From owner-freebsd-current  Wed Oct  9 14:51:02 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id OAA09471
          for current-outgoing; Wed, 9 Oct 1996 14:51:02 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id OAA09456
          for <current@FreeBSD.org>; Wed, 9 Oct 1996 14:50:57 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.7.6/8.6.9) id HAA25087; Thu, 10 Oct 1996 07:48:45 +1000
Date: Thu, 10 Oct 1996 07:48:45 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199610092148.HAA25087@godzilla.zeta.org.au>
To: gibbs@freefall.freebsd.org, nate@mt.sri.com
Subject: Re: SIO problems
Cc: current@FreeBSD.org
Sender: owner-current@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

>I thought this was a problem with the VM system sitting at splhigh()
>too much.  I used to get these a long time ago, but they vanished

It can't be that, because splhigh() doesn't mask "fast" interrupts.

BTW, there is a problem with the PCI interrupt demultiplexor not
supporting "fast" interrupts.  I tried running the cy driver with
"slow" interrupts (a one line change to delate the line that sets
RI_FAST).  The (12 byte) fifo was overrun by a factor of about 4
for every syscons vt switch (because busy waiting is used to set
the LEDs and the LEDs are set at spltty()).

Bruce