linux-kernel - Re: 2.6.19-rc3-git7: BUG: mutex warning sysfs_dir

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Mon, 06 Nov 2006 16:13:50 +0100
From:	Takashi Iwai <tiwai@...e.de>
To:	Linus Torvalds <torvalds@...l.org>
Cc:	Mark Lord <lkml@....ca>,
	Linux Kernel <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...l.org>,
	Frederik Deweerdt <deweerdt@...e.fr>
Subject: Re: 2.6.19-rc3-git7:  BUG: mutex warning sysfs_dir_open

At Tue, 31 Oct 2006 09:53:35 -0800 (PST),
Linus Torvalds wrote:
> 
> 
> 
> On Tue, 31 Oct 2006, Mark Lord wrote:
> > 
> > I found this in my syslog just now:
> > 
> > kernel: BUG: warning at kernel/mutex.c:132/__mutex_lock_common()
> 
> That's the "spin_lock_mutex()" debugging (kernel/mutex-debug.h), and it's 
> one of either
> 
> 	DEBUG_LOCKS_WARN_ON(in_interrupt());
> or
> 	DEBUG_LOCKS_WARN_ON(l->magic != l);
> 
> where the much more likely one is the "l->magic" one triggering, because 
> your bug that follows is:
> 
> > kernel: BUG: unable to handle kernel paging request at virtual address 6e726574
> 
> 64726574 is "tern", for whatever that's worth, and that EIP is in 
> "__list_add()", which comes from list_add_tail(), and so it looks like the 
> mutex "lock->lock_list" is also totally buggered (which would certainly 
> fit with the lock "magic" field being crud too).
> 
> > kernel: Call Trace:
> > kernel:  [<c035d76e>] __mutex_lock_slowpath+0x8e/0x260
> > kernel:  [<c035d961>] mutex_lock+0x21/0x30
> > kernel:  [<c01a0286>] sysfs_dir_open+0x26/0x60
> 
> This would be
> 
> 	mutex_lock(&dentry->d_inode->i_mutex);
> 
> in sysfs_dir_open(), so you would seem to have an inode that is 
> overwritten by crud (or it could be a dentry with a bogus d_inode 
> pointer, of course).
> 
> There's a ton of stuff with "tern" in it, mostly "Internal", so sadly even 
> the nice string pattern doesn't seem to be all that helpful.
> 
> Is this somewhat reproducible? 
> 
> HOWEVER, google does find a clue. We've had reports of something somewhat 
> similar before: googling for "__mutex_lock_slowpath" and "sysfs_dir_open" 
> shows for example
> 
> 	http://lkml.org/lkml/2006/8/18/72
> 
> where the thread ended up first blaming sound, but then deciding that 
> maybe it was DRM-related. However, you don't seem to have any AGP or DRM 
> support at all, so maybe it really _is_ a sound problem.
> 
> Mark - can you verify that you don't have any DRM support in your kernel?
> 
> Frederic, Takashi, did that bug-report ever get resolved?

IIRC, in Frederik's case, it turned out to be irrelevant from the
sound stuff since the bug occured even before loading ALSA modules.

Of course, it still _might_ be a sound problem :)

It'd be nice if we can see what sysfs entry triggers bugs...


Takashi
-
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/