2004-06-04 09:50:29

by Andrew Walrond

[permalink] [raw]
Subject: [ACPI] Asus SK8N (x86_64) motherboard ata1 DMA timeout (Promise SATA)

Using linux-2.6.7-rc2 (32bit), two SATA drives on Promise TX2plus. The bios
has been upgraded to latest version 1007.

Without any kernel parameters, the i/o locks after a short time copying files.
Copied by hand from dmesg after i/o freeze:

ata1 DMA timeout
scsi0: ERROR on channel 0, id 0, lun 0, CDB: 0x2a 00 04 ae a2 f2 00 00 08 00
Current sda: sense = 70 3
ASC= c ASCQ= 2
Raw sense data: 0x70 0x00 0x03 0x00 0x00 0x00 0x00 0x06 0x00 0x00 0x00 0x00
0x0c 0x02
end_request: I/O error, dev sda, sector 78553842

I guess this is interrupt related?

From dmesg without kernel params (broken case)

ata1: SATA max UDMA/133 cmd 0xF8ABF200 ctl 0xF8ABF238 bmdma 0x0 irq 19
ata2: SATA max UDMA/133 cmd 0xF8ABF280 ctl 0xF8ABF2B8 bmdma 0x0 irq 19

From dmesg with acpi=off (working fine)

ata1: SATA max UDMA/133 cmd 0xF8AB6200 ctl 0xF8AB6238 bmdma 0x0 irq 11
ata2: SATA max UDMA/133 cmd 0xF8AB6280 ctl 0xF8AB62B8 bmdma 0x0 irq 11

Full dmesg's are attached for working and broken cases. Let me know if I can
provide any more information.

Andrew Walrond


Attachments:
(No filename) (1.03 kB)
dmesg.no_kernel_params (14.71 kB)
dmesg.acpi_off (12.32 kB)
Download all attachments

2004-06-04 17:17:43

by Rafał J. Wysocki

[permalink] [raw]
Subject: Re: [ACPI] Asus SK8N (x86_64) motherboard ata1 DMA timeout (Promise SATA)

On Friday 04 of June 2004 11:45, Andrew Walrond wrote:
> Using linux-2.6.7-rc2 (32bit), two SATA drives on Promise TX2plus. The bios
> has been upgraded to latest version 1007.
>
> Without any kernel parameters, the i/o locks after a short time copying
> files. Copied by hand from dmesg after i/o freeze:
>
> ata1 DMA timeout
> scsi0: ERROR on channel 0, id 0, lun 0, CDB: 0x2a 00 04 ae a2 f2 00 00 08
> 00 Current sda: sense = 70 3
> ASC= c ASCQ= 2
> Raw sense data: 0x70 0x00 0x03 0x00 0x00 0x00 0x00 0x06 0x00 0x00 0x00 0x00
> 0x0c 0x02
> end_request: I/O error, dev sda, sector 78553842

I saw something like this on a SiI 3114 controller, but then I thought it was
because of a failing drive, since the second one that was attached to another
SATA channel worked just fine at the same time. I didn't try it with the
acpi=off setting (the system was a dual Opteron w/ AMD chipset, and the
kernel was a 2.6.7-rc1, AFAIR).

rjw