diff options
author | Rik van Riel <riel@redhat.com> | 2014-09-12 09:12:15 -0400 |
---|---|---|
committer | Ingo Molnar <mingo@kernel.org> | 2014-09-19 06:35:17 -0400 |
commit | 9c368b5b6eccce1cbd7f68142106b3b4ddb1c5b5 (patch) | |
tree | 2d74483c5bccdf6f513a763971fe419191afffd8 | |
parent | ef8ac06359ddf95431cf6bb04ad2b36fff562328 (diff) |
sched, time: Fix lock inversion in thread_group_cputime()
The sig->stats_lock nests inside the tasklist_lock and the
sighand->siglock in __exit_signal and wait_task_zombie.
However, both of those locks can be taken from irq context,
which means we need to use the interrupt safe variant of
read_seqbegin_or_lock. This blocks interrupts when the "lock"
branch is taken (seq is odd), preventing the lock inversion.
On the first (lockless) pass through the loop, irqs are not
blocked.
Reported-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: prarit@redhat.com
Cc: oleg@redhat.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410527535-9814-3-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-rw-r--r-- | kernel/sched/cputime.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 2b57031afc19..64492dff8a81 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c | |||
@@ -289,13 +289,14 @@ void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times) | |||
289 | cputime_t utime, stime; | 289 | cputime_t utime, stime; |
290 | struct task_struct *t; | 290 | struct task_struct *t; |
291 | unsigned int seq, nextseq; | 291 | unsigned int seq, nextseq; |
292 | unsigned long flags; | ||
292 | 293 | ||
293 | rcu_read_lock(); | 294 | rcu_read_lock(); |
294 | /* Attempt a lockless read on the first round. */ | 295 | /* Attempt a lockless read on the first round. */ |
295 | nextseq = 0; | 296 | nextseq = 0; |
296 | do { | 297 | do { |
297 | seq = nextseq; | 298 | seq = nextseq; |
298 | read_seqbegin_or_lock(&sig->stats_lock, &seq); | 299 | flags = read_seqbegin_or_lock_irqsave(&sig->stats_lock, &seq); |
299 | times->utime = sig->utime; | 300 | times->utime = sig->utime; |
300 | times->stime = sig->stime; | 301 | times->stime = sig->stime; |
301 | times->sum_exec_runtime = sig->sum_sched_runtime; | 302 | times->sum_exec_runtime = sig->sum_sched_runtime; |
@@ -309,7 +310,7 @@ void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times) | |||
309 | /* If lockless access failed, take the lock. */ | 310 | /* If lockless access failed, take the lock. */ |
310 | nextseq = 1; | 311 | nextseq = 1; |
311 | } while (need_seqretry(&sig->stats_lock, seq)); | 312 | } while (need_seqretry(&sig->stats_lock, seq)); |
312 | done_seqretry(&sig->stats_lock, seq); | 313 | done_seqretry_irqrestore(&sig->stats_lock, seq, flags); |
313 | rcu_read_unlock(); | 314 | rcu_read_unlock(); |
314 | } | 315 | } |
315 | 316 | ||