aboutsummaryrefslogtreecommitdiff
path: root/arch/x86/kernel/nmi.c
blob: ac6d51222e7d3562abb1e7b5e8bd4a16a1c1c1f4 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
/*
 *  NMI watchdog support on APIC systems
 *
 *  Started by Ingo Molnar <mingo@redhat.com>
 *
 *  Fixes:
 *  Mikael Pettersson	: AMD K7 support for local APIC NMI watchdog.
 *  Mikael Pettersson	: Power Management for local APIC NMI watchdog.
 *  Mikael Pettersson	: Pentium 4 support for local APIC NMI watchdog.
 *  Pavel Machek and
 *  Mikael Pettersson	: PM converted to driver model. Disable/enable API.
 */

#include <asm/apic.h>

#include <linux/nmi.h>
#include <linux/mm.h>
#include <linux/delay.h>
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/sysdev.h>
#include <linux/sysctl.h>
#include <linux/percpu.h>
#include <linux/kprobes.h>
#include <linux/cpumask.h>
#include <linux/kernel_stat.h>
#include <linux/kdebug.h>
#include <linux/smp.h>

#include <asm/i8259.h>
#include <asm/io_apic.h>
#include <asm/smp.h>
#include <asm/nmi.h>
#include <asm/proto.h>
#include <asm/timer.h>

#include <asm/mce.h>

#include <mach_traps.h>

int unknown_nmi_panic;
int nmi_watchdog_enabled;

static cpumask_t backtrace_mask = CPU_MASK_NONE;

/* nmi_active:
 * >0: the lapic NMI watchdog is active, but can be disabled
 * <0: the lapic NMI watchdog has not been set up, and cannot
 *     be enabled
 *  0: the lapic NMI watchdog is disabled, but can be enabled
 */
atomic_t nmi_active = ATOMIC_INIT(0);		/* oprofile uses this */
EXPORT_SYMBOL(nmi_active);

unsigned int nmi_watchdog = NMI_NONE;
EXPORT_SYMBOL(nmi_watchdog);

static int panic_on_timeout;

static unsigned int nmi_hz = HZ;
static DEFINE_PER_CPU(short, wd_enabled);
static int endflag __initdata;

static inline unsigned int get_nmi_count(int cpu)
{
#ifdef CONFIG_X86_64
	return cpu_pda(cpu)->__nmi_count;
#else
	return nmi_count(cpu);
#endif
}

static inline int mce_in_progress(void)
{
#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE)
	return atomic_read(&mce_entry) > 0;
#endif
	return 0;
}

/*
 * Take the local apic timer and PIT/HPET into account. We don't
 * know which one is active, when we have highres/dyntick on
 */
static inline unsigned int get_timer_irqs(int cpu)
{
#ifdef CONFIG_X86_64
	return read_pda(apic_timer_irqs) + read_pda(irq0_irqs);
#else
	return per_cpu(irq_stat, cpu).apic_timer_irqs +
		per_cpu(irq_stat, cpu).irq0_irqs;
#endif
}

#ifdef CONFIG_SMP
/*
 * The performance counters used by NMI_LOCAL_APIC don't trigger when
 * the CPU is idle. To make sure the NMI watchdog really ticks on all
 * CPUs during the test make them busy.
 */
static __init void nmi_cpu_busy(void *data)
{
	local_irq_enable_in_hardirq();
	/*
	 * Intentionally don't use cpu_relax here. This is
	 * to make sure that the performance counter really ticks,
	 * even if there is a simulator or similar that catches the
	 * pause instruction. On a real HT machine this is fine because
	 * all other CPUs are busy with "useless" delay loops and don't
	 * care if they get somewhat less cycles.
	 */
	while (endflag == 0)
		mb();
}
#endif

int __init check_nmi_watchdog(void)
{
	unsigned int *prev_nmi_count;
	int cpu;

	if (!nmi_watchdog_active() || !atomic_read(&nmi_active))
		return 0;

	prev_nmi_count = kmalloc(nr_cpu_ids * sizeof(int), GFP_KERNEL);
	if (!prev_nmi_count)
		goto error;

	printk(KERN_INFO "Testing NMI watchdog ... ");

#ifdef CONFIG_SMP
	if (nmi_watchdog == NMI_LOCAL_APIC)
		smp_call_function(nmi_cpu_busy, (void *)&endflag, 0);
#endif

	for_each_possible_cpu(cpu)
		prev_nmi_count[cpu] = get_nmi_count(cpu);
	local_irq_enable();
	mdelay((20 * 1000) / nmi_hz); /* wait 20 ticks */

	for_each_online_cpu(cpu) {
		if (!per_cpu(wd_enabled, cpu))
			continue;
		if (get_nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
			printk(KERN_WARNING "WARNING: CPU#%d: NMI "
				"appears to be stuck (%d->%d)!\n",
				cpu,
				prev_nmi_count[cpu],
				get_nmi_count(cpu));
			per_cpu(wd_enabled, cpu) = 0;
			atomic_dec(&nmi_active);
		}
	}
	endflag = 1;
	if (!atomic_read(&nmi_active)) {
		kfree(prev_nmi_count);
		atomic_set(&nmi_active, -1);
		goto error;
	}
	printk("OK.\n");

	/*
	 * now that we know it works we can reduce NMI frequency to
	 * something more reasonable; makes a difference in some configs
	 */
	if (nmi_watchdog == NMI_LOCAL_APIC)
		nmi_hz = lapic_adjust_nmi_hz(1);

	kfree(prev_nmi_count);
	return 0;
error:
	if (nmi_watchdog == NMI_IO_APIC && !timer_through_8259)
		disable_8259A_irq(0);
#ifdef CONFIG_X86_32
	timer_ack = 0;
#endif
	return -1;
}

static int __init setup_nmi_watchdog(char *str)
{
	unsigned int nmi;

	if (!strncmp(str, "panic", 5)) {
		panic_on_timeout = 1;
		str = strchr(str, ',');
		if (!str)
			return 1;
		++str;
	}

	get_option(&str, &nmi);

	if (nmi >= NMI_INVALID)
		return 0;

	nmi_watchdog = nmi;
	return 1;
}
__setup("nmi_watchdog=", setup_nmi_watchdog);

/*
 * Suspend/resume support
 */
#ifdef CONFIG_PM

static int nmi_pm_active; /* nmi_active before suspend */

static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state)
{
	/* only CPU0 goes here, other CPUs should be offline */
	nmi_pm_active = atomic_read(&nmi_active);
	stop_apic_nmi_watchdog(NULL);
	BUG_ON(atomic_read(&nmi_active) != 0);
	return 0;
}

static int lapic_nmi_resume(struct sys_device *dev)
{
	/* only CPU0 goes here, other CPUs should be offline */
	if (nmi_pm_active > 0) {
		setup_apic_nmi_watchdog(NULL);
		touch_nmi_watchdog();
	}
	return 0;
}

static struct sysdev_class nmi_sysclass = {
	.name		= "lapic_nmi",
	.resume		= lapic_nmi_resume,
	.suspend	= lapic_nmi_suspend,
};

static struct sys_device device_lapic_nmi = {
	.id	= 0,
	.cls	= &nmi_sysclass,
};

static int __init init_lapic_nmi_sysfs(void)
{
	int error;

	/*
	 * should really be a BUG_ON but b/c this is an
	 * init call, it just doesn't work.  -dcz
	 */
	if (nmi_watchdog != NMI_LOCAL_APIC)
		return 0;

	if (atomic_read(&nmi_active) < 0)
		return 0;

	error = sysdev_class_register(&nmi_sysclass);
	if (!error)
		error = sysdev_register(&device_lapic_nmi);
	return error;
}

/* must come after the local APIC's device_initcall() */
late_initcall(init_lapic_nmi_sysfs);

#endif	/* CONFIG_PM */

static void __acpi_nmi_enable(void *__unused)
{
	apic_write(APIC_LVT0, APIC_DM_NMI);
}

/*
 * Enable timer based NMIs on all CPUs:
 */
void acpi_nmi_enable(void)
{
	if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC)
		on_each_cpu(__acpi_nmi_enable, NULL, 1);
}

static void __acpi_nmi_disable(void *__unused)
{
	apic_write(APIC_LVT0, APIC_DM_NMI | APIC_LVT_MASKED);
}

/*
 * Disable timer based NMIs on all CPUs:
 */
void acpi_nmi_disable(void)
{
	if (atomic_read(&nmi_active) && nmi_watchdog == NMI_IO_APIC)
		on_each_cpu(__acpi_nmi_disable, NULL, 1);
}

void setup_apic_nmi_watchdog(void *unused)
{
	if (__get_cpu_var(wd_enabled))
		return;

	/* cheap hack to support suspend/resume */
	/* if cpu0 is not active neither should the other cpus */
	if (smp_processor_id() != 0 && atomic_read(&nmi_active) <= 0)
		return;

	switch (nmi_watchdog) {
	case NMI_LOCAL_APIC:
		 /* enable it before to avoid race with handler */
		__get_cpu_var(wd_enabled) = 1;
		if (lapic_watchdog_init(nmi_hz) < 0) {
			__get_cpu_var(wd_enabled) = 0;
			return;
		}
		/* FALL THROUGH */
	case NMI_IO_APIC:
		__get_cpu_var(wd_enabled) = 1;
		atomic_inc(&nmi_active);
	}
}

void stop_apic_nmi_watchdog(void *unused)
{
	/* only support LOCAL and IO APICs for now */
	if (!nmi_watchdog_active())
		return;
	if (__get_cpu_var(wd_enabled) == 0)
		return;
	if (nmi_watchdog == NMI_LOCAL_APIC)
		lapic_watchdog_stop();
	__get_cpu_var(wd_enabled) = 0;
	atomic_dec(&nmi_active);
}

/*
 * the best way to detect whether a CPU has a 'hard lockup' problem
 * is to check it's local APIC timer IRQ counts. If they are not
 * changing then that CPU has some problem.
 *
 * as these watchdog NMI IRQs are generated on every CPU, we only
 * have to check the current processor.
 *
 * since NMIs don't listen to _any_ locks, we have to be extremely
 * careful not to rely on unsafe variables. The printk might lock
 * up though, so we have to break up any console locks first ...
 * [when there will be more tty-related locks, break them up here too!]
 */

static DEFINE_PER_CPU(unsigned, last_irq_sum);
static DEFINE_PER_CPU(local_t, alert_counter);
static DEFINE_PER_CPU(int, nmi_touch);

void touch_nmi_watchdog(void)
{
	if (nmi_watchdog_active()) {
		unsigned cpu;

		/*
		 * Tell other CPUs to reset their alert counters. We cannot
		 * do it ourselves because the alert count increase is not
		 * atomic.
		 */
		for_each_present_cpu(cpu) {
			if (per_cpu(nmi_touch, cpu) != 1)
				per_cpu(nmi_touch, cpu) = 1;
		}
	}

	/*
	 * Tickle the softlockup detector too:
	 */
	touch_softlockup_watchdog();
}
EXPORT_SYMBOL(touch_nmi_watchdog);

notrace __kprobes int
nmi_watchdog_tick(struct pt_regs *regs, unsigned reason)
{
	/*
	 * Since current_thread_info()-> is always on the stack, and we
	 * always switch the stack NMI-atomically, it's safe to use
	 * smp_processor_id().
	 */
	unsigned int sum;
	int touched = 0;
	int cpu = smp_processor_id();
	int rc = 0;

	/* check for other users first */
	if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT)
			== NOTIFY_STOP) {
		rc = 1;
		touched = 1;
	}

	sum = get_timer_irqs(cpu);

	if (__get_cpu_var(nmi_touch)) {
		__get_cpu_var(nmi_touch) = 0;
		touched = 1;
	}

	if (cpu_isset(cpu, backtrace_mask)) {
		static DEFINE_SPINLOCK(lock);	/* Serialise the printks */

		spin_lock(&lock);
		printk(KERN_WARNING "NMI backtrace for cpu %d\n", cpu);
		dump_stack();
		spin_unlock(&lock);
		cpu_clear(cpu, backtrace_mask);
	}

	/* Could check oops_in_progress here too, but it's safer not to */
	if (mce_in_progress())
		touched = 1;

	/* if the none of the timers isn't firing, this cpu isn't doing much */
	if (!touched && __get_cpu_var(last_irq_sum) == sum) {
		/*
		 * Ayiee, looks like this CPU is stuck ...
		 * wait a few IRQs (5 seconds) before doing the oops ...
		 */
		local_inc(&__get_cpu_var(alert_counter));
		if (local_read(&__get_cpu_var(alert_counter)) == 5 * nmi_hz)
			/*
			 * die_nmi will return ONLY if NOTIFY_STOP happens..
			 */
			die_nmi("BUG: NMI Watchdog detected LOCKUP",
				regs, panic_on_timeout);
	} else {
		__get_cpu_var(last_irq_sum) = sum;
		local_set(&__get_cpu_var(alert_counter), 0);
	}

	/* see if the nmi watchdog went off */
	if (!__get_cpu_var(wd_enabled))
		return rc;
	switch (nmi_watchdog) {
	case NMI_LOCAL_APIC:
		rc |= lapic_wd_event(nmi_hz);
		break;
	case NMI_IO_APIC:
		/*
		 * don't know how to accurately check for this.
		 * just assume it was a watchdog timer interrupt
		 * This matches the old behaviour.
		 */
		rc = 1;
		break;
	}
	return rc;
}

#ifdef CONFIG_SYSCTL

static int __init setup_unknown_nmi_panic(char *str)
{
	unknown_nmi_panic = 1;
	return 1;
}
__setup("unknown_nmi_panic", setup_unknown_nmi_panic);

static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu)
{
	unsigned char reason = get_nmi_reason();
	char buf[64];

	sprintf(buf, "NMI received for unknown reason %02x\n", reason);
	die_nmi(buf, regs, 1); /* Always panic here */
	return 0;
}

/*
 * proc handler for /proc/sys/kernel/nmi
 */
int proc_nmi_enabled(struct ctl_table *table, int write, struct file *file,
			void __user *buffer, size_t *length, loff_t *ppos)
{
	int old_state;

	nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0;
	old_state = nmi_watchdog_enabled;
	proc_dointvec(table, write, file, buffer, length, ppos);
	if (!!old_state == !!nmi_watchdog_enabled)
		return 0;

	if (atomic_read(&nmi_active) < 0 || !nmi_watchdog_active()) {
		printk(KERN_WARNING
			"NMI watchdog is permanently disabled\n");
		return -EIO;
	}

	if (nmi_watchdog == NMI_LOCAL_APIC) {
		if (nmi_watchdog_enabled)
			enable_lapic_nmi_watchdog();
		else
			disable_lapic_nmi_watchdog();
	} else {
		printk(KERN_WARNING
			"NMI watchdog doesn't know what hardware to touch\n");
		return -EIO;
	}
	return 0;
}

#endif /* CONFIG_SYSCTL */

int do_nmi_callback(struct pt_regs *regs, int cpu)
{
#ifdef CONFIG_SYSCTL
	if (unknown_nmi_panic)
		return unknown_nmi_panic_callback(regs, cpu);
#endif
	return 0;
}

void __trigger_all_cpu_backtrace(void)
{
	int i;

	backtrace_mask = cpu_online_map;
	/* Wait for up to 10 seconds for all CPUs to do the backtrace */
	for (i = 0; i < 10 * 1000; i++) {
		if (cpus_empty(backtrace_mask))
			break;
		mdelay(1);
	}
}