From owner-freebsd-current  Thu Jul  1 23:54:30 1999
Delivered-To: freebsd-current@freebsd.org
Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2])
	by hub.freebsd.org (Postfix) with ESMTP id C5E8F150CB
	for <current@FreeBSD.ORG>; Thu,  1 Jul 1999 23:54:28 -0700 (PDT)
	(envelope-from dillon@apollo.backplane.com)
Received: (from dillon@localhost)
	by apollo.backplane.com (8.9.3/8.9.1) id XAA53043;
	Thu, 1 Jul 1999 23:54:28 -0700 (PDT)
	(envelope-from dillon)
Date: Thu, 1 Jul 1999 23:54:28 -0700 (PDT)
From: Matthew Dillon <dillon@apollo.backplane.com>
Message-Id: <199907020654.XAA53043@apollo.backplane.com>
To: current@FreeBSD.ORG
Subject: Re: Something weird happening w/ SMP -current
References:  <199907020634.XAA52990@apollo.backplane.com>
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

:    interrupts such as clk and rtc continue to work just fine.
:
:    When the interrupts stop, dd is stuck in a disk-wait state.  I'm sure it
:    isn't the SCSI bus:  I've got three SCSI busses and it occurs on all three.
:    Plus it does not occur on a UP kernel, only an SMP kernel.
:
:    If I run an endless loop in another window (e.g. systat shows 50% idle
:    on the 2x SMP box), the interrupts appear to work for longer periods of
:    time before crapping out.
:
:    It is very weird.  I have no idea what is going on.
:
:						-Matt
:

    Here's some more info.  If the dd is stuck and systat -vm 1 is showing
    no interrupts occuring on, for example, ahc2 (irq17), and I then do 
    something that causes an interrupt to occur on mux (irq19), which I
    guess is ahc1, ahc2 then starts working... until it gets stuck again,
    that is.  I can also get ahc2 going again by issuing another dd on ahc2.

    For example:

	dd if=/dev/rda1d bs=32k of=/dev/null
		... let this one run ...
		... it gets stuck ...
	dd if=/dev/rda1d bs=32k of=/dev/null count=1
		... this unsticks the first dd ...
		... the first dd runs again ...
		... the first dd gets stuck again ...
		repeat...

    All I can think of is that something is causing the system to lose an
    occassional interrupt.  There's a race condition somewhere.

					-Matt
					Matthew Dillon 
					<dillon@backplane.com>


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