ATA driver races with interrupts

[Kenneth Culver]

Quoting Ville-Pertti Keinonen <will+freebsd-current at will.iki.fi>:

> My previous patch makes things more reliable, but it's not good enough...
>
> Attached is an updated version of the patch that clears ATA_EXPECT_INTR
> earlier (in the interrupt handler), which seems to stabilize things better.
>
> I'm definitely not convinced of its correctness, since I'm not sure
> which part of the code is racing against interrupts despite the previous
> version of the patch, but I can no longer cause any failures.
>
> Repeating the problem should be trivial with an unpatched -current with
> PREEMPTION enabled and hardware similar to mine - ASUS K8V Deluxe, two
> SATA disks on the VIA 6420:
>
I have this exact same motherboard, but I'm using the onboard promise 
controller
set up with 2 disks as a raid0, and I'm not having any problems with this
setup. I used the promise controller because from what I've been told, the
promise controllers are very fast in FreeBSD compared to other controllers.

Ken