From owner-freebsd-bugs  Sun May 12 20:02:40 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id UAA06733
          for bugs-outgoing; Sun, 12 May 1996 20:02:40 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id UAA06728
          for <freebsd-bugs@freebsd.org>; Sun, 12 May 1996 20:02:36 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id MAA26611; Mon, 13 May 1996 12:56:35 +1000
Date: Mon, 13 May 1996 12:56:35 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199605130256.MAA26611@godzilla.zeta.org.au>
To: coredump@nervosa.com, freebsd-bugs@freebsd.org, mrm@Mole.ORG
Subject: Re: List of open Problem Reports
Cc: jkh@time.cdrom.com
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>> > [1995/06/17] kern/528      slow 386 reports excessive interrupt-level buffer ov
>>
>> Isn't this natural for a slow 386?

>Possibly, but it also happens with a 486DX4/120, 2.1.0R, 16550AFN, kernel ppp,
>296 mtu or 1500 mtu.

Running the vmdaemon process entirely at splhigh() would have caused it.
Interrupts were sometimes masked for 60ms on a P133, but the interrupt-level
sio buffers are "only" large enough for 22ms worth of input at 115200 bps.
This bug lived from 1994/11/25 to 1996/04/11 so it affected all releases
except 2.0.

Overruns will occur if softtty interrupts are blocked or prempted for
longer than 22ms.  This is most likely for slow systems connected to
fast (concurrently) active device(s) through slow polled interfaces.

Bruce