]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/commitdiff
sched/core: Reset task's lockless wake-queues on fork()
authorSebastian Andrzej Siewior <bigeasy@linutronix.de>
Mon, 21 Dec 2015 17:17:10 +0000 (18:17 +0100)
committerIngo Molnar <mingo@kernel.org>
Wed, 6 Jan 2016 10:01:07 +0000 (11:01 +0100)
In the following commit:

  7675104990ed ("sched: Implement lockless wake-queues")

we gained lockless wake-queues.

The -RT kernel managed to lockup itself with those. There could be multiple
attempts for task X to enqueue it for a wakeup _even_ if task X is already
running.

The reason is that task X could be runnable but not yet on CPU. The the
task performing the wakeup did not leave the CPU it could performe
multiple wakeups.

With the proper timming task X could be running and enqueued for a
wakeup. If this happens while X is performing a fork() then its its
child will have a !NULL `wake_q` member copied.

This is not a problem as long as the child task does not participate in
lockless wakeups :)

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Davidlohr Bueso <dbueso@suse.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 7675104990ed ("sched: Implement lockless wake-queues")
Link: http://lkml.kernel.org/r/20151221171710.GA5499@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/fork.c

index fce002ee3ddffbab7613d0f9ec390683901f3c67..1155eac616878b5494ae94f2f1d0294812ddda72 100644 (file)
@@ -380,6 +380,7 @@ static struct task_struct *dup_task_struct(struct task_struct *orig)
 #endif
        tsk->splice_pipe = NULL;
        tsk->task_frag.page = NULL;
+       tsk->wake_q.next = NULL;
 
        account_kernel_stack(ti, 1);