Merge branch 'x86-ras-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 RAS update from Ingo Molnar:
 "The changes in this tree are:

   - ACPI APEI (ACPI Platform Error Interface) improvements, by Chen
     Gong
   - misc MCE fixes/cleanups"

* 'x86-ras-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mce: Update MCE severity condition check
  mce: acpi/apei: Add comments to clarify usage of the various bitfields in the MCA subsystem
  ACPI/APEI: Update einj documentation for param1/param2
  ACPI/APEI: Add parameter check before error injection
  ACPI, APEI, EINJ: Fix error return code in einj_init()
  x86, mce: Fix "braodcast" typo

Documentation/acpi/apei/einj.txt

@@ -47,11 +47,16 @@ directory apei/einj. The following files are provided.
 
 - param1
   This file is used to set the first error parameter value. Effect of
-  parameter depends on error_type specified.
+  parameter depends on error_type specified. For example, if error
+  type is memory related type, the param1 should be a valid physical
+  memory address.
 
 - param2
   This file is used to set the second error parameter value. Effect of
-  parameter depends on error_type specified.
+  parameter depends on error_type specified. For example, if error
+  type is memory related type, the param2 should be a physical memory
+  address mask. Linux requires page or narrower granularity, say,
+  0xfffffffffffff000.
 
 - notrigger
   The EINJ mechanism is a two step process. First inject the error, then

arch/x86/include/asm/mce.h

@@ -61,7 +61,7 @@
 #define MCJ_CTX_IRQ		0x2  /* inject context: IRQ */
 #define MCJ_NMI_BROADCAST	0x4  /* do NMI broadcasting */
 #define MCJ_EXCEPTION		0x8  /* raise as exception */
-#define MCJ_IRQ_BRAODCAST	0x10 /* do IRQ broadcasting */
+#define MCJ_IRQ_BROADCAST	0x10 /* do IRQ broadcasting */
 
 #define MCE_OVERFLOW 0		/* bit 0 in flags means overflow */
 

arch/x86/kernel/cpu/mcheck/mce-inject.c

@@ -153,7 +153,7 @@ static void raise_mce(struct mce *m)
 		return;
 
 #ifdef CONFIG_X86_LOCAL_APIC
-	if (m->inject_flags & (MCJ_IRQ_BRAODCAST | MCJ_NMI_BROADCAST)) {
+	if (m->inject_flags & (MCJ_IRQ_BROADCAST | MCJ_NMI_BROADCAST)) {
 		unsigned long start;
 		int cpu;
 
@@ -167,7 +167,7 @@ static void raise_mce(struct mce *m)
 				cpumask_clear_cpu(cpu, mce_inject_cpumask);
 		}
 		if (!cpumask_empty(mce_inject_cpumask)) {
-			if (m->inject_flags & MCJ_IRQ_BRAODCAST) {
+			if (m->inject_flags & MCJ_IRQ_BROADCAST) {
 				/*
 				 * don't wait because mce_irq_ipi is necessary
 				 * to be sync with following raise_local

arch/x86/kernel/cpu/mcheck/mce-severity.c

@@ -110,22 +110,17 @@ static struct severity {
 	/* 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|MCACOD_DATA),
-		MCGMASK(MCG_STATUS_EIPV, 0)
+		KEEP, "Action required but unaffected thread is continuable",
+		SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR),
+		MCGMASK(MCG_STATUS_RIPV, MCG_STATUS_RIPV)
 		),
 	MCESEV(
-		AR, "Action required: data load error",
+		AR, "Action required: data load error in a user process",
 		SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|MCACOD_DATA),
 		USER
 		),
 	MCESEV(
-		KEEP, "HT thread notices Action required: instruction fetch error",
-		SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|MCACOD_INSTR),
-		MCGMASK(MCG_STATUS_EIPV, 0)
-		),
-	MCESEV(
-		AR, "Action required: instruction fetch error",
+		AR, "Action required: instruction fetch error in a user process",
 		SER, MASK(MCI_STATUS_OVER|MCI_UC_SAR|MCI_ADDR|MCACOD, MCI_UC_SAR|MCI_ADDR|MCACOD_INSTR),
 		USER
 		),

arch/x86/kernel/cpu/mcheck/mce.c

@@ -89,7 +89,10 @@ static DECLARE_WAIT_QUEUE_HEAD(mce_chrdev_wait);
 static DEFINE_PER_CPU(struct mce, mces_seen);
 static int			cpu_missing;
 
-/* MCA banks polled by the period polling timer for corrected events */
+/*
+ * MCA banks polled by the period polling timer for corrected events.
+ * With Intel CMCI, this only has MCA banks which do not support CMCI (if any).
+ */
 DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = {
 	[0 ... BITS_TO_LONGS(MAX_NR_BANKS)-1] = ~0UL
 };

arch/x86/kernel/cpu/mcheck/mce_intel.c

@@ -24,6 +24,18 @@
  * Also supports reliable discovery of shared banks.
  */
 
+/*
+ * CMCI can be delivered to multiple cpus that share a machine check bank
+ * so we need to designate a single cpu to process errors logged in each bank
+ * in the interrupt handler (otherwise we would have many races and potential
+ * double reporting of the same error).
+ * Note that this can change when a cpu is offlined or brought online since
+ * some MCA banks are shared across cpus. When a cpu is offlined, cmci_clear()
+ * disables CMCI on all banks owned by the cpu and clears this bitfield. At
+ * this point, cmci_rediscover() kicks in and a different cpu may end up
+ * taking ownership of some of the shared MCA banks that were previously
+ * owned by the offlined cpu.
+ */
 static DEFINE_PER_CPU(mce_banks_t, mce_banks_owned);
 
 /*

drivers/acpi/apei/einj.c

@@ -32,6 +32,7 @@
 #include <linux/seq_file.h>
 #include <linux/nmi.h>
 #include <linux/delay.h>
+#include <linux/mm.h>
 #include <acpi/acpi.h>
 
 #include "apei-internal.h"
@@ -41,6 +42,10 @@
 #define SPIN_UNIT		100			/* 100ns */
 /* Firmware should respond within 1 milliseconds */
 #define FIRMWARE_TIMEOUT	(1 * NSEC_PER_MSEC)
+#define ACPI5_VENDOR_BIT	BIT(31)
+#define MEM_ERROR_MASK		(ACPI_EINJ_MEMORY_CORRECTABLE | \
+				ACPI_EINJ_MEMORY_UNCORRECTABLE | \
+				ACPI_EINJ_MEMORY_FATAL)
 
 /*
  * ACPI version 5 provides a SET_ERROR_TYPE_WITH_ADDRESS action.
@@ -367,7 +372,7 @@ static int __einj_error_trigger(u64 trigger_paddr, u32 type,
 	 * This will cause resource conflict with regular memory.  So
 	 * remove it from trigger table resources.
 	 */
-	if ((param_extension || acpi5) && (type & 0x0038) && param2) {
+	if ((param_extension || acpi5) && (type & MEM_ERROR_MASK) && param2) {
 		struct apei_resources addr_resources;
 		apei_resources_init(&addr_resources);
 		trigger_param_region = einj_get_trigger_parameter_region(
@@ -427,7 +432,7 @@ static int __einj_error_inject(u32 type, u64 param1, u64 param2)
 		struct set_error_type_with_address *v5param = einj_param;
 
 		v5param->type = type;
-		if (type & 0x80000000) {
+		if (type & ACPI5_VENDOR_BIT) {
 			switch (vendor_flags) {
 			case SETWA_FLAGS_APICID:
 				v5param->apicid = param1;
@@ -512,7 +517,34 @@ static int __einj_error_inject(u32 type, u64 param1, u64 param2)
 static int einj_error_inject(u32 type, u64 param1, u64 param2)
 {
 	int rc;
+	unsigned long pfn;
 
+	/*
+	 * We need extra sanity checks for memory errors.
+	 * Other types leap directly to injection.
+	 */
+
+	/* ensure param1/param2 existed */
+	if (!(param_extension || acpi5))
+		goto inject;
+
+	/* ensure injection is memory related */
+	if (type & ACPI5_VENDOR_BIT) {
+		if (vendor_flags != SETWA_FLAGS_MEM)
+			goto inject;
+	} else if (!(type & MEM_ERROR_MASK))
+		goto inject;
+
+	/*
+	 * Disallow crazy address masks that give BIOS leeway to pick
+	 * injection address almost anywhere. Insist on page or
+	 * better granularity and that target address is normal RAM.
+	 */
+	pfn = PFN_DOWN(param1 & param2);
+	if (!page_is_ram(pfn) || ((param2 & PAGE_MASK) != PAGE_MASK))
+		return -EINVAL;
+
+inject:
 	mutex_lock(&einj_mutex);
 	rc = __einj_error_inject(type, param1, param2);
 	mutex_unlock(&einj_mutex);
@@ -590,7 +622,7 @@ static int error_type_set(void *data, u64 val)
 	 * Vendor defined types have 0x80000000 bit set, and
 	 * are not enumerated by ACPI_EINJ_GET_ERROR_TYPE
 	 */
-	vendor = val & 0x80000000;
+	vendor = val & ACPI5_VENDOR_BIT;
 	tval = val & 0x7fffffff;
 
 	/* Only one error type can be specified */
@@ -694,6 +726,7 @@ static int __init einj_init(void)
 	if (rc)
 		goto err_release;
 
+	rc = -ENOMEM;
 	einj_param = einj_get_parameter_address();
 	if ((param_extension || acpi5) && einj_param) {
 		fentry = debugfs_create_x64("param1", S_IRUSR | S_IWUSR,

kernel/resource.c

@@ -409,6 +409,7 @@ int __weak page_is_ram(unsigned long pfn)
 {
 	return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1;
 }
+EXPORT_SYMBOL_GPL(page_is_ram);
 
 void __weak arch_remove_reservations(struct resource *avail)
 {