From: Andrew Morton <akpm@linux-foundation.org>
To: Michael Buesch <mb@bu3sch.de>
Cc: philb@gnu.org, carsten@sol.wohnheim.uni-ulm.de, renau@acm.org,
tim@cyberelk.demon.co.uk, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] IEEE1284: Use del_timer_sync in parport_wait_event
Date: Tue, 2 Dec 2008 16:30:55 -0800 [thread overview]
Message-ID: <20081202163055.e80e0b81.akpm@linux-foundation.org> (raw)
In-Reply-To: <200811302120.49766.mb@bu3sch.de>
On Sun, 30 Nov 2008 21:20:49 +0100
Michael Buesch <mb@bu3sch.de> wrote:
> Use del_timer_sync() instead of del_timer() to make sure
> the timer won't be running when we return from parport_wait_event(),
> because this would crash due to destruction of timer_list.
>
> Signed-off-by: Michael Buesch <mb@bu3sch.de>
>
> ---
>
> This is completely untested and just based on a code review.
> Just think about the following sequence of events:
> add_timer.
> down_interruptible is interrupted by a signal.
> We enter the timer callback handler on another CPU.
> del_timer, but the timer callback is still running.
> Return from parport_wait_even, which destroys the automatic variable "timer"
> while the callback is running on another CPU.
>
Yes, I agree.
> ===================================================================
> --- linux-2.6.orig/drivers/parport/ieee1284.c 2008-10-24 21:08:12.000000000 +0200
> +++ linux-2.6/drivers/parport/ieee1284.c 2008-11-30 21:09:25.000000000 +0100
> @@ -84,7 +84,7 @@ int parport_wait_event (struct parport *
>
> add_timer (&timer);
> ret = down_interruptible (&port->physport->ieee1284.irq);
> - if (!del_timer (&timer) && !ret)
> + if (!del_timer_sync (&timer) && !ret)
> /* Timed out. */
> ret = 1;
Please fix checkpatch errors when altering code. This function is
already a mismash of correct and incorrect styles - we might as well
increase the correctness density.
What we really want in this function is down_interruptible_timeout().
afaict we have all the infrastructure in place to implement that, but
nobody has typed
static noinline int __sched
__down_interruptible_timeout(struct semaphore *sem, long jiffies)
{
return __down_common(sem, TASK_INTERRUPTIBLE, jiffies);
}
and then tested it.
Or convert the whole thing to mutexes and write
mutex_lock_interruptible_timeout() :)
prev parent reply other threads:[~2008-12-03 0:31 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-11-30 20:20 Michael Buesch
2008-12-03 0:30 ` Andrew Morton [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20081202163055.e80e0b81.akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=carsten@sol.wohnheim.uni-ulm.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mb@bu3sch.de \
--cc=philb@gnu.org \
--cc=renau@acm.org \
--cc=tim@cyberelk.demon.co.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome