Messages in this thread | | | Date | Sun, 11 Sep 2011 20:41:07 +0200 | From | Oleg Nesterov <> | Subject | Re: [PATCH 1/1] freezer: fix wait_event_freezable/__thaw_task races |
| |
On 09/11, Oleg Nesterov wrote: > > On 09/11, Tejun Heo wrote: > > > > On Thu, Sep 08, 2011 at 07:59:26PM +0200, Oleg Nesterov wrote: > > > > > > Yeap, with freezable_with_signal gone, this looks correct & simpler to > > > > me > > > > > > I don't really understand this... set_freezable_with_signal() has a > > > lot of problems, yes... But even if it wasn't removed this fix makes > > > sense anyway, afaics. > > > > > > If freezable_with_signal caller does wait_event_freezable_timeout(), > > > __retval becomes -ERESTARTSYS after freeze_task(). The next iteration > > > will return 0 with the KERN_ERR message from schedule_timeout(). > > > > Hmmm... but with the change, if a kthread gets TIF_SIGPENDING from > > freezer and then thawed, it would not enter try_to_freeze() and thus > > won't clear TIF_SIGPENDING. > > Ah yes, you are right, thanks. But once again, please note that in this > case wait_event_freezable_timeout() is broken anyway.
OTOH, I forgot about wait_event_freezable(). It was racy, but mostly worked. So yes, you are right, without "kill freezable_with_signal" this patch is wrong.
I missed this, thanks.
Oleg.
| |