Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!rutgers!cbmvax!jesup From: jesup@cbmvax.commodore.com (Randell Jesup) Newsgroups: comp.sys.amiga.hardware Subject: Re: More info on A2091 woes Message-ID: <17655@cbmvax.commodore.com> Date: 16 Jan 91 01:53:31 GMT References: <9696@cica.cica.indiana.edu> <1991Jan14.193336.13387@convex.com> <1991Jan15.155835.5230@javelin.es.com> Reply-To: jesup@cbmvax.commodore.com (Randell Jesup) Organization: Commodore, West Chester, PA Lines: 57 In article greg@travis.cica.indiana.edu (Gregory TRAVIS) writes: >blgardne@javelin.es.com (Blaine Gardner) writes: > > -But have you turned off reselection on both drives, as suggested by CBM > -personnel on several occasions? This fixed a similar problem for me on > -the A3000. > >Yes. This alleviates the problem somewhat but is a far cry from 100% >rock-solid. Turning off reselection should totally solve the problem. The problem is caused by the WD controller getting interrupted by a drive doing a reselection while the driver is trying set things up for another command or handling a previous interrupt. The earlier WD chips had a different internal state machine which didn't cause any problems; later chips changed their behavior in this situation (unknown to us). Turning off reselection for ALL drives (and that means resetting where all your partitions are, though if you do it right there's no need to reformat/restore) should make the problem go away totally. Some people have made mistakes when trying to turn off reselection by setting it to NO on the drive definition screen, but not selecting the modified drive type and OK in the drive selection screen (then you go to partition drive and put your partitons back where they were - write it all down first!) Then when that's all done, you select save changes to drive. We're working on a fix (new ROMs), hopefully available in the near future. There has been much discussion of this workaround here on the net in the past. Obviously responses from us here at commodore (on a non- official basis) may include information or suggestions that have not made it out to dealers/repair people yet. Personally, I think the A2091 is a really nice unit, and until I recently switched to an A{model deleted}, I was using an A2091 with 2x200Mb drives, a 50Mb drive, a 40Mb drive, and a SCSI tape unit (most of those are now on the other machine). Worked flawlessly, including doing diskcopies between HD partitions while running compiles, or when validating 4 partitions on each 200 meg drive at the same time (can you say "thrash"?) BTW, Commodore is not the only one that has been bitten by that sort of reselection bug in SCSI drivers. In fact, even people using some NCR SCSI chips have been hit by reselection during a command putting the chip in an undefined state. It seems to be a common problem for SCSI chip designers. (See comp.periphs.scsi.) I'm sorry if you feel you've been given a run-around. I severely doubt it was intentional, and was mostly caused by dealers/repair people being further down the information stream (especially since we only fairly recently (3 months? 6 months? I don't remember) figured out what was causing the problem). We try quite hard to support our users and let them know what's happening, including spending a lot of time giving help on the net and by mail when we're not required to. -- Randell Jesup, Keeper of AmigaDos, Commodore Engineering. {uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.commodore.com BIX: rjesup The compiler runs Like a swift-flowing river I wait in silence. (From "The Zen of Programming") ;-)