Panics after AHCI timeouts

[Alexey Shuvaev]

On Mon, Oct 24, 2011 at 08:27:49PM +0200, C. P. Ghost wrote:
> On Tue, Oct 18, 2011 at 3:13 PM, Alexey Shuvaev
> <shuvaev at physik.uni-wuerzburg.de> wrote:
> > On Tue, Oct 18, 2011 at 06:19:19AM +0800, Adrian Chadd wrote:
> >> On 18 October 2011 03:00, Alexey Shuvaev
> >> <shuvaev at physik.uni-wuerzburg.de> wrote:
> >> > On Sat, Oct 08, 2011 at 10:14:56PM +0200, Alexey Shuvaev wrote:
> >> >> Hello list!
> >> >>
> >> > Errr... Replying to myself... Ping? Should I file a PR and put it
> >> > in the back burner? :)
> >>
> >> I think filing a PR is a good move. Then just be proactive and poke
> >> people about it. It'd be good to get this fixed. :)
> >>
> > Done, kern/161768.
> >
> > Question to the list: does anybody see successful recovery from AHCI
> > timeout an a recent CURRENT? Recent means June 2011 or newer, so 9.0
> > branch counts also. That is, there are some kernel messages like this:
> >
> > ahcich0: Timeout on slot 29 port 0
> > ahcich0: is 00000000 cs 00000000 ss ffffffff rs ffffffff tfd 40 serr 00000000 cmd 0000fc17
> >
> > but then AHCI recovers and the system does not panic?
> 
> I'm seeing these timeouts too on an 8.2-STABLE amd64 r222832
> from June 7. The system hangs partially -- or, more precisely, all
> processes that attempt to access the disk on this channel hang,
> everything else continues as normal.
> 
> I suspect a faulty cable, but I don't have physical access to the system
> to replace parts right now. A panic would be a regression, so I'm holding
> off updates on that server until AHCI becomes more tolerant and somewhat
> self-healing. :(
> 
In a communication not on the list mav has said he has done some tests
not so long ago by injecting artificial failures in the AHCI code.
He has not observed any panics and it is not clear if the problem is
generic / hardware related / or purely local. I would not have any time
to investigate further till November.

So, Your Mileage May Vary :)

0.02$,
Alexey.