2000-11-16 01:01:21

by Karnik, Rahul

[permalink] [raw]
Subject: VIA IDE bug with WD drive?

Hi all,

I get the following error if I try to enable DMA on my Abit KT7 motherboard
with a VIA2C686 chipset:

hdb: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest }
hdb: timeout waiting for DMA
hda: DMA disabled
hdb: DMA disabled
ide0: reset: success

hdb is a Western Digital 136BA 13,6 GB drive and hda is a Maxtor 20GB drive.
I do not get the error when enabling DMA on the Maxtor drive (hda).

I have tried kernel versions 2.2.16-3 (stock RH7), 2.2.17 and 2.4.0-testx.
Is this a known bug? Is it solved by the IDE backport patch?

TIA,
Rahul


2000-11-16 01:24:43

by dep

[permalink] [raw]
Subject: Re: VIA IDE bug with WD drive?

On Wednesday 15 November 2000 19:30, Karnik, Rahul wrote:

| I get the following error if I try to enable DMA on my Abit KT7
| motherboard with a VIA2C686 chipset:
|
| hdb: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest
| } hdb: timeout waiting for DMA
| hda: DMA disabled
| hdb: DMA disabled
| ide0: reset: success

i get the same thing, along with a crc error, over and over on a
20-gig WD IDE drive. alternately puzzling and frightening.
apparently, wd uses some nonstandard goofball error checking thing
that just doesn't work with linux at present. it *seems* to do no
harm.
--
dep
--
Everyone is entitled to his own opinion but not his own facts.
-- Daniel Patrick Moynahan

2000-11-16 15:08:10

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: VIA IDE bug with WD drive?

On Wed, Nov 15, 2000 at 07:56:43PM -0500, dep wrote:
> On Wednesday 15 November 2000 19:30, Karnik, Rahul wrote:
>
> | I get the following error if I try to enable DMA on my Abit KT7
> | motherboard with a VIA2C686 chipset:
> |
> | hdb: irq timeout: status=0x58 { DriveReady SeekComplete DataRequest
> | } hdb: timeout waiting for DMA
> | hda: DMA disabled
> | hdb: DMA disabled
> | ide0: reset: success
>
> i get the same thing, along with a crc error, over and over on a
> 20-gig WD IDE drive. alternately puzzling and frightening.
> apparently, wd uses some nonstandard goofball error checking thing
> that just doesn't work with linux at present. it *seems* to do no
> harm.

Ok, both of you, we can try to track this down.

1) Please try with 2.4.0-latest.
2) Send me the complete dmesg.
3) Send me lspci -vvvxxx
4) Send me /proc/ide/via

I'll see what I can do about the driver.

--
Vojtech Pavlik
SuSE Labs