Re: linux-next: Tree for August 14 (sysfs/acpi errors)

From: Zhang Rui <rui.zhang@intel.com>
To: Andi Kleen <ak@linux.intel.com>
Cc: Greg KH <greg@kroah.com>, Randy Dunlap <randy.dunlap@oracle.com>,
	Stephen Rothwell <sfr@canb.auug.org.au>,
	linux-next@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
	linux-acpi@vger.kernel.org, rusty@rustcorp.com.au
Subject: Re: linux-next: Tree for August 14 (sysfs/acpi errors)
Date: Fri, 15 Aug 2008 11:09:32 +0800	[thread overview]
Message-ID: <1218769772.3209.46.camel@rzhang-dt> (raw)
In-Reply-To: <48A4F05B.9050309@linux.intel.com>

On Fri, 2008-08-15 at 04:56 +0200, Andi Kleen wrote:
> Greg KH wrote:
> > On Thu, Aug 14, 2008 at 08:38:28AM -0700, Randy Dunlap wrote:
> >> calling  param_sysfs_init+0x0/0x1e7
> >> ------------[ cut here ]------------
> >> WARNING: at fs/sysfs/dir.c:463 sysfs_add_one+0x35/0x3d()
> >> sysfs: duplicate filename 'acpi' can not be created
> > 
> > Why would this be a sysfs error, it's acpi doing something foolish,
> > that's why the sysfs layer is warning about it :)
> 
> The NULL pointer reference further down actually looks like something
> foolish in the sysfs layer. But actually ACPI is not doing
> anything wrong here I think. According to the backtrace it happens
> when a kernel param (aka module_param) is set. And creating multiple
> params in the same acpi space is completely legal.
> 
> It looks more like the high level code that sets up these parameters
> broke somehow and starts registering these twice and now ACPI is the
> first one to hit it (maybe because it starts with 'a' :-). 
Perhaps we could try "acpi=off" and see if it still happens.

thanks,
rui