Merge branch 'for-4.5' of git://git.kernel.org/pub/scm/linux/kernel/g…

…it/tj/wq

Pull workqueue update from Tejun Heo:
 "Workqueue changes for v4.5.  One cleanup patch and three to improve
  the debuggability.

  Workqueue now has a stall detector which dumps workqueue state if any
  worker pool hasn't made forward progress over a certain amount of time
  (30s by default) and also triggers a warning if a workqueue which can
  be used in memory reclaim path tries to wait on something which can't
  be.

  These should make workqueue hangs a lot easier to debug."

* 'for-4.5' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: simplify the apply_workqueue_attrs_locked()
  workqueue: implement lockup detector
  watchdog: introduce touch_softlockup_watchdog_sched()
  workqueue: warn if memory reclaim tries to flush !WQ_MEM_RECLAIM workqueue

Documentation/kernel-parameters.txt

@@ -4140,6 +4140,15 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			or other driver-specific files in the
 			Documentation/watchdog/ directory.
 
+	workqueue.watchdog_thresh=
+			If CONFIG_WQ_WATCHDOG is configured, workqueue can
+			warn stall conditions and dump internal state to
+			help debugging.  0 disables workqueue stall
+			detection; otherwise, it's the stall threshold
+			duration in seconds.  The default value is 30 and
+			it can be updated at runtime by writing to the
+			corresponding sysfs file.
+
 	workqueue.disable_numa
 			By default, all work items queued to unbound
 			workqueues are affine to the NUMA nodes they're

include/linux/sched.h

@@ -377,6 +377,7 @@ extern void scheduler_tick(void);
 extern void sched_show_task(struct task_struct *p);
 
 #ifdef CONFIG_LOCKUP_DETECTOR
+extern void touch_softlockup_watchdog_sched(void);
 extern void touch_softlockup_watchdog(void);
 extern void touch_softlockup_watchdog_sync(void);
 extern void touch_all_softlockup_watchdogs(void);
@@ -387,6 +388,9 @@ extern unsigned int  softlockup_panic;
 extern unsigned int  hardlockup_panic;
 void lockup_detector_init(void);
 #else
+static inline void touch_softlockup_watchdog_sched(void)
+{
+}
 static inline void touch_softlockup_watchdog(void)
 {
 }

include/linux/workqueue.h

@@ -618,4 +618,10 @@ static inline int workqueue_sysfs_register(struct workqueue_struct *wq)
 { return 0; }
 #endif	/* CONFIG_SYSFS */
 
+#ifdef CONFIG_WQ_WATCHDOG
+void wq_watchdog_touch(int cpu);
+#else	/* CONFIG_WQ_WATCHDOG */
+static inline void wq_watchdog_touch(int cpu) { }
+#endif	/* CONFIG_WQ_WATCHDOG */
+
 #endif

kernel/sched/clock.c

@@ -354,7 +354,7 @@ void sched_clock_idle_wakeup_event(u64 delta_ns)
 		return;
 
 	sched_clock_tick();
-	touch_softlockup_watchdog();
+	touch_softlockup_watchdog_sched();
 }
 EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event);
 

kernel/time/tick-sched.c

@@ -143,7 +143,7 @@ static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs)
 	 * when we go busy again does not account too much ticks.
 	 */
 	if (ts->tick_stopped) {
-		touch_softlockup_watchdog();
+		touch_softlockup_watchdog_sched();
 		if (is_idle_task(current))
 			ts->idle_jiffies++;
 	}
@@ -430,7 +430,7 @@ static void tick_nohz_update_jiffies(ktime_t now)
 	tick_do_update_jiffies64(now);
 	local_irq_restore(flags);
 
-	touch_softlockup_watchdog();
+	touch_softlockup_watchdog_sched();
 }
 
 /*
@@ -717,7 +717,7 @@ static void tick_nohz_restart_sched_tick(struct tick_sched *ts, ktime_t now, int
 	update_cpu_load_nohz(active);
 
 	calc_load_exit_idle();
-	touch_softlockup_watchdog();
+	touch_softlockup_watchdog_sched();
 	/*
 	 * Cancel the scheduled timer and restore the tick
 	 */

kernel/watchdog.c

@@ -20,6 +20,7 @@
 #include <linux/smpboot.h>
 #include <linux/sched/rt.h>
 #include <linux/tick.h>
+#include <linux/workqueue.h>
 
 #include <asm/irq_regs.h>
 #include <linux/kvm_para.h>
@@ -225,14 +226,28 @@ static void __touch_watchdog(void)
 	__this_cpu_write(watchdog_touch_ts, get_timestamp());
 }
 
-void touch_softlockup_watchdog(void)
+/**
+ * touch_softlockup_watchdog_sched - touch watchdog on scheduler stalls
+ *
+ * Call when the scheduler may have stalled for legitimate reasons
+ * preventing the watchdog task from executing - e.g. the scheduler
+ * entering idle state.  This should only be used for scheduler events.
+ * Use touch_softlockup_watchdog() for everything else.
+ */
+void touch_softlockup_watchdog_sched(void)
 {
 	/*
 	 * Preemption can be enabled.  It doesn't matter which CPU's timestamp
 	 * gets zeroed here, so use the raw_ operation.
 	 */
 	raw_cpu_write(watchdog_touch_ts, 0);
 }
+
+void touch_softlockup_watchdog(void)
+{
+	touch_softlockup_watchdog_sched();
+	wq_watchdog_touch(raw_smp_processor_id());
+}
 EXPORT_SYMBOL(touch_softlockup_watchdog);
 
 void touch_all_softlockup_watchdogs(void)
@@ -246,6 +261,7 @@ void touch_all_softlockup_watchdogs(void)
 	 */
 	for_each_watchdog_cpu(cpu)
 		per_cpu(watchdog_touch_ts, cpu) = 0;
+	wq_watchdog_touch(-1);
 }
 
 #ifdef CONFIG_HARDLOCKUP_DETECTOR