Path: utzoo!yunexus!ists!jarvis.csri.toronto.edu!mailrus!shadooby!samsung!gem.mps.ohio-state.edu!rpi!uwm.edu!genbank!agate!shelby!neon!kaufman From: kaufman@Neon.Stanford.EDU (Marc T. Kaufman) Newsgroups: comp.sys.mac.hardware Subject: Re: Sharing the SCSI bus? Keywords: SCSI, Mac Message-ID: <1989Nov18.162549.10337@Neon.Stanford.EDU> Date: 18 Nov 89 16:25:49 GMT Article-I.D.: Neon.1989Nov18.162549.10337 References: <1240@key.COM> Sender: USENET News System Reply-To: kaufman@Neon.Stanford.EDU (Marc T. Kaufman) Organization: Computer Science Department, Stanford University Lines: 62 In article <1240@key.COM> perry@arkon.key.COM (Perry The Cynic) writes: >On a conventional Mac SCSI bus, the Mac is device #7, and the only one that >initiates communications (talker). All the other devices just listen and >(we hope) obey. I am told that the Mac's SCSI manager is borderline brain-dead >in its restrictions. I haven't looked at it myself (yet); I'm hoping somebody >who has already suffered through that can enlighten me. I have looked through (and rewritten) parts of the SCSI manager. In principle, any SCSI device can act as Bus Master ("Initiator" in SCSI parlance). However, for this to work, all Initiator devices must use a version of the SCSI protocol that allows it. In particular, each Initiator must first go through a "Bus Arbitration" phase. This insures that no one else is on the bus, and that no one else has (successfully) bid for the bus at the same time. Apple's SCSI code DOES NOT perform bus arbitration before seizing the bus. >Let's say I have a CD-ROM reader with a SCSI interface... > ...(We could try to do this with >two Macs, if Apple hadn't hardwired the Mac's SCSI controller as device #7. The "hardwiring" is in the Rom Code. There is no "wiring" as such. However, the ROM code (and patches I have seen) "knows" that device #7 will always win a selection battle, and "knows" that the Mac is device #7, so, shall we say, "takes some shortcuts". >First of all, I'm given to wonder how to terminate this thing. Assuming >both the Mac and PC interfaces are terminated, I could put them at the ends >of the SCSI bus and line up any peripherals in between (without terminators, >of course). Should that work? Does it? Yes. That is how you should do it. >As long as either the PC or the Mac is off, I guess I have a reasonable >chance for the "living" computer to control the devices on the SCSI bus. >Yes, I know file formats are different, and a Mac-formatted disk drive >won't work for the PC and vice versa, and the partition tables are probably >incompatible too. But a CD-ROM drive should work for either (they sell the >same drive with software kits for both Macs and PCs). Another candidate >could be a SyQuest 40MB removable drive. >The interesting question, of course, is whether it'll work if I turn on both >computers. Assuming I use the CD-ROM drive only from one host at a time, >what are my chances? Slim to none, statistically. No problem electrically. > Will the Mac's boot-time scan of the bus confuse it >if it sees a PC host card there? Probably not, since I assume the PC will not respond as a Target Device either. > Can the Mac's SCSI chip and manager (esp. >the manager!) cope with a second talker on the bus? I.e., can the PC card >talk to the CD-ROM player while the Mac talks to a harddrive? Am I expecting >much too much of the poor weak-brained SCSI manager? Yes you are. Why do you think the PC card is any better, BTW? >Perry The Cynic (Peter Kiehtreiber) perry@arkon.key.com Marc Kaufman (kaufman@Neon.stanford.edu)