diego torres wrote:
> Estimado Jim,
> Con fecha domingo, 07 de octubre de 2007, 20:15:26, escribió:
> > I have a very similar disk (fw 3.AAJ) and have not had any problems
> > with offline tests.
> >> Testing causes lots of messages like:
> >> ata2.00: exception Emask 0x2 SAct 0xc SErr 0x0 action 0x2 frozen
> >> ata2.00: (spurious completions during NCQ issue=0x0 SAct=0x8 FIS=004040a1:00000004)
> > Try disabling NCQ (echo 1 > /sys/block/sdX/device/queue_depth) to see
> > if the behavior changes.
> Funny... it is working now (after setting the queue size to 1).
> I was going to add this model to the blacklisted ncq drives and try,
> because the echo command is not working always (sometimes when I
> issue the echo... an invalid argument is returned).
> Are you using this drive with queue_depth != 1?
My controller doesn't support NCQ, so I'm using queue_depth = 1.
Since disabling NCQ does fix the problem, I'd suggest brining this up
on linux-ide to see if it should be blacklisted by default.