Prev: KVM MMU: allow more page become unsync at getting sp time
Next: [PATCH RESEND 1/2] Composite framework: Add suspended sysfs entry
From: Arve Hjønnevåg on 24 Apr 2010 03:30 2010/4/23 Tejun Heo <tj(a)kernel.org>: > Hello, > > On 04/24/2010 12:49 AM, Arve Hj�nnev�g wrote: >> I want the suspend blocker active when the work is pending or running. >> I did not see a way to do this on top of the workqueue api without >> adding additional locking. > > Well, then add additional locking. �suspend_blocker is far away from > being a hot path and there's nothing wrong with additional locking as > long as everything is wrapped inside proper APIs. �Adding stuff to the > core code for things as specific as this is much worse. > OK, I'll try to do this. Do you want it in a separate header file as well? >> If the work is both queued and starts running on another workqueue >> between "get_wq_data(work) == cwq" and "!work_pending(work)", then >> suspend_unblock will be called when it shouldn't. It should work fine >> if I change to it check pending first though, since it cannot move >> back to the current workqueue without locking cwq->lock first. > > The code is more fundamentally broken. �Once work->func has started > executing, the work pointer can't be dereferenced as work callback is > allowed to free or re-purpose the work data structure and in the same > manner you can't check for pending status after execution has > completed. > I only touch the work structure after the callback has returned for suspend blocking work, which does not allow that. >> Or are you talking about the race when the callback is running on >> multiple (cpu) workqueues at the same time. In that case the suspend >> blocker is released when the callback returns from the last workqueue >> is was queued on, not when all the callbacks have returned. On that >> note, is it ever safe to use flush_work and cancel_work_sync for work >> queues on anything other than a single single threaded workqueue? > > flush_work() is guaranteed only to flush from the last queued cpu but > cancel_work_sync() will guarantee that no cpu is executing or holding > onto the work. �So, yeah, as long as the limitations of flush_work() > is understood, it's safe. > Sorry, I didn't see the for_each_cpu(cpu, cpu_map) part of wait_on_work(). cancel_work_sync() does look safe as long as the work has not moved to completely different workqueue. > Going back to the original subject, just add simplistic outside > locking in suspend_blocker_work API (or whatever other name you > prefer). �That will be much cleaner and safer. �Let's think about > moving them into workqueue implementation proper after the number of > the API users grows to hundreds. > OK. -- Arve Hj�nnev�g -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Tejun Heo on 24 Apr 2010 03:50 Hello, On 04/24/2010 09:21 AM, Arve Hj�nnev�g wrote: > OK, I'll try to do this. Do you want it in a separate header file as well? I think it would fit better with the rest of suspend_blocker.h APIs but it's not like I thought about it hard and long, so please take my preference with a grain of salt. > I only touch the work structure after the callback has returned for > suspend blocking work, which does not allow that. Oh, you're right. I missed the current_work_blocks_suspend test. Sorry about that. Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Oleg Nesterov on 26 Apr 2010 10:10 On 04/23, Arve Hj�nnev�g wrote: > > On Fri, Apr 23, 2010 at 5:20 AM, Oleg Nesterov <oleg(a)redhat.com> wrote: > > > The unblock code in run_workqueue() is racy, > > it can unblock after the work was queued on another CPU, cwq->lock can't > > help. > > If the work is both queued and starts running on another workqueue > between "get_wq_data(work) == cwq" and "!work_pending(work)", then > suspend_unblock will be called when it shouldn't. It should work fine > if I change to it check pending first though, since it cannot move > back to the current workqueue without locking cwq->lock first. It can be queued on the same or different workqueue on another CPU, right before run_workqueue() does unblock. But I guess this doesn't matter. You already discussed this all with Tejun while I was sleeping ;) And I agree very much, the new plan looks much better imho. Oleg. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Tejun Heo on 28 Apr 2010 02:50 Hello, > +static void suspend_blocking_work_complete(struct suspend_blocking_work *work) > +{ > + unsigned long flags; > + > + WARN_ON(!work->active); > + spin_lock_irqsave(&work->lock, flags); > + if (!--work->active) > + suspend_unblock(&work->suspend_blocker); > + spin_unlock_irqrestore(&work->lock, flags); > +} Maybe work->active can be an atomic_t and the lock can be removed? > +/** > + * suspend_blocking_work_destroy - Destroy suspend_blocking_work > + * @work: The work item in question > + * > + * If the work was ever queued on more then one workqueue all but the last > + * workqueue must be flushed before calling suspend_blocking_work_destroy. As it's calling cancel_work_sync(), the above is not true. As long as no one is trying to queue it again, suspend_blocking_work_destroy() is safe to call regardless of how the work has been used. > +void suspend_blocking_work_destroy(struct suspend_blocking_work *work) > +{ > + cancel_suspend_blocking_work_sync(work); > + WARN_ON(work->active); > + suspend_blocker_destroy(&work->suspend_blocker); > +} > +EXPORT_SYMBOL(suspend_blocking_work_destroy); Other than the above, it looks good to me. Thanks. -- tejun -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Arve Hjønnevåg on 28 Apr 2010 03:10
2010/4/27 Tejun Heo <tj(a)kernel.org>: > Hello, > >> +static void suspend_blocking_work_complete(struct suspend_blocking_work *work) >> +{ >> + � � unsigned long flags; >> + >> + � � WARN_ON(!work->active); >> + � � spin_lock_irqsave(&work->lock, flags); >> + � � if (!--work->active) >> + � � � � � � suspend_unblock(&work->suspend_blocker); >> + � � spin_unlock_irqrestore(&work->lock, flags); >> +} > > Maybe work->active can be an atomic_t and the lock can be removed? > I need the spinlock to prevent the work from getting re-queued before suspend_unblock. >> +/** >> + * suspend_blocking_work_destroy - Destroy suspend_blocking_work >> + * @work: The work item in question >> + * >> + * If the work was ever queued on more then one workqueue all but the last >> + * workqueue must be flushed before calling suspend_blocking_work_destroy. > > As it's calling cancel_work_sync(), the above is not true. �As long as > no one is trying to queue it again, suspend_blocking_work_destroy() is > safe to call regardless of how the work has been used. > I'm not sure what the best terminology is here, but cancel_work_sync() only waits for work running on all the cpu-workqueues of the last workqueue. So, if the caller queued the work on more than one workqueue, suspend_blocking_work_destroy does not ensure that the suspend_blocking_work structure is not still in use (it should trigger the WARN_ON though). >> +void suspend_blocking_work_destroy(struct suspend_blocking_work *work) >> +{ >> + � � cancel_suspend_blocking_work_sync(work); >> + � � WARN_ON(work->active); >> + � � suspend_blocker_destroy(&work->suspend_blocker); >> +} >> +EXPORT_SYMBOL(suspend_blocking_work_destroy); > > Other than the above, it looks good to me. > > Thanks. > > -- > tejun > -- Arve Hj�nnev�g -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |