From 15d5f8398311f565682959daaca30e3ca7aea600 Mon Sep 17 00:00:00 2001 From: Dmitriy Zavin Date: Tue, 26 Sep 2006 10:52:42 +0200 Subject: [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 Signed-off-by: Andi Kleen --- arch/x86_64/kernel/mce_intel.c | 27 ++++++--------------------- 1 file changed, 6 insertions(+), 21 deletions(-) (limited to 'arch/x86_64/kernel/mce_intel.c') diff --git a/arch/x86_64/kernel/mce_intel.c b/arch/x86_64/kernel/mce_intel.c index 8f533d2c40cb..dec11219e278 100644 --- a/arch/x86_64/kernel/mce_intel.c +++ b/arch/x86_64/kernel/mce_intel.c @@ -11,36 +11,21 @@ #include #include #include - -static DEFINE_PER_CPU(unsigned long, next_check); +#include 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(); } -- cgit v1.2.3 From 3222b36f46c22f46697a0a53fa8804153a32669f Mon Sep 17 00:00:00 2001 From: Dmitriy Zavin Date: Tue, 26 Sep 2006 10:52:42 +0200 Subject: [PATCH] x86: Add a cumulative thermal throttle event counter. The counter is exported to /sys that keeps track of the number of thermal events, such that the user knows how bad the thermal problem might be (since the logging to syslog and mcelog is rate limited). AK: Fixed cpu hotplug locking Signed-off-by: Dmitriy Zavin Signed-off-by: Andi Kleen --- arch/x86_64/kernel/mce_intel.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/x86_64/kernel/mce_intel.c') diff --git a/arch/x86_64/kernel/mce_intel.c b/arch/x86_64/kernel/mce_intel.c index dec11219e278..6551505d8a2c 100644 --- a/arch/x86_64/kernel/mce_intel.c +++ b/arch/x86_64/kernel/mce_intel.c @@ -77,6 +77,9 @@ static void __cpuinit intel_init_thermal(struct cpuinfo_x86 *c) apic_write(APIC_LVTTHMR, l & ~APIC_LVT_MASKED); printk(KERN_INFO "CPU%d: Thermal monitoring enabled (%s)\n", cpu, tm2 ? "TM2" : "TM1"); + + /* enable thermal throttle processing */ + atomic_set(&therm_throt_en, 1); return; } -- cgit v1.2.3