profiling: remove profile=sleep support

The kernel sleep profile is no longer working due to a recursive locking
bug introduced by commit 42a20f8 ("sched: Add wrapper for get_wchan()
to keep task blocked")

Booting with the 'profile=sleep' kernel command line option added or
executing

  # echo -n sleep > /sys/kernel/profiling

after boot causes the system to lock up.

Lockdep reports

  kthreadd/3 is trying to acquire lock:
  ffff93ac82e08d58 (&p->pi_lock){....}-{2:2}, at: get_wchan+0x32/0x70

  but task is already holding lock:
  ffff93ac82e08d58 (&p->pi_lock){....}-{2:2}, at: try_to_wake_up+0x53/0x370

with the call trace being

   lock_acquire+0xc8/0x2f0
   get_wchan+0x32/0x70
   __update_stats_enqueue_sleeper+0x151/0x430
   enqueue_entity+0x4b0/0x520
   enqueue_task_fair+0x92/0x6b0
   ttwu_do_activate+0x73/0x140
   try_to_wake_up+0x213/0x370
   swake_up_locked+0x20/0x50
   complete+0x2f/0x40
   kthread+0xfb/0x180

However, since nobody noticed this regression for more than two years,
let's remove 'profile=sleep' support based on the assumption that nobody
needs this functionality.

Fixes: 42a20f8 ("sched: Add wrapper for get_wchan() to keep task blocked")
Cc: [email protected] # v5.16+
Signed-off-by: Tetsuo Handa <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

Documentation/admin-guide/kernel-parameters.txt

@@ -4798,11 +4798,9 @@
 
 	profile=	[KNL] Enable kernel profiling via /proc/profile
 			Format: [<profiletype>,]<number>
-			Param: <profiletype>: "schedule", "sleep", or "kvm"
+			Param: <profiletype>: "schedule" or "kvm"
 				[defaults to kernel profiling]
 			Param: "schedule" - profile schedule points.
-			Param: "sleep" - profile D-state sleeping (millisecs).
-				Requires CONFIG_SCHEDSTATS
 			Param: "kvm" - profile VM exits.
 			Param: <number> - step/bucket size as a power of 2 for
 				statistical time based profiling.

include/linux/profile.h

@@ -10,7 +10,6 @@
 
 #define CPU_PROFILING	1
 #define SCHED_PROFILING	2
-#define SLEEP_PROFILING	3
 #define KVM_PROFILING	4
 
 struct proc_dir_entry;

kernel/profile.c

@@ -50,20 +50,11 @@ EXPORT_SYMBOL_GPL(prof_on);
 int profile_setup(char *str)
 {
 	static const char schedstr[] = "schedule";
-	static const char sleepstr[] = "sleep";
 	static const char kvmstr[] = "kvm";
 	const char *select = NULL;
 	int par;
 
-	if (!strncmp(str, sleepstr, strlen(sleepstr))) {
-#ifdef CONFIG_SCHEDSTATS
-		force_schedstat_enabled();
-		prof_on = SLEEP_PROFILING;
-		select = sleepstr;
-#else
-		pr_warn("kernel sleep profiling requires CONFIG_SCHEDSTATS\n");
-#endif /* CONFIG_SCHEDSTATS */
-	} else if (!strncmp(str, schedstr, strlen(schedstr))) {
+	if (!strncmp(str, schedstr, strlen(schedstr))) {
 		prof_on = SCHED_PROFILING;
 		select = schedstr;
 	} else if (!strncmp(str, kvmstr, strlen(kvmstr))) {

kernel/sched/stats.c

@@ -92,16 +92,6 @@ void __update_stats_enqueue_sleeper(struct rq *rq, struct task_struct *p,
 
 			trace_sched_stat_blocked(p, delta);
 
-			/*
-			 * Blocking time is in units of nanosecs, so shift by
-			 * 20 to get a milliseconds-range estimation of the
-			 * amount of time that the task spent sleeping:
-			 */
-			if (unlikely(prof_on == SLEEP_PROFILING)) {
-				profile_hits(SLEEP_PROFILING,
-					     (void *)get_wchan(p),
-					     delta >> 20);
-			}
 			account_scheduler_latency(p, delta >> 10, 0);
 		}
 	}