mce.c

linux-2.6.15.6

/*
 * Machine check handler.
 * K8 parts Copyright 2002,2003 Andi Kleen, SuSE Labs.
 * Rest from unknown author(s). 
 * 2004 Andi Kleen. Rewrote most of it. 
 */

#include <linux/init.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/rcupdate.h>
#include <linux/kallsyms.h>
#include <linux/sysdev.h>
#include <linux/miscdevice.h>
#include <linux/fs.h>
#include <linux/cpu.h>
#include <linux/percpu.h>
#include <linux/ctype.h>
#include <asm/processor.h> 
#include <asm/msr.h>
#include <asm/mce.h>
#include <asm/kdebug.h>
#include <asm/uaccess.h>

#define MISC_MCELOG_MINOR 227
#define NR_BANKS 5

static int mce_dont_init;

/* 0: always panic, 1: panic if deadlock possible, 2: try to avoid panic,
   3: never panic or exit (for testing only) */
static int tolerant = 1;
static int banks;
static unsigned long bank[NR_BANKS] = { [0 ... NR_BANKS-1] = ~0UL };
static unsigned long console_logged;
static int notify_user;
static int rip_msr;
static int mce_bootlog = 1;

/*
 * Lockless MCE logging infrastructure.
 * This avoids deadlocks on printk locks without having to break locks. Also
 * separate MCEs from kernel messages to avoid bogus bug reports.
 */

struct mce_log mcelog = { 
	MCE_LOG_SIGNATURE,
	MCE_LOG_LEN,
}; 

void mce_log(struct mce *mce)
{
	unsigned next, entry;
	mce->finished = 0;
	wmb();
	for (;;) {
		entry = rcu_dereference(mcelog.next);
		/* The rmb forces the compiler to reload next in each
		    iteration */
		rmb();
		for (;;) {
			/* When the buffer fills up discard new entries. Assume
			   that the earlier errors are the more interesting. */
			if (entry >= MCE_LOG_LEN) {
				set_bit(MCE_OVERFLOW, &mcelog.flags);
				return;
			}
			/* Old left over entry. Skip. */
			if (mcelog.entry[entry].finished) {
				entry++;
				continue;
			}
			break;
		}
		smp_rmb();
		next = entry + 1;
		if (cmpxchg(&mcelog.next, entry, next) == entry)
			break;
	}
	memcpy(mcelog.entry + entry, mce, sizeof(struct mce));
	wmb();
	mcelog.entry[entry].finished = 1;
	wmb();

	if (!test_and_set_bit(0, &console_logged))
		notify_user = 1;
}

static void print_mce(struct mce *m)
{
	printk(KERN_EMERG "\n"
	       KERN_EMERG
	       "CPU %d: Machine Check Exception: %16Lx Bank %d: %016Lx\n",
	       m->cpu, m->mcgstatus, m->bank, m->status);
	if (m->rip) {
		printk(KERN_EMERG 
		       "RIP%s %02x:<%016Lx> ",
		       !(m->mcgstatus & MCG_STATUS_EIPV) ? " !INEXACT!" : "",
		       m->cs, m->rip);
		if (m->cs == __KERNEL_CS)
			print_symbol("{%s}", m->rip);
		printk("\n");
	}
	printk(KERN_EMERG "TSC %Lx ", m->tsc); 
	if (m->addr)
		printk("ADDR %Lx ", m->addr);
	if (m->misc)
		printk("MISC %Lx ", m->misc); 	
	printk("\n");
}

static void mce_panic(char *msg, struct mce *backup, unsigned long start)
{ 
	int i;
	oops_begin();
	for (i = 0; i < MCE_LOG_LEN; i++) {
		unsigned long tsc = mcelog.entry[i].tsc;
		if (time_before(tsc, start))
			continue;
		print_mce(&mcelog.entry[i]); 
		if (backup && mcelog.entry[i].tsc == backup->tsc)
			backup = NULL;
	}
	if (backup)
		print_mce(backup);
	if (tolerant >= 3)
		printk("Fake panic: %s\n", msg);
	else
		panic(msg);
} 

static int mce_available(struct cpuinfo_x86 *c)
{
	return test_bit(X86_FEATURE_MCE, &c->x86_capability) &&
	       test_bit(X86_FEATURE_MCA, &c->x86_capability);
}

static inline void mce_get_rip(struct mce *m, struct pt_regs *regs)
{
	if (regs && (m->mcgstatus & MCG_STATUS_RIPV)) {
		m->rip = regs->rip;
		m->cs = regs->cs;
	} else {
		m->rip = 0;
		m->cs = 0;
	}
	if (rip_msr) {
		/* Assume the RIP in the MSR is exact. Is this true? */
		m->mcgstatus |= MCG_STATUS_EIPV;
		rdmsrl(rip_msr, m->rip);
		m->cs = 0;
	}
}

/* 
 * The actual machine check handler
 */

void do_machine_check(struct pt_regs * regs, long error_code)
{
	struct mce m, panicm;
	int nowayout = (tolerant < 1); 
	int kill_it = 0;
	u64 mcestart = 0;
	int i;
	int panicm_found = 0;

	if (regs)
		notify_die(DIE_NMI, "machine check", regs, error_code, 255, SIGKILL);
	if (!banks)
		return;

	memset(&m, 0, sizeof(struct mce));
	m.cpu = hard_smp_processor_id();
	rdmsrl(MSR_IA32_MCG_STATUS, m.mcgstatus);
	if (!(m.mcgstatus & MCG_STATUS_RIPV))
		kill_it = 1;
	
	rdtscll(mcestart);
	barrier();

	for (i = 0; i < banks; i++) {
		if (!bank[i])
			continue;
		
		m.misc = 0; 
		m.addr = 0;
		m.bank = i;
		m.tsc = 0;

		rdmsrl(MSR_IA32_MC0_STATUS + i*4, m.status);
		if ((m.status & MCI_STATUS_VAL) == 0)
			continue;

		if (m.status & MCI_STATUS_EN) {
			/* In theory _OVER could be a nowayout too, but
			   assume any overflowed errors were no fatal. */
			nowayout |= !!(m.status & MCI_STATUS_PCC);
			kill_it |= !!(m.status & MCI_STATUS_UC);
		}

		if (m.status & MCI_STATUS_MISCV)
			rdmsrl(MSR_IA32_MC0_MISC + i*4, m.misc);
		if (m.status & MCI_STATUS_ADDRV)
			rdmsrl(MSR_IA32_MC0_ADDR + i*4, m.addr);

		mce_get_rip(&m, regs);
		if (error_code >= 0)
			rdtscll(m.tsc);
		wrmsrl(MSR_IA32_MC0_STATUS + i*4, 0);
		if (error_code != -2)
			mce_log(&m);

		/* Did this bank cause the exception? */
		/* Assume that the bank with uncorrectable errors did it,
		   and that there is only a single one. */
		if ((m.status & MCI_STATUS_UC) && (m.status & MCI_STATUS_EN)) {
			panicm = m;
			panicm_found = 1;
		}

		add_taint(TAINT_MACHINE_CHECK);
	}

	/* Never do anything final in the polling timer */
	if (!regs)
		goto out;

	/* If we didn't find an uncorrectable error, pick
	   the last one (shouldn't happen, just being safe). */
	if (!panicm_found)
		panicm = m;
	if (nowayout)
		mce_panic("Machine check", &panicm, mcestart);
	if (kill_it) {
		int user_space = 0;

		if (m.mcgstatus & MCG_STATUS_RIPV)
			user_space = panicm.rip && (panicm.cs & 3);
		
		/* When the machine was in user space and the CPU didn't get
		   confused it's normally not necessary to panic, unless you 
		   are paranoid (tolerant == 0)

		   RED-PEN could be more tolerant for MCEs in idle,
		   but most likely they occur at boot anyways, where
		   it is best to just halt the machine. */
		if ((!user_space && (panic_on_oops || tolerant < 2)) ||
		    (unsigned)current->pid <= 1)
			mce_panic("Uncorrected machine check", &panicm, mcestart);

		/* do_exit takes an awful lot of locks and has as
		   slight risk of deadlocking. If you don't want that
		   don't set tolerant >= 2 */
		if (tolerant < 3)
			do_exit(SIGBUS);
	}

 out:
	/* Last thing done in the machine check exception to clear state. */
	wrmsrl(MSR_IA32_MCG_STATUS, 0);
}

/*
 * Periodic polling timer for "silent" machine check errors.
 */

static int check_interval = 5 * 60; /* 5 minutes */
static void mcheck_timer(void *data);
static DECLARE_WORK(mcheck_work, mcheck_timer, NULL);

static void mcheck_check_cpu(void *info)
{
	if (mce_available(&current_cpu_data))
		do_machine_check(NULL, 0);
}

static void mcheck_timer(void *data)
{
	on_each_cpu(mcheck_check_cpu, NULL, 1, 1);
	schedule_delayed_work(&mcheck_work, check_interval * HZ);

	/*
	 * It's ok to read stale data here for notify_user and
	 * console_logged as we'll simply get the updated versions
	 * on the next mcheck_timer execution and atomic operations
	 * on console_logged act as synchronization for notify_user
	 * writes.
	 */
	if (notify_user && console_logged) {
		notify_user = 0;
		clear_bit(0, &console_logged);
		printk(KERN_INFO "Machine check events logged\n");
	}
}


static __init int periodic_mcheck_init(void)
{ 
	if (check_interval)
		schedule_delayed_work(&mcheck_work, check_interval*HZ);
	return 0;
} 
__initcall(periodic_mcheck_init);


/* 
 * Initialize Machine Checks for a CPU.
 */
static void mce_init(void *dummy)
{
	u64 cap;
	int i;

	rdmsrl(MSR_IA32_MCG_CAP, cap);
	banks = cap & 0xff;
	if (banks > NR_BANKS) { 
		printk(KERN_INFO "MCE: warning: using only %d banks\n", banks);
		banks = NR_BANKS; 
	}
	/* Use accurate RIP reporting if available. */
	if ((cap & (1<<9)) && ((cap >> 16) & 0xff) >= 9)
		rip_msr = MSR_IA32_MCG_EIP;

	/* Log the machine checks left over from the previous reset.
	   This also clears all registers */
	do_machine_check(NULL, mce_bootlog ? -1 : -2);