2003-06-04 00:40:00

by Andre Hedrick

[permalink] [raw]
Subject: Re: hd?: lost interrupt; at least since 2.4.21-pre7


Basically somewhere a patch fouled the state machine.

The expiry tripped just as an interrupt happens to be reported from the
drive.

I will need to look at the series of innerdiffs when I get time.

On Tue, 3 Jun 2003 [email protected] wrote:

>
> Hi there
>
> hda: dma_timer_expiry: dma status == 0x64
> hda: lost interrupt
> hda: dma_intr: bad DMA status (dma_stat=70)
> hda: dma_intr: status=0x50 { DriveReady SeekComplete }
>
> I've read a lot of people do suffer from that kind of errors.
> Especially people using VIA IDE chipset.
> When using 2.4.21-pre4-aa1 everthing is ok but when I was using .21-pre7 I had to wait
> a few days before starting to see those errors.
> I've just tried 2.4.21-pre6. Errors still happen and that a lot faster.
> 2.4.21-rc2-ac1 was subjected to that kind of errors too.
>
> A few weeks ago, I read people complaining about similar issues.
> Especially people using Local or IO APIC support on uniprocessors and VIA IDE chipsets.
> I'm using both local and IO APIC on uniprocessors.
>
> Is there any patch or solution ? I heard disabling acpi may help.
>
> .config /proc/pci et /proc/dmesg are attached to this mail.
>
> My two harddrives are subjected to those errors. Using 2.4.21-pre4-aa1 there is no error.
> I'm available to do tests.
>
>
>
>

Andre Hedrick
LAD Storage Consulting Group