mce_64.c

linux 内核源代码

		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);

	set_in_cr4(X86_CR4_MCE);

	if (cap & MCG_CTL_P)
		wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff);

	for (i = 0; i < banks; i++) {
		wrmsrl(MSR_IA32_MC0_CTL+4*i, bank[i]);
		wrmsrl(MSR_IA32_MC0_STATUS+4*i, 0);
	}
}

/* Add per CPU specific workarounds here */
static void __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c)
{
	/* This should be disabled by the BIOS, but isn't always */
	if (c->x86_vendor == X86_VENDOR_AMD && c->x86 == 15) {
		/* disable GART TBL walk error reporting, which trips off
		   incorrectly with the IOMMU & 3ware & Cerberus. */
		clear_bit(10, &bank[4]);
		/* Lots of broken BIOS around that don't clear them
		   by default and leave crap in there. Don't log. */
		mce_bootlog = 0;
	}

}

static void __cpuinit mce_cpu_features(struct cpuinfo_x86 *c)
{
	switch (c->x86_vendor) {
	case X86_VENDOR_INTEL:
		mce_intel_feature_init(c);
		break;
	case X86_VENDOR_AMD:
		mce_amd_feature_init(c);
		break;
	default:
		break;
	}
}

/*
 * Called for each booted CPU to set up machine checks.
 * Must be called with preempt off.
 */
void __cpuinit mcheck_init(struct cpuinfo_x86 *c)
{
	static cpumask_t mce_cpus = CPU_MASK_NONE;

	mce_cpu_quirks(c);

	if (mce_dont_init ||
	    cpu_test_and_set(smp_processor_id(), mce_cpus) ||
	    !mce_available(c))
		return;

	mce_init(NULL);
	mce_cpu_features(c);
}

/*
 * Character device to read and clear the MCE log.
 */

static DEFINE_SPINLOCK(mce_state_lock);
static int open_count;	/* #times opened */
static int open_exclu;	/* already open exclusive? */

static int mce_open(struct inode *inode, struct file *file)
{
	spin_lock(&mce_state_lock);

	if (open_exclu || (open_count && (file->f_flags & O_EXCL))) {
		spin_unlock(&mce_state_lock);
		return -EBUSY;
	}

	if (file->f_flags & O_EXCL)
		open_exclu = 1;
	open_count++;

	spin_unlock(&mce_state_lock);

	return nonseekable_open(inode, file);
}

static int mce_release(struct inode *inode, struct file *file)
{
	spin_lock(&mce_state_lock);

	open_count--;
	open_exclu = 0;

	spin_unlock(&mce_state_lock);

	return 0;
}

static void collect_tscs(void *data)
{
	unsigned long *cpu_tsc = (unsigned long *)data;

	rdtscll(cpu_tsc[smp_processor_id()]);
}

static ssize_t mce_read(struct file *filp, char __user *ubuf, size_t usize,
			loff_t *off)
{
	unsigned long *cpu_tsc;
	static DECLARE_MUTEX(mce_read_sem);
	unsigned next;
	char __user *buf = ubuf;
	int i, err;

	cpu_tsc = kmalloc(NR_CPUS * sizeof(long), GFP_KERNEL);
	if (!cpu_tsc)
		return -ENOMEM;

	down(&mce_read_sem);
	next = rcu_dereference(mcelog.next);

	/* Only supports full reads right now */
	if (*off != 0 || usize < MCE_LOG_LEN*sizeof(struct mce)) {
		up(&mce_read_sem);
		kfree(cpu_tsc);
		return -EINVAL;
	}

	err = 0;
	for (i = 0; i < next; i++) {
		unsigned long start = jiffies;

		while (!mcelog.entry[i].finished) {
			if (time_after_eq(jiffies, start + 2)) {
				memset(mcelog.entry + i,0, sizeof(struct mce));
				goto timeout;
			}
			cpu_relax();
		}
		smp_rmb();
		err |= copy_to_user(buf, mcelog.entry + i, sizeof(struct mce));
		buf += sizeof(struct mce);
 timeout:
		;
	}

	memset(mcelog.entry, 0, next * sizeof(struct mce));
	mcelog.next = 0;

	synchronize_sched();

	/*
	 * Collect entries that were still getting written before the
	 * synchronize.
	 */
	on_each_cpu(collect_tscs, cpu_tsc, 1, 1);
	for (i = next; i < MCE_LOG_LEN; i++) {
		if (mcelog.entry[i].finished &&
		    mcelog.entry[i].tsc < cpu_tsc[mcelog.entry[i].cpu]) {
			err |= copy_to_user(buf, mcelog.entry+i,
					    sizeof(struct mce));
			smp_rmb();
			buf += sizeof(struct mce);
			memset(&mcelog.entry[i], 0, sizeof(struct mce));
		}
	}
	up(&mce_read_sem);
	kfree(cpu_tsc);
	return err ? -EFAULT : buf - ubuf;
}

static unsigned int mce_poll(struct file *file, poll_table *wait)
{
	poll_wait(file, &mce_wait, wait);
	if (rcu_dereference(mcelog.next))
		return POLLIN | POLLRDNORM;
	return 0;
}

static int mce_ioctl(struct inode *i, struct file *f,unsigned int cmd,
		     unsigned long arg)
{
	int __user *p = (int __user *)arg;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	switch (cmd) {
	case MCE_GET_RECORD_LEN:
		return put_user(sizeof(struct mce), p);
	case MCE_GET_LOG_LEN:
		return put_user(MCE_LOG_LEN, p);
	case MCE_GETCLEAR_FLAGS: {
		unsigned flags;

		do {
			flags = mcelog.flags;
		} while (cmpxchg(&mcelog.flags, flags, 0) != flags);
		return put_user(flags, p);
	}
	default:
		return -ENOTTY;
	}
}

static const struct file_operations mce_chrdev_ops = {
	.open = mce_open,
	.release = mce_release,
	.read = mce_read,
	.poll = mce_poll,
	.ioctl = mce_ioctl,
};

static struct miscdevice mce_log_device = {
	MISC_MCELOG_MINOR,
	"mcelog",
	&mce_chrdev_ops,
};

static unsigned long old_cr4 __initdata;

void __init stop_mce(void)
{
	old_cr4 = read_cr4();
	clear_in_cr4(X86_CR4_MCE);
}

void __init restart_mce(void)
{
	if (old_cr4 & X86_CR4_MCE)
		set_in_cr4(X86_CR4_MCE);
}

/*
 * Old style boot options parsing. Only for compatibility.
 */
static int __init mcheck_disable(char *str)
{
	mce_dont_init = 1;
	return 1;
}

/* mce=off disables machine check. Note you can re-enable it later
   using sysfs.
   mce=TOLERANCELEVEL (number, see above)
   mce=bootlog Log MCEs from before booting. Disabled by default on AMD.
   mce=nobootlog Don't log MCEs from before booting. */
static int __init mcheck_enable(char *str)
{
	if (!strcmp(str, "off"))
		mce_dont_init = 1;
	else if (!strcmp(str, "bootlog") || !strcmp(str,"nobootlog"))
		mce_bootlog = str[0] == 'b';
	else if (isdigit(str[0]))
		get_option(&str, &tolerant);
	else
		printk("mce= argument %s ignored. Please use /sys", str);
	return 1;
}

__setup("nomce", mcheck_disable);
__setup("mce=", mcheck_enable);

/*
 * Sysfs support
 */

/* On resume clear all MCE state. Don't want to see leftovers from the BIOS.
   Only one CPU is active at this time, the others get readded later using
   CPU hotplug. */
static int mce_resume(struct sys_device *dev)
{
	mce_init(NULL);
	return 0;
}

/* Reinit MCEs after user configuration changes */
static void mce_restart(void)
{
	if (next_interval)
		cancel_delayed_work(&mcheck_work);
	/* Timer race is harmless here */
	on_each_cpu(mce_init, NULL, 1, 1);
	next_interval = check_interval * HZ;
	if (next_interval)
		schedule_delayed_work(&mcheck_work,
				      round_jiffies_relative(next_interval));
}

static struct sysdev_class mce_sysclass = {
	.resume = mce_resume,
	set_kset_name("machinecheck"),
};

DEFINE_PER_CPU(struct sys_device, device_mce);

/* Why are there no generic functions for this? */
#define ACCESSOR(name, var, start) \
	static ssize_t show_ ## name(struct sys_device *s, char *buf) {	\
		return sprintf(buf, "%lx\n", (unsigned long)var);	\
	}								\
	static ssize_t set_ ## name(struct sys_device *s,const char *buf,size_t siz) { \
		char *end;						\
		unsigned long new = simple_strtoul(buf, &end, 0);	\
		if (end == buf) return -EINVAL;				\
		var = new;						\
		start;							\
		return end-buf;						\
	}								\
	static SYSDEV_ATTR(name, 0644, show_ ## name, set_ ## name);

/* TBD should generate these dynamically based on number of available banks */
ACCESSOR(bank0ctl,bank[0],mce_restart())
ACCESSOR(bank1ctl,bank[1],mce_restart())
ACCESSOR(bank2ctl,bank[2],mce_restart())
ACCESSOR(bank3ctl,bank[3],mce_restart())
ACCESSOR(bank4ctl,bank[4],mce_restart())
ACCESSOR(bank5ctl,bank[5],mce_restart())

static ssize_t show_trigger(struct sys_device *s, char *buf)
{
	strcpy(buf, trigger);
	strcat(buf, "\n");
	return strlen(trigger) + 1;
}

static ssize_t set_trigger(struct sys_device *s,const char *buf,size_t siz)
{
	char *p;
	int len;
	strncpy(trigger, buf, sizeof(trigger));
	trigger[sizeof(trigger)-1] = 0;
	len = strlen(trigger);
	p = strchr(trigger, '\n');
	if (*p) *p = 0;
	return len;
}

static SYSDEV_ATTR(trigger, 0644, show_trigger, set_trigger);
ACCESSOR(tolerant,tolerant,)
ACCESSOR(check_interval,check_interval,mce_restart())
static struct sysdev_attribute *mce_attributes[] = {
	&attr_bank0ctl, &attr_bank1ctl, &attr_bank2ctl,
	&attr_bank3ctl, &attr_bank4ctl, &attr_bank5ctl,
	&attr_tolerant, &attr_check_interval, &attr_trigger,
	NULL
};

static cpumask_t mce_device_initialized = CPU_MASK_NONE;

/* Per cpu sysdev init.  All of the cpus still share the same ctl bank */
static __cpuinit int mce_create_device(unsigned int cpu)
{
	int err;
	int i;

	if (!mce_available(&boot_cpu_data))
		return -EIO;

	memset(&per_cpu(device_mce, cpu).kobj, 0, sizeof(struct kobject));
	per_cpu(device_mce,cpu).id = cpu;
	per_cpu(device_mce,cpu).cls = &mce_sysclass;

	err = sysdev_register(&per_cpu(device_mce,cpu));
	if (err)
		return err;

	for (i = 0; mce_attributes[i]; i++) {
		err = sysdev_create_file(&per_cpu(device_mce,cpu),
					 mce_attributes[i]);
		if (err)
			goto error;
	}
	cpu_set(cpu, mce_device_initialized);

	return 0;
error:
	while (i--) {
		sysdev_remove_file(&per_cpu(device_mce,cpu),
				   mce_attributes[i]);
	}
	sysdev_unregister(&per_cpu(device_mce,cpu));

	return err;
}

static void mce_remove_device(unsigned int cpu)
{
	int i;

	if (!cpu_isset(cpu, mce_device_initialized))
		return;

	for (i = 0; mce_attributes[i]; i++)
		sysdev_remove_file(&per_cpu(device_mce,cpu),
			mce_attributes[i]);
	sysdev_unregister(&per_cpu(device_mce,cpu));
	cpu_clear(cpu, mce_device_initialized);
}

/* Get notified when a cpu comes on/off. Be hotplug friendly. */
static int
mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
{
	unsigned int cpu = (unsigned long)hcpu;

	switch (action) {
	case CPU_ONLINE:
	case CPU_ONLINE_FROZEN:
		mce_create_device(cpu);
		break;
	case CPU_DEAD:
	case CPU_DEAD_FROZEN:
		mce_remove_device(cpu);
		break;
	}
	return NOTIFY_OK;
}

static struct notifier_block mce_cpu_notifier = {
	.notifier_call = mce_cpu_callback,
};

static __init int mce_init_device(void)
{
	int err;
	int i = 0;

	if (!mce_available(&boot_cpu_data))
		return -EIO;
	err = sysdev_class_register(&mce_sysclass);
	if (err)
		return err;

	for_each_online_cpu(i) {
		err = mce_create_device(i);
		if (err)
			return err;
	}

	register_hotcpu_notifier(&mce_cpu_notifier);
	misc_register(&mce_log_device);
	return err;
}

device_initcall(mce_init_device);