sched/deadline: Correctly handle active 0-lag timers
authorluca abeni <luca.abeni@santannapisa.it>
Mon, 25 Mar 2019 13:15:30 +0000 (14:15 +0100)
committerIngo Molnar <mingo@kernel.org>
Tue, 16 Apr 2019 14:54:58 +0000 (16:54 +0200)
commit1b02cd6a2d7f3e2a6a5262887d2cb2912083e42f
tree1dfd4308ddb2928cd277a314e6c281766a1cd6c5
parent2e8e19226398db8265a8e675fcc0118b9e80c9e8
sched/deadline: Correctly handle active 0-lag timers

syzbot reported the following warning:

   [ ] WARNING: CPU: 4 PID: 17089 at kernel/sched/deadline.c:255 task_non_contending+0xae0/0x1950

line 255 of deadline.c is:

WARN_ON(hrtimer_active(&dl_se->inactive_timer));

in task_non_contending().

Unfortunately, in some cases (for example, a deadline task
continuosly blocking and waking immediately) it can happen that
a task blocks (and task_non_contending() is called) while the
0-lag timer is still active.

In this case, the safest thing to do is to immediately decrease
the running bandwidth of the task, without trying to re-arm the 0-lag timer.

Signed-off-by: luca abeni <luca.abeni@santannapisa.it>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: chengjian (D) <cj.chengjian@huawei.com>
Link: https://lkml.kernel.org/r/20190325131530.34706-1-luca.abeni@santannapisa.it
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/sched/deadline.c