ATA driver races with interrupts

[Daniel Eriksson]

Søren Schmidt wrote:

> How about not running the SMART monitor ?

After upgrading the system to sources dated 2004.08.11.15.00.00 I rebooted
with the SMART monitor turned off. The system worked fine (apart from having
to run without ACPI due to recent commits) until just 30 minutes ago when I
was turning it off to upgrade kernel again (trying to track down the ACPI
breakage and needed boot -v output for Nate). After all the buffers had been
flushed it reported a timeout on one of the SATA discs, removed it from the
configuration and then went on to a panic (sorry, no log available).
Afterwards the channel was locked up as always when this happens.

It seems the problem is not smartmontools related.

/Daniel Eriksson