2017-07-28 16:39:00

by Benjamin Coddington

Subject: [PATCH] NFSv4.1: Fix a race where CB_NOTIFY_LOCK fails to wake a waiter

nfs4_retry_setlk() sets the task's state to TASK_INTERRUPTIBLE within the
same region protected by the wait_queue's lock after checking for a
notification from CB_NOTIFY_LOCK callback. However, after releasing that
lock, a wakeup for that task may race in before the call to
freezable_schedule_timeout_interruptible() and set TASK_WAKING, then
freezable_schedule_timeout_interruptible() will set the state back to
TASK_INTERRUPTIBLE before the task will sleep. The result is that the task
will sleep for the entire duration of the timeout.

Since we've already set TASK_INTERRUPTIBLE in the locked section, just use
freezable_schedule_timout() instead.

Fixes: a1d617d8f134 ("nfs: allow blocking locks to be awoken by lock callbacks")
Signed-off-by: Benjamin Coddington <[email protected]>
---
fs/nfs/nfs4proc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index dbfa18900e25..f5a7faac39a7 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6441,7 +6441,7 @@ nfs4_retry_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
set_current_state(TASK_INTERRUPTIBLE);
spin_unlock_irqrestore(&q->lock, flags);

- freezable_schedule_timeout_interruptible(NFS4_LOCK_MAXTIMEOUT);
+ freezable_schedule_timeout(NFS4_LOCK_MAXTIMEOUT);
}

finish_wait(q, &wait);
--
2.9.3

2017-07-28 16:49:37

by Jeff Layton

Subject: Re: [PATCH] NFSv4.1: Fix a race where CB_NOTIFY_LOCK fails to wake a waiter

On Fri, 2017-07-28 at 12:33 -0400, Benjamin Coddington wrote:
> nfs4_retry_setlk() sets the task's state to TASK_INTERRUPTIBLE within the
> same region protected by the wait_queue's lock after checking for a
> notification from CB_NOTIFY_LOCK callback. However, after releasing that
> lock, a wakeup for that task may race in before the call to
> freezable_schedule_timeout_interruptible() and set TASK_WAKING, then
> freezable_schedule_timeout_interruptible() will set the state back to
> TASK_INTERRUPTIBLE before the task will sleep. The result is that the task
> will sleep for the entire duration of the timeout.
>
> Since we've already set TASK_INTERRUPTIBLE in the locked section, just use
> freezable_schedule_timout() instead.
>
> Fixes: a1d617d8f134 ("nfs: allow blocking locks to be awoken by lock callbacks")
> Signed-off-by: Benjamin Coddington <[email protected]>
> ---
> fs/nfs/nfs4proc.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> index dbfa18900e25..f5a7faac39a7 100644
> --- a/fs/nfs/nfs4proc.c
> +++ b/fs/nfs/nfs4proc.c
> @@ -6441,7 +6441,7 @@ nfs4_retry_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
> set_current_state(TASK_INTERRUPTIBLE);
> spin_unlock_irqrestore(&q->lock, flags);
>
> - freezable_schedule_timeout_interruptible(NFS4_LOCK_MAXTIMEOUT);
> + freezable_schedule_timeout(NFS4_LOCK_MAXTIMEOUT);
> }
>
> finish_wait(q, &wait);

Nice catch!

Reviewed-by: Jeff Layton <[email protected]>