From 1af0815f9639752409a9c15ba6a3dc0f322fd114 Mon Sep 17 00:00:00 2001 From: Hidetoshi Seto Date: Mon, 15 Jun 2009 17:28:38 +0900 Subject: x86, mce: rename _64.c files which are no longer 64-bit-specific Rename files that are no longer 64bit specific: mce_amd_64.c => mce_amd.c mce_intel_64.c => mce_intel.c Signed-off-by: Hidetoshi Seto Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/mcheck/Makefile | 4 +- arch/x86/kernel/cpu/mcheck/mce_amd.c | 703 ++++++++++++++++++++++++++++++ arch/x86/kernel/cpu/mcheck/mce_amd_64.c | 703 ------------------------------ arch/x86/kernel/cpu/mcheck/mce_intel.c | 225 ++++++++++ arch/x86/kernel/cpu/mcheck/mce_intel_64.c | 225 ---------- 5 files changed, 930 insertions(+), 930 deletions(-) create mode 100644 arch/x86/kernel/cpu/mcheck/mce_amd.c delete mode 100644 arch/x86/kernel/cpu/mcheck/mce_amd_64.c create mode 100644 arch/x86/kernel/cpu/mcheck/mce_intel.c delete mode 100644 arch/x86/kernel/cpu/mcheck/mce_intel_64.c (limited to 'arch/x86') diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile index 659564e5fc0..188a1ca5ad2 100644 --- a/arch/x86/kernel/cpu/mcheck/Makefile +++ b/arch/x86/kernel/cpu/mcheck/Makefile @@ -3,8 +3,8 @@ obj-y = mce.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_64.o -obj-$(CONFIG_X86_MCE_AMD) += mce_amd_64.o +obj-$(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 diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c new file mode 100644 index 00000000000..ddae21620bd --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c @@ -0,0 +1,703 @@ +/* + * (c) 2005, 2006 Advanced Micro Devices, Inc. + * Your use of this code is subject to the terms and conditions of the + * GNU general public license version 2. See "COPYING" or + * http://www.gnu.org/licenses/gpl.html + * + * Written by Jacob Shin - AMD, Inc. + * + * Support : jacob.shin@amd.com + * + * April 2006 + * - added support for AMD Family 0x10 processors + * + * All MC4_MISCi registers are shared between multi-cores + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#define PFX "mce_threshold: " +#define VERSION "version 1.1.1" +#define NR_BANKS 6 +#define NR_BLOCKS 9 +#define THRESHOLD_MAX 0xFFF +#define INT_TYPE_APIC 0x00020000 +#define MASK_VALID_HI 0x80000000 +#define MASK_CNTP_HI 0x40000000 +#define MASK_LOCKED_HI 0x20000000 +#define MASK_LVTOFF_HI 0x00F00000 +#define MASK_COUNT_EN_HI 0x00080000 +#define MASK_INT_TYPE_HI 0x00060000 +#define MASK_OVERFLOW_HI 0x00010000 +#define MASK_ERR_COUNT_HI 0x00000FFF +#define MASK_BLKPTR_LO 0xFF000000 +#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; +}; + +/* defaults used early on boot */ +static struct threshold_block threshold_defaults = { + .interrupt_enable = 0, + .threshold_limit = THRESHOLD_MAX, +}; + +struct threshold_bank { + struct kobject *kobj; + struct threshold_block *blocks; + cpumask_var_t cpus; +}; +static DEFINE_PER_CPU(struct threshold_bank *, threshold_banks[NR_BANKS]); + +#ifdef CONFIG_SMP +static unsigned char shared_bank[NR_BANKS] = { + 0, 0, 0, 0, 1 +}; +#endif + +static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */ + +static void amd_threshold_interrupt(void); + +/* + * CPU Initialization + */ + +struct thresh_restart { + struct threshold_block *b; + int reset; + u16 old_limit; +}; + +/* must be called with correct cpu affinity */ +/* Called via smp_call_function_single() */ +static void threshold_restart_bank(void *_tr) +{ + struct thresh_restart *tr = _tr; + u32 mci_misc_hi, mci_misc_lo; + + rdmsr(tr->b->address, mci_misc_lo, mci_misc_hi); + + if (tr->b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) + tr->reset = 1; /* limit cannot be lower than err count */ + + if (tr->reset) { /* reset err count and overflow bit */ + mci_misc_hi = + (mci_misc_hi & ~(MASK_ERR_COUNT_HI | MASK_OVERFLOW_HI)) | + (THRESHOLD_MAX - tr->b->threshold_limit); + } 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); + } + + tr->b->interrupt_enable ? + (mci_misc_hi = (mci_misc_hi & ~MASK_INT_TYPE_HI) | INT_TYPE_APIC) : + (mci_misc_hi &= ~MASK_INT_TYPE_HI); + + mci_misc_hi |= MASK_COUNT_EN_HI; + wrmsr(tr->b->address, mci_misc_lo, mci_misc_hi); +} + +/* cpu init entry point, called from mce.c with preempt off */ +void mce_amd_feature_init(struct cpuinfo_x86 *c) +{ + unsigned int cpu = smp_processor_id(); + 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) { + if (block == 0) + address = MSR_IA32_MC0_MISC + bank * 4; + else if (block == 1) { + address = (low & MASK_BLKPTR_LO) >> 21; + if (!address) + break; + address += MCG_XBLK_ADDR; + } else + ++address; + + if (rdmsr_safe(address, &low, &high)) + break; + + if (!(high & MASK_VALID_HI)) { + if (block) + continue; + else + break; + } + + if (!(high & MASK_CNTP_HI) || + (high & MASK_LOCKED_HI)) + continue; + + if (!block) + per_cpu(bank_map, cpu) |= (1 << bank); +#ifdef CONFIG_SMP + if (shared_bank[bank] && c->cpu_core_id) + break; +#endif + lvt_off = setup_APIC_eilvt_mce(THRESHOLD_APIC_VECTOR, + APIC_EILVT_MSG_FIX, 0); + + high &= ~MASK_LVTOFF_HI; + high |= lvt_off << 20; + wrmsr(address, low, high); + + threshold_defaults.address = address; + tr.b = &threshold_defaults; + tr.reset = 0; + tr.old_limit = 0; + threshold_restart_bank(&tr); + + mce_threshold_vector = amd_threshold_interrupt; + } + } +} + +/* + * APIC Interrupt Handler + */ + +/* + * threshold interrupt handler will service THRESHOLD_APIC_VECTOR. + * the interrupt goes off when error_count reaches threshold_limit. + * the handler will simply log mcelog w/ software defined bank number. + */ +static void amd_threshold_interrupt(void) +{ + u32 low = 0, high = 0, address = 0; + unsigned int bank, block; + struct mce m; + + mce_setup(&m); + + /* assume first bank caused it */ + for (bank = 0; bank < NR_BANKS; ++bank) { + if (!(per_cpu(bank_map, m.cpu) & (1 << bank))) + continue; + for (block = 0; block < NR_BLOCKS; ++block) { + if (block == 0) { + address = MSR_IA32_MC0_MISC + bank * 4; + } else if (block == 1) { + address = (low & MASK_BLKPTR_LO) >> 21; + if (!address) + break; + address += MCG_XBLK_ADDR; + } else { + ++address; + } + + if (rdmsr_safe(address, &low, &high)) + break; + + if (!(high & MASK_VALID_HI)) { + if (block) + continue; + else + break; + } + + if (!(high & MASK_CNTP_HI) || + (high & MASK_LOCKED_HI)) + continue; + + /* + * Log the machine check that caused the threshold + * event. + */ + machine_check_poll(MCP_TIMESTAMP, + &__get_cpu_var(mce_poll_banks)); + + if (high & MASK_OVERFLOW_HI) { + rdmsrl(address, m.misc); + rdmsrl(MSR_IA32_MC0_STATUS + bank * 4, + m.status); + m.bank = K8_MCE_THRESHOLD_BASE + + bank * NR_BLOCKS + + block; + mce_log(&m); + return; + } + } + } +} + +/* + * Sysfs Interface + */ + +struct threshold_attr { + struct attribute attr; + 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); \ +} +SHOW_FIELDS(interrupt_enable) +SHOW_FIELDS(threshold_limit) + +static ssize_t +store_interrupt_enable(struct threshold_block *b, const char *buf, size_t size) +{ + struct thresh_restart tr; + 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; + + smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); + + return size; +} + +static ssize_t +store_threshold_limit(struct threshold_block *b, const char *buf, size_t size) +{ + struct thresh_restart tr; + 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; + tr.reset = 0; + + smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); + + return size; +} + +struct threshold_block_cross_cpu { + struct threshold_block *tb; + long retval; +}; + +static void local_error_count_handler(void *_tbcc) +{ + struct threshold_block_cross_cpu *tbcc = _tbcc; + struct threshold_block *b = tbcc->tb; + u32 low, high; + + rdmsr(b->address, low, high); + tbcc->retval = (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit); +} + +static ssize_t show_error_count(struct threshold_block *b, char *buf) +{ + struct threshold_block_cross_cpu tbcc = { .tb = b, }; + + smp_call_function_single(b->cpu, local_error_count_handler, &tbcc, 1); + return sprintf(buf, "%lx\n", tbcc.retval); +} + +static ssize_t store_error_count(struct threshold_block *b, + const char *buf, size_t count) +{ + struct thresh_restart tr = { .b = b, .reset = 1, .old_limit = 0 }; + + smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); + return 1; +} + +#define RW_ATTR(val) \ +static struct threshold_attr val = { \ + .attr = {.name = __stringify(val), .mode = 0644 }, \ + .show = show_## val, \ + .store = store_## val, \ +}; + +RW_ATTR(interrupt_enable); +RW_ATTR(threshold_limit); +RW_ATTR(error_count); + +static struct attribute *default_attrs[] = { + &interrupt_enable.attr, + &threshold_limit.attr, + &error_count.attr, + NULL +}; + +#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; +} + +static ssize_t store(struct kobject *kobj, struct attribute *attr, + const char *buf, size_t count) +{ + 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, +}; + +static struct kobj_type threshold_ktype = { + .sysfs_ops = &threshold_ops, + .default_attrs = default_attrs, +}; + +static __cpuinit int allocate_threshold_blocks(unsigned int cpu, + unsigned int bank, + unsigned int block, + u32 address) +{ + struct threshold_block *b = NULL; + u32 low, high; + int err; + + if ((bank >= NR_BANKS) || (block >= NR_BLOCKS)) + return 0; + + if (rdmsr_safe_on_cpu(cpu, address, &low, &high)) + return 0; + + if (!(high & MASK_VALID_HI)) { + if (block) + goto recurse; + else + return 0; + } + + if (!(high & MASK_CNTP_HI) || + (high & MASK_LOCKED_HI)) + goto recurse; + + b = kzalloc(sizeof(struct threshold_block), GFP_KERNEL); + 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; + + INIT_LIST_HEAD(&b->miscj); + + if (per_cpu(threshold_banks, cpu)[bank]->blocks) { + list_add(&b->miscj, + &per_cpu(threshold_banks, cpu)[bank]->blocks->miscj); + } 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, + "misc%i", block); + if (err) + goto out_free; +recurse: + if (!block) { + address = (low & MASK_BLKPTR_LO) >> 21; + if (!address) + return 0; + address += MCG_XBLK_ADDR; + } else { + ++address; + } + + err = allocate_threshold_blocks(cpu, bank, ++block, address); + if (err) + goto out_free; + + if (b) + kobject_uevent(&b->kobj, KOBJ_ADD); + + return err; + +out_free: + if (b) { + kobject_put(&b->kobj); + kfree(b); + } + return err; +} + +static __cpuinit long +local_allocate_threshold_blocks(int cpu, unsigned int bank) +{ + return allocate_threshold_blocks(cpu, bank, 0, + MSR_IA32_MC0_MISC + bank * 4); +} + +/* symlinks sibling shared banks to first core. first core owns dir/files. */ +static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) +{ + int i, err = 0; + struct threshold_bank *b = NULL; + char name[32]; + + sprintf(name, "threshold_bank%i", bank); + +#ifdef CONFIG_SMP + if (cpu_data(cpu).cpu_core_id && shared_bank[bank]) { /* symlink */ + i = cpumask_first(cpu_core_mask(cpu)); + + /* first core not up yet */ + if (cpu_data(i).cpu_core_id) + goto out; + + /* already linked */ + if (per_cpu(threshold_banks, cpu)[bank]) + goto out; + + b = per_cpu(threshold_banks, i)[bank]; + + if (!b) + goto out; + + 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 + + b = kzalloc(sizeof(struct threshold_bank), GFP_KERNEL); + if (!b) { + err = -ENOMEM; + goto out; + } + if (!alloc_cpumask_var(&b->cpus, GFP_KERNEL)) { + kfree(b); + err = -ENOMEM; + goto out; + } + + b->kobj = kobject_create_and_add(name, &per_cpu(mce_dev, cpu).kobj); + if (!b->kobj) + goto out_free; + +#ifndef CONFIG_SMP + cpumask_setall(b->cpus); +#else + cpumask_copy(b->cpus, cpu_core_mask(cpu)); +#endif + + per_cpu(threshold_banks, cpu)[bank] = b; + + err = local_allocate_threshold_blocks(cpu, bank); + if (err) + goto out_free; + + for_each_cpu(i, b->cpus) { + if (i == cpu) + continue; + + err = sysfs_create_link(&per_cpu(mce_dev, i).kobj, + b->kobj, name); + if (err) + goto out; + + per_cpu(threshold_banks, i)[bank] = b; + } + + goto out; + +out_free: + per_cpu(threshold_banks, cpu)[bank] = NULL; + free_cpumask_var(b->cpus); + kfree(b); +out: + return err; +} + +/* create dir/files for all valid threshold banks */ +static __cpuinit int threshold_create_device(unsigned int cpu) +{ + unsigned int bank; + int err = 0; + + for (bank = 0; bank < NR_BANKS; ++bank) { + if (!(per_cpu(bank_map, cpu) & (1 << bank))) + continue; + err = threshold_create_bank(cpu, bank); + if (err) + goto out; + } +out: + return err; +} + +/* + * let's be hotplug friendly. + * in case of multiple core processors, the first core always takes ownership + * of shared sysfs dir/files, and rest of the cores will be symlinked to it. + */ + +static void deallocate_threshold_block(unsigned int cpu, + unsigned int bank) +{ + struct threshold_block *pos = NULL; + struct threshold_block *tmp = NULL; + struct threshold_bank *head = per_cpu(threshold_banks, cpu)[bank]; + + if (!head) + return; + + list_for_each_entry_safe(pos, tmp, &head->blocks->miscj, miscj) { + kobject_put(&pos->kobj); + list_del(&pos->miscj); + kfree(pos); + } + + kfree(per_cpu(threshold_banks, cpu)[bank]->blocks); + per_cpu(threshold_banks, cpu)[bank]->blocks = NULL; +} + +static void threshold_remove_bank(unsigned int cpu, int bank) +{ + 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; + + sprintf(name, "threshold_bank%i", bank); + +#ifdef CONFIG_SMP + /* sibling symlink */ + if (shared_bank[bank] && b->blocks->cpu != cpu) { + sysfs_remove_link(&per_cpu(mce_dev, cpu).kobj, name); + per_cpu(threshold_banks, cpu)[bank] = NULL; + + return; + } +#endif + + /* remove all sibling symlinks before unregistering */ + for_each_cpu(i, b->cpus) { + if (i == cpu) + continue; + + sysfs_remove_link(&per_cpu(mce_dev, i).kobj, name); + per_cpu(threshold_banks, i)[bank] = NULL; + } + + deallocate_threshold_block(cpu, bank); + +free_out: + kobject_del(b->kobj); + kobject_put(b->kobj); + free_cpumask_var(b->cpus); + kfree(b); + per_cpu(threshold_banks, cpu)[bank] = NULL; +} + +static void threshold_remove_device(unsigned int cpu) +{ + unsigned int bank; + + for (bank = 0; bank < NR_BANKS; ++bank) { + if (!(per_cpu(bank_map, cpu) & (1 << bank))) + continue; + threshold_remove_bank(cpu, bank); + } +} + +/* get notified when a cpu comes on/off */ +static void __cpuinit +amd_64_threshold_cpu_callback(unsigned long action, unsigned int cpu) +{ + switch (action) { + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + threshold_create_device(cpu); + break; + case CPU_DEAD: + case CPU_DEAD_FROZEN: + threshold_remove_device(cpu); + break; + default: + break; + } +} + +static __init int threshold_init_device(void) +{ + unsigned lcpu = 0; + + /* 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_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c deleted file mode 100644 index ddae21620bd..00000000000 --- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c +++ /dev/null @@ -1,703 +0,0 @@ -/* - * (c) 2005, 2006 Advanced Micro Devices, Inc. - * Your use of this code is subject to the terms and conditions of the - * GNU general public license version 2. See "COPYING" or - * http://www.gnu.org/licenses/gpl.html - * - * Written by Jacob Shin - AMD, Inc. - * - * Support : jacob.shin@amd.com - * - * April 2006 - * - added support for AMD Family 0x10 processors - * - * All MC4_MISCi registers are shared between multi-cores - */ -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#define PFX "mce_threshold: " -#define VERSION "version 1.1.1" -#define NR_BANKS 6 -#define NR_BLOCKS 9 -#define THRESHOLD_MAX 0xFFF -#define INT_TYPE_APIC 0x00020000 -#define MASK_VALID_HI 0x80000000 -#define MASK_CNTP_HI 0x40000000 -#define MASK_LOCKED_HI 0x20000000 -#define MASK_LVTOFF_HI 0x00F00000 -#define MASK_COUNT_EN_HI 0x00080000 -#define MASK_INT_TYPE_HI 0x00060000 -#define MASK_OVERFLOW_HI 0x00010000 -#define MASK_ERR_COUNT_HI 0x00000FFF -#define MASK_BLKPTR_LO 0xFF000000 -#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; -}; - -/* defaults used early on boot */ -static struct threshold_block threshold_defaults = { - .interrupt_enable = 0, - .threshold_limit = THRESHOLD_MAX, -}; - -struct threshold_bank { - struct kobject *kobj; - struct threshold_block *blocks; - cpumask_var_t cpus; -}; -static DEFINE_PER_CPU(struct threshold_bank *, threshold_banks[NR_BANKS]); - -#ifdef CONFIG_SMP -static unsigned char shared_bank[NR_BANKS] = { - 0, 0, 0, 0, 1 -}; -#endif - -static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */ - -static void amd_threshold_interrupt(void); - -/* - * CPU Initialization - */ - -struct thresh_restart { - struct threshold_block *b; - int reset; - u16 old_limit; -}; - -/* must be called with correct cpu affinity */ -/* Called via smp_call_function_single() */ -static void threshold_restart_bank(void *_tr) -{ - struct thresh_restart *tr = _tr; - u32 mci_misc_hi, mci_misc_lo; - - rdmsr(tr->b->address, mci_misc_lo, mci_misc_hi); - - if (tr->b->threshold_limit < (mci_misc_hi & THRESHOLD_MAX)) - tr->reset = 1; /* limit cannot be lower than err count */ - - if (tr->reset) { /* reset err count and overflow bit */ - mci_misc_hi = - (mci_misc_hi & ~(MASK_ERR_COUNT_HI | MASK_OVERFLOW_HI)) | - (THRESHOLD_MAX - tr->b->threshold_limit); - } 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); - } - - tr->b->interrupt_enable ? - (mci_misc_hi = (mci_misc_hi & ~MASK_INT_TYPE_HI) | INT_TYPE_APIC) : - (mci_misc_hi &= ~MASK_INT_TYPE_HI); - - mci_misc_hi |= MASK_COUNT_EN_HI; - wrmsr(tr->b->address, mci_misc_lo, mci_misc_hi); -} - -/* cpu init entry point, called from mce.c with preempt off */ -void mce_amd_feature_init(struct cpuinfo_x86 *c) -{ - unsigned int cpu = smp_processor_id(); - 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) { - if (block == 0) - address = MSR_IA32_MC0_MISC + bank * 4; - else if (block == 1) { - address = (low & MASK_BLKPTR_LO) >> 21; - if (!address) - break; - address += MCG_XBLK_ADDR; - } else - ++address; - - if (rdmsr_safe(address, &low, &high)) - break; - - if (!(high & MASK_VALID_HI)) { - if (block) - continue; - else - break; - } - - if (!(high & MASK_CNTP_HI) || - (high & MASK_LOCKED_HI)) - continue; - - if (!block) - per_cpu(bank_map, cpu) |= (1 << bank); -#ifdef CONFIG_SMP - if (shared_bank[bank] && c->cpu_core_id) - break; -#endif - lvt_off = setup_APIC_eilvt_mce(THRESHOLD_APIC_VECTOR, - APIC_EILVT_MSG_FIX, 0); - - high &= ~MASK_LVTOFF_HI; - high |= lvt_off << 20; - wrmsr(address, low, high); - - threshold_defaults.address = address; - tr.b = &threshold_defaults; - tr.reset = 0; - tr.old_limit = 0; - threshold_restart_bank(&tr); - - mce_threshold_vector = amd_threshold_interrupt; - } - } -} - -/* - * APIC Interrupt Handler - */ - -/* - * threshold interrupt handler will service THRESHOLD_APIC_VECTOR. - * the interrupt goes off when error_count reaches threshold_limit. - * the handler will simply log mcelog w/ software defined bank number. - */ -static void amd_threshold_interrupt(void) -{ - u32 low = 0, high = 0, address = 0; - unsigned int bank, block; - struct mce m; - - mce_setup(&m); - - /* assume first bank caused it */ - for (bank = 0; bank < NR_BANKS; ++bank) { - if (!(per_cpu(bank_map, m.cpu) & (1 << bank))) - continue; - for (block = 0; block < NR_BLOCKS; ++block) { - if (block == 0) { - address = MSR_IA32_MC0_MISC + bank * 4; - } else if (block == 1) { - address = (low & MASK_BLKPTR_LO) >> 21; - if (!address) - break; - address += MCG_XBLK_ADDR; - } else { - ++address; - } - - if (rdmsr_safe(address, &low, &high)) - break; - - if (!(high & MASK_VALID_HI)) { - if (block) - continue; - else - break; - } - - if (!(high & MASK_CNTP_HI) || - (high & MASK_LOCKED_HI)) - continue; - - /* - * Log the machine check that caused the threshold - * event. - */ - machine_check_poll(MCP_TIMESTAMP, - &__get_cpu_var(mce_poll_banks)); - - if (high & MASK_OVERFLOW_HI) { - rdmsrl(address, m.misc); - rdmsrl(MSR_IA32_MC0_STATUS + bank * 4, - m.status); - m.bank = K8_MCE_THRESHOLD_BASE - + bank * NR_BLOCKS - + block; - mce_log(&m); - return; - } - } - } -} - -/* - * Sysfs Interface - */ - -struct threshold_attr { - struct attribute attr; - 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); \ -} -SHOW_FIELDS(interrupt_enable) -SHOW_FIELDS(threshold_limit) - -static ssize_t -store_interrupt_enable(struct threshold_block *b, const char *buf, size_t size) -{ - struct thresh_restart tr; - 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; - - smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); - - return size; -} - -static ssize_t -store_threshold_limit(struct threshold_block *b, const char *buf, size_t size) -{ - struct thresh_restart tr; - 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; - tr.reset = 0; - - smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); - - return size; -} - -struct threshold_block_cross_cpu { - struct threshold_block *tb; - long retval; -}; - -static void local_error_count_handler(void *_tbcc) -{ - struct threshold_block_cross_cpu *tbcc = _tbcc; - struct threshold_block *b = tbcc->tb; - u32 low, high; - - rdmsr(b->address, low, high); - tbcc->retval = (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit); -} - -static ssize_t show_error_count(struct threshold_block *b, char *buf) -{ - struct threshold_block_cross_cpu tbcc = { .tb = b, }; - - smp_call_function_single(b->cpu, local_error_count_handler, &tbcc, 1); - return sprintf(buf, "%lx\n", tbcc.retval); -} - -static ssize_t store_error_count(struct threshold_block *b, - const char *buf, size_t count) -{ - struct thresh_restart tr = { .b = b, .reset = 1, .old_limit = 0 }; - - smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1); - return 1; -} - -#define RW_ATTR(val) \ -static struct threshold_attr val = { \ - .attr = {.name = __stringify(val), .mode = 0644 }, \ - .show = show_## val, \ - .store = store_## val, \ -}; - -RW_ATTR(interrupt_enable); -RW_ATTR(threshold_limit); -RW_ATTR(error_count); - -static struct attribute *default_attrs[] = { - &interrupt_enable.attr, - &threshold_limit.attr, - &error_count.attr, - NULL -}; - -#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; -} - -static ssize_t store(struct kobject *kobj, struct attribute *attr, - const char *buf, size_t count) -{ - 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, -}; - -static struct kobj_type threshold_ktype = { - .sysfs_ops = &threshold_ops, - .default_attrs = default_attrs, -}; - -static __cpuinit int allocate_threshold_blocks(unsigned int cpu, - unsigned int bank, - unsigned int block, - u32 address) -{ - struct threshold_block *b = NULL; - u32 low, high; - int err; - - if ((bank >= NR_BANKS) || (block >= NR_BLOCKS)) - return 0; - - if (rdmsr_safe_on_cpu(cpu, address, &low, &high)) - return 0; - - if (!(high & MASK_VALID_HI)) { - if (block) - goto recurse; - else - return 0; - } - - if (!(high & MASK_CNTP_HI) || - (high & MASK_LOCKED_HI)) - goto recurse; - - b = kzalloc(sizeof(struct threshold_block), GFP_KERNEL); - 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; - - INIT_LIST_HEAD(&b->miscj); - - if (per_cpu(threshold_banks, cpu)[bank]->blocks) { - list_add(&b->miscj, - &per_cpu(threshold_banks, cpu)[bank]->blocks->miscj); - } 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, - "misc%i", block); - if (err) - goto out_free; -recurse: - if (!block) { - address = (low & MASK_BLKPTR_LO) >> 21; - if (!address) - return 0; - address += MCG_XBLK_ADDR; - } else { - ++address; - } - - err = allocate_threshold_blocks(cpu, bank, ++block, address); - if (err) - goto out_free; - - if (b) - kobject_uevent(&b->kobj, KOBJ_ADD); - - return err; - -out_free: - if (b) { - kobject_put(&b->kobj); - kfree(b); - } - return err; -} - -static __cpuinit long -local_allocate_threshold_blocks(int cpu, unsigned int bank) -{ - return allocate_threshold_blocks(cpu, bank, 0, - MSR_IA32_MC0_MISC + bank * 4); -} - -/* symlinks sibling shared banks to first core. first core owns dir/files. */ -static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) -{ - int i, err = 0; - struct threshold_bank *b = NULL; - char name[32]; - - sprintf(name, "threshold_bank%i", bank); - -#ifdef CONFIG_SMP - if (cpu_data(cpu).cpu_core_id && shared_bank[bank]) { /* symlink */ - i = cpumask_first(cpu_core_mask(cpu)); - - /* first core not up yet */ - if (cpu_data(i).cpu_core_id) - goto out; - - /* already linked */ - if (per_cpu(threshold_banks, cpu)[bank]) - goto out; - - b = per_cpu(threshold_banks, i)[bank]; - - if (!b) - goto out; - - 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 - - b = kzalloc(sizeof(struct threshold_bank), GFP_KERNEL); - if (!b) { - err = -ENOMEM; - goto out; - } - if (!alloc_cpumask_var(&b->cpus, GFP_KERNEL)) { - kfree(b); - err = -ENOMEM; - goto out; - } - - b->kobj = kobject_create_and_add(name, &per_cpu(mce_dev, cpu).kobj); - if (!b->kobj) - goto out_free; - -#ifndef CONFIG_SMP - cpumask_setall(b->cpus); -#else - cpumask_copy(b->cpus, cpu_core_mask(cpu)); -#endif - - per_cpu(threshold_banks, cpu)[bank] = b; - - err = local_allocate_threshold_blocks(cpu, bank); - if (err) - goto out_free; - - for_each_cpu(i, b->cpus) { - if (i == cpu) - continue; - - err = sysfs_create_link(&per_cpu(mce_dev, i).kobj, - b->kobj, name); - if (err) - goto out; - - per_cpu(threshold_banks, i)[bank] = b; - } - - goto out; - -out_free: - per_cpu(threshold_banks, cpu)[bank] = NULL; - free_cpumask_var(b->cpus); - kfree(b); -out: - return err; -} - -/* create dir/files for all valid threshold banks */ -static __cpuinit int threshold_create_device(unsigned int cpu) -{ - unsigned int bank; - int err = 0; - - for (bank = 0; bank < NR_BANKS; ++bank) { - if (!(per_cpu(bank_map, cpu) & (1 << bank))) - continue; - err = threshold_create_bank(cpu, bank); - if (err) - goto out; - } -out: - return err; -} - -/* - * let's be hotplug friendly. - * in case of multiple core processors, the first core always takes ownership - * of shared sysfs dir/files, and rest of the cores will be symlinked to it. - */ - -static void deallocate_threshold_block(unsigned int cpu, - unsigned int bank) -{ - struct threshold_block *pos = NULL; - struct threshold_block *tmp = NULL; - struct threshold_bank *head = per_cpu(threshold_banks, cpu)[bank]; - - if (!head) - return; - - list_for_each_entry_safe(pos, tmp, &head->blocks->miscj, miscj) { - kobject_put(&pos->kobj); - list_del(&pos->miscj); - kfree(pos); - } - - kfree(per_cpu(threshold_banks, cpu)[bank]->blocks); - per_cpu(threshold_banks, cpu)[bank]->blocks = NULL; -} - -static void threshold_remove_bank(unsigned int cpu, int bank) -{ - 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; - - sprintf(name, "threshold_bank%i", bank); - -#ifdef CONFIG_SMP - /* sibling symlink */ - if (shared_bank[bank] && b->blocks->cpu != cpu) { - sysfs_remove_link(&per_cpu(mce_dev, cpu).kobj, name); - per_cpu(threshold_banks, cpu)[bank] = NULL; - - return; - } -#endif - - /* remove all sibling symlinks before unregistering */ - for_each_cpu(i, b->cpus) { - if (i == cpu) - continue; - - sysfs_remove_link(&per_cpu(mce_dev, i).kobj, name); - per_cpu(threshold_banks, i)[bank] = NULL; - } - - deallocate_threshold_block(cpu, bank); - -free_out: - kobject_del(b->kobj); - kobject_put(b->kobj); - free_cpumask_var(b->cpus); - kfree(b); - per_cpu(threshold_banks, cpu)[bank] = NULL; -} - -static void threshold_remove_device(unsigned int cpu) -{ - unsigned int bank; - - for (bank = 0; bank < NR_BANKS; ++bank) { - if (!(per_cpu(bank_map, cpu) & (1 << bank))) - continue; - threshold_remove_bank(cpu, bank); - } -} - -/* get notified when a cpu comes on/off */ -static void __cpuinit -amd_64_threshold_cpu_callback(unsigned long action, unsigned int cpu) -{ - switch (action) { - case CPU_ONLINE: - case CPU_ONLINE_FROZEN: - threshold_create_device(cpu); - break; - case CPU_DEAD: - case CPU_DEAD_FROZEN: - threshold_remove_device(cpu); - break; - default: - break; - } -} - -static __init int threshold_init_device(void) -{ - unsigned lcpu = 0; - - /* 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.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c new file mode 100644 index 00000000000..663a88e8b3c --- /dev/null +++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c @@ -0,0 +1,225 @@ +/* + * Intel specific MCE features. + * Copyright 2004 Zwane Mwaikambo + * Copyright (C) 2008, 2009 Intel Corporation + * Author: Andi Kleen + */ + +#include +#include +#include +#include +#include +#include + +/* + * Support for Intel Correct Machine Check Interrupts. This allows + * the CPU to raise an interrupt when a corrected machine check happened. + * Normally we pick those up using a regular polling timer. + * Also supports reliable discovery of shared banks. + */ + +static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned); + +/* + * cmci_discover_lock protects against parallel discovery attempts + * which could race against each other. + */ +static DEFINE_SPINLOCK(cmci_discover_lock); + +#define CMCI_THRESHOLD 1 + +static int cmci_supported(int *banks) +{ + u64 cap; + + if (mce_cmci_disabled || mce_ignore_ce) + return 0; + + /* + * Vendor check is not strictly needed, but the initial + * initialization is vendor keyed and this + * makes sure none of the backdoors are entered otherwise. + */ + if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL) + return 0; + if (!cpu_has_apic || lapic_get_maxlvt() < 6) + return 0; + rdmsrl(MSR_IA32_MCG_CAP, cap); + *banks = min_t(unsigned, MAX_NR_BANKS, cap & 0xff); + return !!(cap & MCG_CMCI_P); +} + +/* + * The interrupt handler. This is called on every event. + * Just call the poller directly to log any events. + * This could in theory increase the threshold under high load, + * but doesn't for now. + */ +static void intel_threshold_interrupt(void) +{ + machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned)); + mce_notify_irq(); +} + +static void print_update(char *type, int *hdr, int num) +{ + if (*hdr == 0) + printk(KERN_INFO "CPU %d MCA banks", smp_processor_id()); + *hdr = 1; + printk(KERN_CONT " %s:%d", type, num); +} + +/* + * Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks + * on this CPU. Use the algorithm recommended in the SDM to discover shared + * banks. + */ +static void cmci_discover(int banks, int boot) +{ + unsigned long *owned = (void *)&__get_cpu_var(mce_banks_owned); + unsigned long flags; + int hdr = 0; + int i; + + spin_lock_irqsave(&cmci_discover_lock, flags); + for (i = 0; i < banks; i++) { + u64 val; + + if (test_bit(i, owned)) + continue; + + rdmsrl(MSR_IA32_MC0_CTL2 + i, val); + + /* Already owned by someone else? */ + if (val & CMCI_EN) { + if (test_and_clear_bit(i, owned) || boot) + print_update("SHD", &hdr, i); + __clear_bit(i, __get_cpu_var(mce_poll_banks)); + continue; + } + + val |= CMCI_EN | CMCI_THRESHOLD; + wrmsrl(MSR_IA32_MC0_CTL2 + i, val); + rdmsrl(MSR_IA32_MC0_CTL2 + i, val); + + /* Did the enable bit stick? -- the bank supports CMCI */ + if (val & CMCI_EN) { + if (!test_and_set_bit(i, owned) || boot) + print_update("CMCI", &hdr, i); + __clear_bit(i, __get_cpu_var(mce_poll_banks)); + } else { + WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks))); + } + } + spin_unlock_irqrestore(&cmci_discover_lock, flags); + if (hdr) + printk(KERN_CONT "\n"); +} + +/* + * Just in case we missed an event during initialization check + * all the CMCI owned banks. + */ +void cmci_recheck(void) +{ + unsigned long flags; + int banks; + + if (!mce_available(¤t_cpu_data) || !cmci_supported(&banks)) + return; + local_irq_save(flags); + machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned)); + local_irq_restore(flags); +} + +/* + * Disable CMCI on this CPU for all banks it owns when it goes down. + * This allows other CPUs to claim the banks on rediscovery. + */ +void cmci_clear(void) +{ + unsigned long flags; + int i; + int banks; + u64 val; + + if (!cmci_supported(&banks)) + return; + spin_lock_irqsave(&cmci_discover_lock, flags); + for (i = 0; i < banks; i++) { + if (!test_bit(i, __get_cpu_var(mce_banks_owned))) + continue; + /* Disable CMCI */ + rdmsrl(MSR_IA32_MC0_CTL2 + i, val); + val &= ~(CMCI_EN|CMCI_THRESHOLD_MASK); + wrmsrl(MSR_IA32_MC0_CTL2 + i, val); + __clear_bit(i, __get_cpu_var(mce_banks_owned)); + } + spin_unlock_irqrestore(&cmci_discover_lock, flags); +} + +/* + * After a CPU went down cycle through all the others and rediscover + * Must run in process context. + */ +void cmci_rediscover(int dying) +{ + int banks; + int cpu; + cpumask_var_t old; + + if (!cmci_supported(&banks)) + return; + if (!alloc_cpumask_var(&old, GFP_KERNEL)) + return; + cpumask_copy(old, ¤t->cpus_allowed); + + for_each_online_cpu(cpu) { + if (cpu == dying) + continue; + if (set_cpus_allowed_ptr(current, cpumask_of(cpu))) + continue; + /* Recheck banks in case CPUs don't all have the same */ + if (cmci_supported(&banks)) + cmci_discover(banks, 0); + } + + set_cpus_allowed_ptr(current, old); + free_cpumask_var(old); +} + +/* + * Reenable CMCI on this CPU in case a CPU down failed. + */ +void cmci_reenable(void) +{ + int banks; + if (cmci_supported(&banks)) + cmci_discover(banks, 0); +} + +static void intel_init_cmci(void) +{ + int banks; + + if (!cmci_supported(&banks)) + return; + + mce_threshold_vector = intel_threshold_interrupt; + cmci_discover(banks, 1); + /* + * For CPU #0 this runs with still disabled APIC, but that's + * ok because only the vector is set up. We still do another + * check for the banks later for CPU #0 just to make sure + * to not miss any events. + */ + apic_write(APIC_LVTCMCI, THRESHOLD_APIC_VECTOR|APIC_DM_FIXED); + cmci_recheck(); +} + +void mce_intel_feature_init(struct cpuinfo_x86 *c) +{ + intel_init_thermal(c); + intel_init_cmci(); +} diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c b/arch/x86/kernel/cpu/mcheck/mce_intel_64.c deleted file mode 100644 index 663a88e8b3c..00000000000 --- a/arch/x86/kernel/cpu/mcheck/mce_intel_64.c +++ /dev/null @@ -1,225 +0,0 @@ -/* - * Intel specific MCE features. - * Copyright 2004 Zwane Mwaikambo - * Copyright (C) 2008, 2009 Intel Corporation - * Author: Andi Kleen - */ - -#include -#include -#include -#include -#include -#include - -/* - * Support for Intel Correct Machine Check Interrupts. This allows - * the CPU to raise an interrupt when a corrected machine check happened. - * Normally we pick those up using a regular polling timer. - * Also supports reliable discovery of shared banks. - */ - -static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned); - -/* - * cmci_discover_lock protects against parallel discovery attempts - * which could race against each other. - */ -static DEFINE_SPINLOCK(cmci_discover_lock); - -#define CMCI_THRESHOLD 1 - -static int cmci_supported(int *banks) -{ - u64 cap; - - if (mce_cmci_disabled || mce_ignore_ce) - return 0; - - /* - * Vendor check is not strictly needed, but the initial - * initialization is vendor keyed and this - * makes sure none of the backdoors are entered otherwise. - */ - if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL) - return 0; - if (!cpu_has_apic || lapic_get_maxlvt() < 6) - return 0; - rdmsrl(MSR_IA32_MCG_CAP, cap); - *banks = min_t(unsigned, MAX_NR_BANKS, cap & 0xff); - return !!(cap & MCG_CMCI_P); -} - -/* - * The interrupt handler. This is called on every event. - * Just call the poller directly to log any events. - * This could in theory increase the threshold under high load, - * but doesn't for now. - */ -static void intel_threshold_interrupt(void) -{ - machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned)); - mce_notify_irq(); -} - -static void print_update(char *type, int *hdr, int num) -{ - if (*hdr == 0) - printk(KERN_INFO "CPU %d MCA banks", smp_processor_id()); - *hdr = 1; - printk(KERN_CONT " %s:%d", type, num); -} - -/* - * Enable CMCI (Corrected Machine Check Interrupt) for available MCE banks - * on this CPU. Use the algorithm recommended in the SDM to discover shared - * banks. - */ -static void cmci_discover(int banks, int boot) -{ - unsigned long *owned = (void *)&__get_cpu_var(mce_banks_owned); - unsigned long flags; - int hdr = 0; - int i; - - spin_lock_irqsave(&cmci_discover_lock, flags); - for (i = 0; i < banks; i++) { - u64 val; - - if (test_bit(i, owned)) - continue; - - rdmsrl(MSR_IA32_MC0_CTL2 + i, val); - - /* Already owned by someone else? */ - if (val & CMCI_EN) { - if (test_and_clear_bit(i, owned) || boot) - print_update("SHD", &hdr, i); - __clear_bit(i, __get_cpu_var(mce_poll_banks)); - continue; - } - - val |= CMCI_EN | CMCI_THRESHOLD; - wrmsrl(MSR_IA32_MC0_CTL2 + i, val); - rdmsrl(MSR_IA32_MC0_CTL2 + i, val); - - /* Did the enable bit stick? -- the bank supports CMCI */ - if (val & CMCI_EN) { - if (!test_and_set_bit(i, owned) || boot) - print_update("CMCI", &hdr, i); - __clear_bit(i, __get_cpu_var(mce_poll_banks)); - } else { - WARN_ON(!test_bit(i, __get_cpu_var(mce_poll_banks))); - } - } - spin_unlock_irqrestore(&cmci_discover_lock, flags); - if (hdr) - printk(KERN_CONT "\n"); -} - -/* - * Just in case we missed an event during initialization check - * all the CMCI owned banks. - */ -void cmci_recheck(void) -{ - unsigned long flags; - int banks; - - if (!mce_available(¤t_cpu_data) || !cmci_supported(&banks)) - return; - local_irq_save(flags); - machine_check_poll(MCP_TIMESTAMP, &__get_cpu_var(mce_banks_owned)); - local_irq_restore(flags); -} - -/* - * Disable CMCI on this CPU for all banks it owns when it goes down. - * This allows other CPUs to claim the banks on rediscovery. - */ -void cmci_clear(void) -{ - unsigned long flags; - int i; - int banks; - u64 val; - - if (!cmci_supported(&banks)) - return; - spin_lock_irqsave(&cmci_discover_lock, flags); - for (i = 0; i < banks; i++) { - if (!test_bit(i, __get_cpu_var(mce_banks_owned))) - continue; - /* Disable CMCI */ - rdmsrl(MSR_IA32_MC0_CTL2 + i, val); - val &= ~(CMCI_EN|CMCI_THRESHOLD_MASK); - wrmsrl(MSR_IA32_MC0_CTL2 + i, val); - __clear_bit(i, __get_cpu_var(mce_banks_owned)); - } - spin_unlock_irqrestore(&cmci_discover_lock, flags); -} - -/* - * After a CPU went down cycle through all the others and rediscover - * Must run in process context. - */ -void cmci_rediscover(int dying) -{ - int banks; - int cpu; - cpumask_var_t old; - - if (!cmci_supported(&banks)) - return; - if (!alloc_cpumask_var(&old, GFP_KERNEL)) - return; - cpumask_copy(old, ¤t->cpus_allowed); - - for_each_online_cpu(cpu) { - if (cpu == dying) - continue; - if (set_cpus_allowed_ptr(current, cpumask_of(cpu))) - continue; - /* Recheck banks in case CPUs don't all have the same */ - if (cmci_supported(&banks)) - cmci_discover(banks, 0); - } - - set_cpus_allowed_ptr(current, old); - free_cpumask_var(old); -} - -/* - * Reenable CMCI on this CPU in case a CPU down failed. - */ -void cmci_reenable(void) -{ - int banks; - if (cmci_supported(&banks)) - cmci_discover(banks, 0); -} - -static void intel_init_cmci(void) -{ - int banks; - - if (!cmci_supported(&banks)) - return; - - mce_threshold_vector = intel_threshold_interrupt; - cmci_discover(banks, 1); - /* - * For CPU #0 this runs with still disabled APIC, but that's - * ok because only the vector is set up. We still do another - * check for the banks later for CPU #0 just to make sure - * to not miss any events. - */ - apic_write(APIC_LVTCMCI, THRESHOLD_APIC_VECTOR|APIC_DM_FIXED); - cmci_recheck(); -} - -void mce_intel_feature_init(struct cpuinfo_x86 *c) -{ - intel_init_thermal(c); - intel_init_cmci(); -} -- cgit v1.2.3