Hello,

On Mon, Jun 6, 2011 at 8:51 PM, Manuel Reimer
[off-list ref] wrote:

I've converted a 1,5TB SATA drive to IDE to run it on a UDMA100 IDE
controller.

Hmmm...

ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
ata2.00: failed command: WRITE DMA EXT
ata2.00: cmd 35/00:00:30:07:8a/00:04:57:00:00/e0 tag 0 dma 524288 out
        res 40/00:ff:00:00:00/00:00:00:00:00/40 Emask 0x4 (timeout)
ata2.00: status: { DRDY }

Looks like genuine transmission problems.  Lengthening timeout won't
help anything.  It will just make it take longer to fail.

What I definetly need is some workaround to stop the kernel from speeding
down until UDMA/25 is reached. This speed is totally unacceptable for such a
big drive... :-(

Currently, there's no way to disable automatic speed down without
patching the code and given that the mechanism works well for many
cases, I'm somewhat reluctant to introduce changes / switch solely for
this case.  You're having a genuine transmission problem and I think
fixing the underlying hardware issue would be the better choice than
mucking with EH behaviors.

Thank you.

-- 
tejun