diff options
| -rw-r--r-- | Documentation/x86/x86_64/boot-options.txt | 5 | ||||
| -rw-r--r-- | arch/x86/include/asm/acpi.h | 2 | ||||
| -rw-r--r-- | arch/x86/include/asm/mce.h | 16 | ||||
| -rw-r--r-- | arch/x86/kernel/acpi/boot.c | 5 | ||||
| -rw-r--r-- | arch/x86/kernel/cpu/mcheck/mce-internal.h | 3 | ||||
| -rw-r--r-- | arch/x86/kernel/cpu/mcheck/mce.c | 28 | ||||
| -rw-r--r-- | arch/x86/kernel/cpu/mcheck/mce_intel.c | 42 | ||||
| -rw-r--r-- | drivers/acpi/apei/erst.c | 51 | ||||
| -rw-r--r-- | drivers/acpi/apei/ghes.c | 38 | ||||
| -rw-r--r-- | drivers/acpi/apei/hest.c | 39 | ||||
| -rw-r--r-- | include/linux/mm.h | 1 | ||||
| -rw-r--r-- | mm/memory-failure.c | 5 |
12 files changed, 185 insertions, 50 deletions
diff --git a/Documentation/x86/x86_64/boot-options.txt b/Documentation/x86/x86_64/boot-options.txt index e9e8ddbbf376..1228b22e142b 100644 --- a/Documentation/x86/x86_64/boot-options.txt +++ b/Documentation/x86/x86_64/boot-options.txt @@ -176,6 +176,11 @@ ACPI acpi=noirq Don't route interrupts + acpi=nocmcff Disable firmware first mode for corrected errors. This + disables parsing the HEST CMC error source to check if + firmware has set the FF flag. This may result in + duplicate corrected error reports. + PCI pci=off Don't use PCI diff --git a/arch/x86/include/asm/acpi.h b/arch/x86/include/asm/acpi.h index 2dfac58f3b11..b1977bad5435 100644 --- a/arch/x86/include/asm/acpi.h +++ b/arch/x86/include/asm/acpi.h @@ -86,6 +86,7 @@ extern int acpi_pci_disabled; extern int acpi_skip_timer_override; extern int acpi_use_timer_override; extern int acpi_fix_pin2_polarity; +extern int acpi_disable_cmcff; extern u8 acpi_sci_flags; extern int acpi_sci_override_gsi; @@ -168,6 +169,7 @@ static inline void arch_acpi_set_pdc_bits(u32 *buf) #define acpi_lapic 0 #define acpi_ioapic 0 +#define acpi_disable_cmcff 0 static inline void acpi_noirq_set(void) { } static inline void acpi_disable_pci(void) { } static inline void disable_acpi(void) { } diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h index 29e3093bbd21..cbe6b9e404ce 100644 --- a/arch/x86/include/asm/mce.h +++ b/arch/x86/include/asm/mce.h @@ -32,11 +32,20 @@ #define MCI_STATUS_PCC (1ULL<<57) /* processor context corrupt */ #define MCI_STATUS_S (1ULL<<56) /* Signaled machine check */ #define MCI_STATUS_AR (1ULL<<55) /* Action required */ -#define MCACOD 0xffff /* MCA Error Code */ + +/* + * Note that the full MCACOD field of IA32_MCi_STATUS MSR is + * bits 15:0. But bit 12 is the 'F' bit, defined for corrected + * errors to indicate that errors are being filtered by hardware. + * We should mask out bit 12 when looking for specific signatures + * of uncorrected errors - so the F bit is deliberately skipped + * in this #define. + */ +#define MCACOD 0xefff /* MCA Error Code */ /* Architecturally defined codes from SDM Vol. 3B Chapter 15 */ #define MCACOD_SCRUB 0x00C0 /* 0xC0-0xCF Memory Scrubbing */ -#define MCACOD_SCRUBMSK 0xfff0 +#define MCACOD_SCRUBMSK 0xeff0 /* Skip bit 12 ('F' bit) */ #define MCACOD_L3WB 0x017A /* L3 Explicit Writeback */ #define MCACOD_DATA 0x0134 /* Data Load */ #define MCACOD_INSTR 0x0150 /* Instruction Fetch */ @@ -188,6 +197,9 @@ extern void register_mce_write_callback(ssize_t (*)(struct file *filp, const char __user *ubuf, size_t usize, loff_t *off)); +/* Disable CMCI/polling for MCA bank claimed by firmware */ +extern void mce_disable_bank(int bank); + /* * Exception handler */ diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 2627a81253ee..fead83292e6e 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -67,6 +67,7 @@ EXPORT_SYMBOL(acpi_pci_disabled); int acpi_lapic; int acpi_ioapic; int acpi_strict; +int acpi_disable_cmcff; u8 acpi_sci_flags __initdata; int acpi_sci_override_gsi __initdata; @@ -1626,6 +1627,10 @@ static int __init parse_acpi(char *arg) /* "acpi=copy_dsdt" copys DSDT */ else if (strcmp(arg, "copy_dsdt") == 0) { acpi_gbl_copy_dsdt_locally = 1; + } + /* "acpi=nocmcff" disables FF mode for corrected errors */ + else if (strcmp(arg, "nocmcff") == 0) { + acpi_disable_cmcff = 1; } else { /* Core will printk when we return error. */ return -EINVAL; diff --git a/arch/x86/kernel/cpu/mcheck/mce-internal.h b/arch/x86/kernel/cpu/mcheck/mce-internal.h index 5b7d4fa5d3b7..09edd0b65fef 100644 --- a/arch/x86/kernel/cpu/mcheck/mce-internal.h +++ b/arch/x86/kernel/cpu/mcheck/mce-internal.h @@ -25,15 +25,18 @@ int mce_severity(struct mce *a, int tolerant, char **msg); struct dentry *mce_get_debugfs_dir(void); extern struct mce_bank *mce_banks; +extern mce_banks_t mce_banks_ce_disabled; #ifdef CONFIG_X86_MCE_INTEL unsigned long mce_intel_adjust_timer(unsigned long interval); void mce_intel_cmci_poll(void); void mce_intel_hcpu_update(unsigned long cpu); +void cmci_disable_bank(int bank); #else # define mce_intel_adjust_timer mce_adjust_timer_default static inline void mce_intel_cmci_poll(void) { } static inline void mce_intel_hcpu_update(unsigned long cpu) { } +static inline void cmci_disable_bank(int bank) { } #endif void mce_timer_kick(unsigned long interval); diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 87a65c939bcd..b3218cdee95f 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -97,6 +97,15 @@ DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = { [0 ... BITS_TO_LONGS(MAX_NR_BANKS)-1] = ~0UL }; +/* + * MCA banks controlled through firmware first for corrected errors. + * This is a global list of banks for which we won't enable CMCI and we + * won't poll. Firmware controls these banks and is responsible for + * reporting corrected errors through GHES. Uncorrected/recoverable + * errors are still notified through a machine check. + */ +mce_banks_t mce_banks_ce_disabled; + static DEFINE_PER_CPU(struct work_struct, mce_work); static void (*quirk_no_way_out)(int bank, struct mce *m, struct pt_regs *regs); @@ -1935,6 +1944,25 @@ static struct miscdevice mce_chrdev_device = { &mce_chrdev_ops, }; +static void __mce_disable_bank(void *arg) +{ + int bank = *((int *)arg); + __clear_bit(bank, __get_cpu_var(mce_poll_banks)); + cmci_disable_bank(bank); +} + +void mce_disable_bank(int bank) +{ + if (bank >= mca_cfg.banks) { + pr_warn(FW_BUG + "Ignoring request to disable invalid MCA bank %d.\n", + bank); + return; + } + set_bit(bank, mce_banks_ce_disabled); + on_each_cpu(__mce_disable_bank, &bank, 1); +} + /* * mce=off Disables machine check * mce=no_cmci Disables CMCI diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c index d56405309dc1..4cfe0458ca66 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_intel.c +++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c @@ -203,6 +203,10 @@ static void cmci_discover(int banks) if (test_bit(i, owned)) continue; + /* Skip banks in firmware first mode */ + if (test_bit(i, mce_banks_ce_disabled)) + continue; + rdmsrl(MSR_IA32_MCx_CTL2(i), val); /* Already owned by someone else? */ @@ -271,6 +275,19 @@ void cmci_recheck(void) local_irq_restore(flags); } +/* Caller must hold the lock on cmci_discover_lock */ +static void __cmci_disable_bank(int bank) +{ + u64 val; + + if (!test_bit(bank, __get_cpu_var(mce_banks_owned))) + return; + rdmsrl(MSR_IA32_MCx_CTL2(bank), val); + val &= ~MCI_CTL2_CMCI_EN; + wrmsrl(MSR_IA32_MCx_CTL2(bank), val); + __clear_bit(bank, __get_cpu_var(mce_banks_owned)); +} + /* * Disable CMCI on this CPU for all banks it owns when it goes down. * This allows other CPUs to claim the banks on rediscovery. @@ -280,20 +297,12 @@ void cmci_clear(void) unsigned long flags; int i; int banks; - u64 val; if (!cmci_supported(&banks)) return; raw_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_MCx_CTL2(i), val); - val &= ~MCI_CTL2_CMCI_EN; - wrmsrl(MSR_IA32_MCx_CTL2(i), val); - __clear_bit(i, __get_cpu_var(mce_banks_owned)); - } + for (i = 0; i < banks; i++) + __cmci_disable_bank(i); raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); } @@ -327,6 +336,19 @@ void cmci_reenable(void) cmci_discover(banks); } +void cmci_disable_bank(int bank) +{ + int banks; + unsigned long flags; + + if (!cmci_supported(&banks)) + return; + + raw_spin_lock_irqsave(&cmci_discover_lock, flags); + __cmci_disable_bank(bank); + raw_spin_unlock_irqrestore(&cmci_discover_lock, flags); +} + static void intel_init_cmci(void) { int banks; diff --git a/drivers/acpi/apei/erst.c b/drivers/acpi/apei/erst.c index 88d0b0f9f92b..40ffcdafbc6f 100644 --- a/drivers/acpi/apei/erst.c +++ b/drivers/acpi/apei/erst.c @@ -39,7 +39,8 @@ #include "apei-internal.h" -#define ERST_PFX "ERST: " +#undef pr_fmt +#define pr_fmt(fmt) "ERST: " fmt /* ERST command status */ #define ERST_STATUS_SUCCESS 0x0 @@ -109,8 +110,7 @@ static inline int erst_errno(int command_status) static int erst_timedout(u64 *t, u64 spin_unit) { if ((s64)*t < spin_unit) { - pr_warning(FW_WARN ERST_PFX - "Firmware does not respond in time\n"); + pr_warn(FW_WARN "Firmware does not respond in time.\n"); return 1; } *t -= spin_unit; @@ -186,8 +186,8 @@ static int erst_exec_stall(struct apei_exec_context *ctx, if (ctx->value > FIRMWARE_MAX_STALL) { if (!in_nmi()) - pr_warning(FW_WARN ERST_PFX - "Too long stall time for stall instruction: %llx.\n", + pr_warn(FW_WARN + "Too long stall time for stall instruction: 0x%llx.\n", ctx->value); stall_time = FIRMWARE_MAX_STALL; } else @@ -206,8 +206,8 @@ static int erst_exec_stall_while_true(struct apei_exec_context *ctx, if (ctx->var1 > FIRMWARE_MAX_STALL) { if (!in_nmi()) - pr_warning(FW_WARN ERST_PFX - "Too long stall time for stall while true instruction: %llx.\n", + pr_warn(FW_WARN + "Too long stall time for stall while true instruction: 0x%llx.\n", ctx->var1); stall_time = FIRMWARE_MAX_STALL; } else @@ -271,8 +271,7 @@ static int erst_exec_move_data(struct apei_exec_context *ctx, /* ioremap does not work in interrupt context */ if (in_interrupt()) { - pr_warning(ERST_PFX - "MOVE_DATA can not be used in interrupt context"); + pr_warn("MOVE_DATA can not be used in interrupt context.\n"); return -EBUSY; } @@ -522,8 +521,7 @@ retry: ERST_RECORD_ID_CACHE_SIZE_MAX); if (new_size <= erst_record_id_cache.size) { if (printk_ratelimit()) - pr_warning(FW_WARN ERST_PFX - "too many record ID!\n"); + pr_warn(FW_WARN "too many record IDs!\n"); return 0; } alloc_size = new_size * sizeof(entries[0]); @@ -759,8 +757,7 @@ static int __erst_clear_from_storage(u64 record_id) static void pr_unimpl_nvram(void) { if (printk_ratelimit()) - pr_warning(ERST_PFX - "NVRAM ERST Log Address Range is not implemented yet\n"); + pr_warn("NVRAM ERST Log Address Range not implemented yet.\n"); } static int __erst_write_to_nvram(const struct cper_record_header *record) @@ -1120,7 +1117,7 @@ static int __init erst_init(void) goto err; if (erst_disable) { - pr_info(ERST_PFX + pr_info( "Error Record Serialization Table (ERST) support is disabled.\n"); goto err; } @@ -1131,14 +1128,14 @@ static int __init erst_init(void) goto err; else if (ACPI_FAILURE(status)) { const char *msg = acpi_format_exception(status); - pr_err(ERST_PFX "Failed to get table, %s\n", msg); + pr_err("Failed to get table, %s\n", msg); rc = -EINVAL; goto err; } rc = erst_check_table(erst_tab); if (rc) { - pr_err(FW_BUG ERST_PFX "ERST table is invalid\n"); + pr_err(FW_BUG "ERST table is invalid.\n"); goto err; } @@ -1156,21 +1153,19 @@ static int __init erst_init(void) rc = erst_get_erange(&erst_erange); if (rc) { if (rc == -ENODEV) - pr_info(ERST_PFX + pr_info( "The corresponding hardware device or firmware implementation " "is not available.\n"); else - pr_err(ERST_PFX - "Failed to get Error Log Address Range.\n"); + pr_err("Failed to get Error Log Address Range.\n"); goto err_unmap_reg; } r = request_mem_region(erst_erange.base, erst_erange.size, "APEI ERST"); if (!r) { - pr_err(ERST_PFX - "Can not request iomem region <0x%16llx-0x%16llx> for ERST.\n", - (unsigned long long)erst_erange.base, - (unsigned long long)erst_erange.base + erst_erange.size); + pr_err("Can not request [mem %#010llx-%#010llx] for ERST.\n", + (unsigned long long)erst_erange.base, + (unsigned long long)erst_erange.base + erst_erange.size - 1); rc = -EIO; goto err_unmap_reg; } @@ -1180,7 +1175,7 @@ static int __init erst_init(void) if (!erst_erange.vaddr) goto err_release_erange; - pr_info(ERST_PFX + pr_info( "Error Record Serialization Table (ERST) support is initialized.\n"); buf = kmalloc(erst_erange.size, GFP_KERNEL); @@ -1192,15 +1187,15 @@ static int __init erst_init(void) rc = pstore_register(&erst_info); if (rc) { if (rc != -EPERM) - pr_info(ERST_PFX - "Could not register with persistent store\n"); + pr_info( + "Could not register with persistent store.\n"); erst_info.buf = NULL; erst_info.bufsize = 0; kfree(buf); } } else - pr_err(ERST_PFX - "Failed to allocate %lld bytes for persistent store error log\n", + pr_err( + "Failed to allocate %lld bytes for persistent store error log.\n", erst_erange.size); return 0; diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c index ec9b57d428a1..8ec37bbdd699 100644 --- a/drivers/acpi/apei/ghes.c +++ b/drivers/acpi/apei/ghes.c @@ -409,6 +409,34 @@ static void ghes_clear_estatus(struct ghes *ghes) ghes->flags &= ~GHES_TO_CLEAR; } +static void ghes_handle_memory_failure(struct acpi_hest_generic_data *gdata, int sev) +{ +#ifdef CONFIG_ACPI_APEI_MEMORY_FAILURE + unsigned long pfn; + int sec_sev = ghes_severity(gdata->error_severity); + struct cper_sec_mem_err *mem_err; + mem_err = (struct cper_sec_mem_err *)(gdata + 1); + + if (sec_sev == GHES_SEV_CORRECTED && + (gdata->flags & CPER_SEC_ERROR_THRESHOLD_EXCEEDED) && + (mem_err->validation_bits & CPER_MEM_VALID_PHYSICAL_ADDRESS)) { + pfn = mem_err->physical_addr >> PAGE_SHIFT; + if (pfn_valid(pfn)) + memory_failure_queue(pfn, 0, MF_SOFT_OFFLINE); + else if (printk_ratelimit()) + pr_warn(FW_WARN GHES_PFX + "Invalid address in generic error data: %#llx\n", + mem_err->physical_addr); + } + if (sev == GHES_SEV_RECOVERABLE && + sec_sev == GHES_SEV_RECOVERABLE && + mem_err->validation_bits & CPER_MEM_VALID_PHYSICAL_ADDRESS) { + pfn = mem_err->physical_addr >> PAGE_SHIFT; + memory_failure_queue(pfn, 0, 0); + } +#endif +} + static void ghes_do_proc(struct ghes *ghes, const struct acpi_hest_generic_status *estatus) { @@ -428,15 +456,7 @@ static void ghes_do_proc(struct ghes *ghes, apei_mce_report_mem_error(sev == GHES_SEV_CORRECTED, mem_err); #endif -#ifdef CONFIG_ACPI_APEI_MEMORY_FAILURE - if (sev == GHES_SEV_RECOVERABLE && - sec_sev == GHES_SEV_RECOVERABLE && - mem_err->validation_bits & CPER_MEM_VALID_PHYSICAL_ADDRESS) { - unsigned long pfn; - pfn = mem_err->physical_addr >> PAGE_SHIFT; - memory_failure_queue(pfn, 0, 0); - } -#endif + ghes_handle_memory_failure(gdata, sev); } #ifdef CONFIG_ACPI_APEI_PCIEAER else if (!uuid_le_cmp(*(uuid_le *)gdata->section_type, diff --git a/drivers/acpi/apei/hest.c b/drivers/acpi/apei/hest.c index f5ef5d54e4ac..f5e37f32c71f 100644 --- a/drivers/acpi/apei/hest.c +++ b/drivers/acpi/apei/hest.c @@ -36,6 +36,7 @@ #include <linux/io.h> #include <linux/platform_device.h> #include <acpi/apei.h> +#include <asm/mce.h> #include "apei-internal.h" @@ -121,6 +122,41 @@ int apei_hest_parse(apei_hest_func_t func, void *data) } EXPORT_SYMBOL_GPL(apei_hest_parse); +/* + * Check if firmware advertises firmware first mode. We need FF bit to be set + * along with a set of MC banks which work in FF mode. + */ +static int __init hest_parse_cmc(struct acpi_hest_header *hest_hdr, void *data) +{ +#ifdef CONFIG_X86_MCE + int i; + struct acpi_hest_ia_corrected *cmc; + struct acpi_hest_ia_error_bank *mc_bank; + + if (hest_hdr->type != ACPI_HEST_TYPE_IA32_CORRECTED_CHECK) + return 0; + + cmc = (struct acpi_hest_ia_corrected *)hest_hdr; + if (!cmc->enabled) + return 0; + + /* + * We expect HEST to provide a list of MC banks that report errors + * in firmware first mode. Otherwise, return non-zero value to + * indicate that we are done parsing HEST. + */ + if (!(cmc->flags & ACPI_HEST_FIRMWARE_FIRST) || !cmc->num_hardware_banks) + return 1; + + pr_info(HEST_PFX "Enabling Firmware First mode for corrected errors.\n"); + + mc_bank = (struct acpi_hest_ia_error_bank *)(cmc + 1); + for (i = 0; i < cmc->num_hardware_banks; i++, mc_bank++) + mce_disable_bank(mc_bank->bank_number); +#endif + return 1; +} + struct ghes_arr { struct platform_device **ghes_devs; unsigned int count; @@ -227,6 +263,9 @@ void __init acpi_hest_init(void) goto err; } + if (!acpi_disable_cmcff) + apei_hest_parse(hest_parse_cmc, NULL); + if (!ghes_disable) { rc = apei_hest_parse(hest_parse_ghes_count, &ghes_count); if (rc) diff --git a/include/linux/mm.h b/include/linux/mm.h index f0224608d15e..d2d59b4149d0 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1798,6 +1798,7 @@ enum mf_flags { MF_COUNT_INCREASED = 1 << 0, MF_ACTION_REQUIRED = 1 << 1, MF_MUST_KILL = 1 << 2, + MF_SOFT_OFFLINE = 1 << 3, }; extern int memory_failure(unsigned long pfn, int trapno, int flags); extern void memory_failure_queue(unsigned long pfn, int trapno, int flags); diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 2c13aa7a0164..55d7c8026ab0 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -1286,7 +1286,10 @@ static void memory_failure_work_func(struct work_struct *work) spin_unlock_irqrestore(&mf_cpu->lock, proc_flags); if (!gotten) break; - memory_failure(entry.pfn, entry.trapno, entry.flags); + if (entry.flags & MF_SOFT_OFFLINE) + soft_offline_page(pfn_to_page(entry.pfn), entry.flags); + else + memory_failure(entry.pfn, entry.trapno, entry.flags); } } |
