[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20131003185630.GJ13318@ZenIV.linux.org.uk>
Date: Thu, 3 Oct 2013 19:56:30 +0100
From: Al Viro <viro@...IV.linux.org.uk>
To: Dave Jones <davej@...hat.com>,
Linux Kernel <linux-kernel@...r.kernel.org>,
ebiederm@...ssion.com,
Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: pipe/cred lockdep warning
On Tue, Oct 01, 2013 at 10:57:55AM -0400, Dave Jones wrote:
> I seem to recall hitting this quite a while ago. Does it look familiar ?
> Either it didn't get fixed, or it's back..
> [ 2836.628506] trinity-child0/30759 is trying to acquire lock:
> [ 2836.628544] (&sig->cred_guard_mutex){+.+.+.}, at: [<ffffffff81239045>] proc_pid_attr_write+0xf5/0x140
> [ 2836.628624]
> but task is already holding lock:
> [ 2836.628664] (&pipe->mutex/1){+.+.+.}, at: [<ffffffff811ccf66>] pipe_lock+0x26/0x30
> [ 2836.628734]
splice to /proc/<pid>/attr/<something> [cred_guard shite under pipe->mutex]
vs.
splice from XFS [pipe->mutex under exclusive iolock]
vs.
read from XFS in execve() [shared iolock under cred_guard]
Note, BTW, that splice to /proc/<pid>/attr/<something> is broken.
proc_pid_attr_write() is *not* supposed to allow partial writes at all.
Frankly, I'd consider adding a ->splice_write() instance that would
simply return -EINVAL there...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists