From owner-freebsd-stable Wed May 3 14:47:48 2000 Delivered-To: freebsd-stable@freebsd.org Received: from mercury.campbell-mithun.com (Mercury.campbell-mithun.com [192.159.32.151]) by hub.freebsd.org (Postfix) with ESMTP id 7847B37BE6A; Wed, 3 May 2000 14:47:44 -0700 (PDT) (envelope-from swb@grasslake.net) Received: from marlowe (Marlowe.campbell-mithun.com [192.159.32.184]) by mercury.campbell-mithun.com (8.9.1a/8.9.1) with SMTP id PAA01218; Wed, 3 May 2000 15:52:20 -0500 Message-ID: <00fe01bfb549$419272e0$b8209fc0@marlowe> From: "Shawn Barnhart" To: "Mike Smith" Cc: References: <200005031949.MAA00506@mass.cdrom.com> Subject: Re: HPDA/DAC960PL errors Date: Wed, 3 May 2000 16:48:01 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Content-Transfer-Encoding: 7bit Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG ----- Original Message ----- From: "Mike Smith" | None taken. Unfortunately the 3.x firmware is probably going to have the | same problem; we cap at 16 S/G segments there as well. Then I won't jump up right now an order the 3.x firmware. In other words, DAC960 firmware 2.x and 3.x on x86 is broken, period. (Again, just checking facts). I'm going to set the box up on a single SCSI drive for now. If you want to make patches available to me when you get them done, I'd do some testing for you with both the 2.73 and the 3.x firmware. | I know what's going on there - I don't support I/O mapped controllers at | all. I'm not sure that I want to, either, yet; there is a performance | cost that extends to other controllers if I do. If you can narrow down | the differences between the controllers that come up I/O mapped and those | that come up memory-mapped, I'd really appreciate it. It's possible that | the I/O mapped ones can be convinced to memory-map if I can work out | what's wrong. Mylex claims that any card with a Mfg. ID of D040-347 or later can get the firmware upgrade chips. The card I have at home is a DAC960PL but is D040-345 -- no firmware upgrade available, even though it sports two flash sockets just like the 960PL I have at work which is D040-354 which *can* get the flash upgrade. The two cards look pretty much identical, although the older one has SCSI controllers on it I don't recognize and the newer one has Symbios Logic controllers. I'm not Sherlock Holmes, but I'd wager that cards -347 and after the have some difference in controller logic that enables memory mapping, and support for memory mapping is manditory in the 3.x firmware. Whether this has anything to do with which SCSI chips get used or not is beyond my knowledge level, but if I was writing a README as to which card the driver supported I wouldn't feel bad about using this logic. It would certainly account for why a -345 card won't memory map and can't get a flash chip upgrade and -347s and above can, even though the -345 sports two flash sockets. I have one other DAC960, but it's an even older, DAC960E-2 card which as I look at it is an EISA card, not PCI. It sports Symbios SCSI chips, but it also sports 50 PIN connectors. I can check this one to see if it does indeed fail to memory map (does the driver even support EISA cards?), but I'd bet it won't memory map as its series -312. It doesn't support flash at all, looks like the firmware is on EEPROM. Anyway, now I know more about Mylex cards than I thought I ever would. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message