From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261539AbTLPMkY (ORCPT ); Tue, 16 Dec 2003 07:40:24 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261552AbTLPMkY (ORCPT ); Tue, 16 Dec 2003 07:40:24 -0500 Received: from hades.mk.cvut.cz ([147.32.96.3]:10937 "EHLO hades.mk.cvut.cz") by vger.kernel.org with ESMTP id S261539AbTLPMkX (ORCPT ); Tue, 16 Dec 2003 07:40:23 -0500 Message-ID: <3FDEFD31.6050405@kmlinux.fjfi.cvut.cz> Date: Tue, 16 Dec 2003 13:40:17 +0100 From: Jindrich Makovicka User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6b) Gecko/20031208 X-Accept-Language: cs, en-us, en MIME-Version: 1.0 To: akpm@osdl.org, linux-kernel@vger.kernel.org Subject: Re:Re: crypto-loop + highmen -> random crashes in -test11 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > Soeren Sonnenburg wrote: >> >> Hi. >> >> I get random crashes/corruption/ init kills when I use cryptoloop on >> this highmem enabled ppc machine. > > People have reported cryptoloop+highmem crashes on ia32 as well. I'm not > sure if that was with -mm though. I tried to apply loop-highmem.patch, loop-highmem-fixes.patch, and loop-remove-blkdev-special-case.patch on a vanilla 2.6.0-test11, and it finally made the cryptoloop work on my machine - an Athlon XP, 1.5G RAM. Without the patch, cryptoloop+highmem didn't work at all - a fresh created ext2 was either impossible to mount, or it mounted corrupted, and cryptoloop overall seemed to act more like /dev/urandom. Regards, -- Jindrich Makovicka