From owner-freebsd-hackers Sun Jan 21 06:30:42 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id GAA17859 for hackers-outgoing; Sun, 21 Jan 1996 06:30:42 -0800 (PST) Received: from hda.com (hda.com [199.232.40.182]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id GAA17854 for ; Sun, 21 Jan 1996 06:30:39 -0800 (PST) Received: (from dufault@localhost) by hda.com (8.6.11/8.6.9) id JAA06531; Sun, 21 Jan 1996 09:29:05 -0500 From: Peter Dufault Message-Id: <199601211429.JAA06531@hda.com> Subject: Re: Whew!!!!!!! (MAJOR sigh of relief!) To: bde@zeta.org.au (Bruce Evans) Date: Sun, 21 Jan 1996 09:29:04 -0500 (EST) Cc: davidg@Root.COM, peter@jhome.DIALix.COM, hackers@FreeBSD.org In-Reply-To: <199601211118.WAA07471@godzilla.zeta.org.au> from "Bruce Evans" at Jan 21, 96 10:18:09 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-hackers@FreeBSD.org Precedence: bulk > > > I think you can make a much more general statement: "With few exceptions, > >all of our SCSI drivers handle timeouts and aborts very badly." > > I think you can make a much more general statement: "With few exceptions, > all of our drivers handle abnormal time-dependent things and abnormal errors > very badly." Sigh. How about "with few exceptions, all software handles abnormal errors very badly". (HDA once had a problem with a client when we insisted that a project should have no errors - all situations should be fully tested, the behavior well defined and part of the acceptance test and all other situations would freeze their machine - sort of a panic and core dump. This was interpreted by some as HD refusing to handle errors) I plan on plugging into the SCSI code with Justin shortly. I am finishing up a new rev of the GPIB driver that supports different boards and modes, and check in and write man pages for three new Data Acq drivers. After that I plan on moving policy up out of the SCSI board drivers and into the common code, hopefully reduce some of the common code in the drivers, and adding common SCSI resource scheduling. I have the 1542 and now the NCR covered (though I'm sure Stefan will help with that) and Justin has the AHC. It will be nice if some others can step forward to pick up the other boards. Given the time we have to spend on this I hope this can be done around May 1. If things change one way or the other I'll let you know. Peter -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267