From mboxrd@z Thu Jan 1 00:00:00 1970 From: "George Spelvin" Subject: Re: metadata_csum + e4defrag seems to cause problems Date: 29 Apr 2013 11:57:15 -0400 Message-ID: <20130429155715.23219.qmail@science.horizon.com> References: <20130426025753.GA21659@gmail.com> Cc: linux-ext4@vger.kernel.org To: gnehzuil.liu@gmail.com, linux@horizon.com Return-path: Received: from science.horizon.com ([71.41.210.146]:63605 "HELO science.horizon.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1757253Ab3D2P5Q (ORCPT ); Mon, 29 Apr 2013 11:57:16 -0400 In-Reply-To: <20130426025753.GA21659@gmail.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: > Yes, metatdata_csum + e4defrag could corrupt ext4 file system. We have > found this bug and it has been fixed by this commit (2656497b, it is in > dev branch of ext4 tree). I am not sure whether you use dev branch. > Could you please tell me your kernel version? If you don't use dev > branch, could you please try dev branch of ext4 tree? Here is the git > link: > https://git.kernel.org/cgit/linux/kernel/git/tytso/ext4.git/log/?h=dev I was running 3.8-rc7 at the time of the corruption; I rebooted to -rc8 as part of the recovery. Yes, I already have the ext4 repository cloned, but that was just for pulling a particular fix Ted recommended a while ago. I've been reluctant to use a branch labelled "dev" more routinely without more reassurance that it's believed to be "brave" rather than "stupid" to use in something other than a dedicated test environment. I'll try v3.9 + ext4/dev. Thank you! (Current head is 0d606e2c9fcc: "ext4: fix type-widening bug in inode table readahead code")