From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751875AbZGFP2S (ORCPT ); Mon, 6 Jul 2009 11:28:18 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751183AbZGFP2L (ORCPT ); Mon, 6 Jul 2009 11:28:11 -0400 Received: from lvps87-230-0-242.dedicated.hosteurope.de ([87.230.0.242]:38933 "EHLO lvps87-230-0-242.dedicated.hosteurope.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751118AbZGFP2K (ORCPT ); Mon, 6 Jul 2009 11:28:10 -0400 Cc: Alan Cox , linux-kernel@vger.kernel.org, "David S. Miller" , netdev@vger.kernel.org Message-Id: <11EC5046-5F30-4CB6-BDAF-0DAA2C69929B@it-loops.com> From: Michael Guntsche To: Herbert Xu In-Reply-To: <20090706083119.GA17782@gondor.apana.org.au> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v935.3) Subject: Re: BUG: scheduling while atomic Date: Mon, 6 Jul 2009 17:28:06 +0200 References: <23AED400-6DA5-462F-94F0-BEE34AC08A7A@it-loops.com> <20090627140643.7140f6a9@lxorguk.ukuu.org.uk> <33C7D336-F062-47B3-9033-D3DC465F793C@it-loops.com> <20090706074333.GA17276@gondor.apana.org.au> <20090706083119.GA17782@gondor.apana.org.au> X-Mailer: Apple Mail (2.935.3) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Jul 6, 2009, at 10:31, Herbert Xu wrote: > On Mon, Jul 06, 2009 at 03:43:33PM +0800, Herbert Xu wrote: >> >> I think the first backtrace is bogus, this is where the real >> problem is. So what see here is that on xmit from the network, >> PPP will call pty_write from BH context which eventually leads >> to tty_throttle and the sleep. > > Here's the stupidest possible fix for this, changing the mutex > to a spin lock. We should also revert the other commit that > tried to fix this, a6540f731d506d9e82444cf0020e716613d4c46c as > those unthrottles are really needed Just to confirm. I reverted the commit and applied the patch on-top of 2.6.31-rc2. Up to now I was not able to reproduce the crashes I had before on my setup here. Kind regards, Michael