Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 07 Apr 2007 12:35:44 +0100
From:      Tom Judge <tom@tomjudge.com>
To:        Kris Kennaway <kris@obsecurity.org>
Cc:        freebsd-stable@freebsd.org
Subject:   Re: Repeatable crash with mkdir causing a divide by zero error
Message-ID:  <46178210.9000809@tomjudge.com>
In-Reply-To: <20070406184200.GA62383@xor.obsecurity.org>
References:  <46161B62.4070505@tomjudge.com> <20070406184200.GA62383@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Kris Kennaway wrote:
> On Fri, Apr 06, 2007 at 11:05:22AM +0100, Tom Judge wrote:
>> Hi,
>>
>> I have seen some problems with a new file system that I created 
>> yesterday in that I could repeatedly get the system to crash in with a 
>> mkdir.
>>
>> Here is the disk information
>> mfid1: <MFI Logical Disk> on mfi1
>> mfid1: 5716992MB (11708399616 sectors) RAID volume 'Images' is optimal
>>
>> I created a new file system tuned for 64k blocks, an average file size 
>> of 1Mb, and 2500 files per directory.
>>
>> newfs -b 65535 -g 1048576 -h 2500 /dev/mfid1p1
>> mount /dev/mfid1p1 /compere
>> mkdir /compere/images
>> mkdir /compere/images/1999
>>
>> (Also tested with mkdir test; mkdir test/1998)
>>
>> The system is and amd64 system running 6.2-RELEASE and the pmap.c patch. 
>>  I have 3 cores cause by 3 different apps (rsync, gmkdir, mkdir) and 
>> can provide any more information if required.  I have attached a back 
>> trace, unfortunatly I cannot do any testing  as the system is now in 
>> testing (newfs -b 65535 -g 1048576 /dev/mfid1p1 was used and seems not 
>> to cause the bug).
> 
> This might be simple to fix, but please file a PR if it does not get
> picked up by someone on this list.
> 
> Kris
> 
<<SNIP>>

For any one that is interested there is now a PR for this problem 
(kern/111352).

Tom




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46178210.9000809>