x86, mce: machine check recovery for applications

index 2af127d..1a08ce5 100644
@@ -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)

index 8272d92..9a92444 100644
@@ -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;
 }
 

index 4baadd1..bcc5234 100644
@@ -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;

index c7fc7fd..cc448bb 100644
@@ -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)

index 74bf193..f5ab745 100644
@@ -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;
 }

64 lines not on this page · raw message

index bcc5234..bf169ca 100644
@@ -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);

100 lines not on this page · raw message

index 1a08ce5..2f1c200 100644
@@ -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.

index 2f1c200..e1579c5 100644
@@ -887,6 +887,49 @@ 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.
  *

index e1579c5..56e4e79 100644
@@ -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);

index 7395d5f..f6c92f9 100644
@@ -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)