[3540] in linux-scsi channel archive
Re: SMP 2.1.90-pre3 SCSI kernel panic
daemon@ATHENA.MIT.EDU (sistema@readysoft.es)
Mon Mar 16 10:18:32 1998
Date: Mon, 16 Mar 1998 16:12:53 +0100 (MET)
From: sistema@readysoft.es
To: dledford@dialnet.net
cc: eowmob@exp-math.uni-essen.de, linux-scsi@vger.rutgers.edu
In-Reply-To: <Pine.LNX.3.95.980316065513.1085B-100000@dledford.dialnet.net>
On 16 Mar, Doug Ledford wrote:
> On Mon, 16 Mar 1998 sistema@readysoft.es wrote:
>
>> Ok. I´ve already activated remapping bits, running 2.0.33.
>> But I´ve been able to backup all scsi devices with no read error many
>> times and keep on working. No SCSI errors.
>>
>> Why does 2.1.89 and over generate so many failures doing exactly the
>> same? It doesn´t look as a hardware problem. Maybe it´s an SMP issue or
>> a scsi problem. I would discard HD errors.
>
> Because 2.1.x is *not* doing exactly the same. If it were, it would be
> 2.0.33 wouldn't it? You're making the assumption that much of the code in
Ok. Not the same.
> the linux kernel is static and does things exactly the same from version
> to version. This assumption is false. Secondly, those errors you posted
I agree. I know they´re very different. I only meant that it was the
same driver.
> were Peripheral Write Fault errors. That's not reading. That also goes
> back to what I mentioned about the possible usage of portions of the
> ext2fs that haven't been used in the past. In short, there are a lot of
> possibilities and to discount anything yet is premature. What we do know,
> is that the HD is saying it can't write something that it accepted as a
> valid command. It wasn't out of range, it suppossedly went to real
> sectors, etc, but the drive says it can't do it. That's the reason that
> I've been bringing up hardware.
I get the errors trying to do a backup and that´s reading, but I suppose
it has to do with updating atime and other file attributes.
I´ll try this evening to get a repeateable failure, with remapping bits
on and 2.1.90pre-3. I´m quiet sure my backup won´t finnish. Maybe you
want me to do some debugging?
Thanks
Pau
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
