From owner-freebsd-smp Thu Jan 9 15:53:07 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.4/8.8.4) id PAA29087 for smp-outgoing; Thu, 9 Jan 1997 15:53:07 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.8.4/8.8.4) with SMTP id PAA29082 for ; Thu, 9 Jan 1997 15:53:04 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id QAA18917; Thu, 9 Jan 1997 16:42:36 -0700 From: Terry Lambert Message-Id: <199701092342.QAA18917@phaeton.artisoft.com> Subject: Re: unusual question To: phk@critter.dk.tfs.com (Poul-Henning Kamp) Date: Thu, 9 Jan 1997 16:42:36 -0700 (MST) Cc: mef@cs.washington.edu, smp@freebsd.org In-Reply-To: <7514.852822663@critter.dk.tfs.com> from "Poul-Henning Kamp" at Jan 9, 97 04:11:03 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > Generally, you can allow device-drivers to be multi-threaded only as > far as you have separate pieces of hardware. Ie, in you average > FreeBSD system you could have one thread/cpu per sio port, but only > one for wdc0+wd0+wd1 because they share registers. > > I would expect the "one-lock per device-driver" to be implemented > in the code that calls into the device drivers, simply to avoid > changing all of them, so most device-drivers will probably not > change much if at all. Initially, anyway. Ideally, there would be top/bottom based drivers to implement code for running above and below a HAL; this would ensure maximum concurrency (mostly a problem for SCSI controllers; most other things won't care, but it's the principle of the thing... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.