linux-kernel - Re: 2.6.24-rc7 lockdep warning when poweroff

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

Message-Id: <1200386536.15103.16.camel@twins>
Date:	Tue, 15 Jan 2008 09:42:16 +0100
From:	Peter Zijlstra <peterz@...radead.org>
To:	Dave Young <hidave.darkstar@...il.com>
Cc:	Johannes Berg <johannes@...solutions.net>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Ingo Molnar <mingo@...e.hu>
Subject: Re: 2.6.24-rc7 lockdep warning when poweroff


On Tue, 2008-01-15 at 09:24 +0800, Dave Young wrote:
> On Tue, Jan 15, 2008 at 08:31:48AM +0800, Dave Young wrote:
> > On Jan 14, 2008 6:51 PM, Johannes Berg <johannes@...solutions.net> wrote:
> > >
> > > > The warning that triggered (lockdep.c:700) means that one class (key)
> > > > was used with more than one name.
> > >
> > > Right.
> > >
> > > > Looking at cleanup_workqueue_thread(), the lock_acquire() there works on
> > > > wq->lockdep_map, and that is only initialized at one spot:
> > > > __create_workqueue_key(), thus it stands to reason that that was
> > > > mis-used.
> > >
> > > Oh ok, yes, makes sense. Maybe something is generating a workqueue with
> > > a name that's passed in but the key is statically from that place. I'll
> > > try to find it.
> > 
> > I add some debug printk and found the names :
> > 
> > block_osm/exec_osm
> > 
> > in drivers/message/i2o
> > 
> > maybe this helps.
> 
> Not sure right or not, the following patch fixed the problem:
> 
> diff -upr linux/include/linux/workqueue.h linux.new/include/linux/workqueue.h
> --- linux/include/linux/workqueue.h	2008-01-15 08:49:08.000000000 +0800
> +++ linux.new/include/linux/workqueue.h	2008-01-15 08:49:42.000000000 +0800
> @@ -154,7 +154,7 @@ __create_workqueue_key(const char *name,
>  #ifdef CONFIG_LOCKDEP
>  #define __create_workqueue(name, singlethread, freezeable)	\
>  ({								\
> -	static struct lock_class_key __key;			\
> +	struct lock_class_key __key;			\
>  								\
>  	__create_workqueue_key((name), (singlethread),		\
>  			       (freezeable), &__key);		\

That didn't get you:

  INFO: trying to register non-static key.

Msgs?

But it sure looks like __create_workqueue() is asking for trouble, if
there is a __create_workqueue() instance that takes a non constant name
we're in trouble.

--
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/