Merge branch 'perf/mce' into perf/core

Merge reason: It's ready for v2.6.33.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

arch/x86/include/asm/hardirq.h

@@ -20,11 +20,11 @@ typedef struct {
 	unsigned int irq_call_count;
 	unsigned int irq_tlb_count;
 #endif
-#ifdef CONFIG_X86_MCE
+#ifdef CONFIG_X86_THERMAL_VECTOR
 	unsigned int irq_thermal_count;
-# ifdef CONFIG_X86_MCE_THRESHOLD
+#endif
+#ifdef CONFIG_X86_MCE_THRESHOLD
 	unsigned int irq_threshold_count;
-# endif
 #endif
 } ____cacheline_aligned irq_cpustat_t;
 

arch/x86/include/asm/mce.h

@@ -108,6 +108,8 @@ struct mce_log {
 #define K8_MCE_THRESHOLD_BANK_5    (MCE_THRESHOLD_BASE + 5 * 9)
 #define K8_MCE_THRESHOLD_DRAM_ECC  (MCE_THRESHOLD_BANK_4 + 0)
 
+extern struct atomic_notifier_head x86_mce_decoder_chain;
+
 #ifdef __KERNEL__
 
 #include <linux/percpu.h>
@@ -118,9 +120,11 @@ extern int mce_disabled;
 extern int mce_p5_enabled;
 
 #ifdef CONFIG_X86_MCE
-void mcheck_init(struct cpuinfo_x86 *c);
+int mcheck_init(void);
+void mcheck_cpu_init(struct cpuinfo_x86 *c);
 #else
-static inline void mcheck_init(struct cpuinfo_x86 *c) {}
+static inline int mcheck_init(void) { return 0; }
+static inline void mcheck_cpu_init(struct cpuinfo_x86 *c) {}
 #endif
 
 #ifdef CONFIG_X86_ANCIENT_MCE
@@ -214,5 +218,11 @@ void intel_init_thermal(struct cpuinfo_x86 *c);
 
 void mce_log_therm_throt_event(__u64 status);
 
+#ifdef CONFIG_X86_THERMAL_VECTOR
+extern void mcheck_intel_therm_init(void);
+#else
+static inline void mcheck_intel_therm_init(void) { }
+#endif
+
 #endif /* __KERNEL__ */
 #endif /* _ASM_X86_MCE_H */

arch/x86/kernel/cpu/common.c

@@ -837,10 +837,8 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c)
 			boot_cpu_data.x86_capability[i] &= c->x86_capability[i];
 	}
 
-#ifdef CONFIG_X86_MCE
 	/* Init Machine Check Exception if available. */
-	mcheck_init(c);
-#endif
+	mcheck_cpu_init(c);
 
 	select_idle_routine(c);
 

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

@@ -46,6 +46,9 @@
 
 #include "mce-internal.h"
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/mce.h>
+
 int mce_disabled __read_mostly;
 
 #define MISC_MCELOG_MINOR	227
@@ -85,18 +88,26 @@ static DECLARE_WAIT_QUEUE_HEAD(mce_wait);
 static DEFINE_PER_CPU(struct mce, mces_seen);
 static int			cpu_missing;
 
-static void default_decode_mce(struct mce *m)
+/*
+ * CPU/chipset specific EDAC code can register a notifier call here to print
+ * MCE errors in a human-readable form.
+ */
+ATOMIC_NOTIFIER_HEAD(x86_mce_decoder_chain);
+EXPORT_SYMBOL_GPL(x86_mce_decoder_chain);
+
+static int default_decode_mce(struct notifier_block *nb, unsigned long val,
+			       void *data)
 {
 	pr_emerg("No human readable MCE decoding support on this CPU type.\n");
 	pr_emerg("Run the message through 'mcelog --ascii' to decode.\n");
+
+	return NOTIFY_STOP;
 }
 
-/*
- * CPU/chipset specific EDAC code can register a callback here to print
- * MCE errors in a human-readable form:
- */
-void (*x86_mce_decode_callback)(struct mce *m) = default_decode_mce;
-EXPORT_SYMBOL(x86_mce_decode_callback);
+static struct notifier_block mce_dec_nb = {
+	.notifier_call = default_decode_mce,
+	.priority      = -1,
+};
 
 /* MCA banks polled by the period polling timer for corrected events */
 DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = {
@@ -141,6 +152,9 @@ void mce_log(struct mce *mce)
 {
 	unsigned next, entry;
 
+	/* Emit the trace record: */
+	trace_mce_record(mce);
+
 	mce->finished = 0;
 	wmb();
 	for (;;) {
@@ -204,9 +218,9 @@ static void print_mce(struct mce *m)
 
 	/*
 	 * Print out human-readable details about the MCE error,
-	 * (if the CPU has an implementation for that):
+	 * (if the CPU has an implementation for that)
 	 */
-	x86_mce_decode_callback(m);
+	atomic_notifier_call_chain(&x86_mce_decoder_chain, 0, m);
 }
 
 static void print_mce_head(void)
@@ -1122,7 +1136,7 @@ static int check_interval = 5 * 60; /* 5 minutes */
 static DEFINE_PER_CPU(int, mce_next_interval); /* in jiffies */
 static DEFINE_PER_CPU(struct timer_list, mce_timer);
 
-static void mcheck_timer(unsigned long data)
+static void mce_start_timer(unsigned long data)
 {
 	struct timer_list *t = &per_cpu(mce_timer, data);
 	int *n;
@@ -1187,7 +1201,7 @@ int mce_notify_irq(void)
 }
 EXPORT_SYMBOL_GPL(mce_notify_irq);
 
-static int mce_banks_init(void)
+static int __cpuinit __mcheck_cpu_mce_banks_init(void)
 {
 	int i;
 
@@ -1206,7 +1220,7 @@ static int mce_banks_init(void)
 /*
  * Initialize Machine Checks for a CPU.
  */
-static int __cpuinit mce_cap_init(void)
+static int __cpuinit __mcheck_cpu_cap_init(void)
 {
 	unsigned b;
 	u64 cap;
@@ -1228,7 +1242,7 @@ static int __cpuinit mce_cap_init(void)
 	WARN_ON(banks != 0 && b != banks);
 	banks = b;
 	if (!mce_banks) {
-		int err = mce_banks_init();
+		int err = __mcheck_cpu_mce_banks_init();
 
 		if (err)
 			return err;
@@ -1244,7 +1258,7 @@ static int __cpuinit mce_cap_init(void)
 	return 0;
 }
 
-static void mce_init(void)
+static void __mcheck_cpu_init_generic(void)
 {
 	mce_banks_t all_banks;
 	u64 cap;
@@ -1273,7 +1287,7 @@ static void mce_init(void)
 }
 
 /* Add per CPU specific workarounds here */
-static int __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
+static int __cpuinit __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c)
 {
 	if (c->x86_vendor == X86_VENDOR_UNKNOWN) {
 		pr_info("MCE: unknown CPU type - not enabling MCE support.\n");
@@ -1341,7 +1355,7 @@ static int __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
 	return 0;
 }
 
-static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c)
+static void __cpuinit __mcheck_cpu_ancient_init(struct cpuinfo_x86 *c)
 {
 	if (c->x86 != 5)
 		return;
@@ -1355,7 +1369,7 @@ static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c)
 	}
 }
 
-static void mce_cpu_features(struct cpuinfo_x86 *c)
+static void __mcheck_cpu_init_vendor(struct cpuinfo_x86 *c)
 {
 	switch (c->x86_vendor) {
 	case X86_VENDOR_INTEL:
@@ -1369,7 +1383,7 @@ static void mce_cpu_features(struct cpuinfo_x86 *c)
 	}
 }
 
-static void mce_init_timer(void)
+static void __mcheck_cpu_init_timer(void)
 {
 	struct timer_list *t = &__get_cpu_var(mce_timer);
 	int *n = &__get_cpu_var(mce_next_interval);
@@ -1380,7 +1394,7 @@ static void mce_init_timer(void)
 	*n = check_interval * HZ;
 	if (!*n)
 		return;
-	setup_timer(t, mcheck_timer, smp_processor_id());
+	setup_timer(t, mce_start_timer, smp_processor_id());
 	t->expires = round_jiffies(jiffies + *n);
 	add_timer_on(t, smp_processor_id());
 }
@@ -1400,27 +1414,28 @@ void (*machine_check_vector)(struct pt_regs *, long error_code) =
  * Called for each booted CPU to set up machine checks.
  * Must be called with preempt off:
  */
-void __cpuinit mcheck_init(struct cpuinfo_x86 *c)
+void __cpuinit mcheck_cpu_init(struct cpuinfo_x86 *c)
 {
 	if (mce_disabled)
 		return;
 
-	mce_ancient_init(c);
+	__mcheck_cpu_ancient_init(c);
 
 	if (!mce_available(c))
 		return;
 
-	if (mce_cap_init() < 0 || mce_cpu_quirks(c) < 0) {
+	if (__mcheck_cpu_cap_init() < 0 || __mcheck_cpu_apply_quirks(c) < 0) {
 		mce_disabled = 1;
 		return;
 	}
 
 	machine_check_vector = do_machine_check;
 
-	mce_init();
-	mce_cpu_features(c);
-	mce_init_timer();
+	__mcheck_cpu_init_generic();
+	__mcheck_cpu_init_vendor(c);
+	__mcheck_cpu_init_timer();
 	INIT_WORK(&__get_cpu_var(mce_work), mce_process_work);
+
 }
 
 /*
@@ -1640,6 +1655,15 @@ static int __init mcheck_enable(char *str)
 }
 __setup("mce", mcheck_enable);
 
+int __init mcheck_init(void)
+{
+	atomic_notifier_chain_register(&x86_mce_decoder_chain, &mce_dec_nb);
+
+	mcheck_intel_therm_init();
+
+	return 0;
+}
+
 /*
  * Sysfs support
  */
@@ -1648,7 +1672,7 @@ __setup("mce", mcheck_enable);
  * Disable machine checks on suspend and shutdown. We can't really handle
  * them later.
  */
-static int mce_disable(void)
+static int mce_disable_error_reporting(void)
 {
 	int i;
 
@@ -1663,12 +1687,12 @@ static int mce_disable(void)
 
 static int mce_suspend(struct sys_device *dev, pm_message_t state)
 {
-	return mce_disable();
+	return mce_disable_error_reporting();
 }
 
 static int mce_shutdown(struct sys_device *dev)
 {
-	return mce_disable();
+	return mce_disable_error_reporting();
 }
 
 /*
@@ -1678,8 +1702,8 @@ static int mce_shutdown(struct sys_device *dev)
  */
 static int mce_resume(struct sys_device *dev)
 {
-	mce_init();
-	mce_cpu_features(&current_cpu_data);
+	__mcheck_cpu_init_generic();
+	__mcheck_cpu_init_vendor(&current_cpu_data);
 
 	return 0;
 }
@@ -1689,8 +1713,8 @@ static void mce_cpu_restart(void *data)
 	del_timer_sync(&__get_cpu_var(mce_timer));
 	if (!mce_available(&current_cpu_data))
 		return;
-	mce_init();
-	mce_init_timer();
+	__mcheck_cpu_init_generic();
+	__mcheck_cpu_init_timer();
 }
 
 /* Reinit MCEs after user configuration changes */
@@ -1716,7 +1740,7 @@ static void mce_enable_ce(void *all)
 	cmci_reenable();
 	cmci_recheck();
 	if (all)
-		mce_init_timer();
+		__mcheck_cpu_init_timer();
 }
 
 static struct sysdev_class mce_sysclass = {
@@ -1929,13 +1953,14 @@ static __cpuinit void mce_remove_device(unsigned int cpu)
 }
 
 /* Make sure there are no machine checks on offlined CPUs. */
-static void mce_disable_cpu(void *h)
+static void __cpuinit mce_disable_cpu(void *h)
 {
 	unsigned long action = *(unsigned long *)h;
 	int i;
 
 	if (!mce_available(&current_cpu_data))
 		return;
+
 	if (!(action & CPU_TASKS_FROZEN))
 		cmci_clear();
 	for (i = 0; i < banks; i++) {
@@ -1946,7 +1971,7 @@ static void mce_disable_cpu(void *h)
 	}
 }
 
-static void mce_reenable_cpu(void *h)
+static void __cpuinit mce_reenable_cpu(void *h)
 {
 	unsigned long action = *(unsigned long *)h;
 	int i;
@@ -2025,7 +2050,7 @@ static __init void mce_init_banks(void)
 	}
 }
 
-static __init int mce_init_device(void)
+static __init int mcheck_init_device(void)
 {
 	int err;
 	int i = 0;
@@ -2053,7 +2078,7 @@ static __init int mce_init_device(void)
 	return err;
 }
 
-device_initcall(mce_init_device);
+device_initcall(mcheck_init_device);
 
 /*
  * Old style boot options parsing. Only for compatibility.
@@ -2101,7 +2126,7 @@ static int fake_panic_set(void *data, u64 val)
 DEFINE_SIMPLE_ATTRIBUTE(fake_panic_fops, fake_panic_get,
 			fake_panic_set, "%llu\n");
 
-static int __init mce_debugfs_init(void)
+static int __init mcheck_debugfs_init(void)
 {
 	struct dentry *dmce, *ffake_panic;
 
@@ -2115,5 +2140,5 @@ static int __init mce_debugfs_init(void)
 
 	return 0;
 }
-late_initcall(mce_debugfs_init);
+late_initcall(mcheck_debugfs_init);
 #endif

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

@@ -49,6 +49,8 @@ static DEFINE_PER_CPU(struct thermal_state, thermal_state);
 
 static atomic_t therm_throt_en	= ATOMIC_INIT(0);
 
+static u32 lvtthmr_init __read_mostly;
+
 #ifdef CONFIG_SYSFS
 #define define_therm_throt_sysdev_one_ro(_name)				\
 	static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL)
@@ -254,6 +256,18 @@ asmlinkage void smp_thermal_interrupt(struct pt_regs *regs)
 	ack_APIC_irq();
 }
 
+void __init mcheck_intel_therm_init(void)
+{
+	/*
+	 * This function is only called on boot CPU. Save the init thermal
+	 * LVT value on BSP and use that value to restore APs' thermal LVT
+	 * entry BIOS programmed later
+	 */
+	if (cpu_has(&boot_cpu_data, X86_FEATURE_ACPI) &&
+		cpu_has(&boot_cpu_data, X86_FEATURE_ACC))
+		lvtthmr_init = apic_read(APIC_LVTTHMR);
+}
+
 void intel_init_thermal(struct cpuinfo_x86 *c)
 {
 	unsigned int cpu = smp_processor_id();
@@ -270,7 +284,20 @@ void intel_init_thermal(struct cpuinfo_x86 *c)
 	 * since it might be delivered via SMI already:
 	 */
 	rdmsr(MSR_IA32_MISC_ENABLE, l, h);
-	h = apic_read(APIC_LVTTHMR);
+
+	/*
+	 * The initial value of thermal LVT entries on all APs always reads
+	 * 0x10000 because APs are woken up by BSP issuing INIT-SIPI-SIPI
+	 * sequence to them and LVT registers are reset to 0s except for
+	 * the mask bits which are set to 1s when APs receive INIT IPI.
+	 * Always restore the value that BIOS has programmed on AP based on
+	 * BSP's info we saved since BIOS is always setting the same value
+	 * for all threads/cores
+	 */
+	apic_write(APIC_LVTTHMR, lvtthmr_init);
+
+	h = lvtthmr_init;
+
 	if ((l & MSR_IA32_MISC_ENABLE_TM1) && (h & APIC_DM_SMI)) {
 		printk(KERN_DEBUG
 		       "CPU%d: Thermal monitoring handled by SMI\n", cpu);

arch/x86/kernel/irq.c

@@ -92,17 +92,17 @@ static int show_other_interrupts(struct seq_file *p, int prec)
 		seq_printf(p, "%10u ", irq_stats(j)->irq_tlb_count);
 	seq_printf(p, "  TLB shootdowns\n");
 #endif
-#ifdef CONFIG_X86_MCE
+#ifdef CONFIG_X86_THERMAL_VECTOR
 	seq_printf(p, "%*s: ", prec, "TRM");
 	for_each_online_cpu(j)
 		seq_printf(p, "%10u ", irq_stats(j)->irq_thermal_count);
 	seq_printf(p, "  Thermal event interrupts\n");
-# ifdef CONFIG_X86_MCE_THRESHOLD
+#endif
+#ifdef CONFIG_X86_MCE_THRESHOLD
 	seq_printf(p, "%*s: ", prec, "THR");
 	for_each_online_cpu(j)
 		seq_printf(p, "%10u ", irq_stats(j)->irq_threshold_count);
 	seq_printf(p, "  Threshold APIC interrupts\n");
-# endif
 #endif
 #ifdef CONFIG_X86_MCE
 	seq_printf(p, "%*s: ", prec, "MCE");
@@ -194,11 +194,11 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
 	sum += irq_stats(cpu)->irq_call_count;
 	sum += irq_stats(cpu)->irq_tlb_count;
 #endif
-#ifdef CONFIG_X86_MCE
+#ifdef CONFIG_X86_THERMAL_VECTOR
 	sum += irq_stats(cpu)->irq_thermal_count;
-# ifdef CONFIG_X86_MCE_THRESHOLD
+#endif
+#ifdef CONFIG_X86_MCE_THRESHOLD
 	sum += irq_stats(cpu)->irq_threshold_count;
-# endif
 #endif
 #ifdef CONFIG_X86_MCE
 	sum += per_cpu(mce_exception_count, cpu);

arch/x86/kernel/setup.c

@@ -109,6 +109,7 @@
 #ifdef CONFIG_X86_64
 #include <asm/numa_64.h>
 #endif
+#include <asm/mce.h>
 
 /*
  * end_pfn only includes RAM, while max_pfn_mapped includes all e820 entries.
@@ -1031,6 +1032,8 @@ void __init setup_arch(char **cmdline_p)
 #endif
 #endif
 	x86_init.oem.banner();
+
+	mcheck_init();
 }
 
 #ifdef CONFIG_X86_32

drivers/edac/edac_mce_amd.c

@@ -3,7 +3,6 @@
 
 static bool report_gart_errors;
 static void (*nb_bus_decoder)(int node_id, struct err_regs *regs);
-static void (*orig_mce_callback)(struct mce *m);
 
 void amd_report_gart_errors(bool v)
 {
@@ -363,8 +362,10 @@ static inline void amd_decode_err_code(unsigned int ec)
 		pr_warning("Huh? Unknown MCE error 0x%x\n", ec);
 }
 
-static void amd_decode_mce(struct mce *m)
+static int amd_decode_mce(struct notifier_block *nb, unsigned long val,
+			   void *data)
 {
+	struct mce *m = (struct mce *)data;
 	struct err_regs regs;
 	int node, ecc;
 
@@ -420,20 +421,22 @@ static void amd_decode_mce(struct mce *m)
 	}
 
 	amd_decode_err_code(m->status & 0xffff);
+
+	return NOTIFY_STOP;
 }
 
+static struct notifier_block amd_mce_dec_nb = {
+	.notifier_call	= amd_decode_mce,
+};
+
 static int __init mce_amd_init(void)
 {
 	/*
 	 * We can decode MCEs for Opteron and later CPUs:
 	 */
 	if ((boot_cpu_data.x86_vendor == X86_VENDOR_AMD) &&
-	    (boot_cpu_data.x86 >= 0xf)) {
-		/* safe the default decode mce callback */
-		orig_mce_callback = x86_mce_decode_callback;
-
-		x86_mce_decode_callback = amd_decode_mce;
-	}
+	    (boot_cpu_data.x86 >= 0xf))
+		atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb);
 
 	return 0;
 }
@@ -442,7 +445,7 @@ early_initcall(mce_amd_init);
 #ifdef MODULE
 static void __exit mce_amd_exit(void)
 {
-	x86_mce_decode_callback = orig_mce_callback;
+	atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb);
 }
 
 MODULE_DESCRIPTION("AMD MCE decoder");

include/trace/events/mce.h

@@ -0,0 +1,69 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mce
+
+#if !defined(_TRACE_MCE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_MCE_H
+
+#include <linux/ktime.h>
+#include <linux/tracepoint.h>
+#include <asm/mce.h>
+
+TRACE_EVENT(mce_record,
+
+	TP_PROTO(struct mce *m),
+
+	TP_ARGS(m),
+
+	TP_STRUCT__entry(
+		__field(	u64,		mcgcap		)
+		__field(	u64,		mcgstatus	)
+		__field(	u8,		bank		)
+		__field(	u64,		status		)
+		__field(	u64,		addr		)
+		__field(	u64,		misc		)
+		__field(	u64,		ip		)
+		__field(	u8,		cs		)
+		__field(	u64,		tsc		)
+		__field(	u64,		walltime	)
+		__field(	u32,		cpu		)
+		__field(	u32,		cpuid		)
+		__field(	u32,		apicid		)
+		__field(	u32,		socketid	)
+		__field(	u8,		cpuvendor	)
+	),
+
+	TP_fast_assign(
+		__entry->mcgcap		= m->mcgcap;
+		__entry->mcgstatus	= m->mcgstatus;
+		__entry->bank		= m->bank;
+		__entry->status		= m->status;
+		__entry->addr		= m->addr;
+		__entry->misc		= m->misc;
+		__entry->ip		= m->ip;
+		__entry->cs		= m->cs;
+		__entry->tsc		= m->tsc;
+		__entry->walltime	= m->time;
+		__entry->cpu		= m->extcpu;
+		__entry->cpuid		= m->cpuid;
+		__entry->apicid		= m->apicid;
+		__entry->socketid	= m->socketid;
+		__entry->cpuvendor	= m->cpuvendor;
+	),
+
+	TP_printk("CPU: %d, MCGc/s: %llx/%llx, MC%d: %016Lx, ADDR/MISC: %016Lx/%016Lx, RIP: %02x:<%016Lx>, TSC: %llx, PROCESSOR: %u:%x, TIME: %llu, SOCKET: %u, APIC: %x",
+		__entry->cpu,
+		__entry->mcgcap, __entry->mcgstatus,
+		__entry->bank, __entry->status,
+		__entry->addr, __entry->misc,
+		__entry->cs, __entry->ip,
+		__entry->tsc,
+		__entry->cpuvendor, __entry->cpuid,
+		__entry->walltime,
+		__entry->socketid,
+		__entry->apicid)
+);
+
+#endif /* _TRACE_MCE_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>