From: Jens Axboe <axboe@xxxxxxxxx>
commit 87df7fb922d18e96992aa5e824aa34b2065fef59 upstream.
When new work is added, io_wqe_enqueue() checks if we need to wake or
create a new worker. But that check is done outside the lock that
otherwise synchronizes us with a worker going to sleep, so we can end
up in the following situation:
CPU0 CPU1
lock
insert work
unlock
atomic_read(nr_running) != 0
lock
atomic_dec(nr_running)
no wakeup needed
Hold the wqe lock around the "need to wakeup" check. Then we can also get
rid of the temporary work_flags variable, as we know the work will remain
valid as long as we hold the lock.
Cc: stable@xxxxxxxxxxxxxxx
Reported-by: Andres Freund <andres@xxxxxxxxxxx>
Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
fs/io-wq.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -798,7 +798,7 @@ append:
static void io_wqe_enqueue(struct io_wqe *wqe, struct io_wq_work *work)
{
struct io_wqe_acct *acct = io_work_get_acct(wqe, work);
- int work_flags;
+ bool do_wake;
unsigned long flags;
/*
@@ -811,14 +811,14 @@ static void io_wqe_enqueue(struct io_wqe
return;
}
- work_flags = work->flags;
raw_spin_lock_irqsave(&wqe->lock, flags);
io_wqe_insert_work(wqe, work);
wqe->flags &= ~IO_WQE_FLAG_STALLED;
+ do_wake = (work->flags & IO_WQ_WORK_CONCURRENT) ||
+ !atomic_read(&acct->nr_running);
raw_spin_unlock_irqrestore(&wqe->lock, flags);
- if ((work_flags & IO_WQ_WORK_CONCURRENT) ||
- !atomic_read(&acct->nr_running))
+ if (do_wake)
io_wqe_wake_worker(wqe, acct);
}
