From owner-freebsd-current Sat Sep 9 12:21:05 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id MAA24459 for current-outgoing; Sat, 9 Sep 1995 12:21:05 -0700 Received: from Sysiphos (Sysiphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.6.12/8.6.6) with SMTP id MAA24453 for ; Sat, 9 Sep 1995 12:21:02 -0700 Received: by Sysiphos id AA16534 (5.67b/IDA-1.5 for current@freebsd.org); Sat, 9 Sep 1995 21:20:50 +0200 Message-Id: <199509091920.AA16534@Sysiphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Sat, 9 Sep 1995 21:20:50 +0200 In-Reply-To: Jean-Marc Zucconi "NCR problem: progress" (Sep 9, 2:16) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: Jean-Marc Zucconi Subject: Re: NCR problem: progress Cc: current@freebsd.org Sender: current-owner@freebsd.org Precedence: bulk On Sep 9, 2:16, Jean-Marc Zucconi wrote: } Subject: NCR problem: progress } Hi, } } I have some new data points :-). I tried several combinations and here } is the result: } } 1 streamer + Micropolis2210 OK } 2 streamer + Quantum OK } 3 streamer + Micropolis2210 + Quantum OK } 4 streamer + Micropolis4110 crash } 5 Micropolis4110 alone crash What are the technical characteristics of the 4110 ? Is it possible that it doesn't implement tagged command queues correctly ? } So it seem that I have a problem with the 4110. However, some of the } crashes I had did not involve that disk (id#0). One of the crashes } occured during a backup of id#4 and the 4110 was not mounted. Is it } possible that this disk generate some noise on the bus? Possible ... Maybe the drive has got an internal terminator that is enabled by some jumper ? } I have connected the 4110 to a macintosh (an old SE), formatted it, } and I currently make reads/writes with this disk. It works without any } trouble (of course that kind of machine will not stress the scsi bus...) } I will try to plug the disk to another machine this week-end (I have a } friend running linux on a 486/33+ 1542B). What if the disk only fails } with the ncr driver? If it works with the 1542B and doesn't with the NCR, that doesn't prove much (except that using the 1542B would give you a working system :) The 1542B is limited to 5MHz and doesn't use tagged commands, AFAIK ... You can disable tagged commands for a single target: # ncrcontrol -t 4 -s tags=0 (sets max. 0 tags (i.e. send no TAG message at all) for target 4). Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/staff/esser/esser.html