cvs commit: src/sys/dev/acpica acpi.c

[Andrea Campi]

On Wed, Oct 22, 2003 at 04:40:15PM -0400, John Baldwin wrote:
> >> By the way, any ideas on why my laptop gets an interrupt storm when it
> >> resumes from S1 and any ideas on how to fix that?
> > 
> > No idea at the moment.  What interrupt and who's on it?
> 
> No idea really.  I showed it to you at BSDCon when I tried to suspend
> and you mentioned the Linux guys had seen something similar.  It seems
> to be related to the 0619 import btw.

FYI: I think I'm seeing this on my Thinkpad 570E on reboot. I get an
interrupt storm on reboot; it's on irq 11, which is what gets assigned
to most devices. However, I've disabled all devices which could possibly
use it, and the only only one left is cbb. I haven't been able yet to
determine when this started, as I noticed it only recently. I tried
comparing a verbose boot after cold boot (working) vs warm (non working)
and they looked the same to a visual inspection, so it doesn't seem to
be a case of irq misrouting.

The boot stops after printing: Mounting root from ufs:/dev/ad2s2a.

Breaking into DDB and using the new 'show intrcnt' command shows I am
indeed getting a lot (thousands per second) interrupt 11. However,
setting a breakpoint in cbb_intr and tracing through it shows that
cbb_get(sc, CBB_SOCKET_EVENT) is returning 0.

I haven't gotten any further than this for now. With my limited knowledge,
this makes me think that either something else is generating irqs, or
cbb is but cbb_get() is giving a wrong answer. It's not clear to how to
determine this, if anybody has a clue I could use it too.


Bye,
	Andrea

-- 
     The three Rs of Microsoft support: Retry, Reboot, Reinstall.