From: "Darrick J. Wong" <darrick.wong@oracle.com>
Subject: Re: superblock completely overwritten
Date: Tue, 13 Dec 2016 10:13:47 -0800
Message-ID: <20161213181346.GA6354@birch.djwong.org>
References: <trinity-fba1c40f-e300-43ee-a4cd-f9d7f48ecef2-1481643547528@3capp-gmx-bs76>
 <7a0b791c-8a9f-23e5-02a2-cd903024ef06@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Heinz Nimmervoll <bt1now@gmx.at>, linux-ext4@vger.kernel.org
To: Eric Sandeen <sandeen@redhat.com>
Return-path: <linux-ext4-owner@vger.kernel.org>
Received: from aserp1040.oracle.com ([141.146.126.69]:20770 "EHLO
        aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S932886AbcLMSNz (ORCPT
        <rfc822;linux-ext4@vger.kernel.org>); Tue, 13 Dec 2016 13:13:55 -0500
Content-Disposition: inline
In-Reply-To: <7a0b791c-8a9f-23e5-02a2-cd903024ef06@redhat.com>
Sender: linux-ext4-owner@vger.kernel.org
List-ID: <linux-ext4.vger.kernel.org>

On Tue, Dec 13, 2016 at 10:15:25AM -0600, Eric Sandeen wrote:
> On 12/13/16 9:39 AM, Heinz Nimmervoll wrote:
> > I still got no answer for my problem thats why I try it here... hopefully you could help me out.
> >  
> > System:
> > 
> > Embedded board with Atmel SAM9x25
> > Debian Wheezy Kernel 3.11.6
> > 32GB Samsung SDHC card with ext4 root- partition (journal activated)
> >  
> > 
> > After system running two weeks or so superblock from rootfs (ext4) at block 0 got overwritten with "trash data".
> > This is happening with like 20% of the embedded devices.
> >  
> > hex comparision between faulty and good superblock starting at byte 1024:
> >  
> > Before (good):
> >  
> > 00000000  00 ee 02 00 00 b8 0b 00  00 96 00 00 ab a9 05 00  |................|
> > 00000010  c3 0c 02 00 00 00 00 00  02 00 00 00 02 00 00 00  |................|
> > 00000020  00 80 00 00 00 80 00 00  40 1f 00 00 9e 68 46 58  |........@....hFX|
> > 00000030  9e 68 46 58 2e 00 64 00  53 ef 01 00 01 00 00 00  |.hFX..d.S.......|
> 
> <snip>
> 
> >  
> > After (corrupted):
> >  
> > 00000000  00 00 00 00 a4 81 00 00  dd 00 00 00 24 8e 5d 54  |............$.]T|
> > 00000010  7e 8e 5d 54 18 a6 9f 41  00 00 00 00 00 00 01 00  |~.]T...A........|
> > 00000020  08 00 00 00 00 00 08 00  01 00 00 00 0a f3 01 00  |................|
> > 00000030  04 00 00 00 00 00 00 00  00 00 00 00 01 00 00 00  |................|

81a4?  f30a?  That looks like an inode block (i_mode 0x81a4) and an
extent tree root (eh_magic 0xf30a), shifted up by 4 bytes before being
written onto sector 0.

> <snip>
> 
> > - How is it possible, that even the magic number (and everything else) got overwritten?
> > - Why could it ever be overwritten?
> 
> I don't think anyone here can tell you what happened, it is almost certainly not
> an ext4 bug.  Could be a driver bug, or an admin running a stray "dd" command,
> or some other utility gone astray, or ... anything, really.

Bad/malicious SDHC card would be my guess?  (Are you doing power fail
testing?)

--D

> As a very long shot, what does "blkid" or "file -s" tell you about the block device
> after it's been overwritten?  Perhaps it will recognize a signature.
> 
> Otherwise, you could do something like a modified kernel to trap any IO to block
> zero on the device and issue a printk about the process which is doing it, filtering
> out any expected ext4 accesses.
> 
> -Eric
>   
> > Thank you so much!
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html