diff options
Diffstat (limited to 'arch/x86/kernel/cpu')
38 files changed, 5295 insertions, 2399 deletions
diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 4e242f9a06e..c1f253dac15 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -1,5 +1,5 @@ # -# Makefile for x86-compatible CPU details and quirks +# Makefile for x86-compatible CPU details, features and quirks # # Don't trace early stages of a secondary CPU boot @@ -7,6 +7,10 @@ ifdef CONFIG_FUNCTION_TRACER CFLAGS_REMOVE_common.o = -pg endif +# Make sure load_percpu_segment has no stackprotector +nostackp := $(call cc-option, -fno-stack-protector) +CFLAGS_common.o := $(nostackp) + obj-y := intel_cacheinfo.o addon_cpuid_features.o obj-y += proc.o capflags.o powerflags.o common.o obj-y += vmware.o hypervisor.o @@ -23,11 +27,13 @@ obj-$(CONFIG_CPU_SUP_CENTAUR) += centaur.o obj-$(CONFIG_CPU_SUP_TRANSMETA_32) += transmeta.o obj-$(CONFIG_CPU_SUP_UMC_32) += umc.o -obj-$(CONFIG_X86_MCE) += mcheck/ -obj-$(CONFIG_MTRR) += mtrr/ -obj-$(CONFIG_CPU_FREQ) += cpufreq/ +obj-$(CONFIG_PERF_COUNTERS) += perf_counter.o + +obj-$(CONFIG_X86_MCE) += mcheck/ +obj-$(CONFIG_MTRR) += mtrr/ +obj-$(CONFIG_CPU_FREQ) += cpufreq/ -obj-$(CONFIG_X86_LOCAL_APIC) += perfctr-watchdog.o +obj-$(CONFIG_X86_LOCAL_APIC) += perfctr-watchdog.o quiet_cmd_mkcapflags = MKCAP $@ cmd_mkcapflags = $(PERL) $(srctree)/$(src)/mkcapflags.pl $< $@ diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 7e4a459daa6..63fddcd082c 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -6,6 +6,7 @@ #include <asm/processor.h> #include <asm/apic.h> #include <asm/cpu.h> +#include <asm/pci-direct.h> #ifdef CONFIG_X86_64 # include <asm/numa_64.h> @@ -257,13 +258,15 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) { #ifdef CONFIG_X86_HT unsigned bits; + int cpu = smp_processor_id(); bits = c->x86_coreid_bits; - /* Low order bits define the core id (index of core in socket) */ c->cpu_core_id = c->initial_apicid & ((1 << bits)-1); /* Convert the initial APIC ID into the socket ID */ c->phys_proc_id = c->initial_apicid >> bits; + /* use socket ID also for last level cache */ + per_cpu(cpu_llc_id, cpu) = c->phys_proc_id; #endif } @@ -272,7 +275,7 @@ static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c) #if defined(CONFIG_NUMA) && defined(CONFIG_X86_64) int cpu = smp_processor_id(); int node; - unsigned apicid = hard_smp_processor_id(); + unsigned apicid = cpu_has_apic ? hard_smp_processor_id() : c->apicid; node = c->phys_proc_id; if (apicid_to_node[apicid] != NUMA_NO_NODE) @@ -351,6 +354,15 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c) (c->x86_model == 8 && c->x86_mask >= 8)) set_cpu_cap(c, X86_FEATURE_K6_MTRR); #endif +#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_PCI) + /* check CPU config space for extended APIC ID */ + if (cpu_has_apic && c->x86 >= 0xf) { + unsigned int val; + val = read_pci_config(0, 24, 0, 0x68); + if ((val & ((1 << 17) | (1 << 18))) == ((1 << 17) | (1 << 18))) + set_cpu_cap(c, X86_FEATURE_EXTD_APICID); + } +#endif } static void __cpuinit init_amd(struct cpuinfo_x86 *c) @@ -388,6 +400,13 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c) level = cpuid_eax(1); if((level >= 0x0f48 && level < 0x0f50) || level >= 0x0f58) set_cpu_cap(c, X86_FEATURE_REP_GOOD); + + /* + * Some BIOSes incorrectly force this feature, but only K8 + * revision D (model = 0x14) and later actually support it. + */ + if (c->x86_model < 0x14) + clear_cpu_cap(c, X86_FEATURE_LAHF_LM); } if (c->x86 == 0x10 || c->x86 == 0x11) set_cpu_cap(c, X86_FEATURE_REP_GOOD); diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 77848d9fca6..5ce60a88027 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -13,6 +13,7 @@ #include <linux/io.h> #include <asm/stackprotector.h> +#include <asm/perf_counter.h> #include <asm/mmu_context.h> #include <asm/hypervisor.h> #include <asm/processor.h> @@ -58,7 +59,30 @@ void __init setup_cpu_local_masks(void) alloc_bootmem_cpumask_var(&cpu_sibling_setup_mask); } -static const struct cpu_dev *this_cpu __cpuinitdata; +static void __cpuinit default_init(struct cpuinfo_x86 *c) +{ +#ifdef CONFIG_X86_64 + display_cacheinfo(c); +#else + /* Not much we can do here... */ + /* Check if at least it has cpuid */ + if (c->cpuid_level == -1) { + /* No cpuid. It must be an ancient CPU */ + if (c->x86 == 4) + strcpy(c->x86_model_id, "486"); + else if (c->x86 == 3) + strcpy(c->x86_model_id, "386"); + } +#endif +} + +static const struct cpu_dev __cpuinitconst default_cpu = { + .c_init = default_init, + .c_vendor = "Unknown", + .c_x86_vendor = X86_VENDOR_UNKNOWN, +}; + +static const struct cpu_dev *this_cpu __cpuinitdata = &default_cpu; DEFINE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page) = { .gdt = { #ifdef CONFIG_X86_64 @@ -107,7 +131,7 @@ DEFINE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page) = { .gdt = { /* data */ [GDT_ENTRY_APMBIOS_BASE+2] = { { { 0x0000ffff, 0x00409200 } } }, - [GDT_ENTRY_ESPFIX_SS] = { { { 0x00000000, 0x00c09200 } } }, + [GDT_ENTRY_ESPFIX_SS] = { { { 0x0000ffff, 0x00cf9200 } } }, [GDT_ENTRY_PERCPU] = { { { 0x0000ffff, 0x00cf9200 } } }, GDT_STACK_CANARY_INIT #endif @@ -299,7 +323,8 @@ static const char *__cpuinit table_lookup_model(struct cpuinfo_x86 *c) return NULL; /* Not found */ } -__u32 cleared_cpu_caps[NCAPINTS] __cpuinitdata; +__u32 cpu_caps_cleared[NCAPINTS] __cpuinitdata; +__u32 cpu_caps_set[NCAPINTS] __cpuinitdata; void load_percpu_segment(int cpu) { @@ -330,29 +355,6 @@ void switch_to_new_gdt(int cpu) static const struct cpu_dev *__cpuinitdata cpu_devs[X86_VENDOR_NUM] = {}; -static void __cpuinit default_init(struct cpuinfo_x86 *c) -{ -#ifdef CONFIG_X86_64 - display_cacheinfo(c); -#else - /* Not much we can do here... */ - /* Check if at least it has cpuid */ - if (c->cpuid_level == -1) { - /* No cpuid. It must be an ancient CPU */ - if (c->x86 == 4) - strcpy(c->x86_model_id, "486"); - else if (c->x86 == 3) - strcpy(c->x86_model_id, "386"); - } -#endif -} - -static const struct cpu_dev __cpuinitconst default_cpu = { - .c_init = default_init, - .c_vendor = "Unknown", - .c_x86_vendor = X86_VENDOR_UNKNOWN, -}; - static void __cpuinit get_model_name(struct cpuinfo_x86 *c) { unsigned int *v; @@ -485,7 +487,6 @@ out: static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) { char *v = c->x86_vendor_id; - static int printed; int i; for (i = 0; i < X86_VENDOR_NUM; i++) { @@ -502,13 +503,9 @@ static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) } } - if (!printed) { - printed++; - printk(KERN_ERR - "CPU: vendor_id '%s' unknown, using generic init.\n", v); - - printk(KERN_ERR "CPU: Your system may be unstable.\n"); - } + printk_once(KERN_ERR + "CPU: vendor_id '%s' unknown, using generic init.\n" \ + "CPU: Your system may be unstable.\n", v); c->x86_vendor = X86_VENDOR_UNKNOWN; this_cpu = &default_cpu; @@ -768,6 +765,12 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) if (this_cpu->c_identify) this_cpu->c_identify(c); + /* Clear/Set all flags overriden by options, after probe */ + for (i = 0; i < NCAPINTS; i++) { + c->x86_capability[i] &= ~cpu_caps_cleared[i]; + c->x86_capability[i] |= cpu_caps_set[i]; + } + #ifdef CONFIG_X86_64 c->apicid = apic->phys_pkg_id(c->initial_apicid, 0); #endif @@ -813,6 +816,16 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) #endif init_hypervisor(c); + + /* + * Clear/Set all flags overriden by options, need do it + * before following smp all cpus cap AND. + */ + for (i = 0; i < NCAPINTS; i++) { + c->x86_capability[i] &= ~cpu_caps_cleared[i]; + c->x86_capability[i] |= cpu_caps_set[i]; + } + /* * On SMP, boot_cpu_data holds the common feature set between * all CPUs; so make sure that we indicate which features are @@ -825,10 +838,6 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) boot_cpu_data.x86_capability[i] &= c->x86_capability[i]; } - /* Clear all flags overriden by options */ - for (i = 0; i < NCAPINTS; i++) - c->x86_capability[i] &= ~cleared_cpu_caps[i]; - #ifdef CONFIG_X86_MCE /* Init Machine Check Exception if available. */ mcheck_init(c); @@ -861,6 +870,7 @@ void __init identify_boot_cpu(void) #else vgetcpu_set_mode(); #endif + init_hw_perf_counters(); } void __cpuinit identify_secondary_cpu(struct cpuinfo_x86 *c) diff --git a/arch/x86/kernel/cpu/cpu_debug.c b/arch/x86/kernel/cpu/cpu_debug.c index 46e29ab96c6..6b2a52dd040 100644 --- a/arch/x86/kernel/cpu/cpu_debug.c +++ b/arch/x86/kernel/cpu/cpu_debug.c @@ -32,9 +32,7 @@ static DEFINE_PER_CPU(struct cpu_cpuX_base, cpu_arr[CPU_REG_ALL_BIT]); static DEFINE_PER_CPU(struct cpu_private *, priv_arr[MAX_CPU_FILES]); -static DEFINE_PER_CPU(unsigned, cpu_modelflag); static DEFINE_PER_CPU(int, cpu_priv_count); -static DEFINE_PER_CPU(unsigned, cpu_model); static DEFINE_MUTEX(cpu_debug_lock); @@ -80,302 +78,102 @@ static struct cpu_file_base cpu_file[] = { { "value", CPU_REG_ALL, 1 }, }; -/* Intel Registers Range */ -static struct cpu_debug_range cpu_intel_range[] = { - { 0x00000000, 0x00000001, CPU_MC, CPU_INTEL_ALL }, - { 0x00000006, 0x00000007, CPU_MONITOR, CPU_CX_AT_XE }, - { 0x00000010, 0x00000010, CPU_TIME, CPU_INTEL_ALL }, - { 0x00000011, 0x00000013, CPU_PMC, CPU_INTEL_PENTIUM }, - { 0x00000017, 0x00000017, CPU_PLATFORM, CPU_PX_CX_AT_XE }, - { 0x0000001B, 0x0000001B, CPU_APIC, CPU_P6_CX_AT_XE }, - - { 0x0000002A, 0x0000002A, CPU_POWERON, CPU_PX_CX_AT_XE }, - { 0x0000002B, 0x0000002B, CPU_POWERON, CPU_INTEL_XEON }, - { 0x0000002C, 0x0000002C, CPU_FREQ, CPU_INTEL_XEON }, - { 0x0000003A, 0x0000003A, CPU_CONTROL, CPU_CX_AT_XE }, - - { 0x00000040, 0x00000043, CPU_LBRANCH, CPU_PM_CX_AT_XE }, - { 0x00000044, 0x00000047, CPU_LBRANCH, CPU_PM_CO_AT }, - { 0x00000060, 0x00000063, CPU_LBRANCH, CPU_C2_AT }, - { 0x00000064, 0x00000067, CPU_LBRANCH, CPU_INTEL_ATOM }, - - { 0x00000079, 0x00000079, CPU_BIOS, CPU_P6_CX_AT_XE }, - { 0x00000088, 0x0000008A, CPU_CACHE, CPU_INTEL_P6 }, - { 0x0000008B, 0x0000008B, CPU_BIOS, CPU_P6_CX_AT_XE }, - { 0x0000009B, 0x0000009B, CPU_MONITOR, CPU_INTEL_XEON }, - - { 0x000000C1, 0x000000C2, CPU_PMC, CPU_P6_CX_AT }, - { 0x000000CD, 0x000000CD, CPU_FREQ, CPU_CX_AT }, - { 0x000000E7, 0x000000E8, CPU_PERF, CPU_CX_AT }, - { 0x000000FE, 0x000000FE, CPU_MTRR, CPU_P6_CX_XE }, - - { 0x00000116, 0x00000116, CPU_CACHE, CPU_INTEL_P6 }, - { 0x00000118, 0x00000118, CPU_CACHE, CPU_INTEL_P6 }, - { 0x00000119, 0x00000119, CPU_CACHE, CPU_INTEL_PX }, - { 0x0000011A, 0x0000011B, CPU_CACHE, CPU_INTEL_P6 }, - { 0x0000011E, 0x0000011E, CPU_CACHE, CPU_PX_CX_AT }, - - { 0x00000174, 0x00000176, CPU_SYSENTER, CPU_P6_CX_AT_XE }, - { 0x00000179, 0x0000017A, CPU_MC, CPU_PX_CX_AT_XE }, - { 0x0000017B, 0x0000017B, CPU_MC, CPU_P6_XE }, - { 0x00000186, 0x00000187, CPU_PMC, CPU_P6_CX_AT }, - { 0x00000198, 0x00000199, CPU_PERF, CPU_PM_CX_AT_XE }, - { 0x0000019A, 0x0000019A, CPU_TIME, CPU_PM_CX_AT_XE }, - { 0x0000019B, 0x0000019D, CPU_THERM, CPU_PM_CX_AT_XE }, - { 0x000001A0, 0x000001A0, CPU_MISC, CPU_PM_CX_AT_XE }, - - { 0x000001C9, 0x000001C9, CPU_LBRANCH, CPU_PM_CX_AT }, - { 0x000001D7, 0x000001D8, CPU_LBRANCH, CPU_INTEL_XEON }, - { 0x000001D9, 0x000001D9, CPU_DEBUG, CPU_CX_AT_XE }, - { 0x000001DA, 0x000001DA, CPU_LBRANCH, CPU_INTEL_XEON }, - { 0x000001DB, 0x000001DB, CPU_LBRANCH, CPU_P6_XE }, - { 0x000001DC, 0x000001DC, CPU_LBRANCH, CPU_INTEL_P6 }, - { 0x000001DD, 0x000001DE, CPU_LBRANCH, CPU_PX_CX_AT_XE }, - { 0x000001E0, 0x000001E0, CPU_LBRANCH, CPU_INTEL_P6 }, - - { 0x00000200, 0x0000020F, CPU_MTRR, CPU_P6_CX_XE }, - { 0x00000250, 0x00000250, CPU_MTRR, CPU_P6_CX_XE }, - { 0x00000258, 0x00000259, CPU_MTRR, CPU_P6_CX_XE }, - { 0x00000268, 0x0000026F, CPU_MTRR, CPU_P6_CX_XE }, - { 0x00000277, 0x00000277, CPU_PAT, CPU_C2_AT_XE }, - { 0x000002FF, 0x000002FF, CPU_MTRR, CPU_P6_CX_XE }, - - { 0x00000300, 0x00000308, CPU_PMC, CPU_INTEL_XEON }, - { 0x00000309, 0x0000030B, CPU_PMC, CPU_C2_AT_XE }, - { 0x0000030C, 0x00000311, CPU_PMC, CPU_INTEL_XEON }, - { 0x00000345, 0x00000345, CPU_PMC, CPU_C2_AT }, - { 0x00000360, 0x00000371, CPU_PMC, CPU_INTEL_XEON }, - { 0x0000038D, 0x00000390, CPU_PMC, CPU_C2_AT }, - { 0x000003A0, 0x000003BE, CPU_PMC, CPU_INTEL_XEON }, - { 0x000003C0, 0x000003CD, CPU_PMC, CPU_INTEL_XEON }, - { 0x000003E0, 0x000003E1, CPU_PMC, CPU_INTEL_XEON }, - { 0x000003F0, 0x000003F0, CPU_PMC, CPU_INTEL_XEON }, - { 0x000003F1, 0x000003F1, CPU_PMC, CPU_C2_AT_XE }, - { 0x000003F2, 0x000003F2, CPU_PMC, CPU_INTEL_XEON }, - - { 0x00000400, 0x00000402, CPU_MC, CPU_PM_CX_AT_XE }, - { 0x00000403, 0x00000403, CPU_MC, CPU_INTEL_XEON }, - { 0x00000404, 0x00000406, CPU_MC, CPU_PM_CX_AT_XE }, - { 0x00000407, 0x00000407, CPU_MC, CPU_INTEL_XEON }, - { 0x00000408, 0x0000040A, CPU_MC, CPU_PM_CX_AT_XE }, - { 0x0000040B, 0x0000040B, CPU_MC, CPU_INTEL_XEON }, - { 0x0000040C, 0x0000040E, CPU_MC, CPU_PM_CX_XE }, - { 0x0000040F, 0x0000040F, CPU_MC, CPU_INTEL_XEON }, - { 0x00000410, 0x00000412, CPU_MC, CPU_PM_CX_AT_XE }, - { 0x00000413, 0x00000417, CPU_MC, CPU_CX_AT_XE }, - { 0x00000480, 0x0000048B, CPU_VMX, CPU_CX_AT_XE }, - - { 0x00000600, 0x00000600, CPU_DEBUG, CPU_PM_CX_AT_XE }, - { 0x00000680, 0x0000068F, CPU_LBRANCH, CPU_INTEL_XEON }, - { 0x000006C0, 0x000006CF, CPU_LBRANCH, CPU_INTEL_XEON }, - - { 0x000107CC, 0x000107D3, CPU_PMC, CPU_INTEL_XEON_MP }, - - { 0xC0000080, 0xC0000080, CPU_FEATURES, CPU_INTEL_XEON }, - { 0xC0000081, 0xC0000082, CPU_CALL, CPU_INTEL_XEON }, - { 0xC0000084, 0xC0000084, CPU_CALL, CPU_INTEL_XEON }, - { 0xC0000100, 0xC0000102, CPU_BASE, CPU_INTEL_XEON }, +/* CPU Registers Range */ +static struct cpu_debug_range cpu_reg_range[] = { + { 0x00000000, 0x00000001, CPU_MC, }, + { 0x00000006, 0x00000007, CPU_MONITOR, }, + { 0x00000010, 0x00000010, CPU_TIME, }, + { 0x00000011, 0x00000013, CPU_PMC, }, + { 0x00000017, 0x00000017, CPU_PLATFORM, }, + { 0x0000001B, 0x0000001B, CPU_APIC, }, + { 0x0000002A, 0x0000002B, CPU_POWERON, }, + { 0x0000002C, 0x0000002C, CPU_FREQ, }, + { 0x0000003A, 0x0000003A, CPU_CONTROL, }, + { 0x00000040, 0x00000047, CPU_LBRANCH, }, + { 0x00000060, 0x00000067, CPU_LBRANCH, }, + { 0x00000079, 0x00000079, CPU_BIOS, }, + { 0x00000088, 0x0000008A, CPU_CACHE, }, + { 0x0000008B, 0x0000008B, CPU_BIOS, }, + { 0x0000009B, 0x0000009B, CPU_MONITOR, }, + { 0x000000C1, 0x000000C4, CPU_PMC, }, + { 0x000000CD, 0x000000CD, CPU_FREQ, }, + { 0x000000E7, 0x000000E8, CPU_PERF, }, + { 0x000000FE, 0x000000FE, CPU_MTRR, }, + + { 0x00000116, 0x0000011E, CPU_CACHE, }, + { 0x00000174, 0x00000176, CPU_SYSENTER, }, + { 0x00000179, 0x0000017B, CPU_MC, }, + { 0x00000186, 0x00000189, CPU_PMC, }, + { 0x00000198, 0x00000199, CPU_PERF, }, + { 0x0000019A, 0x0000019A, CPU_TIME, }, + { 0x0000019B, 0x0000019D, CPU_THERM, }, + { 0x000001A0, 0x000001A0, CPU_MISC, }, + { 0x000001C9, 0x000001C9, CPU_LBRANCH, }, + { 0x000001D7, 0x000001D8, CPU_LBRANCH, }, + { 0x000001D9, 0x000001D9, CPU_DEBUG, }, + { 0x000001DA, 0x000001E0, CPU_LBRANCH, }, + + { 0x00000200, 0x0000020F, CPU_MTRR, }, + { 0x00000250, 0x00000250, CPU_MTRR, }, + { 0x00000258, 0x00000259, CPU_MTRR, }, + { 0x00000268, 0x0000026F, CPU_MTRR, }, + { 0x00000277, 0x00000277, CPU_PAT, }, + { 0x000002FF, 0x000002FF, CPU_MTRR, }, + + { 0x00000300, 0x00000311, CPU_PMC, }, + { 0x00000345, 0x00000345, CPU_PMC, }, + { 0x00000360, 0x00000371, CPU_PMC, }, + { 0x0000038D, 0x00000390, CPU_PMC, }, + { 0x000003A0, 0x000003BE, CPU_PMC, }, + { 0x000003C0, 0x000003CD, CPU_PMC, }, + { 0x000003E0, 0x000003E1, CPU_PMC, }, + { 0x000003F0, 0x000003F2, CPU_PMC, }, + + { 0x00000400, 0x00000417, CPU_MC, }, + { 0x00000480, 0x0000048B, CPU_VMX, }, + + { 0x00000600, 0x00000600, CPU_DEBUG, }, + { 0x00000680, 0x0000068F, CPU_LBRANCH, }, + { 0x000006C0, 0x000006CF, CPU_LBRANCH, }, + + { 0x000107CC, 0x000107D3, CPU_PMC, }, + + { 0xC0000080, 0xC0000080, CPU_FEATURES, }, + { 0xC0000081, 0xC0000084, CPU_CALL, }, + { 0xC0000100, 0xC0000102, CPU_BASE, }, + { 0xC0000103, 0xC0000103, CPU_TIME, }, + + { 0xC0010000, 0xC0010007, CPU_PMC, }, + { 0xC0010010, 0xC0010010, CPU_CONF, }, + { 0xC0010015, 0xC0010015, CPU_CONF, }, + { 0xC0010016, 0xC001001A, CPU_MTRR, }, + { 0xC001001D, 0xC001001D, CPU_MTRR, }, + { 0xC001001F, 0xC001001F, CPU_CONF, }, + { 0xC0010030, 0xC0010035, CPU_BIOS, }, + { 0xC0010044, 0xC0010048, CPU_MC, }, + { 0xC0010050, 0xC0010056, CPU_SMM, }, + { 0xC0010058, 0xC0010058, CPU_CONF, }, + { 0xC0010060, 0xC0010060, CPU_CACHE, }, + { 0xC0010061, 0xC0010068, CPU_SMM, }, + { 0xC0010069, 0xC001006B, CPU_SMM, }, + { 0xC0010070, 0xC0010071, CPU_SMM, }, + { 0xC0010111, 0xC0010113, CPU_SMM, }, + { 0xC0010114, 0xC0010118, CPU_SVM, }, + { 0xC0010140, 0xC0010141, CPU_OSVM, }, + { 0xC0011022, 0xC0011023, CPU_CONF, }, }; -/* AMD Registers Range */ -static struct cpu_debug_range cpu_amd_range[] = { - { 0x00000000, 0x00000001, CPU_MC, CPU_K10_PLUS, }, - { 0x00000010, 0x00000010, CPU_TIME, CPU_K8_PLUS, }, - { 0x0000001B, 0x0000001B, CPU_APIC, CPU_K8_PLUS, }, - { 0x0000002A, 0x0000002A, CPU_POWERON, CPU_K7_PLUS }, - { 0x0000008B, 0x0000008B, CPU_VER, CPU_K8_PLUS }, - { 0x000000FE, 0x000000FE, CPU_MTRR, CPU_K8_PLUS, }, - - { 0x00000174, 0x00000176, CPU_SYSENTER, CPU_K8_PLUS, }, - { 0x00000179, 0x0000017B, CPU_MC, CPU_K8_PLUS, }, - { 0x000001D9, 0x000001D9, CPU_DEBUG, CPU_K8_PLUS, }, - { 0x000001DB, 0x000001DE, CPU_LBRANCH, CPU_K8_PLUS, }, - - { 0x00000200, 0x0000020F, CPU_MTRR, CPU_K8_PLUS, }, - { 0x00000250, 0x00000250, CPU_MTRR, CPU_K8_PLUS, }, - { 0x00000258, 0x00000259, CPU_MTRR, CPU_K8_PLUS, }, - { 0x00000268, 0x0000026F, CPU_MTRR, CPU_K8_PLUS, }, - { 0x00000277, 0x00000277, CPU_PAT, CPU_K8_PLUS, }, - { 0x000002FF, 0x000002FF, CPU_MTRR, CPU_K8_PLUS, }, - - { 0x00000400, 0x00000413, CPU_MC, CPU_K8_PLUS, }, - - { 0xC0000080, 0xC0000080, CPU_FEATURES, CPU_AMD_ALL, }, - { 0xC0000081, 0xC0000084, CPU_CALL, CPU_K8_PLUS, }, - { 0xC0000100, 0xC0000102, CPU_BASE, CPU_K8_PLUS, }, - { 0xC0000103, 0xC0000103, CPU_TIME, CPU_K10_PLUS, }, - - { 0xC0010000, 0xC0010007, CPU_PMC, CPU_K8_PLUS, }, - { 0xC0010010, 0xC0010010, CPU_CONF, CPU_K7_PLUS, }, - { 0xC0010015, 0xC0010015, CPU_CONF, CPU_K7_PLUS, }, - { 0xC0010016, 0xC001001A, CPU_MTRR, CPU_K8_PLUS, }, - { 0xC001001D, 0xC001001D, CPU_MTRR, CPU_K8_PLUS, }, - { 0xC001001F, 0xC001001F, CPU_CONF, CPU_K8_PLUS, }, - { 0xC0010030, 0xC0010035, CPU_BIOS, CPU_K8_PLUS, }, - { 0xC0010044, 0xC0010048, CPU_MC, CPU_K8_PLUS, }, - { 0xC0010050, 0xC0010056, CPU_SMM, CPU_K0F_PLUS, }, - { 0xC0010058, 0xC0010058, CPU_CONF, CPU_K10_PLUS, }, - { 0xC0010060, 0xC0010060, CPU_CACHE, CPU_AMD_11, }, - { 0xC0010061, 0xC0010068, CPU_SMM, CPU_K10_PLUS, }, - { 0xC0010069, 0xC001006B, CPU_SMM, CPU_AMD_11, }, - { 0xC0010070, 0xC0010071, CPU_SMM, CPU_K10_PLUS, }, - { 0xC0010111, 0xC0010113, CPU_SMM, CPU_K8_PLUS, }, - { 0xC0010114, 0xC0010118, CPU_SVM, CPU_K10_PLUS, }, - { 0xC0010140, 0xC0010141, CPU_OSVM, CPU_K10_PLUS, }, - { 0xC0011022, 0xC0011023, CPU_CONF, CPU_K10_PLUS, }, -}; - - -/* Intel */ -static int get_intel_modelflag(unsigned model) -{ - int flag; - - switch (model) { - case 0x0501: - case 0x0502: - case 0x0504: - flag = CPU_INTEL_PENTIUM; - break; - case 0x0601: - case 0x0603: - case 0x0605: - case 0x0607: - case 0x0608: - case 0x060A: - case 0x060B: - flag = CPU_INTEL_P6; - break; - case 0x0609: - case 0x060D: - flag = CPU_INTEL_PENTIUM_M; - break; - case 0x060E: - flag = CPU_INTEL_CORE; - break; - case 0x060F: - case 0x0617: - flag = CPU_INTEL_CORE2; - break; - case 0x061C: - flag = CPU_INTEL_ATOM; - break; - case 0x0F00: - case 0x0F01: - case 0x0F02: - case 0x0F03: - case 0x0F04: - flag = CPU_INTEL_XEON_P4; - break; - case 0x0F06: - flag = CPU_INTEL_XEON_MP; - break; - default: - flag = CPU_NONE; - break; - } - - return flag; -} - -/* AMD */ -static int get_amd_modelflag(unsigned model) -{ - int flag; - - switch (model >> 8) { - case 0x6: - flag = CPU_AMD_K6; - break; - case 0x7: - flag = CPU_AMD_K7; - break; - case 0x8: - flag = CPU_AMD_K8; - break; - case 0xf: - flag = CPU_AMD_0F; - break; - case 0x10: - flag = CPU_AMD_10; - break; - case 0x11: - flag = CPU_AMD_11; - break; - default: - flag = CPU_NONE; - break; - } - - return flag; -} - -static int get_cpu_modelflag(unsigned cpu) -{ - int flag; - - flag = per_cpu(cpu_model, cpu); - - switch (flag >> 16) { - case X86_VENDOR_INTEL: - flag = get_intel_modelflag(flag); - break; - case X86_VENDOR_AMD: - flag = get_amd_modelflag(flag & 0xffff); - break; - default: - flag = CPU_NONE; - break; - } - - return flag; -} - -static int get_cpu_range_count(unsigned cpu) -{ - int index; - - switch (per_cpu(cpu_model, cpu) >> 16) { - case X86_VENDOR_INTEL: - index = ARRAY_SIZE(cpu_intel_range); - break; - case X86_VENDOR_AMD: - index = ARRAY_SIZE(cpu_amd_range); - break; - default: - index = 0; - break; - } - - return index; -} - static int is_typeflag_valid(unsigned cpu, unsigned flag) { - unsigned vendor, modelflag; - int i, index; + int i; /* Standard Registers should be always valid */ if (flag >= CPU_TSS) return 1; - modelflag = per_cpu(cpu_modelflag, cpu); - vendor = per_cpu(cpu_model, cpu) >> 16; - index = get_cpu_range_count(cpu); - - for (i = 0; i < index; i++) { - switch (vendor) { - case X86_VENDOR_INTEL: - if ((cpu_intel_range[i].model & modelflag) && - (cpu_intel_range[i].flag & flag)) - return 1; - break; - case X86_VENDOR_AMD: - if ((cpu_amd_range[i].model & modelflag) && - (cpu_amd_range[i].flag & flag)) - return 1; - break; - } + for (i = 0; i < ARRAY_SIZE(cpu_reg_range); i++) { + if (cpu_reg_range[i].flag == flag) + return 1; } /* Invalid */ @@ -385,26 +183,11 @@ static int is_typeflag_valid(unsigned cpu, unsigned flag) static unsigned get_cpu_range(unsigned cpu, unsigned *min, unsigned *max, int index, unsigned flag) { - unsigned modelflag; - - modelflag = per_cpu(cpu_modelflag, cpu); - *max = 0; - switch (per_cpu(cpu_model, cpu) >> 16) { - case X86_VENDOR_INTEL: - if ((cpu_intel_range[index].model & modelflag) && - (cpu_intel_range[index].flag & flag)) { - *min = cpu_intel_range[index].min; - *max = cpu_intel_range[index].max; - } - break; - case X86_VENDOR_AMD: - if ((cpu_amd_range[index].model & modelflag) && - (cpu_amd_range[index].flag & flag)) { - *min = cpu_amd_range[index].min; - *max = cpu_amd_range[index].max; - } - break; - } + if (cpu_reg_range[index].flag == flag) { + *min = cpu_reg_range[index].min; + *max = cpu_reg_range[index].max; + } else + *max = 0; return *max; } @@ -434,7 +217,7 @@ static void print_msr(struct seq_file *seq, unsigned cpu, unsigned flag) unsigned msr, msr_min, msr_max; struct cpu_private *priv; u32 low, high; - int i, range; + int i; if (seq) { priv = seq->private; @@ -446,9 +229,7 @@ static void print_msr(struct seq_file *seq, unsigned cpu, unsigned flag) } } - range = get_cpu_range_count(cpu); - - for (i = 0; i < range; i++) { + for (i = 0; i < ARRAY_SIZE(cpu_reg_range); i++) { if (!get_cpu_range(cpu, &msr_min, &msr_max, i, flag)) continue; @@ -588,8 +369,20 @@ static void print_apic(void *arg) seq_printf(seq, " TMICT\t\t: %08x\n", apic_read(APIC_TMICT)); seq_printf(seq, " TMCCT\t\t: %08x\n", apic_read(APIC_TMCCT)); seq_printf(seq, " TDCR\t\t: %08x\n", apic_read(APIC_TDCR)); -#endif /* CONFIG_X86_LOCAL_APIC */ + if (boot_cpu_has(X86_FEATURE_EXTAPIC)) { + unsigned int i, v, maxeilvt; + + v = apic_read(APIC_EFEAT); + maxeilvt = (v >> 16) & 0xff; + seq_printf(seq, " EFEAT\t\t: %08x\n", v); + seq_printf(seq, " ECTRL\t\t: %08x\n", apic_read(APIC_ECTRL)); + for (i = 0; i < maxeilvt; i++) { + v = apic_read(APIC_EILVTn(i)); + seq_printf(seq, " EILVT%d\t\t: %08x\n", i, v); + } + } +#endif /* CONFIG_X86_LOCAL_APIC */ seq_printf(seq, "\n MSR\t:\n"); } @@ -788,13 +581,11 @@ static int cpu_init_msr(unsigned cpu, unsigned type, struct dentry *dentry) { struct dentry *cpu_dentry = NULL; unsigned reg, reg_min, reg_max; - int i, range, err = 0; + int i, err = 0; char reg_dir[12]; u32 low, high; - range = get_cpu_range_count(cpu); - - for (i = 0; i < range; i++) { + for (i = 0; i < ARRAY_SIZE(cpu_reg_range); i++) { if (!get_cpu_range(cpu, ®_min, ®_max, i, cpu_base[type].flag)) continue; @@ -850,10 +641,6 @@ static int cpu_init_cpu(void) cpui = &cpu_data(cpu); if (!cpu_has(cpui, X86_FEATURE_MSR)) continue; - per_cpu(cpu_model, cpu) = ((cpui->x86_vendor << 16) | - (cpui->x86 << 8) | - (cpui->x86_model)); - per_cpu(cpu_modelflag, cpu) = get_cpu_modelflag(cpu); sprintf(cpu_dir, "cpu%d", cpu); cpu_dentry = debugfs_create_dir(cpu_dir, cpu_debugfs_dir); diff --git a/arch/x86/kernel/cpu/cpufreq/Kconfig b/arch/x86/kernel/cpu/cpufreq/Kconfig index 52c83987547..f138c6c389b 100644 --- a/arch/x86/kernel/cpu/cpufreq/Kconfig +++ b/arch/x86/kernel/cpu/cpufreq/Kconfig @@ -220,11 +220,14 @@ config X86_LONGHAUL If in doubt, say N. config X86_E_POWERSAVER - tristate "VIA C7 Enhanced PowerSaver" + tristate "VIA C7 Enhanced PowerSaver (DANGEROUS)" select CPU_FREQ_TABLE - depends on X86_32 + depends on X86_32 && EXPERIMENTAL help - This adds the CPUFreq driver for VIA C7 processors. + This adds the CPUFreq driver for VIA C7 processors. However, this driver + does not have any safeguards to prevent operating the CPU out of spec + and is thus considered dangerous. Please use the regular ACPI cpufreq + driver, enabled by CONFIG_X86_ACPI_CPUFREQ. If in doubt, say N. diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 752e8c6b2c7..ae9b503220c 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -90,11 +90,7 @@ static int check_est_cpu(unsigned int cpuid) { struct cpuinfo_x86 *cpu = &cpu_data(cpuid); - if (cpu->x86_vendor != X86_VENDOR_INTEL || - !cpu_has(cpu, X86_FEATURE_EST)) - return 0; - - return 1; + return cpu_has(cpu, X86_FEATURE_EST); } static unsigned extract_io(u32 value, struct acpi_cpufreq_data *data) diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c index cf52215d9eb..2a50ef89100 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c @@ -1,3 +1,4 @@ + /* * (c) 2003-2006 Advanced Micro Devices, Inc. * Your use of this code is subject to the terms and conditions of the @@ -117,20 +118,17 @@ static int query_current_values_with_pending_wait(struct powernow_k8_data *data) u32 i = 0; if (cpu_family == CPU_HW_PSTATE) { - if (data->currpstate == HW_PSTATE_INVALID) { - /* read (initial) hw pstate if not yet set */ - rdmsr(MSR_PSTATE_STATUS, lo, hi); - i = lo & HW_PSTATE_MASK; - - /* - * a workaround for family 11h erratum 311 might cause - * an "out-of-range Pstate if the core is in Pstate-0 - */ - if (i >= data->numps) - data->currpstate = HW_PSTATE_0; - else - data->currpstate = i; - } + rdmsr(MSR_PSTATE_STATUS, lo, hi); + i = lo & HW_PSTATE_MASK; + data->currpstate = i; + + /* + * a workaround for family 11h erratum 311 might cause + * an "out-of-range Pstate if the core is in Pstate-0 + */ + if ((boot_cpu_data.x86 == 0x11) && (i >= data->numps)) + data->currpstate = HW_PSTATE_0; + return 0; } do { @@ -301,7 +299,7 @@ static int transition_pstate(struct powernow_k8_data *data, u32 pstate) static int transition_fid_vid(struct powernow_k8_data *data, u32 reqfid, u32 reqvid) { - if (core_voltage_pre_transition(data, reqvid)) + if (core_voltage_pre_transition(data, reqvid, reqfid)) return 1; if (core_frequency_transition(data, reqfid)) @@ -329,17 +327,20 @@ static int transition_fid_vid(struct powernow_k8_data *data, /* Phase 1 - core voltage transition ... setup voltage */ static int core_voltage_pre_transition(struct powernow_k8_data *data, - u32 reqvid) + u32 reqvid, u32 reqfid) { u32 rvosteps = data->rvo; u32 savefid = data->currfid; - u32 maxvid, lo; + u32 maxvid, lo, rvomult = 1; dprintk("ph1 (cpu%d): start, currfid 0x%x, currvid 0x%x, " "reqvid 0x%x, rvo 0x%x\n", smp_processor_id(), data->currfid, data->currvid, reqvid, data->rvo); + if ((savefid < LO_FID_TABLE_TOP) && (reqfid < LO_FID_TABLE_TOP)) + rvomult = 2; + rvosteps *= rvomult; rdmsr(MSR_FIDVID_STATUS, lo, maxvid); maxvid = 0x1f & (maxvid >> 16); dprintk("ph1 maxvid=0x%x\n", maxvid); @@ -353,7 +354,8 @@ static int core_voltage_pre_transition(struct powernow_k8_data *data, return 1; } - while ((rvosteps > 0) && ((data->rvo + data->currvid) > reqvid)) { + while ((rvosteps > 0) && + ((rvomult * data->rvo + data->currvid) > reqvid)) { if (data->currvid == maxvid) { rvosteps = 0; } else { @@ -386,13 +388,6 @@ static int core_frequency_transition(struct powernow_k8_data *data, u32 reqfid) u32 vcoreqfid, vcocurrfid, vcofiddiff; u32 fid_interval, savevid = data->currvid; - if ((reqfid < HI_FID_TABLE_BOTTOM) && - (data->currfid < HI_FID_TABLE_BOTTOM)) { - printk(KERN_ERR PFX "ph2: illegal lo-lo transition " - "0x%x 0x%x\n", reqfid, data->currfid); - return 1; - } - if (data->currfid == reqfid) { printk(KERN_ERR PFX "ph2 null fid transition 0x%x\n", data->currfid); @@ -409,6 +404,9 @@ static int core_frequency_transition(struct powernow_k8_data *data, u32 reqfid) vcofiddiff = vcocurrfid > vcoreqfid ? vcocurrfid - vcoreqfid : vcoreqfid - vcocurrfid; + if ((reqfid <= LO_FID_TABLE_TOP) && (data->currfid <= LO_FID_TABLE_TOP)) + vcofiddiff = 0; + while (vcofiddiff > 2) { (data->currfid & 1) ? (fid_interval = 1) : (fid_interval = 2); @@ -510,41 +508,34 @@ static int core_voltage_post_transition(struct powernow_k8_data *data, return 0; } -static int check_supported_cpu(unsigned int cpu) +static void check_supported_cpu(void *_rc) { - cpumask_t oldmask; u32 eax, ebx, ecx, edx; - unsigned int rc = 0; + int *rc = _rc; - oldmask = current->cpus_allowed; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)); - - if (smp_processor_id() != cpu) { - printk(KERN_ERR PFX "limiting to cpu %u failed\n", cpu); - goto out; - } + *rc = -ENODEV; if (current_cpu_data.x86_vendor != X86_VENDOR_AMD) - goto out; + return; eax = cpuid_eax(CPUID_PROCESSOR_SIGNATURE); if (((eax & CPUID_XFAM) != CPUID_XFAM_K8) && ((eax & CPUID_XFAM) < CPUID_XFAM_10H)) - goto out; + return; if ((eax & CPUID_XFAM) == CPUID_XFAM_K8) { if (((eax & CPUID_USE_XFAM_XMOD) != CPUID_USE_XFAM_XMOD) || ((eax & CPUID_XMOD) > CPUID_XMOD_REV_MASK)) { printk(KERN_INFO PFX "Processor cpuid %x not supported\n", eax); - goto out; + return; } eax = cpuid_eax(CPUID_GET_MAX_CAPABILITIES); if (eax < CPUID_FREQ_VOLT_CAPABILITIES) { printk(KERN_INFO PFX "No frequency change capabilities detected\n"); - goto out; + return; } cpuid(CPUID_FREQ_VOLT_CAPABILITIES, &eax, &ebx, &ecx, &edx); @@ -552,21 +543,17 @@ static int check_supported_cpu(unsigned int cpu) != P_STATE_TRANSITION_CAPABLE) { printk(KERN_INFO PFX "Power state transitions not supported\n"); - goto out; + return; } } else { /* must be a HW Pstate capable processor */ cpuid(CPUID_FREQ_VOLT_CAPABILITIES, &eax, &ebx, &ecx, &edx); if ((edx & USE_HW_PSTATE) == USE_HW_PSTATE) cpu_family = CPU_HW_PSTATE; else - goto out; + return; } - rc = 1; - -out: - set_cpus_allowed_ptr(current, &oldmask); - return rc; + *rc = 0; } static int check_pst_table(struct powernow_k8_data *data, struct pst_s *pst, @@ -823,13 +810,14 @@ static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, if (!data->acpi_data.state_count || (cpu_family == CPU_HW_PSTATE)) return; - control = data->acpi_data.states[index].control; data->irt = (control - >> IRT_SHIFT) & IRT_MASK; data->rvo = (control >> - RVO_SHIFT) & RVO_MASK; data->exttype = (control - >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK; - data->plllock = (control >> PLL_L_SHIFT) & PLL_L_MASK; data->vidmvs = 1 - << ((control >> MVS_SHIFT) & MVS_MASK); data->vstable = - (control >> VST_SHIFT) & VST_MASK; } + control = data->acpi_data.states[index].control; + data->irt = (control >> IRT_SHIFT) & IRT_MASK; + data->rvo = (control >> RVO_SHIFT) & RVO_MASK; + data->exttype = (control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK; + data->plllock = (control >> PLL_L_SHIFT) & PLL_L_MASK; + data->vidmvs = 1 << ((control >> MVS_SHIFT) & MVS_MASK); + data->vstable = (control >> VST_SHIFT) & VST_MASK; +} static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { @@ -1046,6 +1034,19 @@ static int get_transition_latency(struct powernow_k8_data *data) if (cur_latency > max_latency) max_latency = cur_latency; } + if (max_latency == 0) { + /* + * Fam 11h always returns 0 as transition latency. + * This is intended and means "very fast". While cpufreq core + * and governors currently can handle that gracefully, better + * set it to 1 to avoid problems in the future. + * For all others it's a BIOS bug. + */ + if (!boot_cpu_data.x86 == 0x11) + printk(KERN_ERR FW_WARN PFX "Invalid zero transition " + "latency\n"); + max_latency = 1; + } /* value in usecs, needs to be in nanoseconds */ return 1000 * max_latency; } @@ -1080,20 +1081,12 @@ static int transition_frequency_fidvid(struct powernow_k8_data *data, return 0; } - if ((fid < HI_FID_TABLE_BOTTOM) && - (data->currfid < HI_FID_TABLE_BOTTOM)) { - printk(KERN_ERR PFX - "ignoring illegal change in lo freq table-%x to 0x%x\n", - data->currfid, fid); - return 1; - } - dprintk("cpu %d, changing to fid 0x%x, vid 0x%x\n", smp_processor_id(), fid, vid); freqs.old = find_khz_freq_from_fid(data->currfid); freqs.new = find_khz_freq_from_fid(fid); - for_each_cpu_mask_nr(i, *(data->available_cores)) { + for_each_cpu(i, data->available_cores) { freqs.cpu = i; cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); } @@ -1101,7 +1094,7 @@ static int transition_frequency_fidvid(struct powernow_k8_data *data, res = transition_fid_vid(data, fid, vid); freqs.new = find_khz_freq_from_fid(data->currfid); - for_each_cpu_mask_nr(i, *(data->available_cores)) { + for_each_cpu(i, data->available_cores) { freqs.cpu = i; cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); } @@ -1126,7 +1119,7 @@ static int transition_frequency_pstate(struct powernow_k8_data *data, data->currpstate); freqs.new = find_khz_freq_from_pstate(data->powernow_table, pstate); - for_each_cpu_mask_nr(i, *(data->available_cores)) { + for_each_cpu(i, data->available_cores) { freqs.cpu = i; cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); } @@ -1134,7 +1127,7 @@ static int transition_frequency_pstate(struct powernow_k8_data *data, res = transition_pstate(data, pstate); freqs.new = find_khz_freq_from_pstate(data->powernow_table, pstate); - for_each_cpu_mask_nr(i, *(data->available_cores)) { + for_each_cpu(i, data->available_cores) { freqs.cpu = i; cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); } @@ -1235,21 +1228,47 @@ static int powernowk8_verify(struct cpufreq_policy *pol) return cpufreq_frequency_table_verify(pol, data->powernow_table); } -static const char ACPI_PSS_BIOS_BUG_MSG[] = - KERN_ERR FW_BUG PFX "No compatible ACPI _PSS objects found.\n" - KERN_ERR FW_BUG PFX "Try again with latest BIOS.\n"; +struct init_on_cpu { + struct powernow_k8_data *data; + int rc; +}; + +static void __cpuinit powernowk8_cpu_init_on_cpu(void *_init_on_cpu) +{ + struct init_on_cpu *init_on_cpu = _init_on_cpu; + + if (pending_bit_stuck()) { + printk(KERN_ERR PFX "failing init, change pending bit set\n"); + init_on_cpu->rc = -ENODEV; + return; + } + + if (query_current_values_with_pending_wait(init_on_cpu->data)) { + init_on_cpu->rc = -ENODEV; + return; + } + + if (cpu_family == CPU_OPTERON) + fidvid_msr_init(); + + init_on_cpu->rc = 0; +} /* per CPU init entry point to the driver */ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) { + static const char ACPI_PSS_BIOS_BUG_MSG[] = + KERN_ERR FW_BUG PFX "No compatible ACPI _PSS objects found.\n" + FW_BUG PFX "Try again with latest BIOS.\n"; struct powernow_k8_data *data; - cpumask_t oldmask; + struct init_on_cpu init_on_cpu; int rc; if (!cpu_online(pol->cpu)) return -ENODEV; - if (!check_supported_cpu(pol->cpu)) + smp_call_function_single(pol->cpu, check_supported_cpu, &rc, 1); + if (rc) return -ENODEV; data = kzalloc(sizeof(struct powernow_k8_data), GFP_KERNEL); @@ -1289,27 +1308,12 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) pol->cpuinfo.transition_latency = get_transition_latency(data); /* only run on specific CPU from here on */ - oldmask = current->cpus_allowed; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(pol->cpu)); - - if (smp_processor_id() != pol->cpu) { - printk(KERN_ERR PFX "limiting to cpu %u failed\n", pol->cpu); - goto err_out_unmask; - } - - if (pending_bit_stuck()) { - printk(KERN_ERR PFX "failing init, change pending bit set\n"); - goto err_out_unmask; - } - - if (query_current_values_with_pending_wait(data)) - goto err_out_unmask; - - if (cpu_family == CPU_OPTERON) - fidvid_msr_init(); - - /* run on any CPU again */ - set_cpus_allowed_ptr(current, &oldmask); + init_on_cpu.data = data; + smp_call_function_single(data->cpu, powernowk8_cpu_init_on_cpu, + &init_on_cpu, 1); + rc = init_on_cpu.rc; + if (rc != 0) + goto err_out_exit_acpi; if (cpu_family == CPU_HW_PSTATE) cpumask_copy(pol->cpus, cpumask_of(pol->cpu)); @@ -1346,8 +1350,7 @@ static int __cpuinit powernowk8_cpu_init(struct cpufreq_policy *pol) return 0; -err_out_unmask: - set_cpus_allowed_ptr(current, &oldmask); +err_out_exit_acpi: powernow_k8_cpu_exit_acpi(data); err_out: @@ -1372,28 +1375,25 @@ static int __devexit powernowk8_cpu_exit(struct cpufreq_policy *pol) return 0; } +static void query_values_on_cpu(void *_err) +{ + int *err = _err; + struct powernow_k8_data *data = __get_cpu_var(powernow_data); + + *err = query_current_values_with_pending_wait(data); +} + static unsigned int powernowk8_get(unsigned int cpu) { - struct powernow_k8_data *data; - cpumask_t oldmask = current->cpus_allowed; + struct powernow_k8_data *data = per_cpu(powernow_data, cpu); unsigned int khz = 0; - unsigned int first; - - first = cpumask_first(cpu_core_mask(cpu)); - data = per_cpu(powernow_data, first); + int err; if (!data) return -EINVAL; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)); - if (smp_processor_id() != cpu) { - printk(KERN_ERR PFX - "limiting to CPU %d failed in powernowk8_get\n", cpu); - set_cpus_allowed_ptr(current, &oldmask); - return 0; - } - - if (query_current_values_with_pending_wait(data)) + smp_call_function_single(cpu, query_values_on_cpu, &err, true); + if (err) goto out; if (cpu_family == CPU_HW_PSTATE) @@ -1404,7 +1404,6 @@ static unsigned int powernowk8_get(unsigned int cpu) out: - set_cpus_allowed_ptr(current, &oldmask); return khz; } @@ -1430,7 +1429,9 @@ static int __cpuinit powernowk8_init(void) unsigned int i, supported_cpus = 0; for_each_online_cpu(i) { - if (check_supported_cpu(i)) + int rc; + smp_call_function_single(i, check_supported_cpu, &rc, 1); + if (rc == 0) supported_cpus++; } diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h index 6c6698feade..02ce824073c 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.h +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.h @@ -215,7 +215,8 @@ struct pst_s { #define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "powernow-k8", msg) -static int core_voltage_pre_transition(struct powernow_k8_data *data, u32 reqvid); +static int core_voltage_pre_transition(struct powernow_k8_data *data, + u32 reqvid, u32 regfid); static int core_voltage_post_transition(struct powernow_k8_data *data, u32 reqvid); static int core_frequency_transition(struct powernow_k8_data *data, u32 reqfid); @@ -223,14 +224,3 @@ static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpufreq_frequency_table *powernow_table); static int fill_powernow_table_fidvid(struct powernow_k8_data *data, struct cpufreq_frequency_table *powernow_table); - -#ifdef CONFIG_SMP -static inline void define_siblings(int cpu, cpumask_t cpu_sharedcore_mask[]) -{ -} -#else -static inline void define_siblings(int cpu, cpumask_t cpu_sharedcore_mask[]) -{ - cpu_set(0, cpu_sharedcore_mask[0]); -} -#endif diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c index 55c831ed71c..8d672ef162c 100644 --- a/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c +++ b/arch/x86/kernel/cpu/cpufreq/speedstep-centrino.c @@ -323,14 +323,8 @@ static unsigned int get_cur_freq(unsigned int cpu) { unsigned l, h; unsigned clock_freq; - cpumask_t saved_mask; - saved_mask = current->cpus_allowed; - set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu)); - if (smp_processor_id() != cpu) - return 0; - - rdmsr(MSR_IA32_PERF_STATUS, l, h); + rdmsr_on_cpu(cpu, MSR_IA32_PERF_STATUS, &l, &h); clock_freq = extract_clock(l, cpu, 0); if (unlikely(clock_freq == 0)) { @@ -340,11 +334,9 @@ static unsigned int get_cur_freq(unsigned int cpu) * P-state transition (like TM2). Get the last freq set * in PERF_CTL. */ - rdmsr(MSR_IA32_PERF_CTL, l, h); + rdmsr_on_cpu(cpu, MSR_IA32_PERF_CTL, &l, &h); clock_freq = extract_clock(l, cpu, 1); } - - set_cpus_allowed_ptr(current, &saved_mask); return clock_freq; } @@ -467,15 +459,10 @@ static int centrino_target (struct cpufreq_policy *policy, struct cpufreq_freqs freqs; int retval = 0; unsigned int j, k, first_cpu, tmp; - cpumask_var_t saved_mask, covered_cpus; + cpumask_var_t covered_cpus; - if (unlikely(!alloc_cpumask_var(&saved_mask, GFP_KERNEL))) - return -ENOMEM; - if (unlikely(!zalloc_cpumask_var(&covered_cpus, GFP_KERNEL))) { - free_cpumask_var(saved_mask); + if (unlikely(!zalloc_cpumask_var(&covered_cpus, GFP_KERNEL))) return -ENOMEM; - } - cpumask_copy(saved_mask, ¤t->cpus_allowed); if (unlikely(per_cpu(centrino_model, cpu) == NULL)) { retval = -ENODEV; @@ -493,7 +480,7 @@ static int centrino_target (struct cpufreq_policy *policy, first_cpu = 1; for_each_cpu(j, policy->cpus) { - const struct cpumask *mask; + int good_cpu; /* cpufreq holds the hotplug lock, so we are safe here */ if (!cpu_online(j)) @@ -504,32 +491,30 @@ static int centrino_target (struct cpufreq_policy *policy, * Make sure we are running on CPU that wants to change freq */ if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) - mask = policy->cpus; + good_cpu = cpumask_any_and(policy->cpus, + cpu_online_mask); else - mask = cpumask_of(j); + good_cpu = j; - set_cpus_allowed_ptr(current, mask); - preempt_disable(); - if (unlikely(!cpu_isset(smp_processor_id(), *mask))) { + if (good_cpu >= nr_cpu_ids) { dprintk("couldn't limit to CPUs in this domain\n"); retval = -EAGAIN; if (first_cpu) { /* We haven't started the transition yet. */ - goto migrate_end; + goto out; } - preempt_enable(); break; } msr = per_cpu(centrino_model, cpu)->op_points[newstate].index; if (first_cpu) { - rdmsr(MSR_IA32_PERF_CTL, oldmsr, h); + rdmsr_on_cpu(good_cpu, MSR_IA32_PERF_CTL, &oldmsr, &h); if (msr == (oldmsr & 0xffff)) { dprintk("no change needed - msr was and needs " "to be %x\n", oldmsr); retval = 0; - goto migrate_end; + goto out; } freqs.old = extract_clock(oldmsr, cpu, 0); @@ -553,14 +538,11 @@ static int centrino_target (struct cpufreq_policy *policy, oldmsr |= msr; } - wrmsr(MSR_IA32_PERF_CTL, oldmsr, h); - if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) { - preempt_enable(); + wrmsr_on_cpu(good_cpu, MSR_IA32_PERF_CTL, oldmsr, h); + if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) break; - } - cpu_set(j, *covered_cpus); - preempt_enable(); + cpumask_set_cpu(j, covered_cpus); } for_each_cpu(k, policy->cpus) { @@ -578,10 +560,8 @@ static int centrino_target (struct cpufreq_policy *policy, * Best effort undo.. */ - for_each_cpu_mask_nr(j, *covered_cpus) { - set_cpus_allowed_ptr(current, &cpumask_of_cpu(j)); - wrmsr(MSR_IA32_PERF_CTL, oldmsr, h); - } + for_each_cpu(j, covered_cpus) + wrmsr_on_cpu(j, MSR_IA32_PERF_CTL, oldmsr, h); tmp = freqs.new; freqs.new = freqs.old; @@ -593,15 +573,9 @@ static int centrino_target (struct cpufreq_policy *policy, cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); } } - set_cpus_allowed_ptr(current, saved_mask); retval = 0; - goto out; -migrate_end: - preempt_enable(); - set_cpus_allowed_ptr(current, saved_mask); out: - free_cpumask_var(saved_mask); free_cpumask_var(covered_cpus); return retval; } diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c index 016c1a4fa3f..6911e91fb4f 100644 --- a/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c +++ b/arch/x86/kernel/cpu/cpufreq/speedstep-ich.c @@ -89,7 +89,8 @@ static int speedstep_find_register(void) * speedstep_set_state - set the SpeedStep state * @state: new processor frequency state (SPEEDSTEP_LOW or SPEEDSTEP_HIGH) * - * Tries to change the SpeedStep state. + * Tries to change the SpeedStep state. Can be called from + * smp_call_function_single. */ static void speedstep_set_state(unsigned int state) { @@ -143,6 +144,11 @@ static void speedstep_set_state(unsigned int state) return; } +/* Wrapper for smp_call_function_single. */ +static void _speedstep_set_state(void *_state) +{ + speedstep_set_state(*(unsigned int *)_state); +} /** * speedstep_activate - activate SpeedStep control in the chipset @@ -226,22 +232,28 @@ static unsigned int speedstep_detect_chipset(void) return 0; } -static unsigned int _speedstep_get(const struct cpumask *cpus) -{ +struct get_freq_data { unsigned int speed; - cpumask_t cpus_allowed; - - cpus_allowed = current->cpus_allowed; - set_cpus_allowed_ptr(current, cpus); - speed = speedstep_get_frequency(speedstep_processor); - set_cpus_allowed_ptr(current, &cpus_allowed); - dprintk("detected %u kHz as current frequency\n", speed); - return speed; + unsigned int processor; +}; + +static void get_freq_data(void *_data) +{ + struct get_freq_data *data = _data; + + data->speed = speedstep_get_frequency(data->processor); } static unsigned int speedstep_get(unsigned int cpu) { - return _speedstep_get(cpumask_of(cpu)); + struct get_freq_data data = { .processor = cpu }; + + /* You're supposed to ensure CPU is online. */ + if (smp_call_function_single(cpu, get_freq_data, &data, 1) != 0) + BUG(); + + dprintk("detected %u kHz as current frequency\n", data.speed); + return data.speed; } /** @@ -257,16 +269,16 @@ static int speedstep_target(struct cpufreq_policy *policy, unsigned int target_freq, unsigned int relation) { - unsigned int newstate = 0; + unsigned int newstate = 0, policy_cpu; struct cpufreq_freqs freqs; - cpumask_t cpus_allowed; int i; if (cpufreq_frequency_table_target(policy, &speedstep_freqs[0], target_freq, relation, &newstate)) return -EINVAL; - freqs.old = _speedstep_get(policy->cpus); + policy_cpu = cpumask_any_and(policy->cpus, cpu_online_mask); + freqs.old = speedstep_get(policy_cpu); freqs.new = speedstep_freqs[newstate].frequency; freqs.cpu = policy->cpu; @@ -276,20 +288,13 @@ static int speedstep_target(struct cpufreq_policy *policy, if (freqs.old == freqs.new) return 0; - cpus_allowed = current->cpus_allowed; - for_each_cpu(i, policy->cpus) { freqs.cpu = i; cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); } - /* switch to physical CPU where state is to be changed */ - set_cpus_allowed_ptr(current, policy->cpus); - - speedstep_set_state(newstate); - - /* allow to be run on all CPUs */ - set_cpus_allowed_ptr(current, &cpus_allowed); + smp_call_function_single(policy_cpu, _speedstep_set_state, &newstate, + true); for_each_cpu(i, policy->cpus) { freqs.cpu = i; @@ -312,33 +317,43 @@ static int speedstep_verify(struct cpufreq_policy *policy) return cpufreq_frequency_table_verify(policy, &speedstep_freqs[0]); } +struct get_freqs { + struct cpufreq_policy *policy; + int ret; +}; + +static void get_freqs_on_cpu(void *_get_freqs) +{ + struct get_freqs *get_freqs = _get_freqs; + + get_freqs->ret = + speedstep_get_freqs(speedstep_processor, + &speedstep_freqs[SPEEDSTEP_LOW].frequency, + &speedstep_freqs[SPEEDSTEP_HIGH].frequency, + &get_freqs->policy->cpuinfo.transition_latency, + &speedstep_set_state); +} static int speedstep_cpu_init(struct cpufreq_policy *policy) { - int result = 0; - unsigned int speed; - cpumask_t cpus_allowed; + int result; + unsigned int policy_cpu, speed; + struct get_freqs gf; /* only run on CPU to be set, or on its sibling */ #ifdef CONFIG_SMP cpumask_copy(policy->cpus, cpu_sibling_mask(policy->cpu)); #endif - - cpus_allowed = current->cpus_allowed; - set_cpus_allowed_ptr(current, policy->cpus); + policy_cpu = cpumask_any_and(policy->cpus, cpu_online_mask); /* detect low and high frequency and transition latency */ - result = speedstep_get_freqs(speedstep_processor, - &speedstep_freqs[SPEEDSTEP_LOW].frequency, - &speedstep_freqs[SPEEDSTEP_HIGH].frequency, - &policy->cpuinfo.transition_latency, - &speedstep_set_state); - set_cpus_allowed_ptr(current, &cpus_allowed); - if (result) - return result; + gf.policy = policy; + smp_call_function_single(policy_cpu, get_freqs_on_cpu, &gf, 1); + if (gf.ret) + return gf.ret; /* get current speed setting */ - speed = _speedstep_get(policy->cpus); + speed = speedstep_get(policy_cpu); if (!speed) return -EIO; diff --git a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c index 2e3c6862657..f4c290b8482 100644 --- a/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c +++ b/arch/x86/kernel/cpu/cpufreq/speedstep-lib.c @@ -226,6 +226,7 @@ static unsigned int pentium4_get_frequency(void) } +/* Warning: may get called from smp_call_function_single. */ unsigned int speedstep_get_frequency(unsigned int processor) { switch (processor) { diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 7437fa133c0..3260ab04499 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -86,6 +86,29 @@ static void __cpuinit early_init_intel(struct cpuinfo_x86 *c) */ if (c->x86 == 6 && c->x86_model < 15) clear_cpu_cap(c, X86_FEATURE_PAT); + +#ifdef CONFIG_KMEMCHECK + /* + * P4s have a "fast strings" feature which causes single- + * stepping REP instructions to only generate a #DB on + * cache-line boundaries. + * + * Ingo Molnar reported a Pentium D (model 6) and a Xeon + * (model 2) with the same problem. + */ + if (c->x86 == 15) { + u64 misc_enable; + + rdmsrl(MSR_IA32_MISC_ENABLE, misc_enable); + + if (misc_enable & MSR_IA32_MISC_ENABLE_FAST_STRING) { + printk(KERN_INFO "kmemcheck: Disabling fast string operations\n"); + + misc_enable &= ~MSR_IA32_MISC_ENABLE_FAST_STRING; + wrmsrl(MSR_IA32_MISC_ENABLE, misc_enable); + } + } +#endif } #ifdef CONFIG_X86_32 @@ -229,12 +252,12 @@ static void __cpuinit intel_workarounds(struct cpuinfo_x86 *c) } #endif -static void __cpuinit srat_detect_node(void) +static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c) { #if defined(CONFIG_NUMA) && defined(CONFIG_X86_64) unsigned node; int cpu = smp_processor_id(); - int apicid = hard_smp_processor_id(); + int apicid = cpu_has_apic ? hard_smp_processor_id() : c->apicid; /* Don't do the funky fallback heuristics the AMD version employs for now. */ @@ -400,7 +423,7 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) } /* Work around errata */ - srat_detect_node(); + srat_detect_node(c); if (cpu_has(c, X86_FEATURE_VMX)) detect_vmx_virtcap(c); diff --git a/arch/x86/kernel/cpu/intel_cacheinfo.c b/arch/x86/kernel/cpu/intel_cacheinfo.c index 483eda96e10..789efe217e1 100644 --- a/arch/x86/kernel/cpu/intel_cacheinfo.c +++ b/arch/x86/kernel/cpu/intel_cacheinfo.c @@ -17,6 +17,7 @@ #include <asm/processor.h> #include <asm/smp.h> +#include <asm/k8.h> #define LVL_1_INST 1 #define LVL_1_DATA 2 @@ -159,14 +160,6 @@ struct _cpuid4_info_regs { unsigned long can_disable; }; -#if defined(CONFIG_PCI) && defined(CONFIG_SYSFS) -static struct pci_device_id k8_nb_id[] = { - { PCI_DEVICE(PCI_VENDOR_ID_AMD, 0x1103) }, - { PCI_DEVICE(PCI_VENDOR_ID_AMD, 0x1203) }, - {} -}; -#endif - unsigned short num_cache_leaves; /* AMD doesn't have CPUID4. Emulate it here to report the same @@ -207,10 +200,17 @@ union l3_cache { }; static const unsigned short __cpuinitconst assocs[] = { - [1] = 1, [2] = 2, [4] = 4, [6] = 8, - [8] = 16, [0xa] = 32, [0xb] = 48, + [1] = 1, + [2] = 2, + [4] = 4, + [6] = 8, + [8] = 16, + [0xa] = 32, + [0xb] = 48, [0xc] = 64, - [0xf] = 0xffff // ?? + [0xd] = 96, + [0xe] = 128, + [0xf] = 0xffff /* fully associative - no way to show this currently */ }; static const unsigned char __cpuinitconst levels[] = { 1, 1, 2, 3 }; @@ -271,7 +271,8 @@ amd_cpuid4(int leaf, union _cpuid4_leaf_eax *eax, eax->split.type = types[leaf]; eax->split.level = levels[leaf]; if (leaf == 3) - eax->split.num_threads_sharing = current_cpu_data.x86_max_cores - 1; + eax->split.num_threads_sharing = + current_cpu_data.x86_max_cores - 1; else eax->split.num_threads_sharing = 0; eax->split.num_cores_on_die = current_cpu_data.x86_max_cores - 1; @@ -291,6 +292,14 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) { if (index < 3) return; + + if (boot_cpu_data.x86 == 0x11) + return; + + /* see erratum #382 */ + if ((boot_cpu_data.x86 == 0x10) && (boot_cpu_data.x86_model < 0x8)) + return; + this_leaf->can_disable = 1; } @@ -696,97 +705,75 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) #define to_object(k) container_of(k, struct _index_kobject, kobj) #define to_attr(a) container_of(a, struct _cache_attr, attr) -#ifdef CONFIG_PCI -static struct pci_dev *get_k8_northbridge(int node) -{ - struct pci_dev *dev = NULL; - int i; - - for (i = 0; i <= node; i++) { - do { - dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev); - if (!dev) - break; - } while (!pci_match_id(&k8_nb_id[0], dev)); - if (!dev) - break; - } - return dev; -} -#else -static struct pci_dev *get_k8_northbridge(int node) -{ - return NULL; -} -#endif - -static ssize_t show_cache_disable(struct _cpuid4_info *this_leaf, char *buf) +static ssize_t show_cache_disable(struct _cpuid4_info *this_leaf, char *buf, + unsigned int index) { - const struct cpumask *mask = to_cpumask(this_leaf->shared_cpu_map); - int node = cpu_to_node(cpumask_first(mask)); - struct pci_dev *dev = NULL; - ssize_t ret = 0; - int i; + int cpu = cpumask_first(to_cpumask(this_leaf->shared_cpu_map)); + int node = cpu_to_node(cpu); + struct pci_dev *dev = node_to_k8_nb_misc(node); + unsigned int reg = 0; if (!this_leaf->can_disable) - return sprintf(buf, "Feature not enabled\n"); - - dev = get_k8_northbridge(node); - if (!dev) { - printk(KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n"); return -EINVAL; - } - for (i = 0; i < 2; i++) { - unsigned int reg; + if (!dev) + return -EINVAL; - pci_read_config_dword(dev, 0x1BC + i * 4, ®); + pci_read_config_dword(dev, 0x1BC + index * 4, ®); + return sprintf(buf, "%x\n", reg); +} - ret += sprintf(buf, "%sEntry: %d\n", buf, i); - ret += sprintf(buf, "%sReads: %s\tNew Entries: %s\n", - buf, - reg & 0x80000000 ? "Disabled" : "Allowed", - reg & 0x40000000 ? "Disabled" : "Allowed"); - ret += sprintf(buf, "%sSubCache: %x\tIndex: %x\n", - buf, (reg & 0x30000) >> 16, reg & 0xfff); - } - return ret; +#define SHOW_CACHE_DISABLE(index) \ +static ssize_t \ +show_cache_disable_##index(struct _cpuid4_info *this_leaf, char *buf) \ +{ \ + return show_cache_disable(this_leaf, buf, index); \ } +SHOW_CACHE_DISABLE(0) +SHOW_CACHE_DISABLE(1) -static ssize_t -store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf, - size_t count) +static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf, + const char *buf, size_t count, unsigned int index) { - const struct cpumask *mask = to_cpumask(this_leaf->shared_cpu_map); - int node = cpu_to_node(cpumask_first(mask)); - struct pci_dev *dev = NULL; - unsigned int ret, index, val; + int cpu = cpumask_first(to_cpumask(this_leaf->shared_cpu_map)); + int node = cpu_to_node(cpu); + struct pci_dev *dev = node_to_k8_nb_misc(node); + unsigned long val = 0; + unsigned int scrubber = 0; if (!this_leaf->can_disable) - return 0; - - if (strlen(buf) > 15) return -EINVAL; - ret = sscanf(buf, "%x %x", &index, &val); - if (ret != 2) + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (!dev) return -EINVAL; - if (index > 1) + + if (strict_strtoul(buf, 10, &val) < 0) return -EINVAL; val |= 0xc0000000; - dev = get_k8_northbridge(node); - if (!dev) { - printk(KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n"); - return -EINVAL; - } + + pci_read_config_dword(dev, 0x58, &scrubber); + scrubber &= ~0x1f000000; + pci_write_config_dword(dev, 0x58, scrubber); pci_write_config_dword(dev, 0x1BC + index * 4, val & ~0x40000000); wbinvd(); pci_write_config_dword(dev, 0x1BC + index * 4, val); + return count; +} - return 1; +#define STORE_CACHE_DISABLE(index) \ +static ssize_t \ +store_cache_disable_##index(struct _cpuid4_info *this_leaf, \ + const char *buf, size_t count) \ +{ \ + return store_cache_disable(this_leaf, buf, count, index); \ } +STORE_CACHE_DISABLE(0) +STORE_CACHE_DISABLE(1) struct _cache_attr { struct attribute attr; @@ -808,7 +795,10 @@ define_one_ro(size); define_one_ro(shared_cpu_map); define_one_ro(shared_cpu_list); -static struct _cache_attr cache_disable = __ATTR(cache_disable, 0644, show_cache_disable, store_cache_disable); +static struct _cache_attr cache_disable_0 = __ATTR(cache_disable_0, 0644, + show_cache_disable_0, store_cache_disable_0); +static struct _cache_attr cache_disable_1 = __ATTR(cache_disable_1, 0644, + show_cache_disable_1, store_cache_disable_1); static struct attribute * default_attrs[] = { &type.attr, @@ -820,7 +810,8 @@ static struct attribute * default_attrs[] = { &size.attr, &shared_cpu_map.attr, &shared_cpu_list.attr, - &cache_disable.attr, + &cache_disable_0.attr, + &cache_disable_1.attr, NULL }; diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile index b2f89829bbe..188a1ca5ad2 100644 --- a/arch/x86/kernel/cpu/mcheck/Makefile +++ b/arch/x86/kernel/cpu/mcheck/Makefile @@ -1,7 +1,12 @@ -obj-y = mce_$(BITS).o therm_throt.o +obj-y = mce.o -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_NEW_MCE) += mce-severity.o +obj-$(CONFIG_X86_OLD_MCE) += k7.o p4.o p6.o +obj-$(CONFIG_X86_ANCIENT_MCE) += winchip.o p5.o +obj-$(CONFIG_X86_MCE_INTEL) += mce_intel.o +obj-$(CONFIG_X86_MCE_AMD) += mce_amd.o obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o obj-$(CONFIG_X86_MCE_THRESHOLD) += threshold.o +obj-$(CONFIG_X86_MCE_INJECT) += mce-inject.o + +obj-$(CONFIG_X86_THERMAL_VECTOR) += therm_throt.o diff --git a/arch/x86/kernel/cpu/mcheck/k7.c b/arch/x86/kernel/cpu/mcheck/k7.c index dd3af6e7b39..b945d5dbc60 100644 --- a/arch/x86/kernel/cpu/mcheck/k7.c +++ b/arch/x86/kernel/cpu/mcheck/k7.c @@ -2,25 +2,23 @@ * Athlon specific Machine Check Exception Reporting * (C) Copyright 2002 Dave Jones <davej@redhat.com> */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> #include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/init.h> #include <linux/smp.h> #include <asm/processor.h> #include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include "mce.h" - -/* Machine Check Handler For AMD Athlon/Duron */ +/* Machine Check Handler For AMD Athlon/Duron: */ static void k7_machine_check(struct pt_regs *regs, long error_code) { - int recover = 1; u32 alow, ahigh, high, low; u32 mcgstl, mcgsth; + int recover = 1; int i; rdmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); @@ -32,15 +30,19 @@ static void k7_machine_check(struct pt_regs *regs, long error_code) for (i = 1; i < nr_mce_banks; i++) { rdmsr(MSR_IA32_MC0_STATUS+i*4, low, high); - if (high&(1<<31)) { + if (high & (1<<31)) { char misc[20]; char addr[24]; - misc[0] = addr[0] = '\0'; + + misc[0] = '\0'; + addr[0] = '\0'; + if (high & (1<<29)) recover |= 1; if (high & (1<<25)) recover |= 2; high &= ~(1<<31); + if (high & (1<<27)) { rdmsr(MSR_IA32_MC0_MISC+i*4, alow, ahigh); snprintf(misc, 20, "[%08x%08x]", ahigh, alow); @@ -49,27 +51,31 @@ static void k7_machine_check(struct pt_regs *regs, long error_code) rdmsr(MSR_IA32_MC0_ADDR+i*4, alow, ahigh); snprintf(addr, 24, " at %08x%08x", ahigh, alow); } + printk(KERN_EMERG "CPU %d: Bank %d: %08x%08x%s%s\n", smp_processor_id(), i, high, low, misc, addr); - /* Clear it */ + + /* Clear it: */ wrmsr(MSR_IA32_MC0_STATUS+i*4, 0UL, 0UL); - /* Serialize */ + /* Serialize: */ wmb(); add_taint(TAINT_MACHINE_CHECK); } } - if (recover&2) + if (recover & 2) panic("CPU context corrupt"); - if (recover&1) + if (recover & 1) panic("Unable to continue"); + printk(KERN_EMERG "Attempting to continue.\n"); + mcgstl &= ~(1<<2); wrmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); } -/* AMD K7 machine check is Intel like */ +/* AMD K7 machine check is Intel like: */ void amd_mcheck_init(struct cpuinfo_x86 *c) { u32 l, h; @@ -79,21 +85,26 @@ void amd_mcheck_init(struct cpuinfo_x86 *c) return; machine_check_vector = k7_machine_check; + /* Make sure the vector pointer is visible before we enable MCEs: */ wmb(); printk(KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr(MSR_IA32_MCG_CAP, l, h); if (l & (1<<8)) /* Control register present ? */ wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); nr_mce_banks = l & 0xff; - /* Clear status for MC index 0 separately, we don't touch CTL, - * as some K7 Athlons cause spurious MCEs when its enabled. */ + /* + * Clear status for MC index 0 separately, we don't touch CTL, + * as some K7 Athlons cause spurious MCEs when its enabled: + */ if (boot_cpu_data.x86 == 6) { wrmsr(MSR_IA32_MC0_STATUS, 0x0, 0x0); i = 1; } else i = 0; + for (; i < nr_mce_banks; i++) { wrmsr(MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); wrmsr(MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); diff --git a/arch/x86/kernel/cpu/mcheck/mce-inject.c b/arch/x86/kernel/cpu/mcheck/mce-inject.c new file mode 100644 index 00000000000..a3a235a53f0 --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce-inject.c @@ -0,0 +1,127 @@ +/* + * Machine check injection support. + * Copyright 2008 Intel Corporation. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; version 2 + * of the License. + * + * Authors: + * Andi Kleen + * Ying Huang + */ +#include <linux/uaccess.h> +#include <linux/module.h> +#include <linux/timer.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/fs.h> +#include <linux/smp.h> +#include <asm/mce.h> + +/* Update fake mce registers on current CPU. */ +static void inject_mce(struct mce *m) +{ + struct mce *i = &per_cpu(injectm, m->extcpu); + + /* Make sure noone reads partially written injectm */ + i->finished = 0; + mb(); + m->finished = 0; + /* First set the fields after finished */ + i->extcpu = m->extcpu; + mb(); + /* Now write record in order, finished last (except above) */ + memcpy(i, m, sizeof(struct mce)); + /* Finally activate it */ + mb(); + i->finished = 1; +} + +struct delayed_mce { + struct timer_list timer; + struct mce m; +}; + +/* Inject mce on current CPU */ +static void raise_mce(unsigned long data) +{ + struct delayed_mce *dm = (struct delayed_mce *)data; + struct mce *m = &dm->m; + int cpu = m->extcpu; + + inject_mce(m); + if (m->status & MCI_STATUS_UC) { + struct pt_regs regs; + memset(®s, 0, sizeof(struct pt_regs)); + regs.ip = m->ip; + regs.cs = m->cs; + printk(KERN_INFO "Triggering MCE exception on CPU %d\n", cpu); + do_machine_check(®s, 0); + printk(KERN_INFO "MCE exception done on CPU %d\n", cpu); + } else { + mce_banks_t b; + memset(&b, 0xff, sizeof(mce_banks_t)); + printk(KERN_INFO "Starting machine check poll CPU %d\n", cpu); + machine_check_poll(0, &b); + mce_notify_irq(); + printk(KERN_INFO "Finished machine check poll on CPU %d\n", + cpu); + } + kfree(dm); +} + +/* Error injection interface */ +static ssize_t mce_write(struct file *filp, const char __user *ubuf, + size_t usize, loff_t *off) +{ + struct delayed_mce *dm; + struct mce m; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + /* + * There are some cases where real MSR reads could slip + * through. + */ + if (!boot_cpu_has(X86_FEATURE_MCE) || !boot_cpu_has(X86_FEATURE_MCA)) + return -EIO; + + if ((unsigned long)usize > sizeof(struct mce)) + usize = sizeof(struct mce); + if (copy_from_user(&m, ubuf, usize)) + return -EFAULT; + + if (m.extcpu >= num_possible_cpus() || !cpu_online(m.extcpu)) + return -EINVAL; + + dm = kmalloc(sizeof(struct delayed_mce), GFP_KERNEL); + if (!dm) + return -ENOMEM; + + /* + * Need to give user space some time to set everything up, + * so do it a jiffie or two later everywhere. + * Should we use a hrtimer here for better synchronization? + */ + memcpy(&dm->m, &m, sizeof(struct mce)); + setup_timer(&dm->timer, raise_mce, (unsigned long)dm); + dm->timer.expires = jiffies + 2; + add_timer_on(&dm->timer, m.extcpu); + return usize; +} + +static int inject_init(void) +{ + printk(KERN_INFO "Machine check injector initialized\n"); + mce_chrdev_ops.write = mce_write; + return 0; +} + +module_init(inject_init); +/* + * Cannot tolerate unloading currently because we cannot + * guarantee all openers of mce_chrdev will get a reference to us. + */ +MODULE_LICENSE("GPL"); diff --git a/arch/x86/kernel/cpu/mcheck/mce-internal.h b/arch/x86/kernel/cpu/mcheck/mce-internal.h new file mode 100644 index 00000000000..54dcb8ff12e --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce-internal.h @@ -0,0 +1,15 @@ +#include <asm/mce.h> + +enum severity_level { + MCE_NO_SEVERITY, + MCE_KEEP_SEVERITY, + MCE_SOME_SEVERITY, + MCE_AO_SEVERITY, + MCE_UC_SEVERITY, + MCE_AR_SEVERITY, + MCE_PANIC_SEVERITY, +}; + +int mce_severity(struct mce *a, int tolerant, char **msg); + +extern int mce_ser; diff --git a/arch/x86/kernel/cpu/mcheck/mce-severity.c b/arch/x86/kernel/cpu/mcheck/mce-severity.c new file mode 100644 index 00000000000..ff0807f9705 --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce-severity.c @@ -0,0 +1,218 @@ +/* + * MCE grading rules. + * Copyright 2008, 2009 Intel Corporation. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; version 2 + * of the License. + * + * Author: Andi Kleen + */ +#include <linux/kernel.h> +#include <linux/seq_file.h> +#include <linux/init.h> +#include <linux/debugfs.h> +#include <asm/mce.h> + +#include "mce-internal.h" + +/* + * Grade an mce by severity. In general the most severe ones are processed + * first. Since there are quite a lot of combinations test the bits in a + * table-driven way. The rules are simply processed in order, first + * match wins. + * + * Note this is only used for machine check exceptions, the corrected + * errors use much simpler rules. The exceptions still check for the corrected + * errors, but only to leave them alone for the CMCI handler (except for + * panic situations) + */ + +enum context { IN_KERNEL = 1, IN_USER = 2 }; +enum ser { SER_REQUIRED = 1, NO_SER = 2 }; + +static struct severity { + u64 mask; + u64 result; + unsigned char sev; + unsigned char mcgmask; + unsigned char mcgres; + unsigned char ser; + unsigned char context; + unsigned char covered; + char *msg; +} severities[] = { +#define KERNEL .context = IN_KERNEL +#define USER .context = IN_USER +#define SER .ser = SER_REQUIRED +#define NOSER .ser = NO_SER +#define SEV(s) .sev = MCE_ ## s ## _SEVERITY +#define BITCLR(x, s, m, r...) { .mask = x, .result = 0, SEV(s), .msg = m, ## r } +#define BITSET(x, s, m, r...) { .mask = x, .result = x, SEV(s), .msg = m, ## r } +#define MCGMASK(x, res, s, m, r...) \ + { .mcgmask = x, .mcgres = res, SEV(s), .msg = m, ## r } +#define MASK(x, y, s, m, r...) \ + { .mask = x, .result = y, SEV(s), .msg = m, ## r } +#define MCI_UC_S (MCI_STATUS_UC|MCI_STATUS_S) +#define MCI_UC_SAR (MCI_STATUS_UC|MCI_STATUS_S|MCI_STATUS_AR) +#define MCACOD 0xffff + + BITCLR(MCI_STATUS_VAL, NO, "Invalid"), + BITCLR(MCI_STATUS_EN, NO, "Not enabled"), + BITSET(MCI_STATUS_PCC, PANIC, "Processor context corrupt"), + /* When MCIP is not set something is very confused */ + MCGMASK(MCG_STATUS_MCIP, 0, PANIC, "MCIP not set in MCA handler"), + /* Neither return not error IP -- no chance to recover -> PANIC */ + MCGMASK(MCG_STATUS_RIPV|MCG_STATUS_EIPV, 0, PANIC, + "Neither restart nor error IP"), + MCGMASK(MCG_STATUS_RIPV, 0, PANIC, "In kernel and no restart IP", + KERNEL), + BITCLR(MCI_STATUS_UC, KEEP, "Corrected error", NOSER), + MASK(MCI_STATUS_OVER|MCI_STATUS_UC|MCI_STATUS_EN, MCI_STATUS_UC, SOME, + "Spurious not enabled", SER), + + /* ignore OVER for UCNA */ + MASK(MCI_UC_SAR, MCI_STATUS_UC, KEEP, + "Uncorrected no action required", SER), + MASK(MCI_STATUS_OVER|MCI_UC_SAR, MCI_STATUS_UC|MCI_STATUS_AR, PANIC, + "Illegal combination (UCNA with AR=1)", SER), + MASK(MCI_STATUS_S, 0, KEEP, "Non signalled machine check", SER), + + /* AR add known MCACODs here */ + MASK(MCI_STATUS_OVER|MCI_UC_SAR, MCI_STATUS_OVER|MCI_UC_SAR, PANIC, + "Action required with lost events", SER), + MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCACOD, MCI_UC_SAR, PANIC, + "Action required; unknown MCACOD", SER), + + /* known AO MCACODs: */ + MASK(MCI_UC_SAR|MCI_STATUS_OVER|0xfff0, MCI_UC_S|0xc0, AO, + "Action optional: memory scrubbing error", SER), + MASK(MCI_UC_SAR|MCI_STATUS_OVER|MCACOD, MCI_UC_S|0x17a, AO, + "Action optional: last level cache writeback error", SER), + + MASK(MCI_STATUS_OVER|MCI_UC_SAR, MCI_UC_S, SOME, + "Action optional unknown MCACOD", SER), + MASK(MCI_STATUS_OVER|MCI_UC_SAR, MCI_UC_S|MCI_STATUS_OVER, SOME, + "Action optional with lost events", SER), + BITSET(MCI_STATUS_UC|MCI_STATUS_OVER, PANIC, "Overflowed uncorrected"), + BITSET(MCI_STATUS_UC, UC, "Uncorrected"), + BITSET(0, SOME, "No match") /* always matches. keep at end */ +}; + +/* + * If the EIPV bit is set, it means the saved IP is the + * instruction which caused the MCE. + */ +static int error_context(struct mce *m) +{ + if (m->mcgstatus & MCG_STATUS_EIPV) + return (m->ip && (m->cs & 3) == 3) ? IN_USER : IN_KERNEL; + /* Unknown, assume kernel */ + return IN_KERNEL; +} + +int mce_severity(struct mce *a, int tolerant, char **msg) +{ + enum context ctx = error_context(a); + struct severity *s; + + for (s = severities;; s++) { + if ((a->status & s->mask) != s->result) + continue; + if ((a->mcgstatus & s->mcgmask) != s->mcgres) + continue; + if (s->ser == SER_REQUIRED && !mce_ser) + continue; + if (s->ser == NO_SER && mce_ser) + continue; + if (s->context && ctx != s->context) + continue; + if (msg) + *msg = s->msg; + s->covered = 1; + if (s->sev >= MCE_UC_SEVERITY && ctx == IN_KERNEL) { + if (panic_on_oops || tolerant < 1) + return MCE_PANIC_SEVERITY; + } + return s->sev; + } +} + +static void *s_start(struct seq_file *f, loff_t *pos) +{ + if (*pos >= ARRAY_SIZE(severities)) + return NULL; + return &severities[*pos]; +} + +static void *s_next(struct seq_file *f, void *data, loff_t *pos) +{ + if (++(*pos) >= ARRAY_SIZE(severities)) + return NULL; + return &severities[*pos]; +} + +static void s_stop(struct seq_file *f, void *data) +{ +} + +static int s_show(struct seq_file *f, void *data) +{ + struct severity *ser = data; + seq_printf(f, "%d\t%s\n", ser->covered, ser->msg); + return 0; +} + +static const struct seq_operations severities_seq_ops = { + .start = s_start, + .next = s_next, + .stop = s_stop, + .show = s_show, +}; + +static int severities_coverage_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &severities_seq_ops); +} + +static ssize_t severities_coverage_write(struct file *file, + const char __user *ubuf, + size_t count, loff_t *ppos) +{ + int i; + for (i = 0; i < ARRAY_SIZE(severities); i++) + severities[i].covered = 0; + return count; +} + +static const struct file_operations severities_coverage_fops = { + .open = severities_coverage_open, + .release = seq_release, + .read = seq_read, + .write = severities_coverage_write, +}; + +static int __init severities_debugfs_init(void) +{ + struct dentry *dmce = NULL, *fseverities_coverage = NULL; + + dmce = debugfs_create_dir("mce", NULL); + if (dmce == NULL) + goto err_out; + fseverities_coverage = debugfs_create_file("severities-coverage", + 0444, dmce, NULL, + &severities_coverage_fops); + if (fseverities_coverage == NULL) + goto err_out; + + return 0; + +err_out: + if (fseverities_coverage) + debugfs_remove(fseverities_coverage); + if (dmce) + debugfs_remove(dmce); + return -ENOMEM; +} +late_initcall(severities_debugfs_init); diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c new file mode 100644 index 00000000000..01213048f62 --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -0,0 +1,2060 @@ +/* + * Machine check handler. + * + * 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/thread_info.h> +#include <linux/capability.h> +#include <linux/miscdevice.h> +#include <linux/interrupt.h> +#include <linux/ratelimit.h> +#include <linux/kallsyms.h> +#include <linux/rcupdate.h> +#include <linux/kobject.h> +#include <linux/uaccess.h> +#include <linux/kdebug.h> +#include <linux/kernel.h> +#include <linux/percpu.h> +#include <linux/string.h> +#include <linux/sysdev.h> +#include <linux/delay.h> +#include <linux/ctype.h> +#include <linux/sched.h> +#include <linux/sysfs.h> +#include <linux/types.h> +#include <linux/init.h> +#include <linux/kmod.h> +#include <linux/poll.h> +#include <linux/nmi.h> +#include <linux/cpu.h> +#include <linux/smp.h> +#include <linux/fs.h> +#include <linux/mm.h> + +#include <asm/processor.h> +#include <asm/hw_irq.h> +#include <asm/apic.h> +#include <asm/idle.h> +#include <asm/ipi.h> +#include <asm/mce.h> +#include <asm/msr.h> + +#include "mce-internal.h" + +/* Handle unconfigured int18 (should never happen) */ +static void unexpected_machine_check(struct pt_regs *regs, long error_code) +{ + printk(KERN_ERR "CPU#%d: Unexpected int18 (Machine Check).\n", + smp_processor_id()); +} + +/* Call the installed machine check handler for this CPU setup. */ +void (*machine_check_vector)(struct pt_regs *, long error_code) = + unexpected_machine_check; + +int mce_disabled __read_mostly; + +#ifdef CONFIG_X86_NEW_MCE + +#define MISC_MCELOG_MINOR 227 + +#define SPINUNIT 100 /* 100ns */ + +atomic_t mce_entry; + +DEFINE_PER_CPU(unsigned, mce_exception_count); + +/* + * Tolerant levels: + * 0: always panic on uncorrected errors, log corrected errors + * 1: panic or SIGBUS on uncorrected errors, log corrected errors + * 2: SIGBUS or log uncorrected errors (if possible), log corrected errors + * 3: never panic or SIGBUS, log all errors (for testing only) + */ +static int tolerant __read_mostly = 1; +static int banks __read_mostly; +static u64 *bank __read_mostly; +static int rip_msr __read_mostly; +static int mce_bootlog __read_mostly = -1; +static int monarch_timeout __read_mostly = -1; +static int mce_panic_timeout __read_mostly; +static int mce_dont_log_ce __read_mostly; +int mce_cmci_disabled __read_mostly; +int mce_ignore_ce __read_mostly; +int mce_ser __read_mostly; + +/* User mode helper program triggered by machine check event */ +static unsigned long mce_need_notify; +static char mce_helper[128]; +static char *mce_helper_argv[2] = { mce_helper, NULL }; + +static unsigned long dont_init_banks; + +static DECLARE_WAIT_QUEUE_HEAD(mce_wait); +static DEFINE_PER_CPU(struct mce, mces_seen); +static int cpu_missing; + + +/* 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 +}; + +static inline int skip_bank_init(int i) +{ + return i < BITS_PER_LONG && test_bit(i, &dont_init_banks); +} + +static DEFINE_PER_CPU(struct work_struct, mce_work); + +/* Do initial initialization of a struct mce */ +void mce_setup(struct mce *m) +{ + memset(m, 0, sizeof(struct mce)); + m->cpu = m->extcpu = smp_processor_id(); + rdtscll(m->tsc); + /* We hope get_seconds stays lockless */ + m->time = get_seconds(); + m->cpuvendor = boot_cpu_data.x86_vendor; + m->cpuid = cpuid_eax(1); +#ifdef CONFIG_SMP + m->socketid = cpu_data(m->extcpu).phys_proc_id; +#endif + m->apicid = cpu_data(m->extcpu).initial_apicid; + rdmsrl(MSR_IA32_MCG_CAP, m->mcgcap); +} + +DEFINE_PER_CPU(struct mce, injectm); +EXPORT_PER_CPU_SYMBOL_GPL(injectm); + +/* + * Lockless MCE logging infrastructure. + * This avoids deadlocks on printk locks without having to break locks. Also + * separate MCEs from kernel messages to avoid bogus bug reports. + */ + +static struct mce_log mcelog = { + .signature = MCE_LOG_SIGNATURE, + .len = MCE_LOG_LEN, + .recordlen = sizeof(struct mce), +}; + +void mce_log(struct mce *mce) +{ + unsigned next, entry; + + mce->finished = 0; + wmb(); + for (;;) { + entry = rcu_dereference(mcelog.next); + for (;;) { + /* + * When the buffer fills up discard new entries. + * Assume that the earlier errors are the more + * interesting ones: + */ + if (entry >= MCE_LOG_LEN) { + set_bit(MCE_OVERFLOW, + (unsigned long *)&mcelog.flags); + return; + } + /* Old left over entry. Skip: */ + if (mcelog.entry[entry].finished) { + entry++; + continue; + } + break; + } + smp_rmb(); + next = entry + 1; + if (cmpxchg(&mcelog.next, entry, next) == entry) + break; + } + memcpy(mcelog.entry + entry, mce, sizeof(struct mce)); + wmb(); + mcelog.entry[entry].finished = 1; + wmb(); + + mce->finished = 1; + set_bit(0, &mce_need_notify); +} + +static void print_mce(struct mce *m) +{ + printk(KERN_EMERG + "CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n", + m->extcpu, m->mcgstatus, m->bank, m->status); + if (m->ip) { + printk(KERN_EMERG "RIP%s %02x:<%016Lx> ", + !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "", + m->cs, m->ip); + if (m->cs == __KERNEL_CS) + print_symbol("{%s}", m->ip); + printk(KERN_CONT "\n"); + } + printk(KERN_EMERG "TSC %llx ", m->tsc); + if (m->addr) + printk(KERN_CONT "ADDR %llx ", m->addr); + if (m->misc) + printk(KERN_CONT "MISC %llx ", m->misc); + printk(KERN_CONT "\n"); + printk(KERN_EMERG "PROCESSOR %u:%x TIME %llu SOCKET %u APIC %x\n", + m->cpuvendor, m->cpuid, m->time, m->socketid, + m->apicid); +} + +static void print_mce_head(void) +{ + printk(KERN_EMERG "\nHARDWARE ERROR\n"); +} + +static void print_mce_tail(void) +{ + printk(KERN_EMERG "This is not a software problem!\n" + "Run through mcelog --ascii to decode and contact your hardware vendor\n"); +} + +#define PANIC_TIMEOUT 5 /* 5 seconds */ + +static atomic_t mce_paniced; + +/* Panic in progress. Enable interrupts and wait for final IPI */ +static void wait_for_panic(void) +{ + long timeout = PANIC_TIMEOUT*USEC_PER_SEC; + preempt_disable(); + local_irq_enable(); + while (timeout-- > 0) + udelay(1); + if (panic_timeout == 0) + panic_timeout = mce_panic_timeout; + panic("Panicing machine check CPU died"); +} + +static void mce_panic(char *msg, struct mce *final, char *exp) +{ + int i; + + /* + * Make sure only one CPU runs in machine check panic + */ + if (atomic_add_return(1, &mce_paniced) > 1) + wait_for_panic(); + barrier(); + + bust_spinlocks(1); + console_verbose(); + print_mce_head(); + /* First print corrected ones that are still unlogged */ + for (i = 0; i < MCE_LOG_LEN; i++) { + struct mce *m = &mcelog.entry[i]; + if (!(m->status & MCI_STATUS_VAL)) + continue; + if (!(m->status & MCI_STATUS_UC)) + print_mce(m); + } + /* Now print uncorrected but with the final one last */ + for (i = 0; i < MCE_LOG_LEN; i++) { + struct mce *m = &mcelog.entry[i]; + if (!(m->status & MCI_STATUS_VAL)) + continue; + if (!(m->status & MCI_STATUS_UC)) + continue; + if (!final || memcmp(m, final, sizeof(struct mce))) + print_mce(m); + } + if (final) + print_mce(final); + if (cpu_missing) + printk(KERN_EMERG "Some CPUs didn't answer in synchronization\n"); + print_mce_tail(); + if (exp) + printk(KERN_EMERG "Machine check: %s\n", exp); + if (panic_timeout == 0) + panic_timeout = mce_panic_timeout; + panic(msg); +} + +/* Support code for software error injection */ + +static int msr_to_offset(u32 msr) +{ + unsigned bank = __get_cpu_var(injectm.bank); + if (msr == rip_msr) + return offsetof(struct mce, ip); + if (msr == MSR_IA32_MC0_STATUS + bank*4) + return offsetof(struct mce, status); + if (msr == MSR_IA32_MC0_ADDR + bank*4) + return offsetof(struct mce, addr); + if (msr == MSR_IA32_MC0_MISC + bank*4) + return offsetof(struct mce, misc); + if (msr == MSR_IA32_MCG_STATUS) + return offsetof(struct mce, mcgstatus); + return -1; +} + +/* MSR access wrappers used for error injection */ +static u64 mce_rdmsrl(u32 msr) +{ + u64 v; + if (__get_cpu_var(injectm).finished) { + int offset = msr_to_offset(msr); + if (offset < 0) + return 0; + return *(u64 *)((char *)&__get_cpu_var(injectm) + offset); + } + rdmsrl(msr, v); + return v; +} + +static void mce_wrmsrl(u32 msr, u64 v) +{ + if (__get_cpu_var(injectm).finished) { + int offset = msr_to_offset(msr); + if (offset >= 0) + *(u64 *)((char *)&__get_cpu_var(injectm) + offset) = v; + return; + } + wrmsrl(msr, v); +} + +/* + * Simple lockless ring to communicate PFNs from the exception handler with the + * process context work function. This is vastly simplified because there's + * only a single reader and a single writer. + */ +#define MCE_RING_SIZE 16 /* we use one entry less */ + +struct mce_ring { + unsigned short start; + unsigned short end; + unsigned long ring[MCE_RING_SIZE]; +}; +static DEFINE_PER_CPU(struct mce_ring, mce_ring); + +/* Runs with CPU affinity in workqueue */ +static int mce_ring_empty(void) +{ + struct mce_ring *r = &__get_cpu_var(mce_ring); + + return r->start == r->end; +} + +static int mce_ring_get(unsigned long *pfn) +{ + struct mce_ring *r; + int ret = 0; + + *pfn = 0; + get_cpu(); + r = &__get_cpu_var(mce_ring); + if (r->start == r->end) + goto out; + *pfn = r->ring[r->start]; + r->start = (r->start + 1) % MCE_RING_SIZE; + ret = 1; +out: + put_cpu(); + return ret; +} + +/* Always runs in MCE context with preempt off */ +static int mce_ring_add(unsigned long pfn) +{ + struct mce_ring *r = &__get_cpu_var(mce_ring); + unsigned next; + + next = (r->end + 1) % MCE_RING_SIZE; + if (next == r->start) + return -1; + r->ring[r->end] = pfn; + wmb(); + r->end = next; + return 0; +} + +int mce_available(struct cpuinfo_x86 *c) +{ + if (mce_disabled) + return 0; + return cpu_has(c, X86_FEATURE_MCE) && cpu_has(c, X86_FEATURE_MCA); +} + +static void mce_schedule_work(void) +{ + if (!mce_ring_empty()) { + struct work_struct *work = &__get_cpu_var(mce_work); + if (!work_pending(work)) + schedule_work(work); + } +} + +/* + * Get the address of the instruction at the time of the machine check + * error. + */ +static inline void mce_get_rip(struct mce *m, struct pt_regs *regs) +{ + + if (regs && (m->mcgstatus & (MCG_STATUS_RIPV|MCG_STATUS_EIPV))) { + m->ip = regs->ip; + m->cs = regs->cs; + } else { + m->ip = 0; + m->cs = 0; + } + if (rip_msr) + m->ip = mce_rdmsrl(rip_msr); +} + +#ifdef CONFIG_X86_LOCAL_APIC +/* + * Called after interrupts have been reenabled again + * when a MCE happened during an interrupts off region + * in the kernel. + */ +asmlinkage void smp_mce_self_interrupt(struct pt_regs *regs) +{ + ack_APIC_irq(); + exit_idle(); + irq_enter(); + mce_notify_irq(); + mce_schedule_work(); + irq_exit(); +} +#endif + +static void mce_report_event(struct pt_regs *regs) +{ + if (regs->flags & (X86_VM_MASK|X86_EFLAGS_IF)) { + mce_notify_irq(); + /* + * Triggering the work queue here is just an insurance + * policy in case the syscall exit notify handler + * doesn't run soon enough or ends up running on the + * wrong CPU (can happen when audit sleeps) + */ + mce_schedule_work(); + return; + } + +#ifdef CONFIG_X86_LOCAL_APIC + /* + * Without APIC do not notify. The event will be picked + * up eventually. + */ + if (!cpu_has_apic) + return; + + /* + * When interrupts are disabled we cannot use + * kernel services safely. Trigger an self interrupt + * through the APIC to instead do the notification + * after interrupts are reenabled again. + */ + apic->send_IPI_self(MCE_SELF_VECTOR); + + /* + * Wait for idle afterwards again so that we don't leave the + * APIC in a non idle state because the normal APIC writes + * cannot exclude us. + */ + apic_wait_icr_idle(); +#endif +} + +DEFINE_PER_CPU(unsigned, mce_poll_count); + +/* + * Poll for corrected events or events that happened before reset. + * Those are just logged through /dev/mcelog. + * + * This is executed in standard interrupt context. + * + * Note: spec recommends to panic for fatal unsignalled + * errors here. However this would be quite problematic -- + * we would need to reimplement the Monarch handling and + * it would mess up the exclusion between exception handler + * and poll hander -- * so we skip this for now. + * These cases should not happen anyways, or only when the CPU + * is already totally * confused. In this case it's likely it will + * not fully execute the machine check handler either. + */ +void machine_check_poll(enum mcp_flags flags, mce_banks_t *b) +{ + struct mce m; + int i; + + __get_cpu_var(mce_poll_count)++; + + mce_setup(&m); + + m.mcgstatus = mce_rdmsrl(MSR_IA32_MCG_STATUS); + 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(); + m.status = mce_rdmsrl(MSR_IA32_MC0_STATUS + i*4); + if (!(m.status & MCI_STATUS_VAL)) + continue; + + /* + * Uncorrected or signalled events are handled by the exception + * handler when it is enabled, so don't process those here. + * + * TBD do the same check for MCI_STATUS_EN here? + */ + if (!(flags & MCP_UC) && + (m.status & (mce_ser ? MCI_STATUS_S : MCI_STATUS_UC))) + continue; + + if (m.status & MCI_STATUS_MISCV) + m.misc = mce_rdmsrl(MSR_IA32_MC0_MISC + i*4); + if (m.status & MCI_STATUS_ADDRV) + m.addr = mce_rdmsrl(MSR_IA32_MC0_ADDR + i*4); + + 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. + */ + if (!(flags & MCP_DONTLOG) && !mce_dont_log_ce) { + mce_log(&m); + add_taint(TAINT_MACHINE_CHECK); + } + + /* + * Clear state for this bank. + */ + mce_wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0); + } + + /* + * Don't clear MCG_STATUS here because it's only defined for + * exceptions. + */ + + sync_core(); +} +EXPORT_SYMBOL_GPL(machine_check_poll); + +/* + * Do a quick check if any of the events requires a panic. + * This decides if we keep the events around or clear them. + */ +static int mce_no_way_out(struct mce *m, char **msg) +{ + int i; + + for (i = 0; i < banks; i++) { + m->status = mce_rdmsrl(MSR_IA32_MC0_STATUS + i*4); + if (mce_severity(m, tolerant, msg) >= MCE_PANIC_SEVERITY) + return 1; + } + return 0; +} + +/* + * Variable to establish order between CPUs while scanning. + * Each CPU spins initially until executing is equal its number. + */ +static atomic_t mce_executing; + +/* + * Defines order of CPUs on entry. First CPU becomes Monarch. + */ +static atomic_t mce_callin; + +/* + * Check if a timeout waiting for other CPUs happened. + */ +static int mce_timed_out(u64 *t) +{ + /* + * The others already did panic for some reason. + * Bail out like in a timeout. + * rmb() to tell the compiler that system_state + * might have been modified by someone else. + */ + rmb(); + if (atomic_read(&mce_paniced)) + wait_for_panic(); + if (!monarch_timeout) + goto out; + if ((s64)*t < SPINUNIT) { + /* CHECKME: Make panic default for 1 too? */ + if (tolerant < 1) + mce_panic("Timeout synchronizing machine check over CPUs", + NULL, NULL); + cpu_missing = 1; + return 1; + } + *t -= SPINUNIT; +out: + touch_nmi_watchdog(); + return 0; +} + +/* + * The Monarch's reign. The Monarch is the CPU who entered + * the machine check handler first. It waits for the others to + * raise the exception too and then grades them. When any + * error is fatal panic. Only then let the others continue. + * + * The other CPUs entering the MCE handler will be controlled by the + * Monarch. They are called Subjects. + * + * This way we prevent any potential data corruption in a unrecoverable case + * and also makes sure always all CPU's errors are examined. + * + * Also this detects the case of an machine check event coming from outer + * space (not detected by any CPUs) In this case some external agent wants + * us to shut down, so panic too. + * + * The other CPUs might still decide to panic if the handler happens + * in a unrecoverable place, but in this case the system is in a semi-stable + * state and won't corrupt anything by itself. It's ok to let the others + * continue for a bit first. + * + * All the spin loops have timeouts; when a timeout happens a CPU + * typically elects itself to be Monarch. + */ +static void mce_reign(void) +{ + int cpu; + struct mce *m = NULL; + int global_worst = 0; + char *msg = NULL; + char *nmsg = NULL; + + /* + * This CPU is the Monarch and the other CPUs have run + * through their handlers. + * Grade the severity of the errors of all the CPUs. + */ + for_each_possible_cpu(cpu) { + int severity = mce_severity(&per_cpu(mces_seen, cpu), tolerant, + &nmsg); + if (severity > global_worst) { + msg = nmsg; + global_worst = severity; + m = &per_cpu(mces_seen, cpu); + } + } + + /* + * Cannot recover? Panic here then. + * This dumps all the mces in the log buffer and stops the + * other CPUs. + */ + if (m && global_worst >= MCE_PANIC_SEVERITY && tolerant < 3) + mce_panic("Fatal Machine check", m, msg); + + /* + * For UC somewhere we let the CPU who detects it handle it. + * Also must let continue the others, otherwise the handling + * CPU could deadlock on a lock. + */ + + /* + * No machine check event found. Must be some external + * source or one CPU is hung. Panic. + */ + if (!m && tolerant < 3) + mce_panic("Machine check from unknown source", NULL, NULL); + + /* + * Now clear all the mces_seen so that they don't reappear on + * the next mce. + */ + for_each_possible_cpu(cpu) + memset(&per_cpu(mces_seen, cpu), 0, sizeof(struct mce)); +} + +static atomic_t global_nwo; + +/* + * Start of Monarch synchronization. This waits until all CPUs have + * entered the exception handler and then determines if any of them + * saw a fatal event that requires panic. Then it executes them + * in the entry order. + * TBD double check parallel CPU hotunplug + */ +static int mce_start(int *no_way_out) +{ + int order; + int cpus = num_online_cpus(); + u64 timeout = (u64)monarch_timeout * NSEC_PER_USEC; + + if (!timeout) + return -1; + + atomic_add(*no_way_out, &global_nwo); + /* + * global_nwo should be updated before mce_callin + */ + smp_wmb(); + order = atomic_add_return(1, &mce_callin); + + /* + * Wait for everyone. + */ + while (atomic_read(&mce_callin) != cpus) { + if (mce_timed_out(&timeout)) { + atomic_set(&global_nwo, 0); + return -1; + } + ndelay(SPINUNIT); + } + + /* + * mce_callin should be read before global_nwo + */ + smp_rmb(); + + if (order == 1) { + /* + * Monarch: Starts executing now, the others wait. + */ + atomic_set(&mce_executing, 1); + } else { + /* + * Subject: Now start the scanning loop one by one in + * the original callin order. + * This way when there are any shared banks it will be + * only seen by one CPU before cleared, avoiding duplicates. + */ + while (atomic_read(&mce_executing) < order) { + if (mce_timed_out(&timeout)) { + atomic_set(&global_nwo, 0); + return -1; + } + ndelay(SPINUNIT); + } + } + + /* + * Cache the global no_way_out state. + */ + *no_way_out = atomic_read(&global_nwo); + + return order; +} + +/* + * Synchronize between CPUs after main scanning loop. + * This invokes the bulk of the Monarch processing. + */ +static int mce_end(int order) +{ + int ret = -1; + u64 timeout = (u64)monarch_timeout * NSEC_PER_USEC; + + if (!timeout) + goto reset; + if (order < 0) + goto reset; + + /* + * Allow others to run. + */ + atomic_inc(&mce_executing); + + if (order == 1) { + /* CHECKME: Can this race with a parallel hotplug? */ + int cpus = num_online_cpus(); + + /* + * Monarch: Wait for everyone to go through their scanning + * loops. + */ + while (atomic_read(&mce_executing) <= cpus) { + if (mce_timed_out(&timeout)) + goto reset; + ndelay(SPINUNIT); + } + + mce_reign(); + barrier(); + ret = 0; + } else { + /* + * Subject: Wait for Monarch to finish. + */ + while (atomic_read(&mce_executing) != 0) { + if (mce_timed_out(&timeout)) + goto reset; + ndelay(SPINUNIT); + } + + /* + * Don't reset anything. That's done by the Monarch. + */ + return 0; + } + + /* + * Reset all global state. + */ +reset: + atomic_set(&global_nwo, 0); + atomic_set(&mce_callin, 0); + barrier(); + + /* + * Let others run again. + */ + atomic_set(&mce_executing, 0); + return ret; +} + +/* + * Check if the address reported by the CPU is in a format we can parse. + * It would be possible to add code for most other cases, but all would + * be somewhat complicated (e.g. segment offset would require an instruction + * parser). So only support physical addresses upto page granuality for now. + */ +static int mce_usable_address(struct mce *m) +{ + if (!(m->status & MCI_STATUS_MISCV) || !(m->status & MCI_STATUS_ADDRV)) + return 0; + if ((m->misc & 0x3f) > PAGE_SHIFT) + return 0; + if (((m->misc >> 6) & 7) != MCM_ADDR_PHYS) + return 0; + return 1; +} + +static void mce_clear_state(unsigned long *toclear) +{ + int i; + + for (i = 0; i < banks; i++) { + if (test_bit(i, toclear)) + mce_wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0); + } +} + +/* + * 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! + * + * On Intel systems this is entered on all CPUs in parallel through + * MCE broadcast. However some CPUs might be broken beyond repair, + * so be always careful when synchronizing with others. + */ +void do_machine_check(struct pt_regs *regs, long error_code) +{ + struct mce m, *final; + int i; + int worst = 0; + int severity; + /* + * Establish sequential order between the CPUs entering the machine + * check handler. + */ + int order; + /* + * If no_way_out gets set, there is no safe way to recover from this + * MCE. If tolerant is cranked up, we'll try anyway. + */ + int no_way_out = 0; + /* + * If kill_it gets set, there might be a way to recover from this + * error. + */ + int kill_it = 0; + DECLARE_BITMAP(toclear, MAX_NR_BANKS); + char *msg = "Unknown"; + + atomic_inc(&mce_entry); + + __get_cpu_var(mce_exception_count)++; + + if (notify_die(DIE_NMI, "machine check", regs, error_code, + 18, SIGKILL) == NOTIFY_STOP) + goto out; + if (!banks) + goto out; + + mce_setup(&m); + + m.mcgstatus = mce_rdmsrl(MSR_IA32_MCG_STATUS); + no_way_out = mce_no_way_out(&m, &msg); + + final = &__get_cpu_var(mces_seen); + *final = m; + + barrier(); + + /* + * When no restart IP must always kill or panic. + */ + if (!(m.mcgstatus & MCG_STATUS_RIPV)) + kill_it = 1; + + /* + * Go through all the banks in exclusion of the other CPUs. + * This way we don't report duplicated events on shared banks + * because the first one to see it will clear it. + */ + order = mce_start(&no_way_out); + for (i = 0; i < banks; i++) { + __clear_bit(i, toclear); + if (!bank[i]) + continue; + + m.misc = 0; + m.addr = 0; + m.bank = i; + + m.status = mce_rdmsrl(MSR_IA32_MC0_STATUS + i*4); + if ((m.status & MCI_STATUS_VAL) == 0) + continue; + + /* + * Non uncorrected or non signaled errors are handled by + * machine_check_poll. Leave them alone, unless this panics. + */ + if (!(m.status & (mce_ser ? MCI_STATUS_S : MCI_STATUS_UC)) && + !no_way_out) + continue; + + /* + * Set taint even when machine check was not enabled. + */ + add_taint(TAINT_MACHINE_CHECK); + + severity = mce_severity(&m, tolerant, NULL); + + /* + * When machine check was for corrected handler don't touch, + * unless we're panicing. + */ + if (severity == MCE_KEEP_SEVERITY && !no_way_out) + continue; + __set_bit(i, toclear); + if (severity == MCE_NO_SEVERITY) { + /* + * Machine check event was not enabled. Clear, but + * ignore. + */ + continue; + } + + /* + * Kill on action required. + */ + if (severity == MCE_AR_SEVERITY) + kill_it = 1; + + if (m.status & MCI_STATUS_MISCV) + m.misc = mce_rdmsrl(MSR_IA32_MC0_MISC + i*4); + if (m.status & MCI_STATUS_ADDRV) + m.addr = mce_rdmsrl(MSR_IA32_MC0_ADDR + i*4); + + /* + * Action optional error. Queue address for later processing. + * When the ring overflows we just ignore the AO error. + * RED-PEN add some logging mechanism when + * usable_address or mce_add_ring fails. + * RED-PEN don't ignore overflow for tolerant == 0 + */ + if (severity == MCE_AO_SEVERITY && mce_usable_address(&m)) + mce_ring_add(m.addr >> PAGE_SHIFT); + + mce_get_rip(&m, regs); + mce_log(&m); + + if (severity > worst) { + *final = m; + worst = severity; + } + } + + if (!no_way_out) + mce_clear_state(toclear); + + /* + * Do most of the synchronization with other CPUs. + * When there's any problem use only local no_way_out state. + */ + if (mce_end(order) < 0) + no_way_out = worst >= MCE_PANIC_SEVERITY; + + /* + * If we have decided that we just CAN'T continue, and the user + * has not set tolerant to an insane level, give up and die. + * + * This is mainly used in the case when the system doesn't + * support MCE broadcasting or it has been disabled. + */ + if (no_way_out && tolerant < 3) + mce_panic("Fatal machine check on current CPU", final, msg); + + /* + * If the error seems to be unrecoverable, something should be + * done. Try to kill as little as possible. If we can kill just + * one task, do that. If the user has set the tolerance very + * high, don't try to do anything at all. + */ + + if (kill_it && tolerant < 3) + force_sig(SIGBUS, current); + + /* notify userspace ASAP */ + set_thread_flag(TIF_MCE_NOTIFY); + + if (worst > 0) + mce_report_event(regs); + mce_wrmsrl(MSR_IA32_MCG_STATUS, 0); +out: + atomic_dec(&mce_entry); + sync_core(); +} +EXPORT_SYMBOL_GPL(do_machine_check); + +/* dummy to break dependency. actual code is in mm/memory-failure.c */ +void __attribute__((weak)) memory_failure(unsigned long pfn, int vector) +{ + printk(KERN_ERR "Action optional memory failure at %lx ignored\n", pfn); +} + +/* + * Called after mce notification in process context. This code + * is allowed to sleep. Call the high level VM handler to process + * any corrupted pages. + * Assume that the work queue code only calls this one at a time + * per CPU. + * Note we don't disable preemption, so this code might run on the wrong + * CPU. In this case the event is picked up by the scheduled work queue. + * This is merely a fast path to expedite processing in some common + * cases. + */ +void mce_notify_process(void) +{ + unsigned long pfn; + mce_notify_irq(); + while (mce_ring_get(&pfn)) + memory_failure(pfn, MCE_VECTOR); +} + +static void mce_process_work(struct work_struct *dummy) +{ + mce_notify_process(); +} + +#ifdef CONFIG_X86_MCE_INTEL +/*** + * mce_log_therm_throt_event - Logs the thermal throttling event to mcelog + * @cpu: The CPU on which the event occurred. + * @status: Event status information + * + * This function should be called by the thermal interrupt after the + * event has been processed and the decision was made to log the event + * further. + * + * The status parameter will be saved to the 'status' field of 'struct mce' + * and historically has been the register value of the + * MSR_IA32_THERMAL_STATUS (Intel) msr. + */ +void mce_log_therm_throt_event(__u64 status) +{ + struct mce m; + + mce_setup(&m); + m.bank = MCE_THERMAL_BANK; + m.status = status; + mce_log(&m); +} +#endif /* CONFIG_X86_MCE_INTEL */ + +/* + * Periodic polling timer for "silent" machine check errors. If the + * poller finds an MCE, poll 2x faster. When the poller finds no more + * errors, poll 2x slower (up to check_interval seconds). + */ +static int check_interval = 5 * 60; /* 5 minutes */ + +static DEFINE_PER_CPU(int, next_interval); /* in jiffies */ +static DEFINE_PER_CPU(struct timer_list, mce_timer); + +static void mcheck_timer(unsigned long data) +{ + struct timer_list *t = &per_cpu(mce_timer, data); + int *n; + + WARN_ON(smp_processor_id() != data); + + if (mce_available(¤t_cpu_data)) { + machine_check_poll(MCP_TIMESTAMP, + &__get_cpu_var(mce_poll_banks)); + } + + /* + * Alert userspace if needed. If we logged an MCE, reduce the + * polling interval, otherwise increase the polling interval. + */ + n = &__get_cpu_var(next_interval); + if (mce_notify_irq()) + *n = max(*n/2, HZ/100); + else + *n = min(*n*2, (int)round_jiffies_relative(check_interval*HZ)); + + t->expires = jiffies + *n; + add_timer_on(t, smp_processor_id()); +} + +static void mce_do_trigger(struct work_struct *work) +{ + call_usermodehelper(mce_helper, mce_helper_argv, NULL, UMH_NO_WAIT); +} + +static DECLARE_WORK(mce_trigger_work, mce_do_trigger); + +/* + * 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_irq(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, &mce_need_notify)) { + wake_up_interruptible(&mce_wait); + + /* + * There is no risk of missing notifications because + * work_pending is always cleared before the function is + * executed. + */ + if (mce_helper[0] && !work_pending(&mce_trigger_work)) + schedule_work(&mce_trigger_work); + + if (__ratelimit(&ratelimit)) + printk(KERN_INFO "Machine check events logged\n"); + + return 1; + } + return 0; +} +EXPORT_SYMBOL_GPL(mce_notify_irq); + +/* + * Initialize Machine Checks for a CPU. + */ +static int mce_cap_init(void) +{ + unsigned b; + u64 cap; + + rdmsrl(MSR_IA32_MCG_CAP, cap); + + b = cap & MCG_BANKCNT_MASK; + printk(KERN_INFO "mce: CPU supports %d MCE banks\n", b); + + 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 & MCG_EXT_P) && MCG_EXT_CNT(cap) >= 9) + rip_msr = MSR_IA32_MCG_EIP; + + if (cap & MCG_SER_P) + mce_ser = 1; + + return 0; +} + +static void mce_init(void) +{ + mce_banks_t all_banks; + u64 cap; + int i; + + /* + * Log the machine checks left over from the previous reset. + */ + bitmap_fill(all_banks, MAX_NR_BANKS); + machine_check_poll(MCP_UC|(!mce_bootlog ? MCP_DONTLOG : 0), &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 (skip_bank_init(i)) + continue; + wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]); + wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0); + } +} + +/* Add per CPU specific workarounds here */ +static int mce_cpu_quirks(struct cpuinfo_x86 *c) +{ + if (c->x86_vendor == X86_VENDOR_UNKNOWN) { + pr_info("MCE: unknown CPU type - not enabling MCE support.\n"); + return -EOPNOTSUPP; + } + + /* This should be disabled by the BIOS, but isn't always */ + if (c->x86_vendor == X86_VENDOR_AMD) { + if (c->x86 == 15 && banks > 4) { + /* + * disable GART TBL walk error reporting, which + * trips off incorrectly with the IOMMU & 3ware + * & Cerberus: + */ + 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: + */ + mce_bootlog = 0; + } + /* + * Various K7s with broken bank 0 around. Always disable + * by default. + */ + if (c->x86 == 6 && banks > 0) + bank[0] = 0; + } + + if (c->x86_vendor == X86_VENDOR_INTEL) { + /* + * SDM documents that on family 6 bank 0 should not be written + * because it aliases to another special BIOS controlled + * register. + * But it's not aliased anymore on model 0x1a+ + * Don't ignore bank 0 completely because there could be a + * valid event later, merely don't write CTL0. + */ + + if (c->x86 == 6 && c->x86_model < 0x1A) + __set_bit(0, &dont_init_banks); + + /* + * All newer Intel systems support MCE broadcasting. Enable + * synchronization with a one second timeout. + */ + if ((c->x86 > 6 || (c->x86 == 6 && c->x86_model >= 0xe)) && + monarch_timeout < 0) + monarch_timeout = USEC_PER_SEC; + + /* + * There are also broken BIOSes on some Pentium M and + * earlier systems: + */ + if (c->x86 == 6 && c->x86_model <= 13 && mce_bootlog < 0) + mce_bootlog = 0; + } + if (monarch_timeout < 0) + monarch_timeout = 0; + if (mce_bootlog != 0) + mce_panic_timeout = 30; + + return 0; +} + +static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c) +{ + if (c->x86 != 5) + return; + switch (c->x86_vendor) { + case X86_VENDOR_INTEL: + intel_p5_mcheck_init(c); + break; + case X86_VENDOR_CENTAUR: + winchip_mcheck_init(c); + break; + } +} + +static void mce_cpu_features(struct cpuinfo_x86 *c) +{ + switch (c->x86_vendor) { + case X86_VENDOR_INTEL: + mce_intel_feature_init(c); + break; + case X86_VENDOR_AMD: + mce_amd_feature_init(c); + break; + default: + break; + } +} + +static void mce_init_timer(void) +{ + struct timer_list *t = &__get_cpu_var(mce_timer); + int *n = &__get_cpu_var(next_interval); + + if (mce_ignore_ce) + return; + + *n = check_interval * HZ; + if (!*n) + return; + setup_timer(t, mcheck_timer, smp_processor_id()); + t->expires = round_jiffies(jiffies + *n); + add_timer_on(t, smp_processor_id()); +} + +/* + * Called for each booted CPU to set up machine checks. + * Must be called with preempt off: + */ +void __cpuinit mcheck_init(struct cpuinfo_x86 *c) +{ + if (mce_disabled) + return; + + mce_ancient_init(c); + + if (!mce_available(c)) + return; + + if (mce_cap_init() < 0 || mce_cpu_quirks(c) < 0) { + mce_disabled = 1; + return; + } + + machine_check_vector = do_machine_check; + + mce_init(); + mce_cpu_features(c); + mce_init_timer(); + INIT_WORK(&__get_cpu_var(mce_work), mce_process_work); +} + +/* + * Character device to read and clear the MCE log. + */ + +static DEFINE_SPINLOCK(mce_state_lock); +static int open_count; /* #times opened */ +static int open_exclu; /* already open exclusive? */ + +static int mce_open(struct inode *inode, struct file *file) +{ + spin_lock(&mce_state_lock); + + if (open_exclu || (open_count && (file->f_flags & O_EXCL))) { + spin_unlock(&mce_state_lock); + + return -EBUSY; + } + + if (file->f_flags & O_EXCL) + open_exclu = 1; + open_count++; + + spin_unlock(&mce_state_lock); + + return nonseekable_open(inode, file); +} + +static int mce_release(struct inode *inode, struct file *file) +{ + spin_lock(&mce_state_lock); + + open_count--; + open_exclu = 0; + + spin_unlock(&mce_state_lock); + + return 0; +} + +static void collect_tscs(void *data) +{ + unsigned long *cpu_tsc = (unsigned long *)data; + + rdtscll(cpu_tsc[smp_processor_id()]); +} + +static DEFINE_MUTEX(mce_read_mutex); + +static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize, + loff_t *off) +{ + char __user *buf = ubuf; + unsigned long *cpu_tsc; + unsigned prev, next; + int i, err; + + cpu_tsc = kmalloc(nr_cpu_ids * sizeof(long), GFP_KERNEL); + if (!cpu_tsc) + return -ENOMEM; + + mutex_lock(&mce_read_mutex); + next = rcu_dereference(mcelog.next); + + /* Only supports full reads right now */ + if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce)) { + mutex_unlock(&mce_read_mutex); + kfree(cpu_tsc); + + return -EINVAL; + } + + err = 0; + 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(); + } + smp_rmb(); + err |= copy_to_user(buf, mcelog.entry + i, + sizeof(struct mce)); + buf += sizeof(struct mce); +timeout: + ; + } + + memset(mcelog.entry + prev, 0, + (next - prev) * sizeof(struct mce)); + prev = next; + next = cmpxchg(&mcelog.next, prev, 0); + } while (next != prev); + + synchronize_sched(); + + /* + * Collect entries that were still getting written before the + * synchronize. + */ + on_each_cpu(collect_tscs, cpu_tsc, 1); + + for (i = next; i < MCE_LOG_LEN; i++) { + if (mcelog.entry[i].finished && + mcelog.entry[i].tsc < cpu_tsc[mcelog.entry[i].cpu]) { + err |= copy_to_user(buf, mcelog.entry+i, + sizeof(struct mce)); + smp_rmb(); + buf += sizeof(struct mce); + memset(&mcelog.entry[i], 0, sizeof(struct mce)); + } + } + mutex_unlock(&mce_read_mutex); + kfree(cpu_tsc); + + return err ? -EFAULT : buf - ubuf; +} + +static unsigned int mce_poll(struct file *file, poll_table *wait) +{ + poll_wait(file, &mce_wait, wait); + if (rcu_dereference(mcelog.next)) + return POLLIN | POLLRDNORM; + return 0; +} + +static long mce_ioctl(struct file *f, unsigned int cmd, unsigned long arg) +{ + int __user *p = (int __user *)arg; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + switch (cmd) { + case MCE_GET_RECORD_LEN: + return put_user(sizeof(struct mce), p); + case MCE_GET_LOG_LEN: + return put_user(MCE_LOG_LEN, p); + case MCE_GETCLEAR_FLAGS: { + unsigned flags; + + do { + flags = mcelog.flags; + } while (cmpxchg(&mcelog.flags, flags, 0) != flags); + + return put_user(flags, p); + } + default: + return -ENOTTY; + } +} + +/* Modified in mce-inject.c, so not static or const */ +struct file_operations mce_chrdev_ops = { + .open = mce_open, + .release = mce_release, + .read = mce_read, + .poll = mce_poll, + .unlocked_ioctl = mce_ioctl, +}; +EXPORT_SYMBOL_GPL(mce_chrdev_ops); + +static struct miscdevice mce_log_device = { + MISC_MCELOG_MINOR, + "mcelog", + &mce_chrdev_ops, +}; + +/* + * mce=off Disables machine check + * mce=no_cmci Disables CMCI + * mce=dont_log_ce Clears corrected events silently, no log created for CEs. + * mce=ignore_ce Disables polling and CMCI, corrected events are not cleared. + * mce=TOLERANCELEVEL[,monarchtimeout] (number, see above) + * monarchtimeout is how long to wait for other CPUs on machine + * check, or 0 to not wait + * mce=bootlog Log MCEs from before booting. Disabled by default on AMD. + * mce=nobootlog Don't log MCEs from before booting. + */ +static int __init mcheck_enable(char *str) +{ + if (*str == 0) + enable_p5_mce(); + if (*str == '=') + str++; + if (!strcmp(str, "off")) + mce_disabled = 1; + else if (!strcmp(str, "no_cmci")) + mce_cmci_disabled = 1; + else if (!strcmp(str, "dont_log_ce")) + mce_dont_log_ce = 1; + else if (!strcmp(str, "ignore_ce")) + mce_ignore_ce = 1; + else if (!strcmp(str, "bootlog") || !strcmp(str, "nobootlog")) + mce_bootlog = (str[0] == 'b'); + else if (isdigit(str[0])) { + get_option(&str, &tolerant); + if (*str == ',') { + ++str; + get_option(&str, &monarch_timeout); + } + } else { + printk(KERN_INFO "mce argument %s ignored. Please use /sys\n", + str); + return 0; + } + return 1; +} +__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++) { + if (!skip_bank_init(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 re-added later using + * CPU hotplug: + */ +static int mce_resume(struct sys_device *dev) +{ + mce_init(); + mce_cpu_features(¤t_cpu_data); + + return 0; +} + +static void mce_cpu_restart(void *data) +{ + del_timer_sync(&__get_cpu_var(mce_timer)); + if (!mce_available(¤t_cpu_data)) + return; + mce_init(); + mce_init_timer(); +} + +/* Reinit MCEs after user configuration changes */ +static void mce_restart(void) +{ + on_each_cpu(mce_cpu_restart, NULL, 1); +} + +/* Toggle features for corrected errors */ +static void mce_disable_ce(void *all) +{ + if (!mce_available(¤t_cpu_data)) + return; + if (all) + del_timer_sync(&__get_cpu_var(mce_timer)); + cmci_clear(); +} + +static void mce_enable_ce(void *all) +{ + if (!mce_available(¤t_cpu_data)) + return; + cmci_reenable(); + cmci_recheck(); + if (all) + mce_init_timer(); +} + +static struct sysdev_class mce_sysclass = { + .suspend = mce_suspend, + .shutdown = mce_shutdown, + .resume = mce_resume, + .name = "machinecheck", +}; + +DEFINE_PER_CPU(struct sys_device, mce_dev); + +__cpuinitdata +void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu); + +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 size) +{ + u64 new; + + if (strict_strtoull(buf, 0, &new) < 0) + return -EINVAL; + + bank[attr - bank_attrs] = new; + mce_restart(); + + return size; +} + +static ssize_t +show_trigger(struct sys_device *s, struct sysdev_attribute *attr, char *buf) +{ + strcpy(buf, mce_helper); + strcat(buf, "\n"); + return strlen(mce_helper) + 1; +} + +static ssize_t set_trigger(struct sys_device *s, struct sysdev_attribute *attr, + const char *buf, size_t siz) +{ + char *p; + + strncpy(mce_helper, buf, sizeof(mce_helper)); + mce_helper[sizeof(mce_helper)-1] = 0; + p = strchr(mce_helper, '\n'); + + if (p) + *p = 0; + + return strlen(mce_helper) + !!p; +} + +static ssize_t set_ignore_ce(struct sys_device *s, + struct sysdev_attribute *attr, + const char *buf, size_t size) +{ + u64 new; + + if (strict_strtoull(buf, 0, &new) < 0) + return -EINVAL; + + if (mce_ignore_ce ^ !!new) { + if (new) { + /* disable ce features */ + on_each_cpu(mce_disable_ce, (void *)1, 1); + mce_ignore_ce = 1; + } else { + /* enable ce features */ + mce_ignore_ce = 0; + on_each_cpu(mce_enable_ce, (void *)1, 1); + } + } + return size; +} + +static ssize_t set_cmci_disabled(struct sys_device *s, + struct sysdev_attribute *attr, + const char *buf, size_t size) +{ + u64 new; + + if (strict_strtoull(buf, 0, &new) < 0) + return -EINVAL; + + if (mce_cmci_disabled ^ !!new) { + if (new) { + /* disable cmci */ + on_each_cpu(mce_disable_ce, NULL, 1); + mce_cmci_disabled = 1; + } else { + /* enable cmci */ + mce_cmci_disabled = 0; + on_each_cpu(mce_enable_ce, NULL, 1); + } + } + return size; +} + +static ssize_t store_int_with_restart(struct sys_device *s, + struct sysdev_attribute *attr, + const char *buf, size_t size) +{ + ssize_t ret = sysdev_store_int(s, attr, buf, size); + mce_restart(); + return ret; +} + +static SYSDEV_ATTR(trigger, 0644, show_trigger, set_trigger); +static SYSDEV_INT_ATTR(tolerant, 0644, tolerant); +static SYSDEV_INT_ATTR(monarch_timeout, 0644, monarch_timeout); +static SYSDEV_INT_ATTR(dont_log_ce, 0644, mce_dont_log_ce); + +static struct sysdev_ext_attribute attr_check_interval = { + _SYSDEV_ATTR(check_interval, 0644, sysdev_show_int, + store_int_with_restart), + &check_interval +}; + +static struct sysdev_ext_attribute attr_ignore_ce = { + _SYSDEV_ATTR(ignore_ce, 0644, sysdev_show_int, set_ignore_ce), + &mce_ignore_ce +}; + +static struct sysdev_ext_attribute attr_cmci_disabled = { + _SYSDEV_ATTR(cmci_disabled, 0644, sysdev_show_int, set_cmci_disabled), + &mce_cmci_disabled +}; + +static struct sysdev_attribute *mce_attrs[] = { + &attr_tolerant.attr, + &attr_check_interval.attr, + &attr_trigger, + &attr_monarch_timeout.attr, + &attr_dont_log_ce.attr, + &attr_ignore_ce.attr, + &attr_cmci_disabled.attr, + NULL +}; + +static cpumask_var_t mce_dev_initialized; + +/* Per cpu sysdev init. All of the cpus still share the same ctrl bank: */ +static __cpuinit int mce_create_device(unsigned int cpu) +{ + int err; + int i, j; + + if (!mce_available(&boot_cpu_data)) + return -EIO; + + memset(&per_cpu(mce_dev, cpu).kobj, 0, sizeof(struct kobject)); + per_cpu(mce_dev, cpu).id = cpu; + per_cpu(mce_dev, cpu).cls = &mce_sysclass; + + err = sysdev_register(&per_cpu(mce_dev, cpu)); + if (err) + return err; + + for (i = 0; mce_attrs[i]; i++) { + err = sysdev_create_file(&per_cpu(mce_dev, cpu), mce_attrs[i]); + if (err) + goto error; + } + for (j = 0; j < banks; j++) { + err = sysdev_create_file(&per_cpu(mce_dev, cpu), + &bank_attrs[j]); + if (err) + goto error2; + } + cpumask_set_cpu(cpu, mce_dev_initialized); + + return 0; +error2: + while (--j >= 0) + sysdev_remove_file(&per_cpu(mce_dev, cpu), &bank_attrs[j]); +error: + while (--i >= 0) + sysdev_remove_file(&per_cpu(mce_dev, cpu), mce_attrs[i]); + + sysdev_unregister(&per_cpu(mce_dev, cpu)); + + return err; +} + +static __cpuinit void mce_remove_device(unsigned int cpu) +{ + int i; + + if (!cpumask_test_cpu(cpu, mce_dev_initialized)) + return; + + for (i = 0; mce_attrs[i]; i++) + sysdev_remove_file(&per_cpu(mce_dev, cpu), mce_attrs[i]); + + for (i = 0; i < banks; i++) + sysdev_remove_file(&per_cpu(mce_dev, cpu), &bank_attrs[i]); + + sysdev_unregister(&per_cpu(mce_dev, cpu)); + cpumask_clear_cpu(cpu, mce_dev_initialized); +} + +/* Make sure there are no machine checks on offlined CPUs. */ +static void mce_disable_cpu(void *h) +{ + unsigned long action = *(unsigned long *)h; + int i; + + if (!mce_available(¤t_cpu_data)) + return; + if (!(action & CPU_TASKS_FROZEN)) + cmci_clear(); + for (i = 0; i < banks; i++) { + if (!skip_bank_init(i)) + wrmsrl(MSR_IA32_MC0_CTL + i*4, 0); + } +} + +static void mce_reenable_cpu(void *h) +{ + unsigned long action = *(unsigned long *)h; + int i; + + if (!mce_available(¤t_cpu_data)) + return; + + if (!(action & CPU_TASKS_FROZEN)) + cmci_reenable(); + for (i = 0; i < banks; i++) { + if (!skip_bank_init(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: + case CPU_ONLINE_FROZEN: + mce_create_device(cpu); + if (threshold_cpu_callback) + threshold_cpu_callback(action, cpu); + break; + case CPU_DEAD: + case CPU_DEAD_FROZEN: + if (threshold_cpu_callback) + 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(jiffies + + __get_cpu_var(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; +} + +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; + int i = 0; + + if (!mce_available(&boot_cpu_data)) + return -EIO; + + zalloc_cpumask_var(&mce_dev_initialized, GFP_KERNEL); + + err = mce_init_banks(); + if (err) + return err; + + err = sysdev_class_register(&mce_sysclass); + if (err) + return err; + + for_each_online_cpu(i) { + err = mce_create_device(i); + if (err) + return err; + } + + register_hotcpu_notifier(&mce_cpu_notifier); + misc_register(&mce_log_device); + + return err; +} + +device_initcall(mce_init_device); + +#else /* CONFIG_X86_OLD_MCE: */ + +int nr_mce_banks; +EXPORT_SYMBOL_GPL(nr_mce_banks); /* non-fatal.o */ + +/* This has to be run for each processor */ +void mcheck_init(struct cpuinfo_x86 *c) +{ + if (mce_disabled) + return; + + switch (c->x86_vendor) { + case X86_VENDOR_AMD: + amd_mcheck_init(c); + break; + + case X86_VENDOR_INTEL: + if (c->x86 == 5) + intel_p5_mcheck_init(c); + if (c->x86 == 6) + intel_p6_mcheck_init(c); + if (c->x86 == 15) + intel_p4_mcheck_init(c); + break; + + case X86_VENDOR_CENTAUR: + if (c->x86 == 5) + winchip_mcheck_init(c); + break; + + default: + break; + } + printk(KERN_INFO "mce: CPU supports %d MCE banks\n", nr_mce_banks); +} + +static int __init mcheck_enable(char *str) +{ + mce_p5_enabled = 1; + return 1; +} +__setup("mce", mcheck_enable); + +#endif /* CONFIG_X86_OLD_MCE */ + +/* + * Old style boot options parsing. Only for compatibility. + */ +static int __init mcheck_disable(char *str) +{ + mce_disabled = 1; + return 1; +} +__setup("nomce", mcheck_disable); diff --git a/arch/x86/kernel/cpu/mcheck/mce.h b/arch/x86/kernel/cpu/mcheck/mce.h deleted file mode 100644 index ae9f628838f..00000000000 --- a/arch/x86/kernel/cpu/mcheck/mce.h +++ /dev/null @@ -1,14 +0,0 @@ -#include <linux/init.h> -#include <asm/mce.h> - -void amd_mcheck_init(struct cpuinfo_x86 *c); -void intel_p4_mcheck_init(struct cpuinfo_x86 *c); -void intel_p5_mcheck_init(struct cpuinfo_x86 *c); -void intel_p6_mcheck_init(struct cpuinfo_x86 *c); -void winchip_mcheck_init(struct cpuinfo_x86 *c); - -/* Call the installed machine check handler for this CPU setup. */ -extern void (*machine_check_vector)(struct pt_regs *, long error_code); - -extern int nr_mce_banks; - diff --git a/arch/x86/kernel/cpu/mcheck/mce_32.c b/arch/x86/kernel/cpu/mcheck/mce_32.c deleted file mode 100644 index 3552119b091..00000000000 --- a/arch/x86/kernel/cpu/mcheck/mce_32.c +++ /dev/null @@ -1,76 +0,0 @@ -/* - * mce.c - x86 Machine Check Exception Reporting - * (c) 2002 Alan Cox <alan@lxorguk.ukuu.org.uk>, Dave Jones <davej@redhat.com> - */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/module.h> -#include <linux/smp.h> -#include <linux/thread_info.h> - -#include <asm/processor.h> -#include <asm/system.h> -#include <asm/mce.h> - -#include "mce.h" - -int mce_disabled; -int nr_mce_banks; - -EXPORT_SYMBOL_GPL(nr_mce_banks); /* non-fatal.o */ - -/* Handle unconfigured int18 (should never happen) */ -static void unexpected_machine_check(struct pt_regs *regs, long error_code) -{ - printk(KERN_ERR "CPU#%d: Unexpected int18 (Machine Check).\n", smp_processor_id()); -} - -/* Call the installed machine check handler for this CPU setup. */ -void (*machine_check_vector)(struct pt_regs *, long error_code) = unexpected_machine_check; - -/* This has to be run for each processor */ -void mcheck_init(struct cpuinfo_x86 *c) -{ - if (mce_disabled == 1) - return; - - switch (c->x86_vendor) { - case X86_VENDOR_AMD: - amd_mcheck_init(c); - break; - - case X86_VENDOR_INTEL: - if (c->x86 == 5) - intel_p5_mcheck_init(c); - if (c->x86 == 6) - intel_p6_mcheck_init(c); - if (c->x86 == 15) - intel_p4_mcheck_init(c); - break; - - case X86_VENDOR_CENTAUR: - if (c->x86 == 5) - winchip_mcheck_init(c); - break; - - default: - break; - } -} - -static int __init mcheck_disable(char *str) -{ - mce_disabled = 1; - return 1; -} - -static int __init mcheck_enable(char *str) -{ - mce_disabled = -1; - return 1; -} - -__setup("nomce", mcheck_disable); -__setup("mce", mcheck_enable); diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c deleted file mode 100644 index 09dd1d414fc..00000000000 --- a/arch/x86/kernel/cpu/mcheck/mce_64.c +++ /dev/null @@ -1,1187 +0,0 @@ -/* - * Machine check handler. - * 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> -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/sched.h> -#include <linux/smp_lock.h> -#include <linux/string.h> -#include <linux/rcupdate.h> -#include <linux/kallsyms.h> -#include <linux/sysdev.h> -#include <linux/miscdevice.h> -#include <linux/fs.h> -#include <linux/capability.h> -#include <linux/cpu.h> -#include <linux/percpu.h> -#include <linux/poll.h> -#include <linux/thread_info.h> -#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> -#include <asm/uaccess.h> -#include <asm/smp.h> -#include <asm/idle.h> - -#define MISC_MCELOG_MINOR 227 - -atomic_t mce_entry; - -static int mce_dont_init; - -/* - * Tolerant levels: - * 0: always panic on uncorrected errors, log corrected errors - * 1: panic or SIGBUS on uncorrected errors, log corrected errors - * 2: SIGBUS or log uncorrected errors (if possible), log corrected errors - * 3: never panic or SIGBUS, log all errors (for testing only) - */ -static int tolerant = 1; -static int banks; -static u64 *bank; -static unsigned long notify_user; -static int rip_msr; -static int mce_bootlog = -1; -static atomic_t mce_events; - -static char trigger[128]; -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 - * separate MCEs from kernel messages to avoid bogus bug reports. - */ - -static struct mce_log mcelog = { - MCE_LOG_SIGNATURE, - MCE_LOG_LEN, -}; - -void mce_log(struct mce *mce) -{ - unsigned next, entry; - atomic_inc(&mce_events); - mce->finished = 0; - wmb(); - for (;;) { - entry = rcu_dereference(mcelog.next); - for (;;) { - /* When the buffer fills up discard new entries. Assume - that the earlier errors are the more interesting. */ - if (entry >= MCE_LOG_LEN) { - set_bit(MCE_OVERFLOW, (unsigned long *)&mcelog.flags); - return; - } - /* Old left over entry. Skip. */ - if (mcelog.entry[entry].finished) { - entry++; - continue; - } - break; - } - smp_rmb(); - next = entry + 1; - if (cmpxchg(&mcelog.next, entry, next) == entry) - break; - } - memcpy(mcelog.entry + entry, mce, sizeof(struct mce)); - wmb(); - mcelog.entry[entry].finished = 1; - wmb(); - - set_bit(0, ¬ify_user); -} - -static void print_mce(struct mce *m) -{ - printk(KERN_EMERG "\n" - KERN_EMERG "HARDWARE ERROR\n" - KERN_EMERG - "CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n", - m->cpu, m->mcgstatus, m->bank, m->status); - if (m->ip) { - printk(KERN_EMERG "RIP%s %02x:<%016Lx> ", - !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "", - m->cs, m->ip); - if (m->cs == __KERNEL_CS) - print_symbol("{%s}", m->ip); - printk("\n"); - } - printk(KERN_EMERG "TSC %llx ", m->tsc); - if (m->addr) - printk("ADDR %llx ", m->addr); - if (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 " - "and contact your hardware vendor\n"); -} - -static void mce_panic(char *msg, struct mce *backup, unsigned long start) -{ - int i; - - oops_begin(); - for (i = 0; i < MCE_LOG_LEN; i++) { - unsigned long tsc = mcelog.entry[i].tsc; - - if (time_before(tsc, start)) - continue; - print_mce(&mcelog.entry[i]); - if (backup && mcelog.entry[i].tsc == backup->tsc) - backup = NULL; - } - if (backup) - print_mce(backup); - panic(msg); -} - -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); -} - -static inline void mce_get_rip(struct mce *m, struct pt_regs *regs) -{ - if (regs && (m->mcgstatus & MCG_STATUS_RIPV)) { - m->ip = regs->ip; - m->cs = regs->cs; - } else { - m->ip = 0; - m->cs = 0; - } - if (rip_msr) { - /* Assume the RIP in the MSR is exact. Is this true? */ - m->mcgstatus |= MCG_STATUS_EIPV; - rdmsrl(rip_msr, m->ip); - m->cs = 0; - } -} - -/* - * 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. - */ - if (!(flags & MCP_DONTLOG)) { - 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) -{ - struct mce m, panicm; - u64 mcestart = 0; - int i; - int panicm_found = 0; - /* - * If no_way_out gets set, there is no safe way to recover from this - * MCE. If tolerant is cranked up, we'll try anyway. - */ - int no_way_out = 0; - /* - * If kill_it gets set, there might be a way to recover from this - * error. - */ - int kill_it = 0; - DECLARE_BITMAP(toclear, MAX_NR_BANKS); - - atomic_inc(&mce_entry); - - if (notify_die(DIE_NMI, "machine check", regs, error_code, - 18, SIGKILL) == NOTIFY_STOP) - goto out2; - if (!banks) - goto out2; - - 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)) - no_way_out = 1; - - rdtscll(mcestart); - barrier(); - - for (i = 0; i < banks; i++) { - __clear_bit(i, toclear); - if (!bank[i]) - continue; - - m.misc = 0; - m.addr = 0; - m.bank = i; - - 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); - /* - * If this error was uncorrectable and there was - * an overflow, we're in trouble. If no overflow, - * we might get away with just killing a task. - */ - if (m.status & MCI_STATUS_UC) { - if (tolerant < 1 || m.status & MCI_STATUS_OVER) - no_way_out = 1; - kill_it = 1; - } - } else { - /* - * Machine check event was not enabled. Clear, but - * ignore. - */ - 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); - - mce_get_rip(&m, regs); - mce_log(&m); - - /* Did this bank cause the exception? */ - /* Assume that the bank with uncorrectable errors did it, - and that there is only a single one. */ - if ((m.status & MCI_STATUS_UC) && (m.status & MCI_STATUS_EN)) { - panicm = m; - panicm_found = 1; - } - } - - /* If we didn't find an uncorrectable error, pick - the last one (shouldn't happen, just being safe). */ - if (!panicm_found) - panicm = m; - - /* - * If we have decided that we just CAN'T continue, and the user - * has not set tolerant to an insane level, give up and die. - */ - if (no_way_out && tolerant < 3) - mce_panic("Machine check", &panicm, mcestart); - - /* - * If the error seems to be unrecoverable, something should be - * done. Try to kill as little as possible. If we can kill just - * one task, do that. If the user has set the tolerance very - * high, don't try to do anything at all. - */ - if (kill_it && tolerant < 3) { - int user_space = 0; - - /* - * If the EIPV bit is set, it means the saved IP is the - * instruction which caused the MCE. - */ - if (m.mcgstatus & MCG_STATUS_EIPV) - user_space = panicm.ip && (panicm.cs & 3); - - /* - * If we know that the error was in user space, send a - * SIGBUS. Otherwise, panic if tolerance is low. - * - * force_sig() takes an awful lot of locks and has a slight - * risk of deadlocking. - */ - if (user_space) { - force_sig(SIGBUS, current); - } else if (panic_on_oops || tolerant < 2) { - mce_panic("Uncorrected machine check", - &panicm, mcestart); - } - } - - /* notify userspace ASAP */ - set_thread_flag(TIF_MCE_NOTIFY); - - /* the last thing we do is clear state */ - 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); -} - -#ifdef CONFIG_X86_MCE_INTEL -/*** - * mce_log_therm_throt_event - Logs the thermal throttling event to mcelog - * @cpu: The CPU on which the event occurred. - * @status: Event status information - * - * This function should be called by the thermal interrupt after the - * event has been processed and the decision was made to log the event - * further. - * - * The status parameter will be saved to the 'status' field of 'struct mce' - * and historically has been the register value of the - * MSR_IA32_THERMAL_STATUS (Intel) msr. - */ -void mce_log_therm_throt_event(__u64 status) -{ - struct mce m; - - mce_setup(&m); - m.bank = MCE_THERMAL_BANK; - m.status = status; - mce_log(&m); -} -#endif /* CONFIG_X86_MCE_INTEL */ - -/* - * Periodic polling timer for "silent" machine check errors. If the - * poller finds an MCE, poll 2x faster. When the poller finds no more - * errors, poll 2x slower (up to check_interval seconds). - */ - -static int check_interval = 5 * 60; /* 5 minutes */ -static DEFINE_PER_CPU(int, next_interval); /* in jiffies */ -static void mcheck_timer(unsigned long); -static DEFINE_PER_CPU(struct timer_list, mce_timer); - -static void mcheck_timer(unsigned long data) -{ - struct timer_list *t = &per_cpu(mce_timer, data); - int *n; - - WARN_ON(smp_processor_id() != data); - - if (mce_available(¤t_cpu_data)) - machine_check_poll(MCP_TIMESTAMP, - &__get_cpu_var(mce_poll_banks)); - - /* - * Alert userspace if needed. If we logged an MCE, reduce the - * polling interval, otherwise increase the polling interval. - */ - n = &__get_cpu_var(next_interval); - if (mce_notify_user()) { - *n = max(*n/2, HZ/100); - } else { - *n = min(*n*2, (int)round_jiffies_relative(check_interval*HZ)); - } - - t->expires = jiffies + *n; - 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); - -/* - * 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, ¬ify_user)) { - wake_up_interruptible(&mce_wait); - - /* - * 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; - } - return 0; -} - -/* see if the idle task needs to notify userspace */ -static int -mce_idle_callback(struct notifier_block *nfb, unsigned long action, void *junk) -{ - /* IDLE_END should be safe - interrupts are back on */ - if (action == IDLE_END && test_thread_flag(TIF_MCE_NOTIFY)) - mce_notify_user(); - - return NOTIFY_OK; -} - -static struct notifier_block mce_idle_notifier = { - .notifier_call = mce_idle_callback, -}; - -static __init int periodic_mcheck_init(void) -{ - idle_notifier_register(&mce_idle_notifier); - return 0; -} -__initcall(periodic_mcheck_init); - -/* - * Initialize Machine Checks for a CPU. - */ -static int mce_cap_init(void) -{ - u64 cap; - unsigned b; - - rdmsrl(MSR_IA32_MCG_CAP, cap); - 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; - - 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|(!mce_bootlog ? MCP_DONTLOG : 0), &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++) { - 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 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 && banks > 4) - /* disable GART TBL walk error reporting, which trips off - incorrectly with the IOMMU & 3ware & Cerberus. */ - 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. */ - mce_bootlog = 0; - } - -} - -static void mce_cpu_features(struct cpuinfo_x86 *c) -{ - switch (c->x86_vendor) { - case X86_VENDOR_INTEL: - mce_intel_feature_init(c); - break; - case X86_VENDOR_AMD: - mce_amd_feature_init(c); - break; - default: - break; - } -} - -static void mce_init_timer(void) -{ - struct timer_list *t = &__get_cpu_var(mce_timer); - int *n = &__get_cpu_var(next_interval); - - *n = check_interval * HZ; - if (!*n) - return; - setup_timer(t, mcheck_timer, smp_processor_id()); - t->expires = round_jiffies(jiffies + *n); - 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) -{ - if (!mce_available(c)) - return; - - if (mce_cap_init() < 0) { - mce_dont_init = 1; - return; - } - mce_cpu_quirks(c); - - mce_init(NULL); - mce_cpu_features(c); - mce_init_timer(); -} - -/* - * Character device to read and clear the MCE log. - */ - -static DEFINE_SPINLOCK(mce_state_lock); -static int open_count; /* #times opened */ -static int open_exclu; /* already open exclusive? */ - -static int mce_open(struct inode *inode, struct file *file) -{ - lock_kernel(); - spin_lock(&mce_state_lock); - - if (open_exclu || (open_count && (file->f_flags & O_EXCL))) { - spin_unlock(&mce_state_lock); - unlock_kernel(); - return -EBUSY; - } - - if (file->f_flags & O_EXCL) - open_exclu = 1; - open_count++; - - spin_unlock(&mce_state_lock); - unlock_kernel(); - - return nonseekable_open(inode, file); -} - -static int mce_release(struct inode *inode, struct file *file) -{ - spin_lock(&mce_state_lock); - - open_count--; - open_exclu = 0; - - spin_unlock(&mce_state_lock); - - return 0; -} - -static void collect_tscs(void *data) -{ - unsigned long *cpu_tsc = (unsigned long *)data; - - rdtscll(cpu_tsc[smp_processor_id()]); -} - -static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize, - loff_t *off) -{ - unsigned long *cpu_tsc; - static DEFINE_MUTEX(mce_read_mutex); - unsigned prev, next; - char __user *buf = ubuf; - int i, err; - - cpu_tsc = kmalloc(nr_cpu_ids * sizeof(long), GFP_KERNEL); - if (!cpu_tsc) - return -ENOMEM; - - mutex_lock(&mce_read_mutex); - next = rcu_dereference(mcelog.next); - - /* Only supports full reads right now */ - if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce)) { - mutex_unlock(&mce_read_mutex); - kfree(cpu_tsc); - return -EINVAL; - } - - err = 0; - 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(); - } - smp_rmb(); - err |= copy_to_user(buf, mcelog.entry + i, - sizeof(struct mce)); - buf += sizeof(struct mce); -timeout: - ; - } - - memset(mcelog.entry + prev, 0, - (next - prev) * sizeof(struct mce)); - prev = next; - next = cmpxchg(&mcelog.next, prev, 0); - } while (next != prev); - - synchronize_sched(); - - /* - * Collect entries that were still getting written before the - * synchronize. - */ - on_each_cpu(collect_tscs, cpu_tsc, 1); - for (i = next; i < MCE_LOG_LEN; i++) { - if (mcelog.entry[i].finished && - mcelog.entry[i].tsc < cpu_tsc[mcelog.entry[i].cpu]) { - err |= copy_to_user(buf, mcelog.entry+i, - sizeof(struct mce)); - smp_rmb(); - buf += sizeof(struct mce); - memset(&mcelog.entry[i], 0, sizeof(struct mce)); - } - } - mutex_unlock(&mce_read_mutex); - kfree(cpu_tsc); - return err ? -EFAULT : buf - ubuf; -} - -static unsigned int mce_poll(struct file *file, poll_table *wait) -{ - poll_wait(file, &mce_wait, wait); - if (rcu_dereference(mcelog.next)) - return POLLIN | POLLRDNORM; - return 0; -} - -static long mce_ioctl(struct file *f, unsigned int cmd, unsigned long arg) -{ - int __user *p = (int __user *)arg; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - switch (cmd) { - case MCE_GET_RECORD_LEN: - return put_user(sizeof(struct mce), p); - case MCE_GET_LOG_LEN: - return put_user(MCE_LOG_LEN, p); - case MCE_GETCLEAR_FLAGS: { - unsigned flags; - - do { - flags = mcelog.flags; - } while (cmpxchg(&mcelog.flags, flags, 0) != flags); - return put_user(flags, p); - } - default: - return -ENOTTY; - } -} - -static const struct file_operations mce_chrdev_ops = { - .open = mce_open, - .release = mce_release, - .read = mce_read, - .poll = mce_poll, - .unlocked_ioctl = mce_ioctl, -}; - -static struct miscdevice mce_log_device = { - MISC_MCELOG_MINOR, - "mcelog", - &mce_chrdev_ops, -}; - -/* - * Old style boot options parsing. Only for compatibility. - */ -static int __init mcheck_disable(char *str) -{ - mce_dont_init = 1; - return 1; -} - -/* 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. */ -static int __init mcheck_enable(char *str) -{ - if (!strcmp(str, "off")) - mce_dont_init = 1; - else if (!strcmp(str, "bootlog") || !strcmp(str,"nobootlog")) - mce_bootlog = str[0] == 'b'; - else if (isdigit(str[0])) - get_option(&str, &tolerant); - else - printk("mce= argument %s ignored. Please use /sys", str); - return 1; -} - -__setup("nomce", mcheck_disable); -__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. */ -static int mce_resume(struct sys_device *dev) -{ - mce_init(NULL); - mce_cpu_features(¤t_cpu_data); - return 0; -} - -static void mce_cpu_restart(void *data) -{ - del_timer_sync(&__get_cpu_var(mce_timer)); - if (mce_available(¤t_cpu_data)) - mce_init(NULL); - mce_init_timer(); -} - -/* Reinit MCEs after user configuration changes */ -static void mce_restart(void) -{ - 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", -}; - -DEFINE_PER_CPU(struct sys_device, device_mce); -void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu) __cpuinitdata; - -/* Why are there no generic functions for this? */ -#define ACCESSOR(name, var, start) \ - static ssize_t show_ ## name(struct sys_device *s, \ - struct sysdev_attribute *attr, \ - char *buf) { \ - return sprintf(buf, "%lx\n", (unsigned long)var); \ - } \ - static ssize_t set_ ## name(struct sys_device *s, \ - struct sysdev_attribute *attr, \ - const char *buf, size_t siz) { \ - char *end; \ - unsigned long new = simple_strtoul(buf, &end, 0); \ - if (end == buf) return -EINVAL; \ - var = new; \ - start; \ - return end-buf; \ - } \ - static SYSDEV_ATTR(name, 0644, show_ ## name, set_ ## name); - -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) -{ - strcpy(buf, trigger); - strcat(buf, "\n"); - return strlen(trigger) + 1; -} - -static ssize_t set_trigger(struct sys_device *s, struct sysdev_attribute *attr, - const char *buf,size_t siz) -{ - char *p; - int len; - strncpy(trigger, buf, sizeof(trigger)); - trigger[sizeof(trigger)-1] = 0; - len = strlen(trigger); - p = strchr(trigger, '\n'); - if (*p) *p = 0; - return len; -} - -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_tolerant.attr, &attr_check_interval, &attr_trigger, - NULL -}; - -static cpumask_var_t mce_device_initialized; - -/* Per cpu sysdev init. All of the cpus still share the same ctl bank */ -static __cpuinit int mce_create_device(unsigned int cpu) -{ - int err; - int i; - - if (!mce_available(&boot_cpu_data)) - return -EIO; - - memset(&per_cpu(device_mce, cpu).kobj, 0, sizeof(struct kobject)); - per_cpu(device_mce,cpu).id = cpu; - per_cpu(device_mce,cpu).cls = &mce_sysclass; - - err = sysdev_register(&per_cpu(device_mce,cpu)); - if (err) - return err; - - for (i = 0; mce_attributes[i]; i++) { - err = sysdev_create_file(&per_cpu(device_mce,cpu), - mce_attributes[i]); - 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; - } - cpumask_set_cpu(cpu, mce_device_initialized); - - return 0; -error2: - while (--i >= 0) { - sysdev_remove_file(&per_cpu(device_mce, cpu), - &bank_attrs[i]); - } -error: - while (--i >= 0) { - sysdev_remove_file(&per_cpu(device_mce,cpu), - mce_attributes[i]); - } - sysdev_unregister(&per_cpu(device_mce,cpu)); - - return err; -} - -static __cpuinit void mce_remove_device(unsigned int cpu) -{ - int i; - - if (!cpumask_test_cpu(cpu, mce_device_initialized)) - return; - - 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)); - cpumask_clear_cpu(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(¤t_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(¤t_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: - case CPU_ONLINE_FROZEN: - mce_create_device(cpu); - if (threshold_cpu_callback) - threshold_cpu_callback(action, cpu); - break; - case CPU_DEAD: - case CPU_DEAD_FROZEN: - if (threshold_cpu_callback) - 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(jiffies + - __get_cpu_var(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; -} - -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; - int i = 0; - - if (!mce_available(&boot_cpu_data)) - return -EIO; - - zalloc_cpumask_var(&mce_device_initialized, GFP_KERNEL); - - err = mce_init_banks(); - if (err) - return err; - - err = sysdev_class_register(&mce_sysclass); - if (err) - return err; - - for_each_online_cpu(i) { - err = mce_create_device(i); - if (err) - return err; - } - - register_hotcpu_notifier(&mce_cpu_notifier); - misc_register(&mce_log_device); - return err; -} - -device_initcall(mce_init_device); diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c index 56dde9c4bc9..ddae21620bd 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c @@ -13,22 +13,22 @@ * * All MC4_MISCi registers are shared between multi-cores */ - -#include <linux/cpu.h> -#include <linux/errno.h> -#include <linux/init.h> #include <linux/interrupt.h> -#include <linux/kobject.h> #include <linux/notifier.h> -#include <linux/sched.h> -#include <linux/smp.h> +#include <linux/kobject.h> +#include <linux/percpu.h> #include <linux/sysdev.h> +#include <linux/errno.h> +#include <linux/sched.h> #include <linux/sysfs.h> +#include <linux/init.h> +#include <linux/cpu.h> +#include <linux/smp.h> + #include <asm/apic.h> +#include <asm/idle.h> #include <asm/mce.h> #include <asm/msr.h> -#include <asm/percpu.h> -#include <asm/idle.h> #define PFX "mce_threshold: " #define VERSION "version 1.1.1" @@ -48,26 +48,26 @@ #define MCG_XBLK_ADDR 0xC0000400 struct threshold_block { - unsigned int block; - unsigned int bank; - unsigned int cpu; - u32 address; - u16 interrupt_enable; - u16 threshold_limit; - struct kobject kobj; - struct list_head miscj; + unsigned int block; + unsigned int bank; + unsigned int cpu; + u32 address; + u16 interrupt_enable; + u16 threshold_limit; + struct kobject kobj; + struct list_head miscj; }; /* defaults used early on boot */ static struct threshold_block threshold_defaults = { - .interrupt_enable = 0, - .threshold_limit = THRESHOLD_MAX, + .interrupt_enable = 0, + .threshold_limit = THRESHOLD_MAX, }; struct threshold_bank { - struct kobject *kobj; - struct threshold_block *blocks; - cpumask_var_t cpus; + struct kobject *kobj; + struct threshold_block *blocks; + cpumask_var_t cpus; }; static DEFINE_PER_CPU(struct threshold_bank *, threshold_banks[NR_BANKS]); @@ -86,9 +86,9 @@ static void amd_threshold_interrupt(void); */ struct thresh_restart { - struct threshold_block *b; - int reset; - u16 old_limit; + struct threshold_block *b; + int reset; + u16 old_limit; }; /* must be called with correct cpu affinity */ @@ -110,6 +110,7 @@ static void threshold_restart_bank(void *_tr) } else if (tr->old_limit) { /* change limit w/o reset */ int new_count = (mci_misc_hi & THRESHOLD_MAX) + (tr->old_limit - tr->b->threshold_limit); + mci_misc_hi = (mci_misc_hi & ~MASK_ERR_COUNT_HI) | (new_count & THRESHOLD_MAX); } @@ -125,11 +126,11 @@ static void threshold_restart_bank(void *_tr) /* cpu init entry point, called from mce.c with preempt off */ void mce_amd_feature_init(struct cpuinfo_x86 *c) { - unsigned int bank, block; unsigned int cpu = smp_processor_id(); - u8 lvt_off; u32 low = 0, high = 0, address = 0; + unsigned int bank, block; struct thresh_restart tr; + u8 lvt_off; for (bank = 0; bank < NR_BANKS; ++bank) { for (block = 0; block < NR_BLOCKS; ++block) { @@ -140,8 +141,7 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c) if (!address) break; address += MCG_XBLK_ADDR; - } - else + } else ++address; if (rdmsr_safe(address, &low, &high)) @@ -193,9 +193,9 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c) */ static void amd_threshold_interrupt(void) { + u32 low = 0, high = 0, address = 0; unsigned int bank, block; struct mce m; - u32 low = 0, high = 0, address = 0; mce_setup(&m); @@ -204,16 +204,16 @@ static void amd_threshold_interrupt(void) if (!(per_cpu(bank_map, m.cpu) & (1 << bank))) continue; for (block = 0; block < NR_BLOCKS; ++block) { - if (block == 0) + if (block == 0) { address = MSR_IA32_MC0_MISC + bank * 4; - else if (block == 1) { + } else if (block == 1) { address = (low & MASK_BLKPTR_LO) >> 21; if (!address) break; address += MCG_XBLK_ADDR; - } - else + } else { ++address; + } if (rdmsr_safe(address, &low, &high)) break; @@ -229,8 +229,10 @@ static void amd_threshold_interrupt(void) (high & MASK_LOCKED_HI)) continue; - /* Log the machine check that caused the threshold - event. */ + /* + * Log the machine check that caused the threshold + * event. + */ machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_poll_banks)); @@ -254,48 +256,52 @@ static void amd_threshold_interrupt(void) struct threshold_attr { struct attribute attr; - ssize_t(*show) (struct threshold_block *, char *); - ssize_t(*store) (struct threshold_block *, const char *, size_t count); + ssize_t (*show) (struct threshold_block *, char *); + ssize_t (*store) (struct threshold_block *, const char *, size_t count); }; -#define SHOW_FIELDS(name) \ -static ssize_t show_ ## name(struct threshold_block * b, char *buf) \ -{ \ - return sprintf(buf, "%lx\n", (unsigned long) b->name); \ +#define SHOW_FIELDS(name) \ +static ssize_t show_ ## name(struct threshold_block *b, char *buf) \ +{ \ + return sprintf(buf, "%lx\n", (unsigned long) b->name); \ } SHOW_FIELDS(interrupt_enable) SHOW_FIELDS(threshold_limit) -static ssize_t store_interrupt_enable(struct threshold_block *b, - const char *buf, size_t count) +static ssize_t +store_interrupt_enable(struct threshold_block *b, const char *buf, size_t size) { - char *end; struct thresh_restart tr; - unsigned long new = simple_strtoul(buf, &end, 0); - if (end == buf) + unsigned long new; + + if (strict_strtoul(buf, 0, &new) < 0) return -EINVAL; + b->interrupt_enable = !!new; - tr.b = b; - tr.reset = 0; - tr.old_limit = 0; + tr.b = b; + tr.reset = 0; + tr.old_limit = 0; + smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); - return end - buf; + return size; } -static ssize_t store_threshold_limit(struct threshold_block *b, - const char *buf, size_t count) +static ssize_t +store_threshold_limit(struct threshold_block *b, const char *buf, size_t size) { - char *end; struct thresh_restart tr; - unsigned long new = simple_strtoul(buf, &end, 0); - if (end == buf) + unsigned long new; + + if (strict_strtoul(buf, 0, &new) < 0) return -EINVAL; + if (new > THRESHOLD_MAX) new = THRESHOLD_MAX; if (new < 1) new = 1; + tr.old_limit = b->threshold_limit; b->threshold_limit = new; tr.b = b; @@ -303,12 +309,12 @@ static ssize_t store_threshold_limit(struct threshold_block *b, smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); - return end - buf; + return size; } struct threshold_block_cross_cpu { - struct threshold_block *tb; - long retval; + struct threshold_block *tb; + long retval; }; static void local_error_count_handler(void *_tbcc) @@ -338,16 +344,13 @@ static ssize_t store_error_count(struct threshold_block *b, return 1; } -#define THRESHOLD_ATTR(_name,_mode,_show,_store) { \ - .attr = {.name = __stringify(_name), .mode = _mode }, \ - .show = _show, \ - .store = _store, \ +#define RW_ATTR(val) \ +static struct threshold_attr val = { \ + .attr = {.name = __stringify(val), .mode = 0644 }, \ + .show = show_## val, \ + .store = store_## val, \ }; -#define RW_ATTR(name) \ -static struct threshold_attr name = \ - THRESHOLD_ATTR(name, 0644, show_## name, store_## name) - RW_ATTR(interrupt_enable); RW_ATTR(threshold_limit); RW_ATTR(error_count); @@ -359,15 +362,17 @@ static struct attribute *default_attrs[] = { NULL }; -#define to_block(k) container_of(k, struct threshold_block, kobj) -#define to_attr(a) container_of(a, struct threshold_attr, attr) +#define to_block(k) container_of(k, struct threshold_block, kobj) +#define to_attr(a) container_of(a, struct threshold_attr, attr) static ssize_t show(struct kobject *kobj, struct attribute *attr, char *buf) { struct threshold_block *b = to_block(kobj); struct threshold_attr *a = to_attr(attr); ssize_t ret; + ret = a->show ? a->show(b, buf) : -EIO; + return ret; } @@ -377,18 +382,20 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr, struct threshold_block *b = to_block(kobj); struct threshold_attr *a = to_attr(attr); ssize_t ret; + ret = a->store ? a->store(b, buf, count) : -EIO; + return ret; } static struct sysfs_ops threshold_ops = { - .show = show, - .store = store, + .show = show, + .store = store, }; static struct kobj_type threshold_ktype = { - .sysfs_ops = &threshold_ops, - .default_attrs = default_attrs, + .sysfs_ops = &threshold_ops, + .default_attrs = default_attrs, }; static __cpuinit int allocate_threshold_blocks(unsigned int cpu, @@ -396,9 +403,9 @@ static __cpuinit int allocate_threshold_blocks(unsigned int cpu, unsigned int block, u32 address) { - int err; - u32 low, high; struct threshold_block *b = NULL; + u32 low, high; + int err; if ((bank >= NR_BANKS) || (block >= NR_BLOCKS)) return 0; @@ -421,20 +428,21 @@ static __cpuinit int allocate_threshold_blocks(unsigned int cpu, if (!b) return -ENOMEM; - b->block = block; - b->bank = bank; - b->cpu = cpu; - b->address = address; - b->interrupt_enable = 0; - b->threshold_limit = THRESHOLD_MAX; + b->block = block; + b->bank = bank; + b->cpu = cpu; + b->address = address; + b->interrupt_enable = 0; + b->threshold_limit = THRESHOLD_MAX; INIT_LIST_HEAD(&b->miscj); - if (per_cpu(threshold_banks, cpu)[bank]->blocks) + if (per_cpu(threshold_banks, cpu)[bank]->blocks) { list_add(&b->miscj, &per_cpu(threshold_banks, cpu)[bank]->blocks->miscj); - else + } else { per_cpu(threshold_banks, cpu)[bank]->blocks = b; + } err = kobject_init_and_add(&b->kobj, &threshold_ktype, per_cpu(threshold_banks, cpu)[bank]->kobj, @@ -447,8 +455,9 @@ recurse: if (!address) return 0; address += MCG_XBLK_ADDR; - } else + } else { ++address; + } err = allocate_threshold_blocks(cpu, bank, ++block, address); if (err) @@ -500,13 +509,14 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) if (!b) goto out; - err = sysfs_create_link(&per_cpu(device_mce, cpu).kobj, + err = sysfs_create_link(&per_cpu(mce_dev, cpu).kobj, b->kobj, name); if (err) goto out; cpumask_copy(b->cpus, cpu_core_mask(cpu)); per_cpu(threshold_banks, cpu)[bank] = b; + goto out; } #endif @@ -522,7 +532,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) goto out; } - b->kobj = kobject_create_and_add(name, &per_cpu(device_mce, cpu).kobj); + b->kobj = kobject_create_and_add(name, &per_cpu(mce_dev, cpu).kobj); if (!b->kobj) goto out_free; @@ -542,7 +552,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) if (i == cpu) continue; - err = sysfs_create_link(&per_cpu(device_mce, i).kobj, + err = sysfs_create_link(&per_cpu(mce_dev, i).kobj, b->kobj, name); if (err) goto out; @@ -605,15 +615,13 @@ static void deallocate_threshold_block(unsigned int cpu, static void threshold_remove_bank(unsigned int cpu, int bank) { - int i = 0; struct threshold_bank *b; char name[32]; + int i = 0; b = per_cpu(threshold_banks, cpu)[bank]; - if (!b) return; - if (!b->blocks) goto free_out; @@ -622,8 +630,9 @@ static void threshold_remove_bank(unsigned int cpu, int bank) #ifdef CONFIG_SMP /* sibling symlink */ if (shared_bank[bank] && b->blocks->cpu != cpu) { - sysfs_remove_link(&per_cpu(device_mce, cpu).kobj, name); + sysfs_remove_link(&per_cpu(mce_dev, cpu).kobj, name); per_cpu(threshold_banks, cpu)[bank] = NULL; + return; } #endif @@ -633,7 +642,7 @@ static void threshold_remove_bank(unsigned int cpu, int bank) if (i == cpu) continue; - sysfs_remove_link(&per_cpu(device_mce, i).kobj, name); + sysfs_remove_link(&per_cpu(mce_dev, i).kobj, name); per_cpu(threshold_banks, i)[bank] = NULL; } @@ -659,12 +668,9 @@ static void threshold_remove_device(unsigned int cpu) } /* get notified when a cpu comes on/off */ -static void __cpuinit amd_64_threshold_cpu_callback(unsigned long action, - unsigned int cpu) +static void __cpuinit +amd_64_threshold_cpu_callback(unsigned long action, unsigned int cpu) { - if (cpu >= NR_CPUS) - return; - switch (action) { case CPU_ONLINE: case CPU_ONLINE_FROZEN: @@ -686,11 +692,12 @@ static __init int threshold_init_device(void) /* to hit CPUs online before the notifier is up */ for_each_online_cpu(lcpu) { int err = threshold_create_device(lcpu); + if (err) return err; } threshold_cpu_callback = amd_64_threshold_cpu_callback; + return 0; } - device_initcall(threshold_init_device); diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c index cef3ee30744..e1acec0f7a3 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c +++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c @@ -8,85 +8,10 @@ #include <linux/init.h> #include <linux/interrupt.h> #include <linux/percpu.h> -#include <asm/processor.h> #include <asm/apic.h> +#include <asm/processor.h> #include <asm/msr.h> #include <asm/mce.h> -#include <asm/hw_irq.h> -#include <asm/idle.h> -#include <asm/therm_throt.h> -#include <asm/apic.h> - -asmlinkage void smp_thermal_interrupt(void) -{ - __u64 msr_val; - - ack_APIC_irq(); - - exit_idle(); - irq_enter(); - - rdmsrl(MSR_IA32_THERM_STATUS, msr_val); - if (therm_throt_process(msr_val & 1)) - mce_log_therm_throt_event(msr_val); - - inc_irq_stat(irq_thermal_count); - irq_exit(); -} - -static void intel_init_thermal(struct cpuinfo_x86 *c) -{ - u32 l, h; - int tm2 = 0; - unsigned int cpu = smp_processor_id(); - - if (!cpu_has(c, X86_FEATURE_ACPI)) - return; - - if (!cpu_has(c, X86_FEATURE_ACC)) - return; - - /* first check if TM1 is already enabled by the BIOS, in which - * case there might be some SMM goo which handles it, so we can't even - * put a handler since it might be delivered via SMI already. - */ - rdmsr(MSR_IA32_MISC_ENABLE, l, h); - h = apic_read(APIC_LVTTHMR); - if ((l & MSR_IA32_MISC_ENABLE_TM1) && (h & APIC_DM_SMI)) { - printk(KERN_DEBUG - "CPU%d: Thermal monitoring handled by SMI\n", cpu); - return; - } - - if (cpu_has(c, X86_FEATURE_TM2) && (l & MSR_IA32_MISC_ENABLE_TM2)) - tm2 = 1; - - if (h & APIC_VECTOR_MASK) { - printk(KERN_DEBUG - "CPU%d: Thermal LVT vector (%#x) already " - "installed\n", cpu, (h & APIC_VECTOR_MASK)); - return; - } - - h = THERMAL_APIC_VECTOR; - h |= (APIC_DM_FIXED | APIC_LVT_MASKED); - apic_write(APIC_LVTTHMR, h); - - rdmsr(MSR_IA32_THERM_INTERRUPT, l, h); - wrmsr(MSR_IA32_THERM_INTERRUPT, l | 0x03, h); - - rdmsr(MSR_IA32_MISC_ENABLE, l, h); - wrmsr(MSR_IA32_MISC_ENABLE, l | MSR_IA32_MISC_ENABLE_TM1, h); - - l = apic_read(APIC_LVTTHMR); - apic_write(APIC_LVTTHMR, l & ~APIC_LVT_MASKED); - printk(KERN_INFO "CPU%d: Thermal monitoring enabled (%s)\n", - cpu, tm2 ? "TM2" : "TM1"); - - /* enable thermal throttle processing */ - atomic_set(&therm_throt_en, 1); - return; -} /* * Support for Intel Correct Machine Check Interrupts. This allows @@ -109,6 +34,9 @@ static int cmci_supported(int *banks) { u64 cap; + if (mce_cmci_disabled || mce_ignore_ce) + return 0; + /* * Vendor check is not strictly needed, but the initial * initialization is vendor keyed and this @@ -132,7 +60,7 @@ static int cmci_supported(int *banks) static void intel_threshold_interrupt(void) { machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned)); - mce_notify_user(); + mce_notify_irq(); } static void print_update(char *type, int *hdr, int num) @@ -248,7 +176,7 @@ void cmci_rediscover(int dying) return; cpumask_copy(old, ¤t->cpus_allowed); - for_each_online_cpu (cpu) { + for_each_online_cpu(cpu) { if (cpu == dying) continue; if (set_cpus_allowed_ptr(current, cpumask_of(cpu))) diff --git a/arch/x86/kernel/cpu/mcheck/non-fatal.c b/arch/x86/kernel/cpu/mcheck/non-fatal.c index a74af128efc..f5f2d6f71fb 100644 --- a/arch/x86/kernel/cpu/mcheck/non-fatal.c +++ b/arch/x86/kernel/cpu/mcheck/non-fatal.c @@ -6,25 +6,23 @@ * This file contains routines to check for non-fatal MCEs every 15s * */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/jiffies.h> -#include <linux/workqueue.h> #include <linux/interrupt.h> -#include <linux/smp.h> +#include <linux/workqueue.h> +#include <linux/jiffies.h> +#include <linux/kernel.h> #include <linux/module.h> +#include <linux/types.h> +#include <linux/init.h> +#include <linux/smp.h> #include <asm/processor.h> #include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include "mce.h" +static int firstbank; -static int firstbank; - -#define MCE_RATE 15*HZ /* timer rate is 15s */ +#define MCE_RATE (15*HZ) /* timer rate is 15s */ static void mce_checkregs(void *info) { @@ -34,23 +32,24 @@ static void mce_checkregs(void *info) for (i = firstbank; i < nr_mce_banks; i++) { rdmsr(MSR_IA32_MC0_STATUS+i*4, low, high); - if (high & (1<<31)) { - printk(KERN_INFO "MCE: The hardware reports a non " - "fatal, correctable incident occurred on " - "CPU %d.\n", + if (!(high & (1<<31))) + continue; + + printk(KERN_INFO "MCE: The hardware reports a non fatal, " + "correctable incident occurred on CPU %d.\n", smp_processor_id()); - printk(KERN_INFO "Bank %d: %08x%08x\n", i, high, low); - - /* - * Scrub the error so we don't pick it up in MCE_RATE - * seconds time. - */ - wrmsr(MSR_IA32_MC0_STATUS+i*4, 0UL, 0UL); - - /* Serialize */ - wmb(); - add_taint(TAINT_MACHINE_CHECK); - } + + printk(KERN_INFO "Bank %d: %08x%08x\n", i, high, low); + + /* + * Scrub the error so we don't pick it up in MCE_RATE + * seconds time: + */ + wrmsr(MSR_IA32_MC0_STATUS+i*4, 0UL, 0UL); + + /* Serialize: */ + wmb(); + add_taint(TAINT_MACHINE_CHECK); } } @@ -77,16 +76,17 @@ static int __init init_nonfatal_mce_checker(void) /* Some Athlons misbehave when we frob bank 0 */ if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD && - boot_cpu_data.x86 == 6) - firstbank = 1; + boot_cpu_data.x86 == 6) + firstbank = 1; else - firstbank = 0; + firstbank = 0; /* * Check for non-fatal errors every MCE_RATE s */ schedule_delayed_work(&mce_work, round_jiffies_relative(MCE_RATE)); printk(KERN_INFO "Machine check exception polling timer started.\n"); + return 0; } module_init(init_nonfatal_mce_checker); diff --git a/arch/x86/kernel/cpu/mcheck/p4.c b/arch/x86/kernel/cpu/mcheck/p4.c index f53bdcbaf38..4482aea9aa2 100644 --- a/arch/x86/kernel/cpu/mcheck/p4.c +++ b/arch/x86/kernel/cpu/mcheck/p4.c @@ -1,21 +1,14 @@ /* * P4 specific Machine Check Exception Reporting */ - -#include <linux/init.h> -#include <linux/types.h> #include <linux/kernel.h> -#include <linux/interrupt.h> +#include <linux/types.h> +#include <linux/init.h> #include <linux/smp.h> #include <asm/processor.h> -#include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include <asm/apic.h> - -#include <asm/therm_throt.h> - -#include "mce.h" /* as supported by the P4/Xeon family */ struct intel_mce_extended_msrs { @@ -34,98 +27,8 @@ struct intel_mce_extended_msrs { static int mce_num_extended_msrs; - -#ifdef CONFIG_X86_MCE_P4THERMAL -static void unexpected_thermal_interrupt(struct pt_regs *regs) -{ - printk(KERN_ERR "CPU%d: Unexpected LVT TMR interrupt!\n", - smp_processor_id()); - add_taint(TAINT_MACHINE_CHECK); -} - -/* P4/Xeon Thermal transition interrupt handler */ -static void intel_thermal_interrupt(struct pt_regs *regs) -{ - __u64 msr_val; - - ack_APIC_irq(); - - rdmsrl(MSR_IA32_THERM_STATUS, msr_val); - therm_throt_process(msr_val & 0x1); -} - -/* Thermal interrupt handler for this CPU setup */ -static void (*vendor_thermal_interrupt)(struct pt_regs *regs) = unexpected_thermal_interrupt; - -void smp_thermal_interrupt(struct pt_regs *regs) -{ - irq_enter(); - vendor_thermal_interrupt(regs); - __get_cpu_var(irq_stat).irq_thermal_count++; - irq_exit(); -} - -/* P4/Xeon Thermal regulation detect and init */ -static void intel_init_thermal(struct cpuinfo_x86 *c) -{ - u32 l, h; - unsigned int cpu = smp_processor_id(); - - /* Thermal monitoring */ - if (!cpu_has(c, X86_FEATURE_ACPI)) - return; /* -ENODEV */ - - /* Clock modulation */ - if (!cpu_has(c, X86_FEATURE_ACC)) - return; /* -ENODEV */ - - /* first check if its enabled already, in which case there might - * be some SMM goo which handles it, so we can't even put a handler - * since it might be delivered via SMI already -zwanem. - */ - rdmsr(MSR_IA32_MISC_ENABLE, l, h); - h = apic_read(APIC_LVTTHMR); - if ((l & MSR_IA32_MISC_ENABLE_TM1) && (h & APIC_DM_SMI)) { - printk(KERN_DEBUG "CPU%d: Thermal monitoring handled by SMI\n", - cpu); - return; /* -EBUSY */ - } - - /* check whether a vector already exists, temporarily masked? */ - if (h & APIC_VECTOR_MASK) { - printk(KERN_DEBUG "CPU%d: Thermal LVT vector (%#x) already " - "installed\n", - cpu, (h & APIC_VECTOR_MASK)); - return; /* -EBUSY */ - } - - /* The temperature transition interrupt handler setup */ - h = THERMAL_APIC_VECTOR; /* our delivery vector */ - h |= (APIC_DM_FIXED | APIC_LVT_MASKED); /* we'll mask till we're ready */ - apic_write(APIC_LVTTHMR, h); - - rdmsr(MSR_IA32_THERM_INTERRUPT, l, h); - wrmsr(MSR_IA32_THERM_INTERRUPT, l | 0x03 , h); - - /* ok we're good to go... */ - vendor_thermal_interrupt = intel_thermal_interrupt; - - rdmsr(MSR_IA32_MISC_ENABLE, l, h); - wrmsr(MSR_IA32_MISC_ENABLE, l | MSR_IA32_MISC_ENABLE_TM1, h); - - l = apic_read(APIC_LVTTHMR); - apic_write(APIC_LVTTHMR, l & ~APIC_LVT_MASKED); - printk(KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu); - - /* enable thermal throttle processing */ - atomic_set(&therm_throt_en, 1); - return; -} -#endif /* CONFIG_X86_MCE_P4THERMAL */ - - /* P4/Xeon Extended MCE MSR retrieval, return 0 if unsupported */ -static inline void intel_get_extended_msrs(struct intel_mce_extended_msrs *r) +static void intel_get_extended_msrs(struct intel_mce_extended_msrs *r) { u32 h; @@ -143,9 +46,9 @@ static inline void intel_get_extended_msrs(struct intel_mce_extended_msrs *r) static void intel_machine_check(struct pt_regs *regs, long error_code) { - int recover = 1; u32 alow, ahigh, high, low; u32 mcgstl, mcgsth; + int recover = 1; int i; rdmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); @@ -157,7 +60,9 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) if (mce_num_extended_msrs > 0) { struct intel_mce_extended_msrs dbg; + intel_get_extended_msrs(&dbg); + printk(KERN_DEBUG "CPU %d: EIP: %08x EFLAGS: %08x\n" "\teax: %08x ebx: %08x ecx: %08x edx: %08x\n" "\tesi: %08x edi: %08x ebp: %08x esp: %08x\n", @@ -171,6 +76,7 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) if (high & (1<<31)) { char misc[20]; char addr[24]; + misc[0] = addr[0] = '\0'; if (high & (1<<29)) recover |= 1; @@ -196,6 +102,7 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) panic("Unable to continue"); printk(KERN_EMERG "Attempting to continue.\n"); + /* * Do not clear the MSR_IA32_MCi_STATUS if the error is not * recoverable/continuable.This will allow BIOS to look at the MSRs @@ -217,7 +124,6 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) wrmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); } - void intel_p4_mcheck_init(struct cpuinfo_x86 *c) { u32 l, h; diff --git a/arch/x86/kernel/cpu/mcheck/p5.c b/arch/x86/kernel/cpu/mcheck/p5.c index c9f77ea69ed..5c0e6533d9b 100644 --- a/arch/x86/kernel/cpu/mcheck/p5.c +++ b/arch/x86/kernel/cpu/mcheck/p5.c @@ -2,52 +2,67 @@ * P5 specific Machine Check Exception Reporting * (C) Copyright 2002 Alan Cox <alan@lxorguk.ukuu.org.uk> */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> #include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/init.h> #include <linux/smp.h> #include <asm/processor.h> #include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include "mce.h" +/* By default disabled */ +int mce_p5_enabled __read_mostly; -/* Machine check handler for Pentium class Intel */ +/* Machine check handler for Pentium class Intel CPUs: */ static void pentium_machine_check(struct pt_regs *regs, long error_code) { u32 loaddr, hi, lotype; + rdmsr(MSR_IA32_P5_MC_ADDR, loaddr, hi); rdmsr(MSR_IA32_P5_MC_TYPE, lotype, hi); - printk(KERN_EMERG "CPU#%d: Machine Check Exception: 0x%8X (type 0x%8X).\n", smp_processor_id(), loaddr, lotype); - if (lotype&(1<<5)) - printk(KERN_EMERG "CPU#%d: Possible thermal failure (CPU on fire ?).\n", smp_processor_id()); + + printk(KERN_EMERG + "CPU#%d: Machine Check Exception: 0x%8X (type 0x%8X).\n", + smp_processor_id(), loaddr, lotype); + + if (lotype & (1<<5)) { + printk(KERN_EMERG + "CPU#%d: Possible thermal failure (CPU on fire ?).\n", + smp_processor_id()); + } + add_taint(TAINT_MACHINE_CHECK); } -/* Set up machine check reporting for processors with Intel style MCE */ +/* Set up machine check reporting for processors with Intel style MCE: */ void intel_p5_mcheck_init(struct cpuinfo_x86 *c) { u32 l, h; - /*Check for MCE support */ - if (!cpu_has(c, X86_FEATURE_MCE)) + /* Default P5 to off as its often misconnected: */ + if (!mce_p5_enabled) return; - /* Default P5 to off as its often misconnected */ - if (mce_disabled != -1) + /* Check for MCE support: */ + if (!cpu_has(c, X86_FEATURE_MCE)) return; + machine_check_vector = pentium_machine_check; + /* Make sure the vector pointer is visible before we enable MCEs: */ wmb(); - /* Read registers before enabling */ + /* Read registers before enabling: */ rdmsr(MSR_IA32_P5_MC_ADDR, l, h); rdmsr(MSR_IA32_P5_MC_TYPE, l, h); - printk(KERN_INFO "Intel old style machine check architecture supported.\n"); + printk(KERN_INFO + "Intel old style machine check architecture supported.\n"); - /* Enable MCE */ + /* Enable MCE: */ set_in_cr4(X86_CR4_MCE); - printk(KERN_INFO "Intel old style machine check reporting enabled on CPU#%d.\n", smp_processor_id()); + printk(KERN_INFO + "Intel old style machine check reporting enabled on CPU#%d.\n", + smp_processor_id()); } diff --git a/arch/x86/kernel/cpu/mcheck/p6.c b/arch/x86/kernel/cpu/mcheck/p6.c index 2ac52d7b434..01e4f817818 100644 --- a/arch/x86/kernel/cpu/mcheck/p6.c +++ b/arch/x86/kernel/cpu/mcheck/p6.c @@ -2,25 +2,23 @@ * P6 specific Machine Check Exception Reporting * (C) Copyright 2002 Alan Cox <alan@lxorguk.ukuu.org.uk> */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> #include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/init.h> #include <linux/smp.h> #include <asm/processor.h> #include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include "mce.h" - /* Machine Check Handler For PII/PIII */ static void intel_machine_check(struct pt_regs *regs, long error_code) { - int recover = 1; u32 alow, ahigh, high, low; u32 mcgstl, mcgsth; + int recover = 1; int i; rdmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); @@ -35,12 +33,16 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) if (high & (1<<31)) { char misc[20]; char addr[24]; - misc[0] = addr[0] = '\0'; + + misc[0] = '\0'; + addr[0] = '\0'; + if (high & (1<<29)) recover |= 1; if (high & (1<<25)) recover |= 2; high &= ~(1<<31); + if (high & (1<<27)) { rdmsr(MSR_IA32_MC0_MISC+i*4, alow, ahigh); snprintf(misc, 20, "[%08x%08x]", ahigh, alow); @@ -49,6 +51,7 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) rdmsr(MSR_IA32_MC0_ADDR+i*4, alow, ahigh); snprintf(addr, 24, " at %08x%08x", ahigh, alow); } + printk(KERN_EMERG "CPU %d: Bank %d: %08x%08x%s%s\n", smp_processor_id(), i, high, low, misc, addr); } @@ -63,16 +66,17 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) /* * Do not clear the MSR_IA32_MCi_STATUS if the error is not * recoverable/continuable.This will allow BIOS to look at the MSRs - * for errors if the OS could not log the error. + * for errors if the OS could not log the error: */ for (i = 0; i < nr_mce_banks; i++) { unsigned int msr; + msr = MSR_IA32_MC0_STATUS+i*4; rdmsr(msr, low, high); if (high & (1<<31)) { - /* Clear it */ + /* Clear it: */ wrmsr(msr, 0UL, 0UL); - /* Serialize */ + /* Serialize: */ wmb(); add_taint(TAINT_MACHINE_CHECK); } @@ -81,7 +85,7 @@ static void intel_machine_check(struct pt_regs *regs, long error_code) wrmsr(MSR_IA32_MCG_STATUS, mcgstl, mcgsth); } -/* Set up machine check reporting for processors with Intel style MCE */ +/* Set up machine check reporting for processors with Intel style MCE: */ void intel_p6_mcheck_init(struct cpuinfo_x86 *c) { u32 l, h; @@ -97,6 +101,7 @@ void intel_p6_mcheck_init(struct cpuinfo_x86 *c) /* Ok machine check is available */ machine_check_vector = intel_machine_check; + /* Make sure the vector pointer is visible before we enable MCEs: */ wmb(); printk(KERN_INFO "Intel machine check architecture supported.\n"); diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c index d5ae2243f0b..5957a93e517 100644 --- a/arch/x86/kernel/cpu/mcheck/therm_throt.c +++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c @@ -1,7 +1,7 @@ /* - * * Thermal throttle event support code (such as syslog messaging and rate * limiting) that was factored out from x86_64 (mce_intel.c) and i386 (p4.c). + * * This allows consistent reporting of CPU thermal throttle events. * * Maintains a counter in /sys that keeps track of the number of thermal @@ -13,43 +13,54 @@ * Credits: Adapted from Zwane Mwaikambo's original code in mce_intel.c. * Inspired by Ross Biro's and Al Borchers' counter code. */ - +#include <linux/interrupt.h> +#include <linux/notifier.h> +#include <linux/jiffies.h> +#include <linux/kernel.h> #include <linux/percpu.h> #include <linux/sysdev.h> +#include <linux/types.h> +#include <linux/init.h> +#include <linux/smp.h> #include <linux/cpu.h> -#include <asm/cpu.h> -#include <linux/notifier.h> -#include <linux/jiffies.h> -#include <asm/therm_throt.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/apic.h> +#include <asm/idle.h> +#include <asm/mce.h> +#include <asm/msr.h> /* How long to wait between reporting thermal events */ -#define CHECK_INTERVAL (300 * HZ) +#define CHECK_INTERVAL (300 * HZ) static DEFINE_PER_CPU(__u64, next_check) = INITIAL_JIFFIES; static DEFINE_PER_CPU(unsigned long, thermal_throttle_count); -atomic_t therm_throt_en = ATOMIC_INIT(0); +static DEFINE_PER_CPU(bool, thermal_throttle_active); + +static atomic_t therm_throt_en = ATOMIC_INIT(0); #ifdef CONFIG_SYSFS -#define define_therm_throt_sysdev_one_ro(_name) \ - static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL) - -#define define_therm_throt_sysdev_show_func(name) \ -static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \ - struct sysdev_attribute *attr, \ - char *buf) \ -{ \ - unsigned int cpu = dev->id; \ - ssize_t ret; \ - \ - preempt_disable(); /* CPU hotplug */ \ - if (cpu_online(cpu)) \ - ret = sprintf(buf, "%lu\n", \ - per_cpu(thermal_throttle_##name, cpu)); \ - else \ - ret = 0; \ - preempt_enable(); \ - \ - return ret; \ +#define define_therm_throt_sysdev_one_ro(_name) \ + static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL) + +#define define_therm_throt_sysdev_show_func(name) \ +static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \ + struct sysdev_attribute *attr, \ + char *buf) \ +{ \ + unsigned int cpu = dev->id; \ + ssize_t ret; \ + \ + preempt_disable(); /* CPU hotplug */ \ + if (cpu_online(cpu)) \ + ret = sprintf(buf, "%lu\n", \ + per_cpu(thermal_throttle_##name, cpu)); \ + else \ + ret = 0; \ + preempt_enable(); \ + \ + return ret; \ } define_therm_throt_sysdev_show_func(count); @@ -61,8 +72,8 @@ static struct attribute *thermal_throttle_attrs[] = { }; static struct attribute_group thermal_throttle_attr_group = { - .attrs = thermal_throttle_attrs, - .name = "thermal_throttle" + .attrs = thermal_throttle_attrs, + .name = "thermal_throttle" }; #endif /* CONFIG_SYSFS */ @@ -82,38 +93,45 @@ static struct attribute_group thermal_throttle_attr_group = { * 1 : Event should be logged further, and a message has been * printed to the syslog. */ -int therm_throt_process(int curr) +static int therm_throt_process(int curr) { unsigned int cpu = smp_processor_id(); __u64 tmp_jiffs = get_jiffies_64(); + bool was_throttled = __get_cpu_var(thermal_throttle_active); + bool is_throttled = __get_cpu_var(thermal_throttle_active) = curr; - if (curr) + if (is_throttled) __get_cpu_var(thermal_throttle_count)++; - if (time_before64(tmp_jiffs, __get_cpu_var(next_check))) + if (!(was_throttled ^ is_throttled) && + time_before64(tmp_jiffs, __get_cpu_var(next_check))) return 0; __get_cpu_var(next_check) = tmp_jiffs + CHECK_INTERVAL; /* if we just entered the thermal event */ - if (curr) { + if (is_throttled) { printk(KERN_CRIT "CPU%d: Temperature above threshold, " - "cpu clock throttled (total events = %lu)\n", cpu, - __get_cpu_var(thermal_throttle_count)); + "cpu clock throttled (total events = %lu)\n", + cpu, __get_cpu_var(thermal_throttle_count)); add_taint(TAINT_MACHINE_CHECK); - } else { - printk(KERN_CRIT "CPU%d: Temperature/speed normal\n", cpu); + return 1; + } + if (was_throttled) { + printk(KERN_INFO "CPU%d: Temperature/speed normal\n", cpu); + return 1; } - return 1; + return 0; } #ifdef CONFIG_SYSFS -/* Add/Remove thermal_throttle interface for CPU device */ +/* Add/Remove thermal_throttle interface for CPU device: */ static __cpuinit int thermal_throttle_add_dev(struct sys_device *sys_dev) { - return sysfs_create_group(&sys_dev->kobj, &thermal_throttle_attr_group); + return sysfs_create_group(&sys_dev->kobj, + &thermal_throttle_attr_group); } static __cpuinit void thermal_throttle_remove_dev(struct sys_device *sys_dev) @@ -121,19 +139,21 @@ static __cpuinit void thermal_throttle_remove_dev(struct sys_device *sys_dev) sysfs_remove_group(&sys_dev->kobj, &thermal_throttle_attr_group); } -/* Mutex protecting device creation against CPU hotplug */ +/* Mutex protecting device creation against CPU hotplug: */ static DEFINE_MUTEX(therm_cpu_lock); /* Get notified when a cpu comes on/off. Be hotplug friendly. */ -static __cpuinit int thermal_throttle_cpu_callback(struct notifier_block *nfb, - unsigned long action, - void *hcpu) +static __cpuinit int +thermal_throttle_cpu_callback(struct notifier_block *nfb, + unsigned long action, + void *hcpu) { unsigned int cpu = (unsigned long)hcpu; struct sys_device *sys_dev; int err = 0; sys_dev = get_cpu_sysdev(cpu); + switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: @@ -183,6 +203,94 @@ static __init int thermal_throttle_init_device(void) return 0; } - device_initcall(thermal_throttle_init_device); + #endif /* CONFIG_SYSFS */ + +/* Thermal transition interrupt handler */ +static void intel_thermal_interrupt(void) +{ + __u64 msr_val; + + rdmsrl(MSR_IA32_THERM_STATUS, msr_val); + if (therm_throt_process(msr_val & THERM_STATUS_PROCHOT)) + mce_log_therm_throt_event(msr_val); +} + +static void unexpected_thermal_interrupt(void) +{ + printk(KERN_ERR "CPU%d: Unexpected LVT TMR interrupt!\n", + smp_processor_id()); + add_taint(TAINT_MACHINE_CHECK); +} + +static void (*smp_thermal_vector)(void) = unexpected_thermal_interrupt; + +asmlinkage void smp_thermal_interrupt(struct pt_regs *regs) +{ + exit_idle(); + irq_enter(); + inc_irq_stat(irq_thermal_count); + smp_thermal_vector(); + irq_exit(); + /* Ack only at the end to avoid potential reentry */ + ack_APIC_irq(); +} + +void intel_init_thermal(struct cpuinfo_x86 *c) +{ + unsigned int cpu = smp_processor_id(); + int tm2 = 0; + u32 l, h; + + /* Thermal monitoring depends on ACPI and clock modulation*/ + if (!cpu_has(c, X86_FEATURE_ACPI) || !cpu_has(c, X86_FEATURE_ACC)) + return; + + /* + * First check if its enabled already, in which case there might + * be some SMM goo which handles it, so we can't even put a handler + * since it might be delivered via SMI already: + */ + rdmsr(MSR_IA32_MISC_ENABLE, l, h); + h = apic_read(APIC_LVTTHMR); + if ((l & MSR_IA32_MISC_ENABLE_TM1) && (h & APIC_DM_SMI)) { + printk(KERN_DEBUG + "CPU%d: Thermal monitoring handled by SMI\n", cpu); + return; + } + + if (cpu_has(c, X86_FEATURE_TM2) && (l & MSR_IA32_MISC_ENABLE_TM2)) + tm2 = 1; + + /* Check whether a vector already exists */ + if (h & APIC_VECTOR_MASK) { + printk(KERN_DEBUG + "CPU%d: Thermal LVT vector (%#x) already installed\n", + cpu, (h & APIC_VECTOR_MASK)); + return; + } + + /* We'll mask the thermal vector in the lapic till we're ready: */ + h = THERMAL_APIC_VECTOR | APIC_DM_FIXED | APIC_LVT_MASKED; + apic_write(APIC_LVTTHMR, h); + + rdmsr(MSR_IA32_THERM_INTERRUPT, l, h); + wrmsr(MSR_IA32_THERM_INTERRUPT, + l | (THERM_INT_LOW_ENABLE | THERM_INT_HIGH_ENABLE), h); + + smp_thermal_vector = intel_thermal_interrupt; + + rdmsr(MSR_IA32_MISC_ENABLE, l, h); + wrmsr(MSR_IA32_MISC_ENABLE, l | MSR_IA32_MISC_ENABLE_TM1, h); + + /* Unmask the thermal vector: */ + l = apic_read(APIC_LVTTHMR); + apic_write(APIC_LVTTHMR, l & ~APIC_LVT_MASKED); + + printk(KERN_INFO "CPU%d: Thermal monitoring enabled (%s)\n", + cpu, tm2 ? "TM2" : "TM1"); + + /* enable thermal throttle processing */ + atomic_set(&therm_throt_en, 1); +} diff --git a/arch/x86/kernel/cpu/mcheck/threshold.c b/arch/x86/kernel/cpu/mcheck/threshold.c index 23ee9e730f7..d746df2909c 100644 --- a/arch/x86/kernel/cpu/mcheck/threshold.c +++ b/arch/x86/kernel/cpu/mcheck/threshold.c @@ -17,7 +17,7 @@ static void default_threshold_interrupt(void) void (*mce_threshold_vector)(void) = default_threshold_interrupt; -asmlinkage void mce_threshold_interrupt(void) +asmlinkage void smp_threshold_interrupt(void) { exit_idle(); irq_enter(); diff --git a/arch/x86/kernel/cpu/mcheck/winchip.c b/arch/x86/kernel/cpu/mcheck/winchip.c index 2a043d89811..54060f56597 100644 --- a/arch/x86/kernel/cpu/mcheck/winchip.c +++ b/arch/x86/kernel/cpu/mcheck/winchip.c @@ -2,19 +2,17 @@ * IDT Winchip specific Machine Check Exception Reporting * (C) Copyright 2002 Alan Cox <alan@lxorguk.ukuu.org.uk> */ - -#include <linux/init.h> -#include <linux/types.h> -#include <linux/kernel.h> #include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/init.h> #include <asm/processor.h> #include <asm/system.h> +#include <asm/mce.h> #include <asm/msr.h> -#include "mce.h" - -/* Machine check handler for WinChip C6 */ +/* Machine check handler for WinChip C6: */ static void winchip_machine_check(struct pt_regs *regs, long error_code) { printk(KERN_EMERG "CPU0: Machine Check Exception.\n"); @@ -25,12 +23,18 @@ static void winchip_machine_check(struct pt_regs *regs, long error_code) void winchip_mcheck_init(struct cpuinfo_x86 *c) { u32 lo, hi; + machine_check_vector = winchip_machine_check; + /* Make sure the vector pointer is visible before we enable MCEs: */ wmb(); + rdmsr(MSR_IDT_FCR1, lo, hi); lo |= (1<<2); /* Enable EIERRINT (int 18 MCE) */ lo &= ~(1<<4); /* Enable MCE */ wrmsr(MSR_IDT_FCR1, lo, hi); + set_in_cr4(X86_CR4_MCE); - printk(KERN_INFO "Winchip machine check reporting enabled on CPU#0.\n"); + + printk(KERN_INFO + "Winchip machine check reporting enabled on CPU#0.\n"); } diff --git a/arch/x86/kernel/cpu/mtrr/cleanup.c b/arch/x86/kernel/cpu/mtrr/cleanup.c index ce0fe4b5c04..1d584a18a50 100644 --- a/arch/x86/kernel/cpu/mtrr/cleanup.c +++ b/arch/x86/kernel/cpu/mtrr/cleanup.c @@ -808,7 +808,7 @@ int __init mtrr_cleanup(unsigned address_bits) if (!is_cpu(INTEL) || enable_mtrr_cleanup < 1) return 0; - rdmsr(MTRRdefType_MSR, def, dummy); + rdmsr(MSR_MTRRdefType, def, dummy); def &= 0xff; if (def != MTRR_TYPE_UNCACHABLE) return 0; @@ -1003,7 +1003,7 @@ int __init mtrr_trim_uncached_memory(unsigned long end_pfn) */ if (!is_cpu(INTEL) || disable_mtrr_trim) return 0; - rdmsr(MTRRdefType_MSR, def, dummy); + rdmsr(MSR_MTRRdefType, def, dummy); def &= 0xff; if (def != MTRR_TYPE_UNCACHABLE) return 0; diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index d21d4fb161f..0543f69f0b2 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -20,9 +20,9 @@ struct fixed_range_block { }; static struct fixed_range_block fixed_range_blocks[] = { - { MTRRfix64K_00000_MSR, 1 }, /* one 64k MTRR */ - { MTRRfix16K_80000_MSR, 2 }, /* two 16k MTRRs */ - { MTRRfix4K_C0000_MSR, 8 }, /* eight 4k MTRRs */ + { MSR_MTRRfix64K_00000, 1 }, /* one 64k MTRR */ + { MSR_MTRRfix16K_80000, 2 }, /* two 16k MTRRs */ + { MSR_MTRRfix4K_C0000, 8 }, /* eight 4k MTRRs */ {} }; @@ -194,12 +194,12 @@ get_fixed_ranges(mtrr_type * frs) k8_check_syscfg_dram_mod_en(); - rdmsr(MTRRfix64K_00000_MSR, p[0], p[1]); + rdmsr(MSR_MTRRfix64K_00000, p[0], p[1]); for (i = 0; i < 2; i++) - rdmsr(MTRRfix16K_80000_MSR + i, p[2 + i * 2], p[3 + i * 2]); + rdmsr(MSR_MTRRfix16K_80000 + i, p[2 + i * 2], p[3 + i * 2]); for (i = 0; i < 8; i++) - rdmsr(MTRRfix4K_C0000_MSR + i, p[6 + i * 2], p[7 + i * 2]); + rdmsr(MSR_MTRRfix4K_C0000 + i, p[6 + i * 2], p[7 + i * 2]); } void mtrr_save_fixed_ranges(void *info) @@ -310,7 +310,7 @@ void __init get_mtrr_state(void) vrs = mtrr_state.var_ranges; - rdmsr(MTRRcap_MSR, lo, dummy); + rdmsr(MSR_MTRRcap, lo, dummy); mtrr_state.have_fixed = (lo >> 8) & 1; for (i = 0; i < num_var_ranges; i++) @@ -318,7 +318,7 @@ void __init get_mtrr_state(void) if (mtrr_state.have_fixed) get_fixed_ranges(mtrr_state.fixed_ranges); - rdmsr(MTRRdefType_MSR, lo, dummy); + rdmsr(MSR_MTRRdefType, lo, dummy); mtrr_state.def_type = (lo & 0xff); mtrr_state.enabled = (lo & 0xc00) >> 10; @@ -583,10 +583,10 @@ static void prepare_set(void) __acquires(set_atomicity_lock) __flush_tlb(); /* Save MTRR state */ - rdmsr(MTRRdefType_MSR, deftype_lo, deftype_hi); + rdmsr(MSR_MTRRdefType, deftype_lo, deftype_hi); /* Disable MTRRs, and set the default type to uncached */ - mtrr_wrmsr(MTRRdefType_MSR, deftype_lo & ~0xcff, deftype_hi); + mtrr_wrmsr(MSR_MTRRdefType, deftype_lo & ~0xcff, deftype_hi); } static void post_set(void) __releases(set_atomicity_lock) @@ -595,7 +595,7 @@ static void post_set(void) __releases(set_atomicity_lock) __flush_tlb(); /* Intel (P6) standard MTRRs */ - mtrr_wrmsr(MTRRdefType_MSR, deftype_lo, deftype_hi); + mtrr_wrmsr(MSR_MTRRdefType, deftype_lo, deftype_hi); /* Enable caches */ write_cr0(read_cr0() & 0xbfffffff); @@ -707,7 +707,7 @@ int generic_validate_add_page(unsigned long base, unsigned long size, unsigned i static int generic_have_wrcomb(void) { unsigned long config, dummy; - rdmsr(MTRRcap_MSR, config, dummy); + rdmsr(MSR_MTRRcap, config, dummy); return (config & (1 << 10)); } diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c index 03cda01f57c..8fc248b5aea 100644 --- a/arch/x86/kernel/cpu/mtrr/main.c +++ b/arch/x86/kernel/cpu/mtrr/main.c @@ -104,7 +104,7 @@ static void __init set_num_var_ranges(void) unsigned long config = 0, dummy; if (use_intel()) { - rdmsr(MTRRcap_MSR, config, dummy); + rdmsr(MSR_MTRRcap, config, dummy); } else if (is_cpu(AMD)) config = 2; else if (is_cpu(CYRIX) || is_cpu(CENTAUR)) diff --git a/arch/x86/kernel/cpu/mtrr/mtrr.h b/arch/x86/kernel/cpu/mtrr/mtrr.h index 77f67f7b347..7538b767f20 100644 --- a/arch/x86/kernel/cpu/mtrr/mtrr.h +++ b/arch/x86/kernel/cpu/mtrr/mtrr.h @@ -5,21 +5,6 @@ #include <linux/types.h> #include <linux/stddef.h> -#define MTRRcap_MSR 0x0fe -#define MTRRdefType_MSR 0x2ff - -#define MTRRfix64K_00000_MSR 0x250 -#define MTRRfix16K_80000_MSR 0x258 -#define MTRRfix16K_A0000_MSR 0x259 -#define MTRRfix4K_C0000_MSR 0x268 -#define MTRRfix4K_C8000_MSR 0x269 -#define MTRRfix4K_D0000_MSR 0x26a -#define MTRRfix4K_D8000_MSR 0x26b -#define MTRRfix4K_E0000_MSR 0x26c -#define MTRRfix4K_E8000_MSR 0x26d -#define MTRRfix4K_F0000_MSR 0x26e -#define MTRRfix4K_F8000_MSR 0x26f - #define MTRR_CHANGE_MASK_FIXED 0x01 #define MTRR_CHANGE_MASK_VARIABLE 0x02 #define MTRR_CHANGE_MASK_DEFTYPE 0x04 diff --git a/arch/x86/kernel/cpu/mtrr/state.c b/arch/x86/kernel/cpu/mtrr/state.c index 7f7e2753685..1f5fb1588d1 100644 --- a/arch/x86/kernel/cpu/mtrr/state.c +++ b/arch/x86/kernel/cpu/mtrr/state.c @@ -35,7 +35,7 @@ void set_mtrr_prepare_save(struct set_mtrr_context *ctxt) if (use_intel()) /* Save MTRR state */ - rdmsr(MTRRdefType_MSR, ctxt->deftype_lo, ctxt->deftype_hi); + rdmsr(MSR_MTRRdefType, ctxt->deftype_lo, ctxt->deftype_hi); else /* Cyrix ARRs - everything else were excluded at the top */ ctxt->ccr3 = getCx86(CX86_CCR3); @@ -46,7 +46,7 @@ void set_mtrr_cache_disable(struct set_mtrr_context *ctxt) { if (use_intel()) /* Disable MTRRs, and set the default type to uncached */ - mtrr_wrmsr(MTRRdefType_MSR, ctxt->deftype_lo & 0xf300UL, + mtrr_wrmsr(MSR_MTRRdefType, ctxt->deftype_lo & 0xf300UL, ctxt->deftype_hi); else if (is_cpu(CYRIX)) /* Cyrix ARRs - everything else were excluded at the top */ @@ -64,7 +64,7 @@ void set_mtrr_done(struct set_mtrr_context *ctxt) /* Restore MTRRdefType */ if (use_intel()) /* Intel (P6) standard MTRRs */ - mtrr_wrmsr(MTRRdefType_MSR, ctxt->deftype_lo, ctxt->deftype_hi); + mtrr_wrmsr(MSR_MTRRdefType, ctxt->deftype_lo, ctxt->deftype_hi); else /* Cyrix ARRs - everything else was excluded at the top */ setCx86(CX86_CCR3, ctxt->ccr3); diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c new file mode 100644 index 00000000000..900332b800f --- /dev/null +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -0,0 +1,1964 @@ +/* + * Performance counter x86 architecture code + * + * Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de> + * Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar + * Copyright (C) 2009 Jaswinder Singh Rajput + * Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter + * Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * For licencing details see kernel-base/COPYING + */ + +#include <linux/perf_counter.h> +#include <linux/capability.h> +#include <linux/notifier.h> +#include <linux/hardirq.h> +#include <linux/kprobes.h> +#include <linux/module.h> +#include <linux/kdebug.h> +#include <linux/sched.h> +#include <linux/uaccess.h> +#include <linux/highmem.h> + +#include <asm/apic.h> +#include <asm/stacktrace.h> +#include <asm/nmi.h> + +static u64 perf_counter_mask __read_mostly; + +struct cpu_hw_counters { + struct perf_counter *counters[X86_PMC_IDX_MAX]; + unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + unsigned long active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + unsigned long interrupts; + int enabled; +}; + +/* + * struct x86_pmu - generic x86 pmu + */ +struct x86_pmu { + const char *name; + int version; + int (*handle_irq)(struct pt_regs *); + void (*disable_all)(void); + void (*enable_all)(void); + void (*enable)(struct hw_perf_counter *, int); + void (*disable)(struct hw_perf_counter *, int); + unsigned eventsel; + unsigned perfctr; + u64 (*event_map)(int); + u64 (*raw_event)(u64); + int max_events; + int num_counters; + int num_counters_fixed; + int counter_bits; + u64 counter_mask; + int apic; + u64 max_period; + u64 intel_ctrl; +}; + +static struct x86_pmu x86_pmu __read_mostly; + +static DEFINE_PER_CPU(struct cpu_hw_counters, cpu_hw_counters) = { + .enabled = 1, +}; + +/* + * Not sure about some of these + */ +static const u64 p6_perfmon_event_map[] = +{ + [PERF_COUNT_HW_CPU_CYCLES] = 0x0079, + [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0, + [PERF_COUNT_HW_CACHE_REFERENCES] = 0x0f2e, + [PERF_COUNT_HW_CACHE_MISSES] = 0x012e, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4, + [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5, + [PERF_COUNT_HW_BUS_CYCLES] = 0x0062, +}; + +static u64 p6_pmu_event_map(int event) +{ + return p6_perfmon_event_map[event]; +} + +/* + * Counter setting that is specified not to count anything. + * We use this to effectively disable a counter. + * + * L2_RQSTS with 0 MESI unit mask. + */ +#define P6_NOP_COUNTER 0x0000002EULL + +static u64 p6_pmu_raw_event(u64 event) +{ +#define P6_EVNTSEL_EVENT_MASK 0x000000FFULL +#define P6_EVNTSEL_UNIT_MASK 0x0000FF00ULL +#define P6_EVNTSEL_EDGE_MASK 0x00040000ULL +#define P6_EVNTSEL_INV_MASK 0x00800000ULL +#define P6_EVNTSEL_COUNTER_MASK 0xFF000000ULL + +#define P6_EVNTSEL_MASK \ + (P6_EVNTSEL_EVENT_MASK | \ + P6_EVNTSEL_UNIT_MASK | \ + P6_EVNTSEL_EDGE_MASK | \ + P6_EVNTSEL_INV_MASK | \ + P6_EVNTSEL_COUNTER_MASK) + + return event & P6_EVNTSEL_MASK; +} + + +/* + * Intel PerfMon v3. Used on Core2 and later. + */ +static const u64 intel_perfmon_event_map[] = +{ + [PERF_COUNT_HW_CPU_CYCLES] = 0x003c, + [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0, + [PERF_COUNT_HW_CACHE_REFERENCES] = 0x4f2e, + [PERF_COUNT_HW_CACHE_MISSES] = 0x412e, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4, + [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5, + [PERF_COUNT_HW_BUS_CYCLES] = 0x013c, +}; + +static u64 intel_pmu_event_map(int event) +{ + return intel_perfmon_event_map[event]; +} + +/* + * Generalized hw caching related event table, filled + * in on a per model basis. A value of 0 means + * 'not supported', -1 means 'event makes no sense on + * this CPU', any other value means the raw event + * ID. + */ + +#define C(x) PERF_COUNT_HW_CACHE_##x + +static u64 __read_mostly hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX]; + +static const u64 nehalem_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI */ + [ C(RESULT_MISS) ] = 0x0140, /* L1D_CACHE_LD.I_STATE */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI */ + [ C(RESULT_MISS) ] = 0x0141, /* L1D_CACHE_ST.I_STATE */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x014e, /* L1D_PREFETCH.REQUESTS */ + [ C(RESULT_MISS) ] = 0x024e, /* L1D_PREFETCH.MISS */ + }, + }, + [ C(L1I ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS */ + [ C(RESULT_MISS) ] = 0x0280, /* L1I.MISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = 0x0, + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0324, /* L2_RQSTS.LOADS */ + [ C(RESULT_MISS) ] = 0x0224, /* L2_RQSTS.LD_MISS */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0c24, /* L2_RQSTS.RFOS */ + [ C(RESULT_MISS) ] = 0x0824, /* L2_RQSTS.RFO_MISS */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x4f2e, /* LLC Reference */ + [ C(RESULT_MISS) ] = 0x412e, /* LLC Misses */ + }, + }, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x0108, /* DTLB_LOAD_MISSES.ANY */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x010c, /* MEM_STORE_RETIRED.DTLB_MISS */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = 0x0, + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x01c0, /* INST_RETIRED.ANY_P */ + [ C(RESULT_MISS) ] = 0x20c8, /* ITLB_MISS_RETIRED */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + [ C(BPU ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ALL_BRANCHES */ + [ C(RESULT_MISS) ] = 0x03e8, /* BPU_CLEARS.ANY */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +static const u64 core2_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI */ + [ C(RESULT_MISS) ] = 0x0140, /* L1D_CACHE_LD.I_STATE */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI */ + [ C(RESULT_MISS) ] = 0x0141, /* L1D_CACHE_ST.I_STATE */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x104e, /* L1D_PREFETCH.REQUESTS */ + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(L1I ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0080, /* L1I.READS */ + [ C(RESULT_MISS) ] = 0x0081, /* L1I.MISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI */ + [ C(RESULT_MISS) ] = 0x4129, /* L2_LD.ISTATE */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI */ + [ C(RESULT_MISS) ] = 0x412A, /* L2_ST.ISTATE */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0f40, /* L1D_CACHE_LD.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x0208, /* DTLB_MISSES.MISS_LD */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0f41, /* L1D_CACHE_ST.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x0808, /* DTLB_MISSES.MISS_ST */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P */ + [ C(RESULT_MISS) ] = 0x1282, /* ITLBMISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + [ C(BPU ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY */ + [ C(RESULT_MISS) ] = 0x00c5, /* BP_INST_RETIRED.MISPRED */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +static const u64 atom_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE.LD */ + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE.ST */ + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(L1I ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0380, /* L1I.READS */ + [ C(RESULT_MISS) ] = 0x0280, /* L1I.MISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x4f29, /* L2_LD.MESI */ + [ C(RESULT_MISS) ] = 0x4129, /* L2_LD.ISTATE */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x4f2A, /* L2_ST.MESI */ + [ C(RESULT_MISS) ] = 0x412A, /* L2_ST.ISTATE */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x2140, /* L1D_CACHE_LD.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x0508, /* DTLB_MISSES.MISS_LD */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x2240, /* L1D_CACHE_ST.MESI (alias) */ + [ C(RESULT_MISS) ] = 0x0608, /* DTLB_MISSES.MISS_ST */ + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c0, /* INST_RETIRED.ANY_P */ + [ C(RESULT_MISS) ] = 0x0282, /* ITLB.MISSES */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + [ C(BPU ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c4, /* BR_INST_RETIRED.ANY */ + [ C(RESULT_MISS) ] = 0x00c5, /* BP_INST_RETIRED.MISPRED */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +static u64 intel_pmu_raw_event(u64 event) +{ +#define CORE_EVNTSEL_EVENT_MASK 0x000000FFULL +#define CORE_EVNTSEL_UNIT_MASK 0x0000FF00ULL +#define CORE_EVNTSEL_EDGE_MASK 0x00040000ULL +#define CORE_EVNTSEL_INV_MASK 0x00800000ULL +#define CORE_EVNTSEL_COUNTER_MASK 0xFF000000ULL + +#define CORE_EVNTSEL_MASK \ + (CORE_EVNTSEL_EVENT_MASK | \ + CORE_EVNTSEL_UNIT_MASK | \ + CORE_EVNTSEL_EDGE_MASK | \ + CORE_EVNTSEL_INV_MASK | \ + CORE_EVNTSEL_COUNTER_MASK) + + return event & CORE_EVNTSEL_MASK; +} + +static const u64 amd_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0040, /* Data Cache Accesses */ + [ C(RESULT_MISS) ] = 0x0041, /* Data Cache Misses */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0142, /* Data Cache Refills :system */ + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x0267, /* Data Prefetcher :attempts */ + [ C(RESULT_MISS) ] = 0x0167, /* Data Prefetcher :cancelled */ + }, + }, + [ C(L1I ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction cache fetches */ + [ C(RESULT_MISS) ] = 0x0081, /* Instruction cache misses */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0x014B, /* Prefetch Instructions :Load */ + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x037D, /* Requests to L2 Cache :IC+DC */ + [ C(RESULT_MISS) ] = 0x037E, /* L2 Cache Misses : IC+DC */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x017F, /* L2 Fill/Writeback */ + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0040, /* Data Cache Accesses */ + [ C(RESULT_MISS) ] = 0x0046, /* L1 DTLB and L2 DLTB Miss */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = 0, + [ C(RESULT_MISS) ] = 0, + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction fecthes */ + [ C(RESULT_MISS) ] = 0x0085, /* Instr. fetch ITLB misses */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, + [ C(BPU ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x00c2, /* Retired Branch Instr. */ + [ C(RESULT_MISS) ] = 0x00c3, /* Retired Mispredicted BI */ + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +/* + * AMD Performance Monitor K7 and later. + */ +static const u64 amd_perfmon_event_map[] = +{ + [PERF_COUNT_HW_CPU_CYCLES] = 0x0076, + [PERF_COUNT_HW_INSTRUCTIONS] = 0x00c0, + [PERF_COUNT_HW_CACHE_REFERENCES] = 0x0080, + [PERF_COUNT_HW_CACHE_MISSES] = 0x0081, + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = 0x00c4, + [PERF_COUNT_HW_BRANCH_MISSES] = 0x00c5, +}; + +static u64 amd_pmu_event_map(int event) +{ + return amd_perfmon_event_map[event]; +} + +static u64 amd_pmu_raw_event(u64 event) +{ +#define K7_EVNTSEL_EVENT_MASK 0x7000000FFULL +#define K7_EVNTSEL_UNIT_MASK 0x00000FF00ULL +#define K7_EVNTSEL_EDGE_MASK 0x000040000ULL +#define K7_EVNTSEL_INV_MASK 0x000800000ULL +#define K7_EVNTSEL_COUNTER_MASK 0x0FF000000ULL + +#define K7_EVNTSEL_MASK \ + (K7_EVNTSEL_EVENT_MASK | \ + K7_EVNTSEL_UNIT_MASK | \ + K7_EVNTSEL_EDGE_MASK | \ + K7_EVNTSEL_INV_MASK | \ + K7_EVNTSEL_COUNTER_MASK) + + return event & K7_EVNTSEL_MASK; +} + +/* + * Propagate counter elapsed time into the generic counter. + * Can only be executed on the CPU where the counter is active. + * Returns the delta events processed. + */ +static u64 +x86_perf_counter_update(struct perf_counter *counter, + struct hw_perf_counter *hwc, int idx) +{ + int shift = 64 - x86_pmu.counter_bits; + u64 prev_raw_count, new_raw_count; + s64 delta; + + /* + * Careful: an NMI might modify the previous counter value. + * + * Our tactic to handle this is to first atomically read and + * exchange a new raw count - then add that new-prev delta + * count to the generic counter atomically: + */ +again: + prev_raw_count = atomic64_read(&hwc->prev_count); + rdmsrl(hwc->counter_base + idx, new_raw_count); + + if (atomic64_cmpxchg(&hwc->prev_count, prev_raw_count, + new_raw_count) != prev_raw_count) + goto again; + + /* + * Now we have the new raw value and have updated the prev + * timestamp already. We can now calculate the elapsed delta + * (counter-)time and add that to the generic counter. + * + * Careful, not all hw sign-extends above the physical width + * of the count. + */ + delta = (new_raw_count << shift) - (prev_raw_count << shift); + delta >>= shift; + + atomic64_add(delta, &counter->count); + atomic64_sub(delta, &hwc->period_left); + + return new_raw_count; +} + +static atomic_t active_counters; +static DEFINE_MUTEX(pmc_reserve_mutex); + +static bool reserve_pmc_hardware(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + int i; + + if (nmi_watchdog == NMI_LOCAL_APIC) + disable_lapic_nmi_watchdog(); + + for (i = 0; i < x86_pmu.num_counters; i++) { + if (!reserve_perfctr_nmi(x86_pmu.perfctr + i)) + goto perfctr_fail; + } + + for (i = 0; i < x86_pmu.num_counters; i++) { + if (!reserve_evntsel_nmi(x86_pmu.eventsel + i)) + goto eventsel_fail; + } +#endif + + return true; + +#ifdef CONFIG_X86_LOCAL_APIC +eventsel_fail: + for (i--; i >= 0; i--) + release_evntsel_nmi(x86_pmu.eventsel + i); + + i = x86_pmu.num_counters; + +perfctr_fail: + for (i--; i >= 0; i--) + release_perfctr_nmi(x86_pmu.perfctr + i); + + if (nmi_watchdog == NMI_LOCAL_APIC) + enable_lapic_nmi_watchdog(); + + return false; +#endif +} + +static void release_pmc_hardware(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + int i; + + for (i = 0; i < x86_pmu.num_counters; i++) { + release_perfctr_nmi(x86_pmu.perfctr + i); + release_evntsel_nmi(x86_pmu.eventsel + i); + } + + if (nmi_watchdog == NMI_LOCAL_APIC) + enable_lapic_nmi_watchdog(); +#endif +} + +static void hw_perf_counter_destroy(struct perf_counter *counter) +{ + if (atomic_dec_and_mutex_lock(&active_counters, &pmc_reserve_mutex)) { + release_pmc_hardware(); + mutex_unlock(&pmc_reserve_mutex); + } +} + +static inline int x86_pmu_initialized(void) +{ + return x86_pmu.handle_irq != NULL; +} + +static inline int +set_ext_hw_attr(struct hw_perf_counter *hwc, struct perf_counter_attr *attr) +{ + unsigned int cache_type, cache_op, cache_result; + u64 config, val; + + config = attr->config; + + cache_type = (config >> 0) & 0xff; + if (cache_type >= PERF_COUNT_HW_CACHE_MAX) + return -EINVAL; + + cache_op = (config >> 8) & 0xff; + if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX) + return -EINVAL; + + cache_result = (config >> 16) & 0xff; + if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX) + return -EINVAL; + + val = hw_cache_event_ids[cache_type][cache_op][cache_result]; + + if (val == 0) + return -ENOENT; + + if (val == -1) + return -EINVAL; + + hwc->config |= val; + + return 0; +} + +/* + * Setup the hardware configuration for a given attr_type + */ +static int __hw_perf_counter_init(struct perf_counter *counter) +{ + struct perf_counter_attr *attr = &counter->attr; + struct hw_perf_counter *hwc = &counter->hw; + u64 config; + int err; + + if (!x86_pmu_initialized()) + return -ENODEV; + + err = 0; + if (!atomic_inc_not_zero(&active_counters)) { + mutex_lock(&pmc_reserve_mutex); + if (atomic_read(&active_counters) == 0 && !reserve_pmc_hardware()) + err = -EBUSY; + else + atomic_inc(&active_counters); + mutex_unlock(&pmc_reserve_mutex); + } + if (err) + return err; + + /* + * Generate PMC IRQs: + * (keep 'enabled' bit clear for now) + */ + hwc->config = ARCH_PERFMON_EVENTSEL_INT; + + /* + * Count user and OS events unless requested not to. + */ + if (!attr->exclude_user) + hwc->config |= ARCH_PERFMON_EVENTSEL_USR; + if (!attr->exclude_kernel) + hwc->config |= ARCH_PERFMON_EVENTSEL_OS; + + if (!hwc->sample_period) { + hwc->sample_period = x86_pmu.max_period; + hwc->last_period = hwc->sample_period; + atomic64_set(&hwc->period_left, hwc->sample_period); + } else { + /* + * If we have a PMU initialized but no APIC + * interrupts, we cannot sample hardware + * counters (user-space has to fall back and + * sample via a hrtimer based software counter): + */ + if (!x86_pmu.apic) + return -EOPNOTSUPP; + } + + counter->destroy = hw_perf_counter_destroy; + + /* + * Raw event type provide the config in the event structure + */ + if (attr->type == PERF_TYPE_RAW) { + hwc->config |= x86_pmu.raw_event(attr->config); + return 0; + } + + if (attr->type == PERF_TYPE_HW_CACHE) + return set_ext_hw_attr(hwc, attr); + + if (attr->config >= x86_pmu.max_events) + return -EINVAL; + + /* + * The generic map: + */ + config = x86_pmu.event_map(attr->config); + + if (config == 0) + return -ENOENT; + + if (config == -1LL) + return -EINVAL; + + hwc->config |= config; + + return 0; +} + +static void p6_pmu_disable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + u64 val; + + if (!cpuc->enabled) + return; + + cpuc->enabled = 0; + barrier(); + + /* p6 only has one enable register */ + rdmsrl(MSR_P6_EVNTSEL0, val); + val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_P6_EVNTSEL0, val); +} + +static void intel_pmu_disable_all(void) +{ + wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0); +} + +static void amd_pmu_disable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + int idx; + + if (!cpuc->enabled) + return; + + cpuc->enabled = 0; + /* + * ensure we write the disable before we start disabling the + * counters proper, so that amd_pmu_enable_counter() does the + * right thing. + */ + barrier(); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + u64 val; + + if (!test_bit(idx, cpuc->active_mask)) + continue; + rdmsrl(MSR_K7_EVNTSEL0 + idx, val); + if (!(val & ARCH_PERFMON_EVENTSEL0_ENABLE)) + continue; + val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_K7_EVNTSEL0 + idx, val); + } +} + +void hw_perf_disable(void) +{ + if (!x86_pmu_initialized()) + return; + return x86_pmu.disable_all(); +} + +static void p6_pmu_enable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + unsigned long val; + + if (cpuc->enabled) + return; + + cpuc->enabled = 1; + barrier(); + + /* p6 only has one enable register */ + rdmsrl(MSR_P6_EVNTSEL0, val); + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_P6_EVNTSEL0, val); +} + +static void intel_pmu_enable_all(void) +{ + wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl); +} + +static void amd_pmu_enable_all(void) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + int idx; + + if (cpuc->enabled) + return; + + cpuc->enabled = 1; + barrier(); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + struct perf_counter *counter = cpuc->counters[idx]; + u64 val; + + if (!test_bit(idx, cpuc->active_mask)) + continue; + + val = counter->hw.config; + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + wrmsrl(MSR_K7_EVNTSEL0 + idx, val); + } +} + +void hw_perf_enable(void) +{ + if (!x86_pmu_initialized()) + return; + x86_pmu.enable_all(); +} + +static inline u64 intel_pmu_get_status(void) +{ + u64 status; + + rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status); + + return status; +} + +static inline void intel_pmu_ack_status(u64 ack) +{ + wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, ack); +} + +static inline void x86_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) +{ + (void)checking_wrmsrl(hwc->config_base + idx, + hwc->config | ARCH_PERFMON_EVENTSEL0_ENABLE); +} + +static inline void x86_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) +{ + (void)checking_wrmsrl(hwc->config_base + idx, hwc->config); +} + +static inline void +intel_pmu_disable_fixed(struct hw_perf_counter *hwc, int __idx) +{ + int idx = __idx - X86_PMC_IDX_FIXED; + u64 ctrl_val, mask; + + mask = 0xfULL << (idx * 4); + + rdmsrl(hwc->config_base, ctrl_val); + ctrl_val &= ~mask; + (void)checking_wrmsrl(hwc->config_base, ctrl_val); +} + +static inline void +p6_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + u64 val = P6_NOP_COUNTER; + + if (cpuc->enabled) + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + + (void)checking_wrmsrl(hwc->config_base + idx, val); +} + +static inline void +intel_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) +{ + if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) { + intel_pmu_disable_fixed(hwc, idx); + return; + } + + x86_pmu_disable_counter(hwc, idx); +} + +static inline void +amd_pmu_disable_counter(struct hw_perf_counter *hwc, int idx) +{ + x86_pmu_disable_counter(hwc, idx); +} + +static DEFINE_PER_CPU(u64, prev_left[X86_PMC_IDX_MAX]); + +/* + * Set the next IRQ period, based on the hwc->period_left value. + * To be called with the counter disabled in hw: + */ +static int +x86_perf_counter_set_period(struct perf_counter *counter, + struct hw_perf_counter *hwc, int idx) +{ + s64 left = atomic64_read(&hwc->period_left); + s64 period = hwc->sample_period; + int err, ret = 0; + + /* + * If we are way outside a reasoable range then just skip forward: + */ + if (unlikely(left <= -period)) { + left = period; + atomic64_set(&hwc->period_left, left); + hwc->last_period = period; + ret = 1; + } + + if (unlikely(left <= 0)) { + left += period; + atomic64_set(&hwc->period_left, left); + hwc->last_period = period; + ret = 1; + } + /* + * Quirk: certain CPUs dont like it if just 1 event is left: + */ + if (unlikely(left < 2)) + left = 2; + + if (left > x86_pmu.max_period) + left = x86_pmu.max_period; + + per_cpu(prev_left[idx], smp_processor_id()) = left; + + /* + * The hw counter starts counting from this counter offset, + * mark it to be able to extra future deltas: + */ + atomic64_set(&hwc->prev_count, (u64)-left); + + err = checking_wrmsrl(hwc->counter_base + idx, + (u64)(-left) & x86_pmu.counter_mask); + + perf_counter_update_userpage(counter); + + return ret; +} + +static inline void +intel_pmu_enable_fixed(struct hw_perf_counter *hwc, int __idx) +{ + int idx = __idx - X86_PMC_IDX_FIXED; + u64 ctrl_val, bits, mask; + int err; + + /* + * Enable IRQ generation (0x8), + * and enable ring-3 counting (0x2) and ring-0 counting (0x1) + * if requested: + */ + bits = 0x8ULL; + if (hwc->config & ARCH_PERFMON_EVENTSEL_USR) + bits |= 0x2; + if (hwc->config & ARCH_PERFMON_EVENTSEL_OS) + bits |= 0x1; + bits <<= (idx * 4); + mask = 0xfULL << (idx * 4); + + rdmsrl(hwc->config_base, ctrl_val); + ctrl_val &= ~mask; + ctrl_val |= bits; + err = checking_wrmsrl(hwc->config_base, ctrl_val); +} + +static void p6_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + u64 val; + + val = hwc->config; + if (cpuc->enabled) + val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + + (void)checking_wrmsrl(hwc->config_base + idx, val); +} + + +static void intel_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) +{ + if (unlikely(hwc->config_base == MSR_ARCH_PERFMON_FIXED_CTR_CTRL)) { + intel_pmu_enable_fixed(hwc, idx); + return; + } + + x86_pmu_enable_counter(hwc, idx); +} + +static void amd_pmu_enable_counter(struct hw_perf_counter *hwc, int idx) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + + if (cpuc->enabled) + x86_pmu_enable_counter(hwc, idx); +} + +static int +fixed_mode_idx(struct perf_counter *counter, struct hw_perf_counter *hwc) +{ + unsigned int event; + + if (!x86_pmu.num_counters_fixed) + return -1; + + event = hwc->config & ARCH_PERFMON_EVENT_MASK; + + if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_INSTRUCTIONS))) + return X86_PMC_IDX_FIXED_INSTRUCTIONS; + if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_CPU_CYCLES))) + return X86_PMC_IDX_FIXED_CPU_CYCLES; + if (unlikely(event == x86_pmu.event_map(PERF_COUNT_HW_BUS_CYCLES))) + return X86_PMC_IDX_FIXED_BUS_CYCLES; + + return -1; +} + +/* + * Find a PMC slot for the freshly enabled / scheduled in counter: + */ +static int x86_pmu_enable(struct perf_counter *counter) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + struct hw_perf_counter *hwc = &counter->hw; + int idx; + + idx = fixed_mode_idx(counter, hwc); + if (idx >= 0) { + /* + * Try to get the fixed counter, if that is already taken + * then try to get a generic counter: + */ + if (test_and_set_bit(idx, cpuc->used_mask)) + goto try_generic; + + hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL; + /* + * We set it so that counter_base + idx in wrmsr/rdmsr maps to + * MSR_ARCH_PERFMON_FIXED_CTR0 ... CTR2: + */ + hwc->counter_base = + MSR_ARCH_PERFMON_FIXED_CTR0 - X86_PMC_IDX_FIXED; + hwc->idx = idx; + } else { + idx = hwc->idx; + /* Try to get the previous generic counter again */ + if (test_and_set_bit(idx, cpuc->used_mask)) { +try_generic: + idx = find_first_zero_bit(cpuc->used_mask, + x86_pmu.num_counters); + if (idx == x86_pmu.num_counters) + return -EAGAIN; + + set_bit(idx, cpuc->used_mask); + hwc->idx = idx; + } + hwc->config_base = x86_pmu.eventsel; + hwc->counter_base = x86_pmu.perfctr; + } + + perf_counters_lapic_init(); + + x86_pmu.disable(hwc, idx); + + cpuc->counters[idx] = counter; + set_bit(idx, cpuc->active_mask); + + x86_perf_counter_set_period(counter, hwc, idx); + x86_pmu.enable(hwc, idx); + + perf_counter_update_userpage(counter); + + return 0; +} + +static void x86_pmu_unthrottle(struct perf_counter *counter) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + struct hw_perf_counter *hwc = &counter->hw; + + if (WARN_ON_ONCE(hwc->idx >= X86_PMC_IDX_MAX || + cpuc->counters[hwc->idx] != counter)) + return; + + x86_pmu.enable(hwc, hwc->idx); +} + +void perf_counter_print_debug(void) +{ + u64 ctrl, status, overflow, pmc_ctrl, pmc_count, prev_left, fixed; + struct cpu_hw_counters *cpuc; + unsigned long flags; + int cpu, idx; + + if (!x86_pmu.num_counters) + return; + + local_irq_save(flags); + + cpu = smp_processor_id(); + cpuc = &per_cpu(cpu_hw_counters, cpu); + + if (x86_pmu.version >= 2) { + rdmsrl(MSR_CORE_PERF_GLOBAL_CTRL, ctrl); + rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status); + rdmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, overflow); + rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR_CTRL, fixed); + + pr_info("\n"); + pr_info("CPU#%d: ctrl: %016llx\n", cpu, ctrl); + pr_info("CPU#%d: status: %016llx\n", cpu, status); + pr_info("CPU#%d: overflow: %016llx\n", cpu, overflow); + pr_info("CPU#%d: fixed: %016llx\n", cpu, fixed); + } + pr_info("CPU#%d: used: %016llx\n", cpu, *(u64 *)cpuc->used_mask); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + rdmsrl(x86_pmu.eventsel + idx, pmc_ctrl); + rdmsrl(x86_pmu.perfctr + idx, pmc_count); + + prev_left = per_cpu(prev_left[idx], cpu); + + pr_info("CPU#%d: gen-PMC%d ctrl: %016llx\n", + cpu, idx, pmc_ctrl); + pr_info("CPU#%d: gen-PMC%d count: %016llx\n", + cpu, idx, pmc_count); + pr_info("CPU#%d: gen-PMC%d left: %016llx\n", + cpu, idx, prev_left); + } + for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++) { + rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, pmc_count); + + pr_info("CPU#%d: fixed-PMC%d count: %016llx\n", + cpu, idx, pmc_count); + } + local_irq_restore(flags); +} + +static void x86_pmu_disable(struct perf_counter *counter) +{ + struct cpu_hw_counters *cpuc = &__get_cpu_var(cpu_hw_counters); + struct hw_perf_counter *hwc = &counter->hw; + int idx = hwc->idx; + + /* + * Must be done before we disable, otherwise the nmi handler + * could reenable again: + */ + clear_bit(idx, cpuc->active_mask); + x86_pmu.disable(hwc, idx); + + /* + * Make sure the cleared pointer becomes visible before we + * (potentially) free the counter: + */ + barrier(); + + /* + * Drain the remaining delta count out of a counter + * that we are disabling: + */ + x86_perf_counter_update(counter, hwc, idx); + cpuc->counters[idx] = NULL; + clear_bit(idx, cpuc->used_mask); + + perf_counter_update_userpage(counter); +} + +/* + * Save and restart an expired counter. Called by NMI contexts, + * so it has to be careful about preempting normal counter ops: + */ +static int intel_pmu_save_and_restart(struct perf_counter *counter) +{ + struct hw_perf_counter *hwc = &counter->hw; + int idx = hwc->idx; + int ret; + + x86_perf_counter_update(counter, hwc, idx); + ret = x86_perf_counter_set_period(counter, hwc, idx); + + if (counter->state == PERF_COUNTER_STATE_ACTIVE) + intel_pmu_enable_counter(hwc, idx); + + return ret; +} + +static void intel_pmu_reset(void) +{ + unsigned long flags; + int idx; + + if (!x86_pmu.num_counters) + return; + + local_irq_save(flags); + + printk("clearing PMU state on CPU#%d\n", smp_processor_id()); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + checking_wrmsrl(x86_pmu.eventsel + idx, 0ull); + checking_wrmsrl(x86_pmu.perfctr + idx, 0ull); + } + for (idx = 0; idx < x86_pmu.num_counters_fixed; idx++) { + checking_wrmsrl(MSR_ARCH_PERFMON_FIXED_CTR0 + idx, 0ull); + } + + local_irq_restore(flags); +} + +static int p6_pmu_handle_irq(struct pt_regs *regs) +{ + struct perf_sample_data data; + struct cpu_hw_counters *cpuc; + struct perf_counter *counter; + struct hw_perf_counter *hwc; + int idx, handled = 0; + u64 val; + + data.regs = regs; + data.addr = 0; + + cpuc = &__get_cpu_var(cpu_hw_counters); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + if (!test_bit(idx, cpuc->active_mask)) + continue; + + counter = cpuc->counters[idx]; + hwc = &counter->hw; + + val = x86_perf_counter_update(counter, hwc, idx); + if (val & (1ULL << (x86_pmu.counter_bits - 1))) + continue; + + /* + * counter overflow + */ + handled = 1; + data.period = counter->hw.last_period; + + if (!x86_perf_counter_set_period(counter, hwc, idx)) + continue; + + if (perf_counter_overflow(counter, 1, &data)) + p6_pmu_disable_counter(hwc, idx); + } + + if (handled) + inc_irq_stat(apic_perf_irqs); + + return handled; +} + +/* + * This handler is triggered by the local APIC, so the APIC IRQ handling + * rules apply: + */ +static int intel_pmu_handle_irq(struct pt_regs *regs) +{ + struct perf_sample_data data; + struct cpu_hw_counters *cpuc; + int bit, loops; + u64 ack, status; + + data.regs = regs; + data.addr = 0; + + cpuc = &__get_cpu_var(cpu_hw_counters); + + perf_disable(); + status = intel_pmu_get_status(); + if (!status) { + perf_enable(); + return 0; + } + + loops = 0; +again: + if (++loops > 100) { + WARN_ONCE(1, "perfcounters: irq loop stuck!\n"); + perf_counter_print_debug(); + intel_pmu_reset(); + perf_enable(); + return 1; + } + + inc_irq_stat(apic_perf_irqs); + ack = status; + for_each_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) { + struct perf_counter *counter = cpuc->counters[bit]; + + clear_bit(bit, (unsigned long *) &status); + if (!test_bit(bit, cpuc->active_mask)) + continue; + + if (!intel_pmu_save_and_restart(counter)) + continue; + + data.period = counter->hw.last_period; + + if (perf_counter_overflow(counter, 1, &data)) + intel_pmu_disable_counter(&counter->hw, bit); + } + + intel_pmu_ack_status(ack); + + /* + * Repeat if there is more work to be done: + */ + status = intel_pmu_get_status(); + if (status) + goto again; + + perf_enable(); + + return 1; +} + +static int amd_pmu_handle_irq(struct pt_regs *regs) +{ + struct perf_sample_data data; + struct cpu_hw_counters *cpuc; + struct perf_counter *counter; + struct hw_perf_counter *hwc; + int idx, handled = 0; + u64 val; + + data.regs = regs; + data.addr = 0; + + cpuc = &__get_cpu_var(cpu_hw_counters); + + for (idx = 0; idx < x86_pmu.num_counters; idx++) { + if (!test_bit(idx, cpuc->active_mask)) + continue; + + counter = cpuc->counters[idx]; + hwc = &counter->hw; + + val = x86_perf_counter_update(counter, hwc, idx); + if (val & (1ULL << (x86_pmu.counter_bits - 1))) + continue; + + /* + * counter overflow + */ + handled = 1; + data.period = counter->hw.last_period; + + if (!x86_perf_counter_set_period(counter, hwc, idx)) + continue; + + if (perf_counter_overflow(counter, 1, &data)) + amd_pmu_disable_counter(hwc, idx); + } + + if (handled) + inc_irq_stat(apic_perf_irqs); + + return handled; +} + +void smp_perf_pending_interrupt(struct pt_regs *regs) +{ + irq_enter(); + ack_APIC_irq(); + inc_irq_stat(apic_pending_irqs); + perf_counter_do_pending(); + irq_exit(); +} + +void set_perf_counter_pending(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + apic->send_IPI_self(LOCAL_PENDING_VECTOR); +#endif +} + +void perf_counters_lapic_init(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + if (!x86_pmu.apic || !x86_pmu_initialized()) + return; + + /* + * Always use NMI for PMU + */ + apic_write(APIC_LVTPC, APIC_DM_NMI); +#endif +} + +static int __kprobes +perf_counter_nmi_handler(struct notifier_block *self, + unsigned long cmd, void *__args) +{ + struct die_args *args = __args; + struct pt_regs *regs; + + if (!atomic_read(&active_counters)) + return NOTIFY_DONE; + + switch (cmd) { + case DIE_NMI: + case DIE_NMI_IPI: + break; + + default: + return NOTIFY_DONE; + } + + regs = args->regs; + +#ifdef CONFIG_X86_LOCAL_APIC + apic_write(APIC_LVTPC, APIC_DM_NMI); +#endif + /* + * Can't rely on the handled return value to say it was our NMI, two + * counters could trigger 'simultaneously' raising two back-to-back NMIs. + * + * If the first NMI handles both, the latter will be empty and daze + * the CPU. + */ + x86_pmu.handle_irq(regs); + + return NOTIFY_STOP; +} + +static __read_mostly struct notifier_block perf_counter_nmi_notifier = { + .notifier_call = perf_counter_nmi_handler, + .next = NULL, + .priority = 1 +}; + +static struct x86_pmu p6_pmu = { + .name = "p6", + .handle_irq = p6_pmu_handle_irq, + .disable_all = p6_pmu_disable_all, + .enable_all = p6_pmu_enable_all, + .enable = p6_pmu_enable_counter, + .disable = p6_pmu_disable_counter, + .eventsel = MSR_P6_EVNTSEL0, + .perfctr = MSR_P6_PERFCTR0, + .event_map = p6_pmu_event_map, + .raw_event = p6_pmu_raw_event, + .max_events = ARRAY_SIZE(p6_perfmon_event_map), + .apic = 1, + .max_period = (1ULL << 31) - 1, + .version = 0, + .num_counters = 2, + /* + * Counters have 40 bits implemented. However they are designed such + * that bits [32-39] are sign extensions of bit 31. As such the + * effective width of a counter for P6-like PMU is 32 bits only. + * + * See IA-32 Intel Architecture Software developer manual Vol 3B + */ + .counter_bits = 32, + .counter_mask = (1ULL << 32) - 1, +}; + +static struct x86_pmu intel_pmu = { + .name = "Intel", + .handle_irq = intel_pmu_handle_irq, + .disable_all = intel_pmu_disable_all, + .enable_all = intel_pmu_enable_all, + .enable = intel_pmu_enable_counter, + .disable = intel_pmu_disable_counter, + .eventsel = MSR_ARCH_PERFMON_EVENTSEL0, + .perfctr = MSR_ARCH_PERFMON_PERFCTR0, + .event_map = intel_pmu_event_map, + .raw_event = intel_pmu_raw_event, + .max_events = ARRAY_SIZE(intel_perfmon_event_map), + .apic = 1, + /* + * Intel PMCs cannot be accessed sanely above 32 bit width, + * so we install an artificial 1<<31 period regardless of + * the generic counter period: + */ + .max_period = (1ULL << 31) - 1, +}; + +static struct x86_pmu amd_pmu = { + .name = "AMD", + .handle_irq = amd_pmu_handle_irq, + .disable_all = amd_pmu_disable_all, + .enable_all = amd_pmu_enable_all, + .enable = amd_pmu_enable_counter, + .disable = amd_pmu_disable_counter, + .eventsel = MSR_K7_EVNTSEL0, + .perfctr = MSR_K7_PERFCTR0, + .event_map = amd_pmu_event_map, + .raw_event = amd_pmu_raw_event, + .max_events = ARRAY_SIZE(amd_perfmon_event_map), + .num_counters = 4, + .counter_bits = 48, + .counter_mask = (1ULL << 48) - 1, + .apic = 1, + /* use highest bit to detect overflow */ + .max_period = (1ULL << 47) - 1, +}; + +static int p6_pmu_init(void) +{ + switch (boot_cpu_data.x86_model) { + case 1: + case 3: /* Pentium Pro */ + case 5: + case 6: /* Pentium II */ + case 7: + case 8: + case 11: /* Pentium III */ + break; + case 9: + case 13: + /* Pentium M */ + break; + default: + pr_cont("unsupported p6 CPU model %d ", + boot_cpu_data.x86_model); + return -ENODEV; + } + + x86_pmu = p6_pmu; + + if (!cpu_has_apic) { + pr_info("no APIC, boot with the \"lapic\" boot parameter to force-enable it.\n"); + pr_info("no hardware sampling interrupt available.\n"); + x86_pmu.apic = 0; + } + + return 0; +} + +static int intel_pmu_init(void) +{ + union cpuid10_edx edx; + union cpuid10_eax eax; + unsigned int unused; + unsigned int ebx; + int version; + + if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) { + /* check for P6 processor family */ + if (boot_cpu_data.x86 == 6) { + return p6_pmu_init(); + } else { + return -ENODEV; + } + } + + /* + * Check whether the Architectural PerfMon supports + * Branch Misses Retired Event or not. + */ + cpuid(10, &eax.full, &ebx, &unused, &edx.full); + if (eax.split.mask_length <= ARCH_PERFMON_BRANCH_MISSES_RETIRED) + return -ENODEV; + + version = eax.split.version_id; + if (version < 2) + return -ENODEV; + + x86_pmu = intel_pmu; + x86_pmu.version = version; + x86_pmu.num_counters = eax.split.num_counters; + x86_pmu.counter_bits = eax.split.bit_width; + x86_pmu.counter_mask = (1ULL << eax.split.bit_width) - 1; + + /* + * Quirk: v2 perfmon does not report fixed-purpose counters, so + * assume at least 3 counters: + */ + x86_pmu.num_counters_fixed = max((int)edx.split.num_counters_fixed, 3); + + /* + * Install the hw-cache-events table: + */ + switch (boot_cpu_data.x86_model) { + case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */ + case 22: /* single-core 65 nm celeron/core2solo "Merom-L"/"Conroe-L" */ + case 23: /* current 45 nm celeron/core2/xeon "Penryn"/"Wolfdale" */ + case 29: /* six-core 45 nm xeon "Dunnington" */ + memcpy(hw_cache_event_ids, core2_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + pr_cont("Core2 events, "); + break; + default: + case 26: + memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + pr_cont("Nehalem/Corei7 events, "); + break; + case 28: + memcpy(hw_cache_event_ids, atom_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + pr_cont("Atom events, "); + break; + } + return 0; +} + +static int amd_pmu_init(void) +{ + /* Performance-monitoring supported from K7 and later: */ + if (boot_cpu_data.x86 < 6) + return -ENODEV; + + x86_pmu = amd_pmu; + + /* Events are common for all AMDs */ + memcpy(hw_cache_event_ids, amd_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + return 0; +} + +void __init init_hw_perf_counters(void) +{ + int err; + + pr_info("Performance Counters: "); + + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_INTEL: + err = intel_pmu_init(); + break; + case X86_VENDOR_AMD: + err = amd_pmu_init(); + break; + default: + return; + } + if (err != 0) { + pr_cont("no PMU driver, software counters only.\n"); + return; + } + + pr_cont("%s PMU driver.\n", x86_pmu.name); + + if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) { + WARN(1, KERN_ERR "hw perf counters %d > max(%d), clipping!", + x86_pmu.num_counters, X86_PMC_MAX_GENERIC); + x86_pmu.num_counters = X86_PMC_MAX_GENERIC; + } + perf_counter_mask = (1 << x86_pmu.num_counters) - 1; + perf_max_counters = x86_pmu.num_counters; + + if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) { + WARN(1, KERN_ERR "hw perf counters fixed %d > max(%d), clipping!", + x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED); + x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED; + } + + perf_counter_mask |= + ((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED; + x86_pmu.intel_ctrl = perf_counter_mask; + + perf_counters_lapic_init(); + register_die_notifier(&perf_counter_nmi_notifier); + + pr_info("... version: %d\n", x86_pmu.version); + pr_info("... bit width: %d\n", x86_pmu.counter_bits); + pr_info("... generic counters: %d\n", x86_pmu.num_counters); + pr_info("... value mask: %016Lx\n", x86_pmu.counter_mask); + pr_info("... max period: %016Lx\n", x86_pmu.max_period); + pr_info("... fixed-purpose counters: %d\n", x86_pmu.num_counters_fixed); + pr_info("... counter mask: %016Lx\n", perf_counter_mask); +} + +static inline void x86_pmu_read(struct perf_counter *counter) +{ + x86_perf_counter_update(counter, &counter->hw, counter->hw.idx); +} + +static const struct pmu pmu = { + .enable = x86_pmu_enable, + .disable = x86_pmu_disable, + .read = x86_pmu_read, + .unthrottle = x86_pmu_unthrottle, +}; + +const struct pmu *hw_perf_counter_init(struct perf_counter *counter) +{ + int err; + + err = __hw_perf_counter_init(counter); + if (err) + return ERR_PTR(err); + + return &pmu; +} + +/* + * callchain support + */ + +static inline +void callchain_store(struct perf_callchain_entry *entry, u64 ip) +{ + if (entry->nr < PERF_MAX_STACK_DEPTH) + entry->ip[entry->nr++] = ip; +} + +static DEFINE_PER_CPU(struct perf_callchain_entry, irq_entry); +static DEFINE_PER_CPU(struct perf_callchain_entry, nmi_entry); +static DEFINE_PER_CPU(int, in_nmi_frame); + + +static void +backtrace_warning_symbol(void *data, char *msg, unsigned long symbol) +{ + /* Ignore warnings */ +} + +static void backtrace_warning(void *data, char *msg) +{ + /* Ignore warnings */ +} + +static int backtrace_stack(void *data, char *name) +{ + per_cpu(in_nmi_frame, smp_processor_id()) = + x86_is_stack_id(NMI_STACK, name); + + return 0; +} + +static void backtrace_address(void *data, unsigned long addr, int reliable) +{ + struct perf_callchain_entry *entry = data; + + if (per_cpu(in_nmi_frame, smp_processor_id())) + return; + + if (reliable) + callchain_store(entry, addr); +} + +static const struct stacktrace_ops backtrace_ops = { + .warning = backtrace_warning, + .warning_symbol = backtrace_warning_symbol, + .stack = backtrace_stack, + .address = backtrace_address, +}; + +#include "../dumpstack.h" + +static void +perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry) +{ + callchain_store(entry, PERF_CONTEXT_KERNEL); + callchain_store(entry, regs->ip); + + dump_trace(NULL, regs, NULL, 0, &backtrace_ops, entry); +} + +/* + * best effort, GUP based copy_from_user() that assumes IRQ or NMI context + */ +static unsigned long +copy_from_user_nmi(void *to, const void __user *from, unsigned long n) +{ + unsigned long offset, addr = (unsigned long)from; + int type = in_nmi() ? KM_NMI : KM_IRQ0; + unsigned long size, len = 0; + struct page *page; + void *map; + int ret; + + do { + ret = __get_user_pages_fast(addr, 1, 0, &page); + if (!ret) + break; + + offset = addr & (PAGE_SIZE - 1); + size = min(PAGE_SIZE - offset, n - len); + + map = kmap_atomic(page, type); + memcpy(to, map+offset, size); + kunmap_atomic(map, type); + put_page(page); + + len += size; + to += size; + addr += size; + + } while (len < n); + + return len; +} + +static int copy_stack_frame(const void __user *fp, struct stack_frame *frame) +{ + unsigned long bytes; + + bytes = copy_from_user_nmi(frame, fp, sizeof(*frame)); + + return bytes == sizeof(*frame); +} + +static void +perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry) +{ + struct stack_frame frame; + const void __user *fp; + + if (!user_mode(regs)) + regs = task_pt_regs(current); + + fp = (void __user *)regs->bp; + + callchain_store(entry, PERF_CONTEXT_USER); + callchain_store(entry, regs->ip); + + while (entry->nr < PERF_MAX_STACK_DEPTH) { + frame.next_frame = NULL; + frame.return_address = 0; + + if (!copy_stack_frame(fp, &frame)) + break; + + if ((unsigned long)fp < regs->sp) + break; + + callchain_store(entry, frame.return_address); + fp = frame.next_frame; + } +} + +static void +perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry) +{ + int is_user; + + if (!regs) + return; + + is_user = user_mode(regs); + + if (!current || current->pid == 0) + return; + + if (is_user && current->state != TASK_RUNNING) + return; + + if (!is_user) + perf_callchain_kernel(regs, entry); + + if (current->mm) + perf_callchain_user(regs, entry); +} + +struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +{ + struct perf_callchain_entry *entry; + + if (in_nmi()) + entry = &__get_cpu_var(nmi_entry); + else + entry = &__get_cpu_var(irq_entry); + + entry->nr = 0; + + perf_do_callchain(regs, entry); + + return entry; +} diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index f6c70a164e3..e60ed740d2b 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -19,8 +19,8 @@ #include <linux/nmi.h> #include <linux/kprobes.h> -#include <asm/genapic.h> -#include <asm/intel_arch_perfmon.h> +#include <asm/apic.h> +#include <asm/perf_counter.h> struct nmi_watchdog_ctlblk { unsigned int cccr_msr; @@ -716,11 +716,15 @@ static void probe_nmi_watchdog(void) wd_ops = &k7_wd_ops; break; case X86_VENDOR_INTEL: - /* - * Work around Core Duo (Yonah) errata AE49 where perfctr1 - * doesn't have a working enable bit. + /* Work around where perfctr1 doesn't have a working enable + * bit as described in the following errata: + * AE49 Core Duo and Intel Core Solo 65 nm + * AN49 Intel Pentium Dual-Core + * AF49 Dual-Core Intel Xeon Processor LV */ - if (boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 14) { + if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 14) || + ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model == 15 && + boot_cpu_data.x86_mask == 4))) { intel_arch_wd_ops.perfctr = MSR_ARCH_PERFMON_PERFCTR0; intel_arch_wd_ops.evntsel = MSR_ARCH_PERFMON_EVENTSEL0; } @@ -799,8 +803,3 @@ int __kprobes lapic_wd_event(unsigned nmi_hz) wd_ops->rearm(wd, nmi_hz); return 1; } - -int lapic_watchdog_ok(void) -{ - return wd_ops != NULL; -} |