From owner-freebsd-bugs Tue Apr 11 21:40:06 1995 Return-Path: bugs-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id VAA26988 for bugs-outgoing; Tue, 11 Apr 1995 21:40:06 -0700 Received: (from gnats@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id VAA26981 ; Tue, 11 Apr 1995 21:40:04 -0700 Date: Tue, 11 Apr 1995 21:40:04 -0700 Message-Id: <199504120440.VAA26981@freefall.cdrom.com> From: Stephen Hocking Reply-To: Stephen Hocking To: freebsd-bugs Subject: i386/337: Seagate code hangs on bootup. In-Reply-To: Your message of Wed, 12 Apr 1995 14:29:44 +1000 <199504120429.EAA12627@netfl15a.devetir.qld.gov.au> Sender: bugs-owner@FreeBSD.org Precedence: bulk >Number: 337 >Category: i386 >Synopsis: seagate controller code hangs on boot. >Confidential: no >Severity: critical >Priority: medium >Responsible: freebsd-bugs (FreeBSD bugs mailing list) >State: open >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Tue Apr 11 21:40:02 1995 >Originator: Stephen Hocking >Organization: DEVETIR >Release: FreeBSD 2.1.0-Development i386 >Environment: FreeBSD-2.1-current src-cur.0500 Hardware - 25MHz 386sx/386sx, 8Mb, S3911, TMC 885 SCSI controller, WREN 6 harddrive "CDC 94191-15 0136" >Description: On boot up, the machine hangs after identifying the controller, while attempting to attach devices. >How-To-Repeat: Get hold of appropriate hardware and boot. >Fix: I had some success by making a bunch of variables volatile. However this was still unstable and would usually start printing out timeouts shortly after boot, if not during. This often had sad consequences for the integrity of the filesystems. The final fix was to drag the source off the 2.0 CDROM and undefine SEA_BLINDTRANSFER and SEA_ASSEMBLER (at line 107 and 108 of /sys/i386/isa/seagate.c). It is slow, but at the moment it is partway through a "make world" without complaining. This controller is at the moment eating upto 50% of the CPU on interrupts, the rest mostly going on system CPU. Why am I running this pile of canine ordure? My beloved Buslogic 542B finally died after many years of service and I was desperate & cashless. The Buslogic card would get by on 3-5% interrupt CPU, some system time (mostly fs code) leaving plenty of idle CPU for other things. It could drive the disk at platter speed, the seagate only manages about 30% (if that). I do not speak for the Worker's Compensation Board of Queensland - They don't pay me enough for that! >Audit-Trail: >Unformatted: