Re: [SLUG] DMA off?

From: Paul M Foster (paulf@quillandmouse.com)
Date: Sat Jun 03 2006 - 00:51:04 EDT


Eben King wrote:
> On Thu, 25 May 2006, Levi Bard wrote:
>
>> I think there's an important difference here:
>>
>> Eben's:
>>> May 23 06:08:21 pc kernel: hdb: dma timeout error: status=0xd0 { Busy }
>>
>> Kwan's:
>>> May 15 17:13:30 helios kernel: hda: dma timeout error: status=0x58 {
>>> DriveReady
>>> SeekComplete DataRequest }
>>
>> I agree that on drives where I start to see the errors with
>> "DriveReady SeekComplete" , failure is imminent. However, I don't
>> think that "Busy" is necessarily an indicator of the same,
>> particularly when doing a huge data dump, as Eben seems to be.
>
> I did some rewiring and repositioning, so now instead of
>
> Primary---hda = HD #1
> |
> +-hdb = HD #2
>
> Secondary---hdc = DVD #1 (R/W)
> |
> +-hdd = DVD #2 (R)
>
> I now have
>
> Primary---hda = HD #1
> |
> +-hdb = nothing
>
> Secondary---hdc = DVD #1 (R/W)
> |
> +-hdd = HD #2
>
> HD #1 -> HD #2 throughput is faster this way (I'm thinking about getting
> a PCI -> IDE card so I can put each device on a separate cable), but I
> still get errors (I wrote a script that watches the log, and re-enables
> DMA if it drops):
>
> Jun 1 06:25:35 pc kernel: hda: dma_timer_expiry: dma status == 0x21
> Jun 1 06:25:45 pc kernel: hda: DMA timeout error
> Jun 1 06:25:45 pc kernel: hda: dma timeout error: status=0xd0 { Busy }
> Jun 1 06:25:45 pc kernel: ide: failed opcode was: unknown
> Jun 1 06:25:45 pc kernel: hda: DMA disabled
> Jun 1 06:25:45 pc kernel: ide0: reset: success
> Jun 1 06:26:40 pc kernel: hda: dma_timer_expiry: dma status == 0x20
> Jun 1 06:26:40 pc kernel: hda: DMA timeout retry
> Jun 1 06:26:40 pc kernel: hda: status timeout: status=0xd0 { Busy }
> Jun 1 06:26:40 pc kernel: ide: failed opcode was: unknown
> Jun 1 06:26:40 pc kernel: ide0: reset: success
> Jun 1 06:27:01 pc kernel: hda: dma_timer_expiry: dma status == 0x21
> Jun 1 06:27:11 pc kernel: hda: DMA timeout error
> Jun 1 06:27:11 pc kernel: hda: dma timeout error: status=0xd0 { Busy }
> Jun 1 06:27:11 pc kernel: ide: failed opcode was: unknown
> Jun 1 06:27:11 pc kernel: hda: DMA disabled
> Jun 1 06:27:11 pc kernel: ide0: reset: success
> Jun 1 06:27:12 pc kernel: hda: set_drive_speed_status: status=0x58 {
> DriveReady SeekComplete DataRequest }
> Jun 1 06:27:12 pc kernel: ide: failed opcode was: unknown
> Jun 1 06:27:12 pc kernel: hda: dma_intr: status=0x58 { DriveReady
> SeekCompleteDataRequest }
> Jun 1 06:27:12 pc kernel: ide: failed opcode was: unknown
> ...
> Jun 2 06:25:42 pc kernel: hda: dma_timer_expiry: dma status == 0x21
> Jun 2 06:25:53 pc kernel: hda: DMA timeout error
> Jun 2 06:25:53 pc kernel: hda: dma timeout error: status=0xd0 { Busy }
> Jun 2 06:25:53 pc kernel: ide: failed opcode was: unknown
> Jun 2 06:25:53 pc kernel: hda: DMA disabled
> Jun 2 06:25:53 pc kernel: ide0: reset: success
>
> Those times are during the backup. The rewiring was a day or two before
> 5/29, so the errors probably don't happen every day. It looks like hda
> is going bad, as I see no errors from hdd (HD #2).
>
> OK, that drive is dying. Time to replace it. Now, if it fails, is it
> likely to prevent me from using DMA on it, or prevent me from accessing
> the drive at all?
>

I haven't followed this closely, but didn't the original errors happen
to hdb? Now they're happening to hda. Something these have in common,
like an IDE channel? Seems unlikely two different hard drives would fail
this way. Or am I missing something?

-- 
Paul M. Foster
-----------------------------------------------------------------------
This list is provided as an unmoderated internet service by Networked
Knowledge Systems (NKS).  Views and opinions expressed in messages
posted are those of the author and do not necessarily reflect the
official policy or position of NKS or any of its employees.



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 20:01:00 EDT