Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!mailrus!purdue!decwrl!shelby!neon!kaufman From: kaufman@Neon.Stanford.EDU (Marc T. Kaufman) Newsgroups: comp.sys.mac.hardware Subject: Re: Sharing the SCSI bus? Message-ID: <1989Nov28.041855.290@Neon.Stanford.EDU> Date: 28 Nov 89 04:18:55 GMT References: <1240@key.COM> <602@telesoft.com> Sender: USENET News System Reply-To: kaufman@Neon.Stanford.EDU (Marc T. Kaufman) Organization: Computer Science Department, Stanford University Lines: 25 In article time@oxtrap.UUCP writes: >In article <602@telesoft.com> mclow@telesoft.com (Marshall Clow @telesoft) writes: - It won't work. As you surmised, the Mac's SCSI manager wants to be the - only "initiator" on the bus. If you want to get technical, - _SCSISelect ( or whatever the trap is called ) will fail if the Mac - cannot get control of the bus. ( i.e, if two other devices are - communicating ). -Sorry, Marshall, but it will work. In fact it is done all the time! -Talk to anyone with an Ensoniq EPS. -Yes, _SCSISelect() can fail, but most programs deal with this. -Further, other initiators on the bus does not necessarily mean that -SCSISelect() will fail. The probability of a timeout on select -increases, but this does not translate into failure. In fact, the -opposite tends to be true, the Select almost never fails. The problem is not that SCSISelect will or will not fail, it is that SCSISelect does not go through the bus arbitration phase required for multi-master SCSI implementation. It is possible for SCSISelect to succeed when, in fact, it has no right to the bus. It may work, most of the time, but I would not want to count on it. Marc Kaufman (kaufman@Neon.stanford.edu)