Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 02 Aug 2004 19:30:48 -0600
From:      Scott Long <scottl@freebsd.org>
To:        Daniel Eriksson <daniel_k_eriksson@telia.com>
Cc:        'freebsd current' <current@freebsd.org>
Subject:   Re: Panic in today's current, UFS problem?
Message-ID:  <410EEAC8.8010900@freebsd.org>
In-Reply-To: <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA0VcX9IoJqUaXPS8MjT1PdsKAAAAQAAAA5PQ1jek5/EqvRvENdXMWUAEAAAAA@telia.com>
References:  <!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAABgAAAAAAAAA0VcX9IoJqUaXPS8MjT1PdsKAAAAQAAAA5PQ1jek5/EqvRvENdXMWUAEAAAAA@telia.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Daniel Eriksson wrote:

> I was trying out the new patch from Ville-Pertti Keinonen that might fix my
> interrupt storm problems with SATA when I ran into a panic. The dmesg is
> attached.
> 
> OS: CURRENT, 2004.08.02.16.00.00
>     + VIA/SATA patch from Ville-Pertti Keinonen
> (http://docs.freebsd.org/cgi/getmsg.cgi?fetch=505643+0+current/freebsd-curre
> nt)
>     + libdisk patch from Lukas Ertl
> (http://docs.freebsd.org/cgi/getmsg.cgi?fetch=526964+0+current/freebsd-curre
> nt)
> 
> 
> What I did:
> 1. fdisk/label two 250GB discs hooked up to the on-board SATA controller
> (using sysinstall)
> 2. copy ~20GB of data to both discs

Is this step necessary in order to produce the panic?

> 3. successfully run the stress-test mentioned in Ville-Pertti Keinonen's
> mail (above)

Is this step necessary also?

> 4. unmount both discs
> 5. run "atacontrol create RAID0 128 ad20 ad22" (the two SATA discs)
> 6. fdisk/label the striped array (using sysinstall, forcing it to UFS1)
> 7. mount the new partition, and create three empty dirs
> 8. *boom*
> 

How reproducable is this, and does it always panic in the same spot? 
There is definitely some sort of buffer corruption going on, and either 
it's happening in the buffer/cache layer or it's a bad DMA from the
drive.

Scott



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?410EEAC8.8010900>