lk/linux-kernel-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

workqueue: fix max_active handling in init_and_link_pwq()

Browse Source 
Since 9e8cd2f589 ("workqueue: implement apply_workqueue_attrs()"),
init_and_link_pwq() may be called to initialize a new pool_workqueue
for a workqueue which is already online, but the function was setting
pwq->max_active to wq->saved_max_active without proper
synchronization.

Fix it by calling pwq_adjust_max_active() under proper locking instead
of manually setting max_active.

Signed-off-by: Tejun Heo <tj@kernel.org>
This commit is contained in:
Tejun Heo
2013-03-13 16:51:35 -07:00
parent 699ce097ef
commit 983ca25e73
1 changed files with 9 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
kernel/workqueue.c
View File

@@ -3771,21 +3771,25 @@ static void init_and_link_pwq(struct pool_workqueue *pwq,
pwq->wq = wq; pwq->wq = wq;
pwq->flush_color = -1; pwq->flush_color = -1;
pwq->refcnt = 1; pwq->refcnt = 1;
pwq->max_active = wq->saved_max_active;
INIT_LIST_HEAD(&pwq->delayed_works); INIT_LIST_HEAD(&pwq->delayed_works);
INIT_LIST_HEAD(&pwq->mayday_node); INIT_LIST_HEAD(&pwq->mayday_node);
INIT_WORK(&pwq->unbound_release_work, pwq_unbound_release_workfn); INIT_WORK(&pwq->unbound_release_work, pwq_unbound_release_workfn);
/*
* Link @pwq and set the matching work_color. This is synchronized
* with flush_mutex to avoid confusing flush_workqueue().
*/
mutex_lock(&wq->flush_mutex); mutex_lock(&wq->flush_mutex);
spin_lock_irq(&workqueue_lock); spin_lock_irq(&workqueue_lock);
/*
* Set the matching work_color. This is synchronized with
* flush_mutex to avoid confusing flush_workqueue().
*/
if (p_last_pwq) if (p_last_pwq)
*p_last_pwq = first_pwq(wq); *p_last_pwq = first_pwq(wq);
pwq->work_color = wq->work_color; pwq->work_color = wq->work_color;
/* sync max_active to the current setting */
pwq_adjust_max_active(pwq);
/* link in @pwq */
list_add_rcu(&pwq->pwqs_node, &wq->pwqs); list_add_rcu(&pwq->pwqs_node, &wq->pwqs);
spin_unlock_irq(&workqueue_lock); spin_unlock_irq(&workqueue_lock);