All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: mingo@kernel.org, oleg@redhat.com, torvalds@linux-foundation.org
Cc: tglx@linutronix.de, ilya.dryomov@inktank.com,
	umgwanakikbuti@gmail.com, linux-kernel@vger.kernel.org,
	Peter Zijlstra <peterz@infradead.org>
Subject: [PATCH 10/11] sched: Debug nested sleeps
Date: Wed, 24 Sep 2014 10:18:55 +0200	[thread overview]
Message-ID: <20140924082242.591637616@infradead.org> (raw)
In-Reply-To: 20140924081845.572814794@infradead.org

[-- Attachment #1: peterz-might_sleep_nesting.patch --]
[-- Type: text/plain, Size: 3739 bytes --]

Validate we call might_sleep() with TASK_RUNNING, which catches places
where we nest blocking primitives, eg. mutex usage in a wait loop.

Since all blocking is arranged through task_struct::state, nesting
this will cause the inner primitive to set TASK_RUNNING and the outer
will thus not block.

Another observed problem is calling a blocking function from
schedule()->sched_submit_work()->blk_schedule_flush_plug() which will
then destroy the task state for the actual __schedule() call that
comes after it.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
 include/linux/sched.h |   46 ++++++++++++++++++++++++++++++++++++++++++++--
 kernel/sched/core.c   |   13 +++++++++++++
 2 files changed, 57 insertions(+), 2 deletions(-)

--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -242,6 +242,43 @@ extern char ___assert_task_state[1 - 2*!
 				((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
 				 (task->flags & PF_FROZEN) == 0)
 
+#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
+
+#define __set_task_state(tsk, state_value)			\
+	do {							\
+		(tsk)->task_state_change = _THIS_IP_;		\
+		(tsk)->state = (state_value);			\
+	} while (0)
+#define set_task_state(tsk, state_value)			\
+	do {							\
+		(tsk)->task_state_change = _THIS_IP_;		\
+		set_mb((tsk)->state, (state_value));		\
+	} while (0)
+
+/*
+ * set_current_state() includes a barrier so that the write of current->state
+ * is correctly serialised wrt the caller's subsequent test of whether to
+ * actually sleep:
+ *
+ *	set_current_state(TASK_UNINTERRUPTIBLE);
+ *	if (do_i_need_to_sleep())
+ *		schedule();
+ *
+ * If the caller does not need such serialisation then use __set_current_state()
+ */
+#define __set_current_state(state_value)			\
+	do {							\
+		current->task_state_change = _THIS_IP_;		\
+		current->state = (state_value);			\
+	} while (0)
+#define set_current_state(state_value)				\
+	do {							\
+		current->task_state_change = _THIS_IP_;		\
+		set_mb(current->state, (state_value));		\
+	} while (0)
+
+#else
+
 #define __set_task_state(tsk, state_value)		\
 	do { (tsk)->state = (state_value); } while (0)
 #define set_task_state(tsk, state_value)		\
@@ -258,11 +295,13 @@ extern char ___assert_task_state[1 - 2*!
  *
  * If the caller does not need such serialisation then use __set_current_state()
  */
-#define __set_current_state(state_value)			\
+#define __set_current_state(state_value)		\
 	do { current->state = (state_value); } while (0)
-#define set_current_state(state_value)		\
+#define set_current_state(state_value)			\
 	set_mb(current->state, (state_value))
 
+#endif
+
 /* Task command name length */
 #define TASK_COMM_LEN 16
 
@@ -1660,6 +1699,9 @@ struct task_struct {
 	unsigned int	sequential_io;
 	unsigned int	sequential_io_avg;
 #endif
+#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
+	unsigned long	task_state_change;
+#endif
 };
 
 /* Future-safe accessor for struct task_struct's cpus_allowed. */
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7143,6 +7143,19 @@ void __might_sleep(const char *file, int
 {
 	static unsigned long prev_jiffy;	/* ratelimiting */
 
+	/*
+	 * Blocking primitives will set (and therefore destroy) current->state,
+	 * since we will exit with TASK_RUNNING make sure we enter with it,
+	 * otherwise we will destroy state.
+	 */
+	if (WARN(current->state != TASK_RUNNING,
+			"do not call blocking ops when !TASK_RUNNING; "
+			"state=%lx set at [<%p>] %pS\n",
+			current->state,
+			(void *)current->task_state_change,
+			(void *)current->task_state_change))
+		__set_current_state(TASK_RUNNING);
+
 	rcu_sleep_check(); /* WARN_ON_ONCE() by default, no rate limit reqd. */
 	if ((preempt_count_equals(preempt_offset) && !irqs_disabled() &&
 	     !is_idle_task(current)) ||



  parent reply	other threads:[~2014-09-24  8:38 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-09-24  8:18 [PATCH 00/11] nested sleeps, fixes and debug infrastructure Peter Zijlstra
2014-09-24  8:18 ` [PATCH 01/11] locking/mutex: Dont assume TASK_RUNNING Peter Zijlstra
2014-10-28 11:09   ` [tip:sched/core] locking/mutex: Don't " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 02/11] wait: Provide infrastructure to deal with nested blocking Peter Zijlstra
2014-09-29 21:02   ` Oleg Nesterov
2014-10-02  7:37     ` Peter Zijlstra
2014-10-02 21:21       ` Oleg Nesterov
2014-10-28 11:09   ` [tip:sched/core] sched/wait: " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 03/11] wait: Add might_sleep() Peter Zijlstra
2014-10-28 11:09   ` [tip:sched/core] sched/wait: Add might_sleep() checks tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 04/11] exit: Deal with nested sleeps Peter Zijlstra
2014-10-28 11:10   ` [tip:sched/core] sched, " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 05/11] inotify: " Peter Zijlstra
2014-10-28 11:10   ` [tip:sched/core] sched, " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 06/11] tty: " Peter Zijlstra
2014-10-28 11:10   ` [tip:sched/core] sched, " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 07/11] smp: Correctly deal " Peter Zijlstra
2014-10-28 11:11   ` [tip:sched/core] sched, " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 08/11] module: Fix nested sleep Peter Zijlstra
2014-09-29 22:18   ` Oleg Nesterov
2014-09-30 13:43     ` Peter Zijlstra
2014-10-28 11:11   ` [tip:sched/core] sched, modules: Fix nested sleep in add_unformed_module() tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 09/11] net: Clean up sk_wait_event() vs might_sleep() Peter Zijlstra
2014-09-24  8:36   ` Peter Zijlstra
2014-10-28 11:11   ` [tip:sched/core] sched, net: Clean up sk_wait_event() vs. might_sleep() tip-bot for Peter Zijlstra
2014-09-24  8:18 ` Peter Zijlstra [this message]
2014-09-29 22:13   ` [PATCH 10/11] sched: Debug nested sleeps Oleg Nesterov
2014-09-30 13:49     ` Peter Zijlstra
2014-09-30 21:47       ` Oleg Nesterov
2014-10-01 16:10         ` Peter Zijlstra
2014-10-01 18:35           ` Oleg Nesterov
2014-10-02  9:07             ` Peter Zijlstra
2014-10-02 21:34               ` Oleg Nesterov
2014-10-28 11:11   ` [tip:sched/core] " tip-bot for Peter Zijlstra
2014-09-24  8:18 ` [PATCH 11/11] sched: Exclude cond_resched() from nested sleep test Peter Zijlstra
2014-10-28 11:12   ` [tip:sched/core] " tip-bot for Peter Zijlstra
2014-09-25  8:30 ` [PATCH 00/11] nested sleeps, fixes and debug infrastructure Mike Galbraith
2014-09-25  9:06   ` Peter Zijlstra
2014-09-25  9:10     ` Mike Galbraith
2014-09-25  9:15     ` Peter Zijlstra
2014-09-25  9:56       ` Mike Galbraith
2014-09-25 13:59         ` BUG: sleeping function called from invalid context at drivers/cpufreq/cpufreq.c:370 Mike Galbraith
2014-09-26  6:24           ` Mike Galbraith
2014-09-26  7:54             ` Mike Galbraith
2014-09-26 14:10               ` Rafael J. Wysocki
2014-09-26 22:44               ` Rafael J. Wysocki
2014-09-27  6:14                 ` Mike Galbraith
2014-09-27 19:57                   ` Rafael J. Wysocki
2014-10-02 10:22       ` [PATCH 00/11] nested sleeps, fixes and debug infrastructure Peter Zijlstra
2014-10-02 12:15         ` Peter Zijlstra
2014-10-27 13:41           ` Peter Zijlstra
2014-10-28  0:07             ` Oleg Nesterov
2014-10-28  8:23               ` Peter Zijlstra
2014-10-29  0:00                 ` Oleg Nesterov
2014-10-29  9:35                   ` Peter Zijlstra
2014-10-29 11:31                     ` Peter Zijlstra
2014-10-29 11:36                       ` Peter Zijlstra
2014-10-29 14:26                   ` Peter Zijlstra
2014-11-04 16:08         ` [tip:sched/core] audit, sched/wait: Fixup kauditd_thread() wait loop tip-bot for Peter Zijlstra

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20140924082242.591637616@infradead.org \
    --to=peterz@infradead.org \
    --cc=ilya.dryomov@inktank.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=oleg@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.org \
    --cc=umgwanakikbuti@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.