[PATCH] x86: Refactor thermal throttle processing

Refactor the event processing (syslog messaging and rate limiting)
into separate file therm_throt.c. This allows consistent reporting
of CPU thermal throttle events.

After ACK'ing the interrupt, if the event is current, the user
(p4.c/mce_intel.c) calls therm_throt_process to log (and rate limit)
the event. If that function returns 1, the user has the option to log
things further (such as to mce_log in x86_64).

AK: minor cleanup
Signed-off-by: Dmitriy Zavin <dmitriyz@google.com>
Signed-off-by: Andi Kleen <ak@suse.de>

arch/i386/kernel/cpu/mcheck/Makefile

-obj-y	=	mce.o k7.o p4.o p5.o p6.o winchip.o
+obj-y	=	mce.o k7.o p4.o p5.o p6.o winchip.o therm_throt.o
 obj-$(CONFIG_X86_MCE_NONFATAL)	+=	non-fatal.o

arch/i386/kernel/cpu/mcheck/p4.c

@@ -13,6 +13,8 @@
 #include <asm/msr.h>
 #include <asm/apic.h>
 
+#include <asm/therm_throt.h>
+
 #include "mce.h"
 
 /* as supported by the P4/Xeon family */
@@ -44,25 +46,12 @@ static void unexpected_thermal_interrupt(struct pt_regs *regs)
 /* P4/Xeon Thermal transition interrupt handler */
 static void intel_thermal_interrupt(struct pt_regs *regs)
 {
-	u32 l, h;
-	unsigned int cpu = smp_processor_id();
-	static unsigned long next[NR_CPUS];
+	__u64 msr_val;
 
 	ack_APIC_irq();
 
-	if (time_after(next[cpu], jiffies))
-		return;
-
-	next[cpu] = jiffies + HZ*5;
-	rdmsr(MSR_IA32_THERM_STATUS, l, h);
-	if (l & 0x1) {
-		printk(KERN_EMERG "CPU%d: Temperature above threshold\n", cpu);
-		printk(KERN_EMERG "CPU%d: Running in modulated clock mode\n",
-				cpu);
-		add_taint(TAINT_MACHINE_CHECK);
-	} else {
-		printk(KERN_INFO "CPU%d: Temperature/speed normal\n", cpu);
-	}
+	rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
+	therm_throt_process(msr_val & 0x1);
 }
 
 /* Thermal interrupt handler for this CPU setup */
@@ -122,7 +111,7 @@ static void intel_init_thermal(struct cpuinfo_x86 *c)
 	
 	rdmsr (MSR_IA32_MISC_ENABLE, l, h);
 	wrmsr (MSR_IA32_MISC_ENABLE, l | (1<<3), h);
-	
+
 	l = apic_read (APIC_LVTTHMR);
 	apic_write_around (APIC_LVTTHMR, l & ~APIC_LVT_MASKED);
 	printk (KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu);

arch/i386/kernel/cpu/mcheck/therm_throt.c

+/*
+ * linux/arch/i386/kerne/cpu/mcheck/therm_throt.c
+ *
+ * Thermal throttle event support code.
+ *
+ * Author: Dmitriy Zavin (dmitriyz@google.com)
+ *
+ * Credits: Adapted from Zwane Mwaikambo's original code in mce_intel.c.
+ *
+ */
+
+#include <linux/percpu.h>
+#include <linux/cpu.h>
+#include <asm/cpu.h>
+#include <linux/notifier.h>
+#include <asm/therm_throt.h>
+
+/* How long to wait between reporting thermal events */
+#define CHECK_INTERVAL              (300 * HZ)
+
+static DEFINE_PER_CPU(unsigned long, next_check);
+
+/***
+ * therm_throt_process - Process thermal throttling event
+ * @curr: Whether the condition is current or not (boolean), since the
+ *        thermal interrupt normally gets called both when the thermal
+ *        event begins and once the event has ended.
+ *
+ * This function is normally called by the thermal interrupt after the
+ * IRQ has been acknowledged.
+ *
+ * It will take care of rate limiting and printing messages to the syslog.
+ *
+ * Returns: 0 : Event should NOT be further logged, i.e. still in
+ *              "timeout" from previous log message.
+ *          1 : Event should be logged further, and a message has been
+ *              printed to the syslog.
+ */
+int therm_throt_process(int curr)
+{
+	unsigned int cpu = smp_processor_id();
+
+	if (time_before(jiffies, __get_cpu_var(next_check)))
+		return 0;
+
+	__get_cpu_var(next_check) = jiffies + CHECK_INTERVAL;
+
+	/* if we just entered the thermal event */
+	if (curr) {
+		printk(KERN_CRIT "CPU%d: Temperature above threshold, "
+		       "cpu clock throttled\n", cpu);
+		add_taint(TAINT_MACHINE_CHECK);
+	} else {
+		printk(KERN_CRIT "CPU%d: Temperature/speed normal\n", cpu);
+	}
+
+	return 1;
+}

arch/x86_64/kernel/Makefile

@@ -11,7 +11,7 @@ obj-y	:= process.o signal.o entry.o traps.o irq.o \
 		pci-dma.o pci-nommu.o alternative.o
 
 obj-$(CONFIG_STACKTRACE)	+= stacktrace.o
-obj-$(CONFIG_X86_MCE)         += mce.o
+obj-$(CONFIG_X86_MCE)		+= mce.o therm_throt.o
 obj-$(CONFIG_X86_MCE_INTEL)	+= mce_intel.o
 obj-$(CONFIG_X86_MCE_AMD)	+= mce_amd.o
 obj-$(CONFIG_MTRR)		+= ../../i386/kernel/cpu/mtrr/
@@ -46,6 +46,7 @@ obj-y				+= intel_cacheinfo.o
 
 CFLAGS_vsyscall.o		:= $(PROFILING) -g0
 
+therm_throt-y                   += ../../i386/kernel/cpu/mcheck/therm_throt.o
 bootflag-y			+= ../../i386/kernel/bootflag.o
 cpuid-$(subst m,y,$(CONFIG_X86_CPUID))  += ../../i386/kernel/cpuid.o
 topology-y                     += ../../i386/kernel/topology.o
@@ -55,4 +56,3 @@ quirks-y			+= ../../i386/kernel/quirks.o
 i8237-y				+= ../../i386/kernel/i8237.o
 msr-$(subst m,y,$(CONFIG_X86_MSR))  += ../../i386/kernel/msr.o
 alternative-y			+= ../../i386/kernel/alternative.o
-

arch/x86_64/kernel/mce.c

@@ -274,6 +274,33 @@ void do_machine_check(struct pt_regs * regs, long error_code)
 	atomic_dec(&mce_entry);
 }
 
+#ifdef CONFIG_X86_MCE_INTEL
+/***
+ * mce_log_therm_throt_event - Logs the thermal throttling event to mcelog
+ * @cpu: The CPU on which the event occured.
+ * @status: Event status information
+ *
+ * This function should be called by the thermal interrupt after the
+ * event has been processed and the decision was made to log the event
+ * further.
+ *
+ * The status parameter will be saved to the 'status' field of 'struct mce'
+ * and historically has been the register value of the
+ * MSR_IA32_THERMAL_STATUS (Intel) msr.
+ */
+void mce_log_therm_throt_event(unsigned int cpu, __u64 status)
+{
+	struct mce m;
+
+	memset(&m, 0, sizeof(m));
+	m.cpu = cpu;
+	m.bank = MCE_THERMAL_BANK;
+	m.status = status;
+	rdtscll(m.tsc);
+	mce_log(&m);
+}
+#endif /* CONFIG_X86_MCE_INTEL */
+
 /*
  * Periodic polling timer for "silent" machine check errors.
  */

arch/x86_64/kernel/mce_intel.c

@@ -11,36 +11,21 @@
 #include <asm/mce.h>
 #include <asm/hw_irq.h>
 #include <asm/idle.h>
-
-static DEFINE_PER_CPU(unsigned long, next_check);
+#include <asm/therm_throt.h>
 
 asmlinkage void smp_thermal_interrupt(void)
 {
-	struct mce m;
+	__u64 msr_val;
 
 	ack_APIC_irq();
 
 	exit_idle();
 	irq_enter();
-	if (time_before(jiffies, __get_cpu_var(next_check)))
-		goto done;
-
-	__get_cpu_var(next_check) = jiffies + HZ*300;
-	memset(&m, 0, sizeof(m));
-	m.cpu = smp_processor_id();
-	m.bank = MCE_THERMAL_BANK;
-	rdtscll(m.tsc);
-	rdmsrl(MSR_IA32_THERM_STATUS, m.status);
-	if (m.status & 0x1) {
-		printk(KERN_EMERG
-			"CPU%d: Temperature above threshold, cpu clock throttled\n", m.cpu);
-		add_taint(TAINT_MACHINE_CHECK);
-	} else {
-		printk(KERN_EMERG "CPU%d: Temperature/speed normal\n", m.cpu);
-	}
 
-	mce_log(&m);
-done:
+	rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
+	if (therm_throt_process(msr_val & 1))
+		mce_log_therm_throt_event(smp_processor_id(), msr_val);
+
 	irq_exit();
 }
 

include/asm-i386/therm_throt.h

+#ifndef __ASM_I386_THERM_THROT_H__
+#define __ASM_I386_THERM_THROT_H__ 1
+
+int therm_throt_process(int curr);
+
+#endif /* __ASM_I386_THERM_THROT_H__ */

include/asm-x86_64/mce.h

@@ -99,6 +99,8 @@ static inline void mce_amd_feature_init(struct cpuinfo_x86 *c)
 }
 #endif
 
+void mce_log_therm_throt_event(unsigned int cpu, __u64 status);
+
 extern atomic_t mce_entry;
 
 #endif

include/asm-x86_64/therm_throt.h

+#include <asm-i386/therm_throt.h>