Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod.mps.ohio-state.edu!usc!apple!portal!cup.portal.com!ts From: ts@cup.portal.com (Tim W Smith) Newsgroups: sci.electronics Subject: Re: SASI = SCSI? Message-ID: <39010@cup.portal.com> Date: 7 Feb 91 12:02:42 GMT References: <86715@unix.cis.pitt.edu> <1991Feb2.222821.12163@zoo.toronto.edu> Organization: The Portal System (TM) Lines: 34 >>SASI is SCSI's grizzled old ancestor. They are nominally still interoperable, >>last I heard. > >Uh, not really. In SASI-land, all you had to do in order to access >the remote device was to assert its SASI ID during the selection phase. >In SCSI-land, you need to assert both the controller's SCSI ID and the >remote device's SCSI ID on the bus during the selection phase. You have fallen for one of the classic blunders. The most famous, of course, is never to get involved in a land war in Asia, but almost as famous is this: never disagree with Henry Spencer on Usenet on factual matters where computers are involved. See section 5.1.3.4 of Revision 17B of the SCSI specification. In single initiator systems where the initiator does not allow the RESELECTION phase, the inititator only needs to assert the target devices's ID bit on the bus during selection. Don't feel bad, though. There are in fact some SCSI chips that will not allow themselves to be selected unless two bits are asserted, so one would have to be rather unconcerned about interoperability to actually take advantage of this. Tim Smith ps: don't ask which chips I am talking about, because I don't remember. I remember that some peripheral firmware I wrote did not seem to respond to selection, and I tracked the problem down to an initiator that was only asserting one bit (due to a broken wire in the #@$%^& cable), and so the selection was being ignored by the SCSI chip in the target. Thus, I remember that there is at least one chip out there that will not recognize such a selection, but I don't remember which one it is.