From owner-freebsd-hackers Fri Jun 6 12:37:04 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id MAA06758 for hackers-outgoing; Fri, 6 Jun 1997 12:37:04 -0700 (PDT) Received: from pluto.plutotech.com (root@pluto100.plutotech.com [206.168.67.137]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id MAA06753 for ; Fri, 6 Jun 1997 12:37:01 -0700 (PDT) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by pluto.plutotech.com (8.8.5/8.8.5) with ESMTP id NAA00243; Fri, 6 Jun 1997 13:36:26 -0600 (MDT) Message-Id: <199706061936.NAA00243@pluto.plutotech.com> X-Mailer: exmh version 2.0beta 12/23/96 To: Bob Bishop cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Extremely poor interactive response under heave SCSI load In-reply-to: Your message of "Fri, 06 Jun 1997 11:38:08 BST." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 06 Jun 1997 14:34:51 -0600 From: "Justin T. Gibbs" Sender: owner-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk >Quick (and nasty) hack: in scsi/sd.c, if the device has been continuously >busy for say > 1sec, just stop feeding it until it has drained. > >Better hack: if the oldest request on the device has been there for say > >1sec, stop feeding until that request has completed. > >None of this would be necessary if sequencers were a bit less totalitarian... Even better hack. Just send an ordered transaction every once in a while which will flush any "starved" simple-tagged transactions. This means that you can always keep the queue full. >-- >Bob Bishop (0118) 977 4017 international code +44 118 >rb@gid.co.uk fax (0118) 989 4254 between 0800 and 1800 UK > > > -- Justin T. Gibbs =========================================== FreeBSD: Turning PCs into workstations ===========================================