From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1754517Ab3A3KbD (ORCPT <rfc822;w@1wt.eu>);
	Wed, 30 Jan 2013 05:31:03 -0500
Received: from cantor2.suse.de ([195.135.220.15]:41588 "EHLO mx2.suse.de"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1753726Ab3A3KbB (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Wed, 30 Jan 2013 05:31:01 -0500
Date: Wed, 30 Jan 2013 11:31:00 +0100
Message-ID: <s5hfw1jrmhn.wl%tiwai@suse.de>
From: Takashi Iwai <tiwai@suse.de>
To: Ming Lei <ming.lei@canonical.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/4] firmware: Refactoring for splitting user-mode helper code
In-Reply-To: <CACVXFVN9JNe6US655hh3cNpKyJFdWrLEYOOOdnk97_5bFmZYxw@mail.gmail.com>
References: <1359470814-7993-1-git-send-email-tiwai@suse.de>
	<1359470814-7993-2-git-send-email-tiwai@suse.de>
	<CACVXFVOf4LZD+qVcqS9yBtvNZCRF=LMpR5BKuXby3Z8dfB+Vzw@mail.gmail.com>
	<s5hpq0nw35i.wl%tiwai@suse.de>
	<CACVXFVN9JNe6US655hh3cNpKyJFdWrLEYOOOdnk97_5bFmZYxw@mail.gmail.com>
User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI/1.14.6 (Maruoka)
 FLIM/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL/10.8 Emacs/24.2
 (x86_64-suse-linux-gnu) MULE/6.0 (HANACHIRUSATO)
MIME-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka")
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

At Wed, 30 Jan 2013 18:25:20 +0800,
Ming Lei wrote:
> 
> On Wed, Jan 30, 2013 at 3:17 PM, Takashi Iwai <tiwai@suse.de> wrote:
> >> The above usermodehelper_read_lock thing may be a functional change,
> >> and looks not what you claimed in commit log, :-). The lock is currently held in
> >> direct loading case, but your patch change the rule. Without holding the lock,
> >> request_firmware() may touch filesystem / storage too early during
> >> kernel boot or system resume in direct loading case.
> >
> > Does it really happen in a real scenario?
> 
> Some crazy drivers might call request_firmware inside resume callback,
> with usermodehelper_read_lock, we can find the mistake easily and log it.

But it's supposed to be cached, no?

> > If so, using usermode helper lock for that purpose sounds like an
> > abuse to be fixed differently or replaced with a better one.
> 
> Might be, but looks not good to introduce this change in a code
> refactoring patch. Or you can do it in another patch for discussion
> if you have better way to handle the situation.

Yes, I already modified the code again.
Will resubmit them soon.


Takashi