From owner-freebsd-scsi Wed Nov 13 14:19:40 1996 Return-Path: owner-freebsd-scsi Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id OAA21063 for freebsd-scsi-outgoing; Wed, 13 Nov 1996 14:19:40 -0800 (PST) Received: from sovcom.kiae.su (sovcom.kiae.su [193.125.152.1]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id OAA21057; Wed, 13 Nov 1996 14:19:28 -0800 (PST) Received: by sovcom.kiae.su id AA27514 (5.65.kiae-1 ); Thu, 14 Nov 1996 01:14:47 +0300 Received: by sovcom.KIAE.su (UUMAIL/2.0); Thu, 14 Nov 96 01:14:46 +0300 Received: (from ache@localhost) by nagual.ru (8.8.2/8.8.2) id BAA00212; Thu, 14 Nov 1996 01:13:25 +0300 (MSK) Message-Id: <199611132213.BAA00212@nagual.ru> Subject: Re: SCB paging is most dangerous option now! In-Reply-To: <199611131606.IAA25507@freefall.freebsd.org> from "Justin T. Gibbs" at "Nov 13, 96 08:06:10 am" To: gibbs@freefall.freebsd.org (Justin T. Gibbs) Date: Thu, 14 Nov 1996 01:13:25 +0300 (MSK) Cc: current@FreeBSD.org, scsi@FreeBSD.org From: "=?KOI8-R?Q?=E1=CE=C4=D2=C5=CA_=FE=C5=D2=CE=CF=D7?=" (Andrey A. Chernov) Organization: self X-Class: Fast X-Mailer: ELM [version 2.4ME+ PL28 (25)] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > >I read only mode I got the almost same result with SCB paging as without it. > >This bug affects _writing_only_, not reading. > >And I can't start 8 dds for each drive in both modes, > >3 dds per drive is enough to cause "Queue Full" in both modes, but > >it is harmless due to read only. > > This is probably what is causing your corruption in the SCB Paging case. > Which drive can only do 3 transactions at a time? I'm sure that your IBM > can do more. It may very well be that one of your drives doesn't like the > fact that the code with SCB paging turned on will basically continually > pound on that drive, causing queue full conditions, until there aren't > enough transactions queued to make this happen any more. The patch I > posted for Andreas' problem may help in this regard. No, _any_ of two drives says "Queue full" after 2-3 dds and goes to "timed out" state after more dds which ends up with "panic: Timed-out command times out again" It happens equally in both SCB paging and not SCB paging modes. BTW, I see it only for dds, I can start f.e. about 20 "ls -lR" safely. Does it means that Tagged Queueing not work well with 2842? > As for your problem happening on the "first write", are you sure this isn't > the "first sync" after boot?? I mean first physical write, of course. It happens just after fsck puts info lines about each drive at boot phase. -- Andrey A. Chernov http://www.nagual.ru/~ache/