In some mail from Cameron Schaus, sie said:
>
> I have a i386 machine running 3.2 following stable (dmesg below). The
> machine was running 2.9 for over a year with no problems. I recently
> upgraded it to 3.2, and the machine crashed after being up for a
> couple of weeks.
>
> I copied the following from the console:
> sd1(ahc:1:0): SCB 0x10 - timed out
> ahc1: Dumping card state while idle at SEQADDR 0x7
> ACCUM=0xe, SINDEX=0x64, DINDEX=0x65, ARG_2=0x2
> ...
> sd1(ahc:1:0): Queuing a BDR SCB
> sd1(ahc:1:0): Bus Device Reset Message Sent
> sd1(ahc:1:0): no longer in timeout, status = 0
> sd1(ahc:1:0): target 1 using async transfers
> sd1(ahc:1:0): Bus Device Reset on A:1 3 SCBs aborted
>
>
> Both drives are sitting on the internal connector, and nothing is
> connected to the external connector.
>
> Is this a problem with my driver, cabling, controller card, or a
> software problem?

FWIW, I've seen a very similar problem with the same driver when
running OpenBSD (I don't recall if it was 3.2 or something earlier.)
Hardware was slightly different - GA-586DX with onboard AIC7880 and
me using the wide SCSI connectors for the disks plus narrow for CD,
etc.

It happened during the quiet of the night when, I presume, the machine
was otherwise idle.

If you can, enable ddb and wait for it to happen again and copy down
the stack trace. I didn't scribe down the output I saw, at the time,
but from memory the driver was recursively calling itself to deal with
a similar problem (SCSI bus had reset.)

But rest assured, you're not the only victim of this "bug".

Darren

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]