From owner-freebsd-scsi Sat Oct 9 12:33:59 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from mail.tvol.com (mail.wgate.com [38.219.83.4]) by hub.freebsd.org (Postfix) with ESMTP id 3152314DB5 for ; Sat, 9 Oct 1999 12:33:56 -0700 (PDT) (envelope-from rjesup@wgate.com) Received: from jesup.eng.tvol.net (jesup.eng.tvol.net [10.32.2.26]) by mail.tvol.com (8.8.8/8.8.3) with ESMTP id PAA01219; Sat, 9 Oct 1999 15:29:12 -0400 (EDT) Reply-To: Randell Jesup To: Gerard Roudier Cc: "Matthew N. Dodd" , scsi@FreeBSD.ORG Subject: Re: Driver for GDT6517RD RAID controller References: From: Randell Jesup Date: 09 Oct 1999 15:29:10 +0000 In-Reply-To: Gerard Roudier's message of "Sat, 9 Oct 1999 14:02:25 +0200 (MET DST)" Message-ID: X-Mailer: Gnus v5.6.43/Emacs 20.4 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Gerard Roudier writes: >On Fri, 8 Oct 1999, Matthew N. Dodd wrote: >> I suspect that all but the very early 53c700 and 710 chips would require >> minimal changes. > >The 53c700 must be considered as having never existed. ;) Quite true. :-) >About the 710 and later 7xx chips, speaking of a driver for this family >does not make sense, since the system BUS stuff allows and requires >proprietary implementations from the system hardware. Since it seems that >machines that use on-board 7xx controllers are old, the needed >documentation may just be no longer available. Quite true; I know the c710 was used on about 5 or 6 different Amiga SCSI boards (including the ones I wrote the software for), but I don't know how much it was used in the PC world, and I don't know if anyone would care. >A driver that will separate system BUS code and SCSI core code may allow >to support some set of 53c7xx based controllers. The 53c7xx support looks >like that under Linux, for example. There's support for c710's in NetBSD as well (at least in the Amiga port). I haven't looked at it. If it's based on the NCR reference driver examples, it can't perform all that well. Their reference drivers for PC's didn't come close to using all the possible performance/power in the c710, especially the ability to execute (really simplistic) code. >Basically, a SIM that supports 53c810 rev 1 can support 53c710 and later >53c7xx chips with minimal changes, provided that the BUS stuff be added or >changed. The stock ncr driver supports the 53c810 rev 1, so starting from >this code should be just fine for some 710 support. Sounds exactly as I'd suspect. >Now, my 0.02 euro opinion about this topic: >I am always been amazed by people who wants latest softwares to support >old hardwares. If such compatibilty constraints had been always applied, >may-be the fastest hardware nowadays would be a Z80 at 5 GigaHz and >graphic would be some incredibly fast 16 colors VGA. ;-) :-) There are some good reasons: a LOT of us over here have older (once-powerful) machines and boards hanging around, and we'd like to get some use out of them. Also, machines used (or donated to) charities and the like are often old or very old, and they want to be able to use them to get something good done without having to buy new $$$ machines every year or two. One doesn't _need_ a PIII/450 w/ 64MB, 20GB disk and a SCSI ultra-wide controller to send email/browse/serve pages (ok, maybe with Microsoft/IE/etc, but not with BSD). Others use them for NAT's, or dedicated firewall/router/proxies, etc. Again, massive horsepower isn't needed. Sometimes it's just for amusement. I have a computer museum in my basement with multiple Sun 2/120's (68010@12MHz), a DEC PDP 11/23, TRS80 model 1, Mac+, Sage, Amigas of all sorts (up to 50MHz 68040 A4000T), Apple II, Pet's, C64's, C128's, etc. >IMO, the 53c7xx parts are now material for history books and IT museums, >and an O/S that does not provide support yet for these chips should not >consider adding such a support, unless some driver magically be provided >by some volonteer and be actually _maintained_. Sure. In this case, if someone had docs on the bus interface for a 710-based card, and people had some real use for a driver for them (I don't know if that's true), it probably wouldn't be hard to modify an '810 driver to support it. I didn't have any real plans to do so when I joined this group, but Matt Dodd asked me to look into how easy it would be, and that's not a bad way to get a handle on how SCSI drivers under FreeBSD work. My main interests lie more in performance, support for weird-ass SCSI devices (I learned a LOT being the primary SCSI person at Commodore, and also in qualifying drives for production), and finding out if FreeBSD supports Section 10 of CAM (target mode), since I was one of the people who wrote that section. If it doesn't, I may look at adding support to (a) SIM for that. I had some of the (NCR) code written to do so for the 710; that code might be applicable, or serve as a guideline. Admittedly, my specialized knowledge of the 33c93(b) and it's myriad bugs and gotcha's probably isn't greatly useful.... ;-) -- Randell Jesup, Worldgate Communications, ex-Scala, ex-Amiga OS team ('88-94) rjesup@wgate.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message