RELENG_5: ata interrupt problems

[Søren Schmidt]

Randy Bush wrote:
>>Since I have a few confirmations that backstepping just ATA doesnt help 
>>at all, I'm not convinced this is an ATA problem, at least not in the 
>>sense that changes to ATA broke it.
> 
> 
> after some decades in swdev, i am never sure of anything until
> i find the cause.  and, even then, i am suspicious.  :-)

Geesh we are alot of oldtimers in here :) and yes there is always just 
one more bug...

Anyhow I committed a fix for the fake slave problem, now we still have 
those "taskqueue stalled" complaints that I'm still scratching my head 
about and that backstepping ATA doesn't solve...

-Søren