Date: Thu, 7 Jun 2007 23:14:04 +0100
From: Alan Cox <alan(a)lxorguk.ukuu.org.uk>
Subject: Re: Odd kernel ATA message on F7
To: For users of Fedora <fedora-list(a)redhat.com>
Cc: abraxis(a)telkomsa.net
Message-ID: <20070607231404.5282f9ec(a)the-village.bc.nu>
Content-Type: text/plain; charset=US-ASCII
On Thu, 7 Jun 2007 20:33:34 +0200
Neil Thompson <abraxis(a)telkomsa.net> wrote:
> Hi all,
>
> I've just seen a message on my logs that I've never seen before -
>
> Jun 7 20:24:34 eeyore kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action
0x2 frozen
> Jun 7 20:24:34 eeyore kernel: ata2.00: (BMDMA stat 0x24)
> Jun 7 20:24:34 eeyore kernel: ata2.00: cmd a0/01:00:00:00:00/00:00:00:00:00/a0 tag
0 cdb 0x4a data 8 in
> Jun 7 20:24:34 eeyore kernel: res 58/00:02:00:08:00/00:00:00:00:00/a0
Emask 0x2 (HSM violation)
That means the drive and controller ended up disagreeing violently about
what the controller was up to.
> Jun 7 20:24:34 eeyore kernel: ata2: soft resetting port
> Jun 7 20:24:34 eeyore kernel: ata2.00: configured for UDMA/33
> Jun 7 20:24:34 eeyore kernel: ata2: EH complete
And we reset it.
> Jun 7 20:24:35 eeyore kernel: VFS: busy inodes on changed media.
Odd
We've seen some drives have problems with NCQ and do that, you also get
HSM errors off some emulations (eg in xen it seems) because of emulation
problems.
Bugzilla it, include dmesg and lspci -vvxxx and assign to alan(a)redhat.com
(me) and I'll take a more detailed look
Was there ever any resolution to this? I'm seeing the same problem on
an F7 x86_64 system with a Promise sata controller.