Re: SMTP 2.0.34pre5 problems

Doug Ledford (dledford@dialnet.net)
Thu, 02 Apr 1998 16:59:03 -0600

Alan Cox wrote:
>
> > I found the machine working perfectly, and it is still up and running.
> >
> > Any idea what could cause such a message ? Anything I can do to track the
> > problem down ?
>
> Its caused when we try and send a message to the other processor but it
> doesnt accept it within 10mS. I've seen hard scsi bus resets cause this
> but not much else (and you'd also have a pile of 'scsi blah error' messages
> too

For 5.0.11 I'll add a loop counter to the aic7xxx_isr routine. Something
like the "Too much work at interrupt" message from some of the NIC drivers.
It shouldn't be a problem in 2.1 because command completion is just queued
by the mid level code, but under 2.0.x the command completion is
synchronous, so that interrupt could take a long time (which is why
interrupts are enabled during those operations). However, I would imagine
that being in an interrupt, even if interrupts are enabled, might be the
cause of this message. So, very heavy disk load could cause this message
with the latest driver although it should be fairly benign in nature if it
isn't a SCSI bus reset.

-- 


 Doug Ledford  <dledford@dialnet.net>
  Opinions expressed are my own, but
     they should be everybody's.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu