Blame view
kernel/watchdog.c
15.6 KB
58687acba lockup_detector: ... |
1 2 3 4 5 |
/* * Detect hard and soft lockups on a system * * started by Don Zickus, Copyright (C) 2010 Red Hat, Inc. * |
86f5e6a7b watchdog: Fix cod... |
6 7 8 |
* Note: Most of this code is borrowed heavily from the original softlockup * detector, so thanks to Ingo for the initial implementation. * Some chunks also taken from the old x86-specific nmi watchdog code, thanks |
58687acba lockup_detector: ... |
9 10 |
* to those contributors as well. */ |
4501980aa kernel/watchdog.c... |
11 |
#define pr_fmt(fmt) "NMI watchdog: " fmt |
58687acba lockup_detector: ... |
12 13 14 15 16 17 18 19 20 21 22 |
#include <linux/mm.h> #include <linux/cpu.h> #include <linux/nmi.h> #include <linux/init.h> #include <linux/delay.h> #include <linux/freezer.h> #include <linux/kthread.h> #include <linux/lockdep.h> #include <linux/notifier.h> #include <linux/module.h> #include <linux/sysctl.h> |
bcd951cf1 watchdog: Use hot... |
23 |
#include <linux/smpboot.h> |
8bd75c77b sched/rt: Move rt... |
24 |
#include <linux/sched/rt.h> |
58687acba lockup_detector: ... |
25 26 |
#include <asm/irq_regs.h> |
5d1c0f4a8 watchdog: add che... |
27 |
#include <linux/kvm_para.h> |
58687acba lockup_detector: ... |
28 |
#include <linux/perf_event.h> |
3c00ea82c watchdog: Rename ... |
29 |
int watchdog_user_enabled = 1; |
4eec42f39 watchdog: Change ... |
30 |
int __read_mostly watchdog_thresh = 10; |
3c00ea82c watchdog: Rename ... |
31 |
static int __read_mostly watchdog_running; |
0f34c4009 watchdog: store t... |
32 |
static u64 __read_mostly sample_period; |
58687acba lockup_detector: ... |
33 34 35 36 37 |
static DEFINE_PER_CPU(unsigned long, watchdog_touch_ts); static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog); static DEFINE_PER_CPU(struct hrtimer, watchdog_hrtimer); static DEFINE_PER_CPU(bool, softlockup_touch_sync); |
58687acba lockup_detector: ... |
38 |
static DEFINE_PER_CPU(bool, soft_watchdog_warn); |
bcd951cf1 watchdog: Use hot... |
39 40 |
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts); static DEFINE_PER_CPU(unsigned long, soft_lockup_hrtimer_cnt); |
23637d477 lockup_detector: ... |
41 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
cafcd80d2 lockup_detector: ... |
42 43 |
static DEFINE_PER_CPU(bool, hard_watchdog_warn); static DEFINE_PER_CPU(bool, watchdog_nmi_touch); |
58687acba lockup_detector: ... |
44 45 46 |
static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); #endif |
58687acba lockup_detector: ... |
47 48 49 50 |
/* boot commands */ /* * Should we panic when a soft-lockup or hard-lockup occurs: */ |
23637d477 lockup_detector: ... |
51 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
fef2c9bc1 kernel/watchdog.c... |
52 53 |
static int hardlockup_panic = CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE; |
58687acba lockup_detector: ... |
54 55 56 57 58 |
static int __init hardlockup_panic_setup(char *str) { if (!strncmp(str, "panic", 5)) hardlockup_panic = 1; |
fef2c9bc1 kernel/watchdog.c... |
59 60 |
else if (!strncmp(str, "nopanic", 7)) hardlockup_panic = 0; |
5dc305587 x86, NMI: Add bac... |
61 |
else if (!strncmp(str, "0", 1)) |
3c00ea82c watchdog: Rename ... |
62 |
watchdog_user_enabled = 0; |
58687acba lockup_detector: ... |
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 |
return 1; } __setup("nmi_watchdog=", hardlockup_panic_setup); #endif unsigned int __read_mostly softlockup_panic = CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE; static int __init softlockup_panic_setup(char *str) { softlockup_panic = simple_strtoul(str, NULL, 0); return 1; } __setup("softlockup_panic=", softlockup_panic_setup); static int __init nowatchdog_setup(char *str) { |
3c00ea82c watchdog: Rename ... |
81 |
watchdog_user_enabled = 0; |
58687acba lockup_detector: ... |
82 83 84 85 86 87 88 |
return 1; } __setup("nowatchdog", nowatchdog_setup); /* deprecated */ static int __init nosoftlockup_setup(char *str) { |
3c00ea82c watchdog: Rename ... |
89 |
watchdog_user_enabled = 0; |
58687acba lockup_detector: ... |
90 91 92 93 |
return 1; } __setup("nosoftlockup", nosoftlockup_setup); /* */ |
4eec42f39 watchdog: Change ... |
94 95 96 97 98 99 100 |
/* * Hard-lockup warnings should be triggered after just a few seconds. Soft- * lockups can have false positives under extreme conditions. So we generally * want a higher threshold for soft lockups than for hard lockups. So we couple * the thresholds with a factor: we make the soft threshold twice the amount of * time the hard threshold is. */ |
6e9101aee watchdog: Fix non... |
101 |
static int get_softlockup_thresh(void) |
4eec42f39 watchdog: Change ... |
102 103 104 |
{ return watchdog_thresh * 2; } |
58687acba lockup_detector: ... |
105 106 107 108 109 110 |
/* * Returns seconds, approximately. We don't need nanosecond * resolution, and we don't need to waste time with a big divide when * 2^30ns == 1.074s. */ |
c06b4f194 watchdog: Use loc... |
111 |
static unsigned long get_timestamp(void) |
58687acba lockup_detector: ... |
112 |
{ |
c06b4f194 watchdog: Use loc... |
113 |
return local_clock() >> 30LL; /* 2^30 ~= 10^9 */ |
58687acba lockup_detector: ... |
114 |
} |
0f34c4009 watchdog: store t... |
115 |
static void set_sample_period(void) |
58687acba lockup_detector: ... |
116 117 |
{ /* |
586692a5a watchdog: Disable... |
118 |
* convert watchdog_thresh from seconds to ns |
86f5e6a7b watchdog: Fix cod... |
119 120 121 122 |
* the divide by 5 is to give hrtimer several chances (two * or three with the current relation between the soft * and hard thresholds) to increment before the * hardlockup detector generates a warning |
58687acba lockup_detector: ... |
123 |
*/ |
0f34c4009 watchdog: store t... |
124 |
sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5); |
58687acba lockup_detector: ... |
125 126 127 128 129 |
} /* Commands for resetting the watchdog */ static void __touch_watchdog(void) { |
c06b4f194 watchdog: Use loc... |
130 |
__this_cpu_write(watchdog_touch_ts, get_timestamp()); |
58687acba lockup_detector: ... |
131 |
} |
332fbdbca lockup_detector: ... |
132 |
void touch_softlockup_watchdog(void) |
58687acba lockup_detector: ... |
133 |
{ |
909ea9646 core: Replace __g... |
134 |
__this_cpu_write(watchdog_touch_ts, 0); |
58687acba lockup_detector: ... |
135 |
} |
0167c7819 watchdog: Export ... |
136 |
EXPORT_SYMBOL(touch_softlockup_watchdog); |
58687acba lockup_detector: ... |
137 |
|
332fbdbca lockup_detector: ... |
138 |
void touch_all_softlockup_watchdogs(void) |
58687acba lockup_detector: ... |
139 140 141 142 143 144 145 146 147 148 149 |
{ int cpu; /* * this is done lockless * do we care if a 0 races with a timestamp? * all it means is the softlock check starts one cycle later */ for_each_online_cpu(cpu) per_cpu(watchdog_touch_ts, cpu) = 0; } |
cafcd80d2 lockup_detector: ... |
150 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
58687acba lockup_detector: ... |
151 152 |
void touch_nmi_watchdog(void) { |
3c00ea82c watchdog: Rename ... |
153 |
if (watchdog_user_enabled) { |
68d3f1d81 lockup_detector: ... |
154 155 156 157 158 159 160 |
unsigned cpu; for_each_present_cpu(cpu) { if (per_cpu(watchdog_nmi_touch, cpu) != true) per_cpu(watchdog_nmi_touch, cpu) = true; } } |
332fbdbca lockup_detector: ... |
161 |
touch_softlockup_watchdog(); |
58687acba lockup_detector: ... |
162 163 |
} EXPORT_SYMBOL(touch_nmi_watchdog); |
cafcd80d2 lockup_detector: ... |
164 |
#endif |
58687acba lockup_detector: ... |
165 166 167 168 169 |
void touch_softlockup_watchdog_sync(void) { __raw_get_cpu_var(softlockup_touch_sync) = true; __raw_get_cpu_var(watchdog_touch_ts) = 0; } |
23637d477 lockup_detector: ... |
170 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
58687acba lockup_detector: ... |
171 |
/* watchdog detector functions */ |
26e09c6ee lockup_detector: ... |
172 |
static int is_hardlockup(void) |
58687acba lockup_detector: ... |
173 |
{ |
909ea9646 core: Replace __g... |
174 |
unsigned long hrint = __this_cpu_read(hrtimer_interrupts); |
58687acba lockup_detector: ... |
175 |
|
909ea9646 core: Replace __g... |
176 |
if (__this_cpu_read(hrtimer_interrupts_saved) == hrint) |
58687acba lockup_detector: ... |
177 |
return 1; |
909ea9646 core: Replace __g... |
178 |
__this_cpu_write(hrtimer_interrupts_saved, hrint); |
58687acba lockup_detector: ... |
179 180 181 |
return 0; } #endif |
26e09c6ee lockup_detector: ... |
182 |
static int is_softlockup(unsigned long touch_ts) |
58687acba lockup_detector: ... |
183 |
{ |
c06b4f194 watchdog: Use loc... |
184 |
unsigned long now = get_timestamp(); |
58687acba lockup_detector: ... |
185 186 |
/* Warn about unreasonable delays: */ |
4eec42f39 watchdog: Change ... |
187 |
if (time_after(now, touch_ts + get_softlockup_thresh())) |
58687acba lockup_detector: ... |
188 189 190 191 |
return now - touch_ts; return 0; } |
23637d477 lockup_detector: ... |
192 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
1880c4ae1 perf, x86: Add hw... |
193 |
|
58687acba lockup_detector: ... |
194 195 196 197 198 199 200 201 202 |
static struct perf_event_attr wd_hw_attr = { .type = PERF_TYPE_HARDWARE, .config = PERF_COUNT_HW_CPU_CYCLES, .size = sizeof(struct perf_event_attr), .pinned = 1, .disabled = 1, }; /* Callback function for perf event subsystem */ |
a8b0ca17b perf: Remove the ... |
203 |
static void watchdog_overflow_callback(struct perf_event *event, |
58687acba lockup_detector: ... |
204 205 206 |
struct perf_sample_data *data, struct pt_regs *regs) { |
c6db67cda watchdog: Don't t... |
207 208 |
/* Ensure the watchdog never gets throttled */ event->hw.interrupts = 0; |
909ea9646 core: Replace __g... |
209 210 |
if (__this_cpu_read(watchdog_nmi_touch) == true) { __this_cpu_write(watchdog_nmi_touch, false); |
58687acba lockup_detector: ... |
211 212 213 214 215 216 217 218 219 |
return; } /* check for a hardlockup * This is done by making sure our timer interrupt * is incrementing. The timer interrupt should have * fired multiple times before we overflow'd. If it hasn't * then this is a good indication the cpu is stuck */ |
26e09c6ee lockup_detector: ... |
220 221 |
if (is_hardlockup()) { int this_cpu = smp_processor_id(); |
58687acba lockup_detector: ... |
222 |
/* only print hardlockups once */ |
909ea9646 core: Replace __g... |
223 |
if (__this_cpu_read(hard_watchdog_warn) == true) |
58687acba lockup_detector: ... |
224 225 226 227 228 229 |
return; if (hardlockup_panic) panic("Watchdog detected hard LOCKUP on cpu %d", this_cpu); else WARN(1, "Watchdog detected hard LOCKUP on cpu %d", this_cpu); |
909ea9646 core: Replace __g... |
230 |
__this_cpu_write(hard_watchdog_warn, true); |
58687acba lockup_detector: ... |
231 232 |
return; } |
909ea9646 core: Replace __g... |
233 |
__this_cpu_write(hard_watchdog_warn, false); |
58687acba lockup_detector: ... |
234 235 |
return; } |
bcd951cf1 watchdog: Use hot... |
236 |
#endif /* CONFIG_HARDLOCKUP_DETECTOR */ |
58687acba lockup_detector: ... |
237 238 |
static void watchdog_interrupt_count(void) { |
909ea9646 core: Replace __g... |
239 |
__this_cpu_inc(hrtimer_interrupts); |
58687acba lockup_detector: ... |
240 |
} |
bcd951cf1 watchdog: Use hot... |
241 242 243 |
static int watchdog_nmi_enable(unsigned int cpu); static void watchdog_nmi_disable(unsigned int cpu); |
58687acba lockup_detector: ... |
244 245 246 247 |
/* watchdog kicker functions */ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer) { |
909ea9646 core: Replace __g... |
248 |
unsigned long touch_ts = __this_cpu_read(watchdog_touch_ts); |
58687acba lockup_detector: ... |
249 250 251 252 253 254 255 |
struct pt_regs *regs = get_irq_regs(); int duration; /* kick the hardlockup detector */ watchdog_interrupt_count(); /* kick the softlockup detector */ |
909ea9646 core: Replace __g... |
256 |
wake_up_process(__this_cpu_read(softlockup_watchdog)); |
58687acba lockup_detector: ... |
257 258 |
/* .. and repeat */ |
0f34c4009 watchdog: store t... |
259 |
hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period)); |
58687acba lockup_detector: ... |
260 261 |
if (touch_ts == 0) { |
909ea9646 core: Replace __g... |
262 |
if (unlikely(__this_cpu_read(softlockup_touch_sync))) { |
58687acba lockup_detector: ... |
263 264 265 266 |
/* * If the time stamp was touched atomically * make sure the scheduler tick is up to date. */ |
909ea9646 core: Replace __g... |
267 |
__this_cpu_write(softlockup_touch_sync, false); |
58687acba lockup_detector: ... |
268 269 |
sched_clock_tick(); } |
5d1c0f4a8 watchdog: add che... |
270 271 272 |
/* Clear the guest paused flag on watchdog reset */ kvm_check_and_clear_guest_paused(); |
58687acba lockup_detector: ... |
273 274 275 276 277 278 279 280 281 282 |
__touch_watchdog(); return HRTIMER_RESTART; } /* check for a softlockup * This is done by making sure a high priority task is * being scheduled. The task touches the watchdog to * indicate it is getting cpu time. If it hasn't then * this is a good indication some task is hogging the cpu */ |
26e09c6ee lockup_detector: ... |
283 |
duration = is_softlockup(touch_ts); |
58687acba lockup_detector: ... |
284 |
if (unlikely(duration)) { |
5d1c0f4a8 watchdog: add che... |
285 286 287 288 289 290 291 |
/* * If a virtual machine is stopped by the host it can look to * the watchdog like a soft lockup, check to see if the host * stopped the vm before we issue the warning */ if (kvm_check_and_clear_guest_paused()) return HRTIMER_RESTART; |
58687acba lockup_detector: ... |
292 |
/* only warn once */ |
909ea9646 core: Replace __g... |
293 |
if (__this_cpu_read(soft_watchdog_warn) == true) |
58687acba lockup_detector: ... |
294 |
return HRTIMER_RESTART; |
b0f4c4b32 bugs, x86: Fix pr... |
295 296 |
printk(KERN_EMERG "BUG: soft lockup - CPU#%d stuck for %us! [%s:%d] ", |
26e09c6ee lockup_detector: ... |
297 |
smp_processor_id(), duration, |
58687acba lockup_detector: ... |
298 299 300 301 302 303 304 305 306 307 |
current->comm, task_pid_nr(current)); print_modules(); print_irqtrace_events(current); if (regs) show_regs(regs); else dump_stack(); if (softlockup_panic) panic("softlockup: hung tasks"); |
909ea9646 core: Replace __g... |
308 |
__this_cpu_write(soft_watchdog_warn, true); |
58687acba lockup_detector: ... |
309 |
} else |
909ea9646 core: Replace __g... |
310 |
__this_cpu_write(soft_watchdog_warn, false); |
58687acba lockup_detector: ... |
311 312 313 |
return HRTIMER_RESTART; } |
bcd951cf1 watchdog: Use hot... |
314 315 316 |
static void watchdog_set_prio(unsigned int policy, unsigned int prio) { struct sched_param param = { .sched_priority = prio }; |
58687acba lockup_detector: ... |
317 |
|
bcd951cf1 watchdog: Use hot... |
318 319 320 321 |
sched_setscheduler(current, policy, ¶m); } static void watchdog_enable(unsigned int cpu) |
58687acba lockup_detector: ... |
322 |
{ |
26e09c6ee lockup_detector: ... |
323 |
struct hrtimer *hrtimer = &__raw_get_cpu_var(watchdog_hrtimer); |
58687acba lockup_detector: ... |
324 |
|
3935e8950 watchdog: Fix dis... |
325 326 327 |
/* kick off the timer for the hardlockup detector */ hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); hrtimer->function = watchdog_timer_fn; |
bcd951cf1 watchdog: Use hot... |
328 329 |
/* Enable the perf event */ watchdog_nmi_enable(cpu); |
58687acba lockup_detector: ... |
330 |
|
58687acba lockup_detector: ... |
331 |
/* done here because hrtimer_start can only pin to smp_processor_id() */ |
0f34c4009 watchdog: store t... |
332 |
hrtimer_start(hrtimer, ns_to_ktime(sample_period), |
58687acba lockup_detector: ... |
333 |
HRTIMER_MODE_REL_PINNED); |
bcd951cf1 watchdog: Use hot... |
334 335 336 337 |
/* initialize timestamp */ watchdog_set_prio(SCHED_FIFO, MAX_RT_PRIO - 1); __touch_watchdog(); } |
58687acba lockup_detector: ... |
338 |
|
bcd951cf1 watchdog: Use hot... |
339 340 341 |
static void watchdog_disable(unsigned int cpu) { struct hrtimer *hrtimer = &__raw_get_cpu_var(watchdog_hrtimer); |
58687acba lockup_detector: ... |
342 |
|
bcd951cf1 watchdog: Use hot... |
343 344 345 346 |
watchdog_set_prio(SCHED_NORMAL, 0); hrtimer_cancel(hrtimer); /* disable the perf event */ watchdog_nmi_disable(cpu); |
58687acba lockup_detector: ... |
347 |
} |
b8900bc02 watchdog: Registe... |
348 349 350 351 |
static void watchdog_cleanup(unsigned int cpu, bool online) { watchdog_disable(cpu); } |
bcd951cf1 watchdog: Use hot... |
352 353 354 355 356 357 358 359 360 |
static int watchdog_should_run(unsigned int cpu) { return __this_cpu_read(hrtimer_interrupts) != __this_cpu_read(soft_lockup_hrtimer_cnt); } /* * The watchdog thread function - touches the timestamp. * |
0f34c4009 watchdog: store t... |
361 |
* It only runs once every sample_period seconds (4 seconds by |
bcd951cf1 watchdog: Use hot... |
362 363 364 365 366 367 368 369 370 371 |
* default) to reset the softlockup timestamp. If this gets delayed * for more than 2*watchdog_thresh seconds then the debug-printout * triggers in watchdog_timer_fn(). */ static void watchdog(unsigned int cpu) { __this_cpu_write(soft_lockup_hrtimer_cnt, __this_cpu_read(hrtimer_interrupts)); __touch_watchdog(); } |
58687acba lockup_detector: ... |
372 |
|
23637d477 lockup_detector: ... |
373 |
#ifdef CONFIG_HARDLOCKUP_DETECTOR |
a70270468 watchdog: Quiet d... |
374 375 376 377 378 379 |
/* * People like the simple clean cpu node info on boot. * Reduce the watchdog noise by only printing messages * that are different from what cpu0 displayed. */ static unsigned long cpu0_err; |
bcd951cf1 watchdog: Use hot... |
380 |
static int watchdog_nmi_enable(unsigned int cpu) |
58687acba lockup_detector: ... |
381 382 383 384 385 386 387 388 389 390 391 |
{ struct perf_event_attr *wd_attr; struct perf_event *event = per_cpu(watchdog_ev, cpu); /* is it already setup and enabled? */ if (event && event->state > PERF_EVENT_STATE_OFF) goto out; /* it is setup but not enabled */ if (event != NULL) goto out_enable; |
58687acba lockup_detector: ... |
392 |
wd_attr = &wd_hw_attr; |
4eec42f39 watchdog: Change ... |
393 |
wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh); |
1880c4ae1 perf, x86: Add hw... |
394 395 |
/* Try to register using hardware perf events */ |
4dc0da869 perf: Add context... |
396 |
event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL); |
a70270468 watchdog: Quiet d... |
397 398 399 400 |
/* save cpu0 error for future comparision */ if (cpu == 0 && IS_ERR(event)) cpu0_err = PTR_ERR(event); |
58687acba lockup_detector: ... |
401 |
if (!IS_ERR(event)) { |
a70270468 watchdog: Quiet d... |
402 403 404 405 |
/* only print for cpu0 or different than cpu0 */ if (cpu == 0 || cpu0_err) pr_info("enabled on all CPUs, permanently consumes one hw-PMU counter. "); |
58687acba lockup_detector: ... |
406 407 |
goto out_save; } |
a70270468 watchdog: Quiet d... |
408 409 410 |
/* skip displaying the same error again */ if (cpu > 0 && (PTR_ERR(event) == cpu0_err)) return PTR_ERR(event); |
5651f7f47 watchdog, nmi: Lo... |
411 412 413 |
/* vary the KERN level based on the returned errno */ if (PTR_ERR(event) == -EOPNOTSUPP) |
4501980aa kernel/watchdog.c... |
414 415 |
pr_info("disabled (cpu%i): not supported (no LAPIC?) ", cpu); |
5651f7f47 watchdog, nmi: Lo... |
416 |
else if (PTR_ERR(event) == -ENOENT) |
4501980aa kernel/watchdog.c... |
417 418 419 |
pr_warning("disabled (cpu%i): hardware events not enabled ", cpu); |
5651f7f47 watchdog, nmi: Lo... |
420 |
else |
4501980aa kernel/watchdog.c... |
421 422 423 |
pr_err("disabled (cpu%i): unable to create perf event: %ld ", cpu, PTR_ERR(event)); |
eac243355 lockup_detector: ... |
424 |
return PTR_ERR(event); |
58687acba lockup_detector: ... |
425 426 427 428 429 430 431 432 433 |
/* success path */ out_save: per_cpu(watchdog_ev, cpu) = event; out_enable: perf_event_enable(per_cpu(watchdog_ev, cpu)); out: return 0; } |
bcd951cf1 watchdog: Use hot... |
434 |
static void watchdog_nmi_disable(unsigned int cpu) |
58687acba lockup_detector: ... |
435 436 437 438 439 440 441 442 443 444 445 446 447 |
{ struct perf_event *event = per_cpu(watchdog_ev, cpu); if (event) { perf_event_disable(event); per_cpu(watchdog_ev, cpu) = NULL; /* should be in cleanup, but blocks oprofile */ perf_event_release_kernel(event); } return; } #else |
bcd951cf1 watchdog: Use hot... |
448 449 |
static int watchdog_nmi_enable(unsigned int cpu) { return 0; } static void watchdog_nmi_disable(unsigned int cpu) { return; } |
23637d477 lockup_detector: ... |
450 |
#endif /* CONFIG_HARDLOCKUP_DETECTOR */ |
58687acba lockup_detector: ... |
451 |
|
b8900bc02 watchdog: Registe... |
452 453 454 455 456 457 458 459 460 461 |
static struct smp_hotplug_thread watchdog_threads = { .store = &softlockup_watchdog, .thread_should_run = watchdog_should_run, .thread_fn = watchdog, .thread_comm = "watchdog/%u", .setup = watchdog_enable, .cleanup = watchdog_cleanup, .park = watchdog_disable, .unpark = watchdog_enable, }; |
9809b18fc watchdog: update ... |
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 |
static void restart_watchdog_hrtimer(void *info) { struct hrtimer *hrtimer = &__raw_get_cpu_var(watchdog_hrtimer); int ret; /* * No need to cancel and restart hrtimer if it is currently executing * because it will reprogram itself with the new period now. * We should never see it unqueued here because we are running per-cpu * with interrupts disabled. */ ret = hrtimer_try_to_cancel(hrtimer); if (ret == 1) hrtimer_start(hrtimer, ns_to_ktime(sample_period), HRTIMER_MODE_REL_PINNED); } static void update_timers(int cpu) { struct call_single_data data = {.func = restart_watchdog_hrtimer}; /* * Make sure that perf event counter will adopt to a new * sampling period. Updating the sampling period directly would * be much nicer but we do not have an API for that now so * let's use a big hammer. * Hrtimer will adopt the new period on the next tick but this * might be late already so we have to restart the timer as well. */ watchdog_nmi_disable(cpu); __smp_call_function_single(cpu, &data, 1); watchdog_nmi_enable(cpu); } static void update_timers_all_cpus(void) { int cpu; get_online_cpus(); preempt_disable(); for_each_online_cpu(cpu) update_timers(cpu); preempt_enable(); put_online_cpus(); } static int watchdog_enable_all_cpus(bool sample_period_changed) |
58687acba lockup_detector: ... |
508 |
{ |
b8900bc02 watchdog: Registe... |
509 |
int err = 0; |
58687acba lockup_detector: ... |
510 |
|
3c00ea82c watchdog: Rename ... |
511 |
if (!watchdog_running) { |
b8900bc02 watchdog: Registe... |
512 513 514 515 516 |
err = smpboot_register_percpu_thread(&watchdog_threads); if (err) pr_err("Failed to create watchdog threads, disabled "); else |
3c00ea82c watchdog: Rename ... |
517 |
watchdog_running = 1; |
9809b18fc watchdog: update ... |
518 519 |
} else if (sample_period_changed) { update_timers_all_cpus(); |
bcd951cf1 watchdog: Use hot... |
520 |
} |
b8900bc02 watchdog: Registe... |
521 522 |
return err; |
58687acba lockup_detector: ... |
523 |
} |
b8900bc02 watchdog: Registe... |
524 525 526 |
/* prepare/enable/disable routines */ /* sysctl functions */ #ifdef CONFIG_SYSCTL |
58687acba lockup_detector: ... |
527 528 |
static void watchdog_disable_all_cpus(void) { |
3c00ea82c watchdog: Rename ... |
529 530 |
if (watchdog_running) { watchdog_running = 0; |
b8900bc02 watchdog: Registe... |
531 |
smpboot_unregister_percpu_thread(&watchdog_threads); |
bcd951cf1 watchdog: Use hot... |
532 |
} |
58687acba lockup_detector: ... |
533 |
} |
58687acba lockup_detector: ... |
534 |
/* |
586692a5a watchdog: Disable... |
535 |
* proc handler for /proc/sys/kernel/nmi_watchdog,watchdog_thresh |
58687acba lockup_detector: ... |
536 |
*/ |
586692a5a watchdog: Disable... |
537 538 |
int proc_dowatchdog(struct ctl_table *table, int write, void __user *buffer, size_t *lenp, loff_t *ppos) |
58687acba lockup_detector: ... |
539 |
{ |
b8900bc02 watchdog: Registe... |
540 |
int err, old_thresh, old_enabled; |
359e6fab6 watchdog: update ... |
541 |
static DEFINE_MUTEX(watchdog_proc_mutex); |
58687acba lockup_detector: ... |
542 |
|
359e6fab6 watchdog: update ... |
543 |
mutex_lock(&watchdog_proc_mutex); |
b8900bc02 watchdog: Registe... |
544 |
old_thresh = ACCESS_ONCE(watchdog_thresh); |
3c00ea82c watchdog: Rename ... |
545 |
old_enabled = ACCESS_ONCE(watchdog_user_enabled); |
bcd951cf1 watchdog: Use hot... |
546 |
|
b8900bc02 watchdog: Registe... |
547 548 |
err = proc_dointvec_minmax(table, write, buffer, lenp, ppos); if (err || !write) |
359e6fab6 watchdog: update ... |
549 |
goto out; |
e04ab2bc4 watchdog: Only di... |
550 |
|
0f34c4009 watchdog: store t... |
551 |
set_sample_period(); |
b66a2356d watchdog: Add com... |
552 553 |
/* * Watchdog threads shouldn't be enabled if they are |
3c00ea82c watchdog: Rename ... |
554 |
* disabled. The 'watchdog_running' variable check in |
b66a2356d watchdog: Add com... |
555 556 |
* watchdog_*_all_cpus() function takes care of this. */ |
3c00ea82c watchdog: Rename ... |
557 |
if (watchdog_user_enabled && watchdog_thresh) |
9809b18fc watchdog: update ... |
558 |
err = watchdog_enable_all_cpus(old_thresh != watchdog_thresh); |
e04ab2bc4 watchdog: Only di... |
559 560 |
else watchdog_disable_all_cpus(); |
b8900bc02 watchdog: Registe... |
561 562 563 |
/* Restore old values on failure */ if (err) { watchdog_thresh = old_thresh; |
3c00ea82c watchdog: Rename ... |
564 |
watchdog_user_enabled = old_enabled; |
b8900bc02 watchdog: Registe... |
565 |
} |
359e6fab6 watchdog: update ... |
566 567 |
out: mutex_unlock(&watchdog_proc_mutex); |
b8900bc02 watchdog: Registe... |
568 |
return err; |
58687acba lockup_detector: ... |
569 |
} |
58687acba lockup_detector: ... |
570 |
#endif /* CONFIG_SYSCTL */ |
004417a6d perf, arch: Clean... |
571 |
void __init lockup_detector_init(void) |
58687acba lockup_detector: ... |
572 |
{ |
0f34c4009 watchdog: store t... |
573 |
set_sample_period(); |
b8900bc02 watchdog: Registe... |
574 |
|
3c00ea82c watchdog: Rename ... |
575 |
if (watchdog_user_enabled) |
9809b18fc watchdog: update ... |
576 |
watchdog_enable_all_cpus(false); |
58687acba lockup_detector: ... |
577 |
} |