| Message ID | 20200619151718.22338-3-john.ogness@linutronix.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | block: remove retry loop | expand |
On Fri, Jun 19, 2020 at 05:23:18PM +0206, John Ogness wrote: > The reverse-order double lock dance in ioc_release_fn() is using a > retry loop. This is a problem on PREEMPT_RT because it could preempt > the task that would release q->queue_lock and thus live lock in the > retry loop. > > RCU is already managing the freeing of the request queue and icq. If > the trylock fails, use RCU to guarantee that the request queue and > icq are not freed and re-acquire the locks in the correct order, > allowing forward progress. > > Signed-off-by: John Ogness <john.ogness@linutronix.de> Again, after starring on it for while and reading up and down, I'd say, it looks good. Also a quick test run with blktests and lockdep enabled didn't produce any warnings. Reviewed-by: Daniel Wagner <dwagner@suse.de>
diff --git a/block/blk-ioc.c b/block/blk-ioc.c index 5dbcfa1b872e..57299f860d41 100644 --- a/block/blk-ioc.c +++ b/block/blk-ioc.c @@ -107,9 +107,23 @@ static void ioc_release_fn(struct work_struct *work) ioc_destroy_icq(icq); spin_unlock(&q->queue_lock); } else { - spin_unlock_irq(&ioc->lock); - cpu_relax(); - spin_lock_irq(&ioc->lock); + /* Make sure q and icq cannot be freed. */ + rcu_read_lock(); + + /* Re-acquire the locks in the correct order. */ + spin_unlock(&ioc->lock); + spin_lock(&q->queue_lock); + spin_lock(&ioc->lock); + + /* + * The icq may have been destroyed when the ioc lock + * was released. + */ + if (!(icq->flags & ICQ_DESTROYED)) + ioc_destroy_icq(icq); + + spin_unlock(&q->queue_lock); + rcu_read_unlock(); } }
The reverse-order double lock dance in ioc_release_fn() is using a retry loop. This is a problem on PREEMPT_RT because it could preempt the task that would release q->queue_lock and thus live lock in the retry loop. RCU is already managing the freeing of the request queue and icq. If the trylock fails, use RCU to guarantee that the request queue and icq are not freed and re-acquire the locks in the correct order, allowing forward progress. Signed-off-by: John Ogness <john.ogness@linutronix.de> --- block/blk-ioc.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-)