one(vaddr);
}
-void
-set_pte_vaddr(unsigned long vaddr, pte_t pteval)
+void set_pte_vaddr(unsigned long vaddr, pte_t pteval)
{
pgd_t *pgd;
pud_t *pud_page;
@@ -216,6 +232,24 @@ set_pte_vaddr(unsigned long vaddr, pte_t pteval)
set_pte_vaddr_pud(pud_page, vaddr, pteval);
}
+pmd_t * __init populate_extra_pmd(unsigned long vaddr)
+{
+ pgd_t *pgd;
+ pud_t *pud;
+
+ pgd = pgd_offset_k(vaddr);
+ pud = fill_pud(pgd, vaddr);
+ return fill_pmd(pud, vaddr);
+}
+
+pte_t * __init populate_extra_pte(unsigned long vaddr)
+{
+ pmd_t *pmd;
+
+ pmd = populate_extra_pmd(vaddr);
+ return fill_pte(pmd, vaddr);
+}
+
/*
* Create large page table mappings for a range of physical addresses.
*/
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index c52f4034c7f..82cd39a6cbd 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -103,7 +103,7 @@ static void xen_vcpu_setup(int cpu)
vcpup = &per_cpu(xen_vcpu_info, cpu);
- info.mfn = virt_to_mfn(vcpup);
+ info.mfn = arbitrary_virt_to_mfn(vcpup);
info.offset = offset_in_page(vcpup);
printk(KERN_DEBUG "trying to map vcpu_info %d at %p, mfn %llx, offset %d\n",
@@ -301,8 +301,10 @@ static void xen_load_gdt(const struct desc_ptr *dtr)
frames = mcs.args;
for (f = 0; va < dtr->address + size; va += PAGE_SIZE, f++) {
- frames[f] = virt_to_mfn(va);
+ frames[f] = arbitrary_virt_to_mfn((void *)va);
+
make_lowmem_page_readonly((void *)va);
+ make_lowmem_page_readonly(mfn_to_virt(frames[f]));
}
MULTI_set_gdt(mcs.mc, frames, size / sizeof(struct desc_struct));
@@ -314,7 +316,7 @@ static void load_TLS_descriptor(struct thread_struct *t,
unsigned int cpu, unsigned int i)
{
struct desc_struct *gdt = get_cpu_gdt_table(cpu);
- xmaddr_t maddr = virt_to_machine(&gdt[GDT_ENTRY_TLS_MIN+i]);
+ xmaddr_t maddr = arbitrary_virt_to_machine(&gdt[GDT_ENTRY_TLS_MIN+i]);
struct multicall_space mc = __xen_mc_entry(0);
MULTI_update_descriptor(mc.mc, maddr.maddr, t->tls_array[i]);
@@ -488,7 +490,7 @@ static void xen_write_gdt_entry(struct desc_struct *dt, int entry,
break;
default: {
- xmaddr_t maddr = virt_to_machine(&dt[entry]);
+ xmaddr_t maddr = arbitrary_virt_to_machine(&dt[entry]);
xen_mc_flush();
if (HYPERVISOR_update_descriptor(maddr.maddr, *(u64 *)desc))
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 319bd40a57c..cb6afa4ec95 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -276,6 +276,13 @@ void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
p2m_top[topidx][idx] = mfn;
}
+unsigned long arbitrary_virt_to_mfn(void *vaddr)
+{
+ xmaddr_t maddr = arbitrary_virt_to_machine(vaddr);
+
+ return PFN_DOWN(maddr.maddr);
+}
+
xmaddr_t arbitrary_virt_to_machine(void *vaddr)
{
unsigned long address = (unsigned long)vaddr;
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index 035582ae815..8d470562ffc 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -219,6 +219,7 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle)
{
struct vcpu_guest_context *ctxt;
struct desc_struct *gdt;
+ unsigned long gdt_mfn;
if (cpumask_test_and_set_cpu(cpu, xen_cpu_initialized_map))
return 0;
@@ -248,9 +249,12 @@ cpu_initialize_context(unsigned int cpu, struct task_struct *idle)
ctxt->ldt_ents = 0;
BUG_ON((unsigned long)gdt & ~PAGE_MASK);
+
+ gdt_mfn = arbitrary_virt_to_mfn(gdt);
make_lowmem_page_readonly(gdt);
+ make_lowmem_page_readonly(mfn_to_virt(gdt_mfn));
- ctxt->gdt_frames[0] = virt_to_mfn(gdt);
+ ctxt->gdt_frames[0] = gdt_mfn;
ctxt->gdt_ents = GDT_ENTRIES;
ctxt->user_regs.cs = __KERNEL_CS;