From owner-freebsd-hackers  Wed Jul 17 16:36:59 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id QAA01116
          for hackers-outgoing; Wed, 17 Jul 1996 16:36:59 -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 SMTP id QAA01111
          for <hackers@freebsd.org>; Wed, 17 Jul 1996 16:36:53 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id JAA03113; Thu, 18 Jul 1996 09:34:18 +1000
Date: Thu, 18 Jul 1996 09:34:18 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199607172334.JAA03113@godzilla.zeta.org.au>
To: dennis@etinc.com, stefan@Promo.DE
Subject: Re: mitsumi CD-ROM
Cc: hackers@freebsd.org
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>If I understand you correctly, you want the disk controller only to =
>consume resources when there is no network activity; I'm not too familiar =
>with PCI's or FBSD's arbitration logic, but I assume that FBSD will try =
>to service disk controllers before network controllers (please correct me =
>on this).

No, the priorities are only parially ordered.  Disk controllers may interrupt
nextwork controllers and vice versa.

>I'm sure one could change that with some effort to give the nic =
>the higher priority.

Tiny effort: `imp_imask |= bio_imask;'

>However, if you're intention is to really make this one router forward =
>packets with lower latency, you might want to try to run it completely =
>off MFS; then there will be no disk activity at all.

There will be, at more-unpredictable times than for a hard disk, when the
MFS fills up.

Bruce