IRQ deadlocks

[Robert G. Brown]

On Sun, 20 Sep 1998, System Administrator wrote:

> anybody get IRQ deadlocks on their motherboard? If so, have you found them
> related to the SCSI drivers? My servers doesn't tell me crap about them?
> Just IRQ DEADLOCK DETECTED BY CPU 0 or CPU 12 (I run SMP) and the syslog
> doesn't say a dang thing.

I don't think that IRQ DEADLOCK... necessarily means that you actually
have an IRQ deadlock anymore.  I got that message several times when
running earlier 5.1.0 revisions (before pre7, but I don't think that
matters) when the system crashed.  The message is, as I recall,
generated when the system doesn't return from an interrupt in a certain
amount of time.  This can (and used to) happen when there was a real
interrupt deadlock, where one CPU held the lock which couldn't process
the its interrupt until a second interrupt was processed, which couldn't
happen because only one CPU could hold the lock and it wasn't the right
one...  It (the timeout) can also happen when a driver tries to
initialize a device and hangs before returning, and I got a deadlock
message on nearly every crash that happened at the wrong place for a
while.

I actually think that the aic7xxx spinlocks are pretty safe, especially
in 2.0.35.  I don't recall anybody really having troubles with "real"
IRQ deadlocks anymore on the linux-smp list -- I think that there may be
one or two more lurking somewhere, but they are in the
one-in-a-blue-moon category of biting and not something that affects
anybody enough to matter (or debug).  2.1.X periodically "breaks" for
some people for a revision or two, but it then restabilizes as well.

SO, this message "probably" means nothing more than that you are hanging
inside the aic7xxx interrupt handler.  Try upgrading to pre10 if you
haven't -- it is much better than pre7.  I think pre11 is around the
corner as well, from some of Doug's messages today.

   rgb

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu




To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe freebsd-aic7xxx" in the body of the message