The big (and only) problem ive had with my bp6 is the fact that it has a
HPT366 built in that i want to use.
e2fsck is now causing me irq_timeouts.
I dont get lockups from reading heavily anymore, but I@$## interrupt
timeouts, their driving me crazy.
Ive given up trying to use *any* sort of striping with lvm or raid, im
using lvm now (with no striping), and still problems.
I try and use my machine normally, after a while i get one or two irq
timout, it tries to disable dma, soemtimes it works,most times it will
sit looping for hours trying to get over a stupid f@#$&* intertupt.
So what can i do, i have to reboot, of course i cannot shutdown
properly, so when it reboots it has to run e2fsck, and i get
irq_timeeouts doing that !
(plus my bp6 isnt shutting properly now, i have to do a cold reboot, new
bios i think causes this)
I feel like im in that move gridlocked, ive tried stable and unstable
kernels going back a while. It doesnt matter where i turn, whenever i
think im safe i get this crap message.
I really want to help to fix this problem, does anyone know of any good
documents describing how to debug the kernel ?
Thanks
Glenn
P.S. I guess i could do the obvious and give up alltogether on using any
of my drives in DMA mode, but thats 3MB/s versus 20MB/s, ill be making
sacrifices to the great hardware gods before i accept that my ide drive
cant do any better than one designed over 10 years ago.
Robert Redelmeier wrote:
>
> Zdenek Kabelac wrote:
>
> > I was able to trace one lock of my BP6 board.
> >
> > Both CPUs were locked in the irq_enter (arch/i386/kernel/irq.h)
> > while loop - one was in the process swapper (<smp_local_timer_interrupt>)
> > and the second process one in (<handle_IRQ_event>)
>
> Interesting. If you have stable hardware (burnP6 & burnBX)
> then this should be reported to the kernel mailing list.
>
> I've beaten very hard on spinlocks and not had them break,
> but all my testing was in userspace, and this is a
> kernel issue.
>
> -- Robert
> --
> =- To unsubscribe, email [EMAIL PROTECTED] with the -=
> =- body of "unsubscribe linux-abit". -=
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
