Path: gmd.de!Germany.EU.net!netmbx.de!zrz.TU-Berlin.DE!zib-berlin.de! xlink.net!howland.reston.ans.net!vixen.cso.uiuc.edu!gateway From: do...@BSDI.COM (Donn Seeley) Newsgroups: info.bsdi.users Subject: Adaptec AHA-2742/2842 news Date: 20 Jan 94 19:44:41 GMT Organization: University of Illinois at Urbana Lines: 72 Approved: Use...@ux1.cso.uiuc.edu Message-ID: <9401201944.AA06391@BSDI.COM> NNTP-Posting-Host: ux1.cso.uiuc.edu Originator: dae...@ux1.cso.uiuc.edu [My apologies for the mangled headers in the previous messages... I guess I'm a bit slow to wake up this morning!] Adaptec has finally provided BSDI with an official response to our request for programming information on their new SCSI parts. We received a document titled 'Programming Directly to AIC-7770-Based Adaptec Products: What Developers Need to Know.' The document mentions that the AHA-174x products will be phased out by June '94, and that new products will be based on the AIC-7770 chip. It then goes on to discuss the AIC-7770 interface. Basically, the AIC-7770 is a 'RISC-Based EISA-to-Fast-SCSI host adapter on a chip.' It is programmable, and in general is only useful if it is running an Adaptec-supplied microcode program. The actual firmware, and the interface to the firmware, will be provided to developers only under non-disclosure. Adaptec will supply source code for downloading the AIC-7770 and driving the interface; this source code is called the HIM or 'Hardware Interface Module'. The HIM source code cannot be redistributed, and developers may not modify the source code except to combine it with their software. Adaptec retains title to the source and object code and grants permission to use it on a 1-year license, which is automatically renewed unless unstated problems arise. All critical documentation about the AIC-7770 interface is 'Adaptec confidential' and is covered by non-disclosure. Adaptec provides an 'HIM Developer Request Form' which requires you to describe the nature of your need for the documentation and firmware, and state who your important customers are and how many units you will ship. If you send in the form and Adaptec accepts it, you get a packet in the mail which has a 'break-the-seal' envelope with a license agreement printed on the outside. (Hope I have summarized this correctly...) Given this information from Adaptec, I see three approaches to developing drivers for the AHA-2742/2842VL adapters. (1) BSDI could try to write its own HIM equivalent, using only freely available documentation such as the AIC-7770 Data Book. The on-chip microcode store is fairly small, holding 448 RISC instructions; this might mean that the microcode normally holds a library of channel programs which manipulate the SCSI state machine for particular operations. To write microcode, we would have to duplicate Adaptec's development efforts in the dark, starting by writing an AIC-7770 micro-assembler, coming up with a set of programs, and going through a lot of work to debug the programs. The resulting driver would be redistributable as source code, however. Just to further discourage folks who might want to try this, the Adaptec material mentions that Rev E of the AIC-7770 will change the HIM, although old versions of the HIM will still work without taking advantage of improved performance. (2) We could sign the non-disclosure and write a host adapter driver which does the work of the HIM without actually including it. The resulting driver might or might not be redistributable in source code form, depending on details of Adaptec's license. If it would indeed be redistributable, then this alternative would be much faster than (1). Keeping up to date with Adaptec's changes could still be a pain, however. (3) BSDI can sign the non-disclosure and not distribute source for the 2742/2842 host adapter driver. We would still ship source code for machine-independent drivers and other host adapter drivers, just not for new Adaptec products. This sucks, but it's better than nothing, and would get a driver to our customers faster than any other alternative. (4) Stop supporting new Adaptec products. I think customers will find this unacceptable, although I'm open to comments. As someone who believes in shipping source code for everything for a reasonable price, I find Adaptec's attitude offensive and short-sighted, but there's probably nothing BSDI can do to change it. If you have comments about any of the alternatives I've listed, send them to me and I will summarize. I really, really hate object-only drivers, Donn
Path: gmd.de!Germany.EU.net!EU.net!ub4b!news.cs.kuleuven.ac.be!blekul11! news.ysu.edu!usenet.ins.cwru.edu!howland.reston.ans.net!vixen.cso.uiuc.edu! gateway From: MDIC...@csi.compuserve.com Newsgroups: info.bsdi.users Subject: Re: Adaptec AHA-2742/2842 news Date: 21 Jan 94 16:28:30 GMT Organization: University of Illinois at Urbana Lines: 22 Approved: Use...@ux1.cso.uiuc.edu Message-ID: <CSI_5991-5377@CompuServe.COM> NNTP-Posting-Host: ux1.cso.uiuc.edu Originator: dae...@ux1.cso.uiuc.edu Donn, There are certainly other high performance cards available. We've been using the Buslogic 747s for a while and I'd planned on doing an enhanced mode driver based on the current eaha code and the copious drive doco I got from Buslogic for free after a phone call. Personally I have a hard time rewarding vendors who treat driver doco as proprietary. We do a fair amount of custom hardware here at CompuServe and the ability to do occasional custom driver work is important to us. I think the real goal is to have a high performance, supported SCSI card with drivers available. Both the Buslogic cards and also competing cards from Ultrastor come to mind. Having sources is important. I may want to look at doing some target mode code in the future an an object only driver is no help at all. Just my 2 cents worth... Mike Dickson CompuServe, Incorporated
Path: gmd.de!Germany.EU.net!EU.net!howland.reston.ans.net!vixen.cso.uiuc.edu! gateway From: bd...@gag.com (Bdale Garbee) Newsgroups: info.bsdi.users Subject: Re: Adaptec AHA-2742/2842 news Date: 21 Jan 94 23:47:31 GMT Organization: University of Illinois at Urbana Lines: 24 Approved: Use...@ux1.cso.uiuc.edu Message-ID: <199401212347.QAA12389@winfree.gag.com> NNTP-Posting-Host: ux1.cso.uiuc.edu Originator: dae...@ux1.cso.uiuc.edu In article <CSI_59...@CompuServe.COM> you wrote: : There are certainly other high performance cards available. We've been : using the Buslogic 747s for a while... Anyone have a list of Buslogic's products and what they do? I've been an Adaptec user (3 1542B's and a 1740A at home) for a couple of years, but will gladly switch to an alternative if driver information is more readily available. A phone number or a pointer to a source of information would be fine. I've seen the cards in ads from time to time, but while I can rattle off Adaptec model numbers and what they do, I have no equivalent in my head for Buslogic. : I think the real goal is to have a high performance, supported SCSI : card with drivers available. Yes. I'd really enjoy having fast-wide-differential support, but that's less important to me than having good ISA/EISA interfaces to 8-bit single-ended hardware... with driver source. So far, the only thing I've been willing to run without complete source on my BSDI system(s) are Xircom PE-2's that are owned by HP... I won't spend personal money on hardware I can't have driver source for. Bdale