From cd42f4a3b2b1c4cbd997363dc57821953d73fd87 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Thu, 15 Dec 2011 10:48:12 -0800 Subject: HWPOISON: Clean up memory_failure() vs. __memory_failure() There is only one caller of memory_failure(), all other users call __memory_failure() and pass in the flags argument explicitly. The lone user of memory_failure() will soon need to pass flags too. Add flags argument to the callsite in mce.c. Delete the old memory_failure() function, and then rename __memory_failure() without the leading "__". Provide clearer message when action optional memory errors are ignored. Acked-by: Borislav Petkov Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce.c | 12 +++++++---- drivers/base/memory.c | 2 +- include/linux/mm.h | 3 +-- mm/hwpoison-inject.c | 4 ++-- mm/madvise.c | 2 +- mm/memory-failure.c | 46 ++++++++++++++++++---------------------- 6 files changed, 34 insertions(+), 35 deletions(-) diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 2af127d4c3d1..1a08ce5f345f 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -1046,11 +1046,15 @@ out: } 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) +#ifndef CONFIG_MEMORY_FAILURE +int memory_failure(unsigned long pfn, int vector, int flags) { - printk(KERN_ERR "Action optional memory failure at %lx ignored\n", pfn); + printk(KERN_ERR "Uncorrected memory error in page 0x%lx ignored\n" + "Rebuild kernel with CONFIG_MEMORY_FAILURE=y for smarter handling\n", pfn); + + return 0; } +#endif /* * Called after mce notification in process context. This code @@ -1068,7 +1072,7 @@ void mce_notify_process(void) unsigned long pfn; mce_notify_irq(); while (mce_ring_get(&pfn)) - memory_failure(pfn, MCE_VECTOR); + memory_failure(pfn, MCE_VECTOR, 0); } static void mce_process_work(struct work_struct *dummy) diff --git a/drivers/base/memory.c b/drivers/base/memory.c index 8272d92d22c0..9a924440053f 100644 --- a/drivers/base/memory.c +++ b/drivers/base/memory.c @@ -474,7 +474,7 @@ store_hard_offline_page(struct class *class, if (strict_strtoull(buf, 0, &pfn) < 0) return -EINVAL; pfn >>= PAGE_SHIFT; - ret = __memory_failure(pfn, 0, 0); + ret = memory_failure(pfn, 0, 0); return ret ? ret : count; } diff --git a/include/linux/mm.h b/include/linux/mm.h index 4baadd18f4ad..bcc523474724 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1607,8 +1607,7 @@ void vmemmap_populate_print_last(void); enum mf_flags { MF_COUNT_INCREASED = 1 << 0, }; -extern void memory_failure(unsigned long pfn, int trapno); -extern int __memory_failure(unsigned long pfn, int trapno, int flags); +extern int memory_failure(unsigned long pfn, int trapno, int flags); extern void memory_failure_queue(unsigned long pfn, int trapno, int flags); extern int unpoison_memory(unsigned long pfn); extern int sysctl_memory_failure_early_kill; diff --git a/mm/hwpoison-inject.c b/mm/hwpoison-inject.c index c7fc7fd00e32..cc448bb983ba 100644 --- a/mm/hwpoison-inject.c +++ b/mm/hwpoison-inject.c @@ -45,7 +45,7 @@ static int hwpoison_inject(void *data, u64 val) * do a racy check with elevated page count, to make sure PG_hwpoison * will only be set for the targeted owner (or on a free page). * We temporarily take page lock for try_get_mem_cgroup_from_page(). - * __memory_failure() will redo the check reliably inside page lock. + * memory_failure() will redo the check reliably inside page lock. */ lock_page(hpage); err = hwpoison_filter(hpage); @@ -55,7 +55,7 @@ static int hwpoison_inject(void *data, u64 val) inject: printk(KERN_INFO "Injecting memory failure at pfn %lx\n", pfn); - return __memory_failure(pfn, 18, MF_COUNT_INCREASED); + return memory_failure(pfn, 18, MF_COUNT_INCREASED); } static int hwpoison_unpoison(void *data, u64 val) diff --git a/mm/madvise.c b/mm/madvise.c index 74bf193eff04..f5ab745672b7 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -251,7 +251,7 @@ static int madvise_hwpoison(int bhv, unsigned long start, unsigned long end) printk(KERN_INFO "Injecting memory failure for page %lx at %lx\n", page_to_pfn(p), start); /* Ignore return value for now */ - __memory_failure(page_to_pfn(p), 0, MF_COUNT_INCREASED); + memory_failure(page_to_pfn(p), 0, MF_COUNT_INCREASED); } return ret; } diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 06d3479513aa..ab259bb0adc5 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -984,7 +984,25 @@ static void clear_page_hwpoison_huge_page(struct page *hpage) ClearPageHWPoison(hpage + i); } -int __memory_failure(unsigned long pfn, int trapno, int flags) +/** + * memory_failure - Handle memory failure of a page. + * @pfn: Page Number of the corrupted page + * @trapno: Trap number reported in the signal to user space. + * @flags: fine tune action taken + * + * This function is called by the low level machine check code + * of an architecture when it detects hardware memory corruption + * of a page. It tries its best to recover, which includes + * dropping pages, killing processes etc. + * + * The function is primarily of use for corruptions that + * happen outside the current execution context (e.g. when + * detected by a background scrubber) + * + * Must run in process context (e.g. a work queue) with interrupts + * enabled and no spinlocks hold. + */ +int memory_failure(unsigned long pfn, int trapno, int flags) { struct page_state *ps; struct page *p; @@ -1156,29 +1174,7 @@ out: unlock_page(hpage); return res; } -EXPORT_SYMBOL_GPL(__memory_failure); - -/** - * memory_failure - Handle memory failure of a page. - * @pfn: Page Number of the corrupted page - * @trapno: Trap number reported in the signal to user space. - * - * This function is called by the low level machine check code - * of an architecture when it detects hardware memory corruption - * of a page. It tries its best to recover, which includes - * dropping pages, killing processes etc. - * - * The function is primarily of use for corruptions that - * happen outside the current execution context (e.g. when - * detected by a background scrubber) - * - * Must run in process context (e.g. a work queue) with interrupts - * enabled and no spinlocks hold. - */ -void memory_failure(unsigned long pfn, int trapno) -{ - __memory_failure(pfn, trapno, 0); -} +EXPORT_SYMBOL_GPL(memory_failure); #define MEMORY_FAILURE_FIFO_ORDER 4 #define MEMORY_FAILURE_FIFO_SIZE (1 << MEMORY_FAILURE_FIFO_ORDER) @@ -1251,7 +1247,7 @@ 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); + memory_failure(entry.pfn, entry.trapno, entry.flags); } } -- cgit v1.2.3-58-ga151 From 7329bbeb92740f35d64a8860ae7837ff4db27fe0 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Tue, 13 Dec 2011 09:27:58 -0800 Subject: HWPOISON: Add code to handle "action required" errors. Add new flag bit "MF_ACTION_REQUIRED" to be used by machine check code to force a signal with si_code = BUS_MCEERR_AR in the case where the error occurs in processor execution context. Pass the flags argument along call chain: memory_failure() hwpoison_user_mappings() kill_procs() kill_proc() Drop the "_ao" suffix from kill_procs_ao() and kill_proc_ao() since they can now handle "action required" as well as "action optional" errors. Acked-by: Borislav Petkov Signed-off-by: Tony Luck --- include/linux/mm.h | 1 + mm/memory-failure.c | 50 +++++++++++++++++++++++++++++--------------------- 2 files changed, 30 insertions(+), 21 deletions(-) diff --git a/include/linux/mm.h b/include/linux/mm.h index bcc523474724..bf169ca69812 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1606,6 +1606,7 @@ void vmemmap_populate_print_last(void); enum mf_flags { MF_COUNT_INCREASED = 1 << 0, + MF_ACTION_REQUIRED = 1 << 1, }; 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 ab259bb0adc5..95fd307ebb30 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -187,33 +187,40 @@ int hwpoison_filter(struct page *p) EXPORT_SYMBOL_GPL(hwpoison_filter); /* - * Send all the processes who have the page mapped an ``action optional'' - * signal. + * Send all the processes who have the page mapped a signal. + * ``action optional'' if they are not immediately affected by the error + * ``action required'' if error happened in current execution context */ -static int kill_proc_ao(struct task_struct *t, unsigned long addr, int trapno, - unsigned long pfn, struct page *page) +static int kill_proc(struct task_struct *t, unsigned long addr, int trapno, + unsigned long pfn, struct page *page, int flags) { struct siginfo si; int ret; printk(KERN_ERR - "MCE %#lx: Killing %s:%d early due to hardware memory corruption\n", + "MCE %#lx: Killing %s:%d due to hardware memory corruption\n", pfn, t->comm, t->pid); si.si_signo = SIGBUS; si.si_errno = 0; - si.si_code = BUS_MCEERR_AO; si.si_addr = (void *)addr; #ifdef __ARCH_SI_TRAPNO si.si_trapno = trapno; #endif si.si_addr_lsb = compound_trans_order(compound_head(page)) + PAGE_SHIFT; - /* - * Don't use force here, it's convenient if the signal - * can be temporarily blocked. - * This could cause a loop when the user sets SIGBUS - * to SIG_IGN, but hopefully no one will do that? - */ - ret = send_sig_info(SIGBUS, &si, t); /* synchronous? */ + + if ((flags & MF_ACTION_REQUIRED) && t == current) { + si.si_code = BUS_MCEERR_AR; + ret = force_sig_info(SIGBUS, &si, t); + } else { + /* + * Don't use force here, it's convenient if the signal + * can be temporarily blocked. + * This could cause a loop when the user sets SIGBUS + * to SIG_IGN, but hopefully no one will do that? + */ + si.si_code = BUS_MCEERR_AO; + ret = send_sig_info(SIGBUS, &si, t); /* synchronous? */ + } if (ret < 0) printk(KERN_INFO "MCE: Error sending signal to %s:%d: %d\n", t->comm, t->pid, ret); @@ -338,8 +345,9 @@ static void add_to_kill(struct task_struct *tsk, struct page *p, * Also when FAIL is set do a force kill because something went * wrong earlier. */ -static void kill_procs_ao(struct list_head *to_kill, int doit, int trapno, - int fail, struct page *page, unsigned long pfn) +static void kill_procs(struct list_head *to_kill, int doit, int trapno, + int fail, struct page *page, unsigned long pfn, + int flags) { struct to_kill *tk, *next; @@ -363,8 +371,8 @@ static void kill_procs_ao(struct list_head *to_kill, int doit, int trapno, * check for that, but we need to tell the * process anyways. */ - else if (kill_proc_ao(tk->tsk, tk->addr, trapno, - pfn, page) < 0) + else if (kill_proc(tk->tsk, tk->addr, trapno, + pfn, page, flags) < 0) printk(KERN_ERR "MCE %#lx: Cannot send advisory machine check signal to %s:%d\n", pfn, tk->tsk->comm, tk->tsk->pid); @@ -844,7 +852,7 @@ static int page_action(struct page_state *ps, struct page *p, * the pages and send SIGBUS to the processes if the data was dirty. */ static int hwpoison_user_mappings(struct page *p, unsigned long pfn, - int trapno) + int trapno, int flags) { enum ttu_flags ttu = TTU_UNMAP | TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS; struct address_space *mapping; @@ -962,8 +970,8 @@ static int hwpoison_user_mappings(struct page *p, unsigned long pfn, * use a more force-full uncatchable kill to prevent * any accesses to the poisoned memory. */ - kill_procs_ao(&tokill, !!PageDirty(ppage), trapno, - ret != SWAP_SUCCESS, p, pfn); + kill_procs(&tokill, !!PageDirty(ppage), trapno, + ret != SWAP_SUCCESS, p, pfn, flags); return ret; } @@ -1148,7 +1156,7 @@ int memory_failure(unsigned long pfn, int trapno, int flags) * Now take care of user space mappings. * Abort on fail: __delete_from_page_cache() assumes unmapped page. */ - if (hwpoison_user_mappings(p, pfn, trapno) != SWAP_SUCCESS) { + if (hwpoison_user_mappings(p, pfn, trapno, flags) != SWAP_SUCCESS) { printk(KERN_ERR "MCE %#lx: cannot unmap page, give up\n", pfn); res = -EBUSY; goto out; -- cgit v1.2.3-58-ga151 From 85f92694affa7dba7f1978666a69552b5dfc628e Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Tue, 13 Dec 2011 09:48:13 -0800 Subject: x86/mce: Create helper function to save addr/misc when needed The MCI_STATUS_MISCV and MCI_STATUS_ADDRV bits in the bank status registers define whether the MISC and ADDR registers respectively contain valid data - provide a helper function to check these bits and read the registers when needed. In addition, processors that support software error recovery (as indicated by the MCG_SER_P bit in the MCG_CAP register) may include some undefined bits in the ADDR register - mask these out. Acked-by: Borislav Petkov Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce.c | 31 +++++++++++++++++++++++-------- 1 file changed, 23 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 1a08ce5f345f..2f1c200f05e6 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -492,6 +492,27 @@ static void mce_report_event(struct pt_regs *regs) irq_work_queue(&__get_cpu_var(mce_irq_work)); } +/* + * Read ADDR and MISC registers. + */ +static void mce_read_aux(struct mce *m, int i) +{ + if (m->status & MCI_STATUS_MISCV) + m->misc = mce_rdmsrl(MSR_IA32_MCx_MISC(i)); + if (m->status & MCI_STATUS_ADDRV) { + m->addr = mce_rdmsrl(MSR_IA32_MCx_ADDR(i)); + + /* + * Mask the reported address by the reported granularity. + */ + if (mce_ser && (m->status & MCI_STATUS_MISCV)) { + u8 shift = MCI_MISC_ADDR_LSB(m->misc); + m->addr >>= shift; + m->addr <<= shift; + } + } +} + DEFINE_PER_CPU(unsigned, mce_poll_count); /* @@ -542,10 +563,7 @@ void machine_check_poll(enum mcp_flags flags, mce_banks_t *b) (m.status & (mce_ser ? MCI_STATUS_S : MCI_STATUS_UC))) continue; - if (m.status & MCI_STATUS_MISCV) - m.misc = mce_rdmsrl(MSR_IA32_MCx_MISC(i)); - if (m.status & MCI_STATUS_ADDRV) - m.addr = mce_rdmsrl(MSR_IA32_MCx_ADDR(i)); + mce_read_aux(&m, i); if (!(flags & MCP_TIMESTAMP)) m.tsc = 0; @@ -981,10 +999,7 @@ void do_machine_check(struct pt_regs *regs, long error_code) if (severity == MCE_AR_SEVERITY) kill_it = 1; - if (m.status & MCI_STATUS_MISCV) - m.misc = mce_rdmsrl(MSR_IA32_MCx_MISC(i)); - if (m.status & MCI_STATUS_ADDRV) - m.addr = mce_rdmsrl(MSR_IA32_MCx_ADDR(i)); + mce_read_aux(&m, i); /* * Action optional error. Queue address for later processing. -- cgit v1.2.3-58-ga151 From af104e394e17e328df85c25a9e21448539725b67 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Wed, 14 Dec 2011 15:55:20 -0800 Subject: x86/mce: Add mechanism to safely save information in MCE handler Machine checks on Intel cpus interrupt execution on all cpus, regardless of interrupt masking. We have a need to save some data about the cause of the machine check (physical address) in the machine check handler that can be retrieved later to attempt recovery in a more flexible execution state. Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce.c | 43 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 43 insertions(+) diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 2f1c200f05e6..e1579c5a71da 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -886,6 +886,49 @@ static void mce_clear_state(unsigned long *toclear) } } +/* + * Need to save faulting physical address associated with a process + * in the machine check handler some place where we can grab it back + * later in mce_notify_process() + */ +#define MCE_INFO_MAX 16 + +struct mce_info { + atomic_t inuse; + struct task_struct *t; + __u64 paddr; +} mce_info[MCE_INFO_MAX]; + +static void mce_save_info(__u64 addr) +{ + struct mce_info *mi; + + for (mi = mce_info; mi < &mce_info[MCE_INFO_MAX]; mi++) { + if (atomic_cmpxchg(&mi->inuse, 0, 1) == 0) { + mi->t = current; + mi->paddr = addr; + return; + } + } + + mce_panic("Too many concurrent recoverable errors", NULL, NULL); +} + +static struct mce_info *mce_find_info(void) +{ + struct mce_info *mi; + + for (mi = mce_info; mi < &mce_info[MCE_INFO_MAX]; mi++) + if (atomic_read(&mi->inuse) && mi->t == current) + return mi; + return NULL; +} + +static void mce_clear_info(struct mce_info *mi) +{ + atomic_set(&mi->inuse, 0); +} + /* * The actual machine check handler. This only handles real * exceptions when something got corrupted coming in through int 18. -- cgit v1.2.3-58-ga151 From a8c321fbf9aeced45519248e5901af8cbc240510 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Tue, 3 Jan 2012 11:45:45 -0800 Subject: x86/mce: Handle "action required" errors All non-urgent actions (reporting low severity errors and handling "action-optional" errors) are now handled by a work queue. This means that TIF_MCE_NOTIFY can be used to block execution for a thread experiencing an "action-required" fault until we get all cpus out of the machine check handler (and the thread that hit the fault into mce_notify_process(). We use the new mce_{save,find,clear}_info() API to get information from do_machine_check() to mce_notify_process(), and then use the newly improved memory_failure(..., MF_ACTION_REQUIRED) to handle the error (possibly signalling the process). Update some comments to make the new code flows clearer. Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce.c | 95 ++++++++++++++++++++++------------------ 1 file changed, 53 insertions(+), 42 deletions(-) diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index e1579c5a71da..56e4e79387c3 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -982,7 +982,9 @@ void do_machine_check(struct pt_regs *regs, long error_code) barrier(); /* - * When no restart IP must always kill or panic. + * When no restart IP might need to kill or panic. + * Assume the worst for now, but if we find the + * severity is MCE_AR_SEVERITY we have other options. */ if (!(m.mcgstatus & MCG_STATUS_RIPV)) kill_it = 1; @@ -1036,12 +1038,6 @@ void do_machine_check(struct pt_regs *regs, long error_code) continue; } - /* - * Kill on action required. - */ - if (severity == MCE_AR_SEVERITY) - kill_it = 1; - mce_read_aux(&m, i); /* @@ -1062,6 +1058,9 @@ void do_machine_check(struct pt_regs *regs, long error_code) } } + /* mce_clear_state will clear *final, save locally for use later */ + m = *final; + if (!no_way_out) mce_clear_state(toclear); @@ -1073,27 +1072,22 @@ void do_machine_check(struct pt_regs *regs, long error_code) 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. + * At insane "tolerant" levels we take no action. Otherwise + * we only die if we have no other choice. For less serious + * issues we try to recover, or limit damage to the current + * process. */ - 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 (tolerant < 3) { + if (no_way_out) + mce_panic("Fatal machine check on current CPU", &m, msg); + if (worst == MCE_AR_SEVERITY) { + /* schedule action before return to userland */ + mce_save_info(m.addr); + set_thread_flag(TIF_MCE_NOTIFY); + } else if (kill_it) { + force_sig(SIGBUS, current); + } + } if (worst > 0) mce_report_event(regs); @@ -1107,6 +1101,8 @@ EXPORT_SYMBOL_GPL(do_machine_check); #ifndef CONFIG_MEMORY_FAILURE int memory_failure(unsigned long pfn, int vector, int flags) { + /* mce_severity() should not hand us an ACTION_REQUIRED error */ + BUG_ON(flags & MF_ACTION_REQUIRED); printk(KERN_ERR "Uncorrected memory error in page 0x%lx ignored\n" "Rebuild kernel with CONFIG_MEMORY_FAILURE=y for smarter handling\n", pfn); @@ -1115,27 +1111,44 @@ int memory_failure(unsigned long pfn, int vector, int flags) #endif /* - * 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. + * Called in process context that interrupted by MCE and marked with + * TIF_MCE_NOTIFY, just before returning to erroneous userland. + * This code is allowed to sleep. + * Attempt possible recovery such as calling the high level VM handler to + * process any corrupted pages, and kill/signal current process if required. + * Action required errors are handled here. */ void mce_notify_process(void) { unsigned long pfn; - mce_notify_irq(); - while (mce_ring_get(&pfn)) - memory_failure(pfn, MCE_VECTOR, 0); + struct mce_info *mi = mce_find_info(); + + if (!mi) + mce_panic("Lost physical address for unconsumed uncorrectable error", NULL, NULL); + pfn = mi->paddr >> PAGE_SHIFT; + + clear_thread_flag(TIF_MCE_NOTIFY); + + pr_err("Uncorrected hardware memory error in user-access at %llx", + mi->paddr); + if (memory_failure(pfn, MCE_VECTOR, MF_ACTION_REQUIRED) < 0) { + pr_err("Memory error not recovered"); + force_sig(SIGBUS, current); + } + mce_clear_info(mi); } +/* + * Action optional processing happens here (picking up + * from the list of faulting pages that do_machine_check() + * placed into the "ring"). + */ static void mce_process_work(struct work_struct *dummy) { - mce_notify_process(); + unsigned long pfn; + + while (mce_ring_get(&pfn)) + memory_failure(pfn, MCE_VECTOR, 0); } #ifdef CONFIG_X86_MCE_INTEL @@ -1225,8 +1238,6 @@ 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 processes polling /dev/mcelog */ wake_up_interruptible(&mce_chrdev_wait); -- cgit v1.2.3-58-ga151 From 5f7b88d51e89771f64c15903b96b5933dd0bc6d8 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Tue, 3 Jan 2012 11:48:04 -0800 Subject: x86/mce: Recognise machine check bank signature for data path error Action required data path signature is defined in table 15-19 of SDM: +-----------------------------------------------------------------------------+ | SRAR Error | Valid | OVER | UC | EN | MISCV | ADDRV | PCC | S | AR | MCACOD | | Data Load | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 1 | 1 | 0x134 | +-----------------------------------------------------------------------------+ Recognise this, and pass MCE_AR_SEVERITY code back to do_machine_check() if we have the action handler configured (CONFIG_MEMORY_FAILURE=y) Acked-by: Borislav Petkov Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce-severity.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/cpu/mcheck/mce-severity.c b/arch/x86/kernel/cpu/mcheck/mce-severity.c index 7395d5f4272d..f6c92f99efa0 100644 --- a/arch/x86/kernel/cpu/mcheck/mce-severity.c +++ b/arch/x86/kernel/cpu/mcheck/mce-severity.c @@ -54,6 +54,7 @@ static struct severity { #define MASK(x, y) .mask = x, .result = y #define MCI_UC_S (MCI_STATUS_UC|MCI_STATUS_S) #define MCI_UC_SAR (MCI_STATUS_UC|MCI_STATUS_S|MCI_STATUS_AR) +#define MCI_ADDR (MCI_STATUS_ADDRV|MCI_STATUS_MISCV) #define MCACOD 0xffff MCESEV( @@ -102,11 +103,24 @@ static struct severity { SER, BITCLR(MCI_STATUS_S) ), - /* AR add known MCACODs here */ MCESEV( PANIC, "Action required with lost events", SER, BITSET(MCI_STATUS_OVER|MCI_UC_SAR) ), + + /* known AR MCACODs: */ +#ifdef CONFIG_MEMORY_FAILURE + MCESEV( + KEEP, "HT thread notices Action required: data load error", + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|0x0134), + MCGMASK(MCG_STATUS_EIPV, 0) + ), + MCESEV( + AR, "Action required: data load error", + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|0x0134), + USER + ), +#endif MCESEV( PANIC, "Action required: unknown MCACOD", SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR, MCI_UC_SAR) -- cgit v1.2.3-58-ga151 From 08dda402d60a721ac94e79efd7646b332be3e3b2 Mon Sep 17 00:00:00 2001 From: Tony Luck Date: Thu, 26 Jan 2012 16:02:22 -0800 Subject: x86/mce: Replace hard coded hex constants with symbolic defines Magic constants like 0x0134 in code just invite questions on where they come from, what they mean, can they be changed. Provide #defines for the architecturally defined MCACOD values with a reference to the Intel Software Developers manual which describes them. Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce-severity.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/arch/x86/kernel/cpu/mcheck/mce-severity.c b/arch/x86/kernel/cpu/mcheck/mce-severity.c index f6c92f99efa0..0c82091b1652 100644 --- a/arch/x86/kernel/cpu/mcheck/mce-severity.c +++ b/arch/x86/kernel/cpu/mcheck/mce-severity.c @@ -56,6 +56,12 @@ static struct severity { #define MCI_UC_SAR (MCI_STATUS_UC|MCI_STATUS_S|MCI_STATUS_AR) #define MCI_ADDR (MCI_STATUS_ADDRV|MCI_STATUS_MISCV) #define MCACOD 0xffff +/* Architecturally defined codes from SDM Vol. 3B Chapter 15 */ +#define MCACOD_SCRUB 0x00C0 /* 0xC0-0xCF Memory Scrubbing */ +#define MCACOD_SCRUBMSK 0xfff0 +#define MCACOD_L3WB 0x017A /* L3 Explicit Writeback */ +#define MCACOD_DATA 0x0134 /* Data Load */ +#define MCACOD_INSTR 0x0150 /* Instruction Fetch */ MCESEV( NO, "Invalid", @@ -112,12 +118,12 @@ static struct severity { #ifdef CONFIG_MEMORY_FAILURE MCESEV( KEEP, "HT thread notices Action required: data load error", - SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|0x0134), + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|MCACOD_DATA), MCGMASK(MCG_STATUS_EIPV, 0) ), MCESEV( AR, "Action required: data load error", - SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|0x0134), + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|MCACOD_DATA), USER ), #endif @@ -129,11 +135,11 @@ static struct severity { /* known AO MCACODs: */ MCESEV( AO, "Action optional: memory scrubbing error", - SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|0xfff0, MCI_UC_S|0x00c0) + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCACOD_SCRUBMSK, MCI_UC_S|MCACOD_SCRUB) ), MCESEV( AO, "Action optional: last level cache writeback error", - SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCACOD, MCI_UC_S|0x017a) + SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCACOD, MCI_UC_S|MCACOD_L3WB) ), MCESEV( SOME, "Action optional: unknown MCACOD", -- cgit v1.2.3-58-ga151 From d6126ef5f31ca54980cb067af659a360dfcca037 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Thu, 26 Jan 2012 15:49:14 -0800 Subject: x86/mce: Convert static array of pointers to per-cpu variables When I previously fixed up the mce_device code, I used a static array of the pointers. It was (rightfully) pointed out to me that I should be using the per_cpu code instead. This patch converts the code over to that structure, moving the variable back into the per_cpu area, like it used to be for 3.2 and earlier. Signed-off-by: Greg Kroah-Hartman Reviewed-by: Srivatsa S. Bhat Link: https://lkml.org/lkml/2012/1/27/165 Signed-off-by: Tony Luck --- arch/x86/include/asm/mce.h | 2 +- arch/x86/kernel/cpu/mcheck/mce.c | 8 ++++---- arch/x86/kernel/cpu/mcheck/mce_amd.c | 9 +++++---- 3 files changed, 10 insertions(+), 9 deletions(-) diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h index 6aefb14cbbc5..441520e4174f 100644 --- a/arch/x86/include/asm/mce.h +++ b/arch/x86/include/asm/mce.h @@ -151,7 +151,7 @@ static inline void enable_p5_mce(void) {} void mce_setup(struct mce *m); void mce_log(struct mce *m); -extern struct device *mce_device[CONFIG_NR_CPUS]; +DECLARE_PER_CPU(struct device *, mce_device); /* * Maximum banks number. diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 5a11ae2e9e91..4979a5dfeba2 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -1859,7 +1859,7 @@ static struct bus_type mce_subsys = { .dev_name = "machinecheck", }; -struct device *mce_device[CONFIG_NR_CPUS]; +DEFINE_PER_CPU(struct device *, mce_device); __cpuinitdata void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu); @@ -2038,7 +2038,7 @@ static __cpuinit int mce_device_create(unsigned int cpu) goto error2; } cpumask_set_cpu(cpu, mce_device_initialized); - mce_device[cpu] = dev; + per_cpu(mce_device, cpu) = dev; return 0; error2: @@ -2055,7 +2055,7 @@ error: static __cpuinit void mce_device_remove(unsigned int cpu) { - struct device *dev = mce_device[cpu]; + struct device *dev = per_cpu(mce_device, cpu); int i; if (!cpumask_test_cpu(cpu, mce_device_initialized)) @@ -2069,7 +2069,7 @@ static __cpuinit void mce_device_remove(unsigned int cpu) device_unregister(dev); cpumask_clear_cpu(cpu, mce_device_initialized); - mce_device[cpu] = NULL; + per_cpu(mce_device, cpu) = NULL; } /* Make sure there are no machine checks on offlined CPUs. */ diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c index 786e76a86322..a4bf9d23cdba 100644 --- a/arch/x86/kernel/cpu/mcheck/mce_amd.c +++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c @@ -523,7 +523,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) { int i, err = 0; struct threshold_bank *b = NULL; - struct device *dev = mce_device[cpu]; + struct device *dev = per_cpu(mce_device, cpu); char name[32]; sprintf(name, "threshold_bank%i", bank); @@ -585,7 +585,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank) if (i == cpu) continue; - dev = mce_device[i]; + dev = per_cpu(mce_device, i); if (dev) err = sysfs_create_link(&dev->kobj,b->kobj, name); if (err) @@ -665,7 +665,8 @@ 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(&mce_device[cpu]->kobj, name); + dev = per_cpu(mce_device, cpu); + sysfs_remove_link(&dev->kobj, name); per_cpu(threshold_banks, cpu)[bank] = NULL; return; @@ -677,7 +678,7 @@ static void threshold_remove_bank(unsigned int cpu, int bank) if (i == cpu) continue; - dev = mce_device[i]; + dev = per_cpu(mce_device, i); if (dev) sysfs_remove_link(&dev->kobj, name); per_cpu(threshold_banks, i)[bank] = NULL; -- cgit v1.2.3-58-ga151 From fadd85f16a8ec3fee8af599e79a209682dc52348 Mon Sep 17 00:00:00 2001 From: Naoya Horiguchi Date: Mon, 23 Jan 2012 15:54:52 -0500 Subject: x86/mce: Fix return value of mce_chrdev_read() when erst is disabled Current kernel MCE code reads ERST at the first reading of /dev/mcelog (maybe in starting mcelogd,) even if the system does not support ERST, which results in a fake "no such device" message (as described in [1].) This problem is not critical, but can confuse system admins. This patch fixes it by filtering the return value from lower (ACPI) layer. [1] http://thread.gmane.org/gmane.linux.kernel/1060250 Reported by: Jon Masters Signed-off-by: Naoya Horiguchi Cc: Andi Kleen Cc: Huang Ying Link: https://lkml.org/lkml/2012/1/23/299 Signed-off-by: Tony Luck --- arch/x86/kernel/cpu/mcheck/mce.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 4979a5dfeba2..87c56ba8080c 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -1541,6 +1541,12 @@ static int __mce_read_apei(char __user **ubuf, size_t usize) /* Error or no more MCE record */ if (rc <= 0) { mce_apei_read_done = 1; + /* + * When ERST is disabled, mce_chrdev_read() should return + * "no record" instead of "no device." + */ + if (rc == -ENODEV) + return 0; return rc; } rc = -EFAULT; -- cgit v1.2.3-58-ga151