From owner-freebsd-hackers Sat Oct 7 12:30:04 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id MAA16462 for hackers-outgoing; Sat, 7 Oct 1995 12:30:04 -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 MAA16451 for ; Sat, 7 Oct 1995 12:29:47 -0700 Received: by Sysiphos id AA11276 (5.67b/IDA-1.5 for hackers@freebsd.org); Sat, 7 Oct 1995 20:29:18 +0100 Message-Id: <199510071929.AA11276@Sysiphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Sat, 7 Oct 1995 20:29:18 +0100 In-Reply-To: -Vince- "machine reboot & kernel maxusers option" (Oct 7, 4:45) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: -Vince- Subject: Re: machine reboot & kernel maxusers option Cc: hackers@freebsd.org Sender: owner-hackers@freebsd.org Precedence: bulk On Oct 7, 4:45, -Vince- wrote: } Subject: machine reboot & kernel maxusers option } Hi everyone, } } I am experiencing the following problem and the machine just } reboots under FreeBSD-current, anyone have any ideas? There have been no functional changes to the driver for nearly a month now ... Did you rebuild your kernel within that time ? } MAND FAILED (4 28) @f0b4aa00. } assertion "cp" failed: file "../../pci/ncr.c", line 5560 } sd0(ncr0:0:0): COMMAND FAILED (4 28) @f0b4aa00. Hmm, the 28 in the message above means "Queue Full", i.e. there were too many commands issued by the generic SCSI driver. But the maximum number of commands to issue is a parameter known to the generic SCSI code, so this should never happen ... In fact, there was a change one month ago, which did change the size of the command queue. But in fact, the queue size was increased ... You may want to try changing back the line (ncr.c:132): #define MAX_START (MAX_TARGET + 7 * SCSI_NCR_MAX_TAGS) into its previous form: #define MAX_START (7 * SCSI_NCR_MAX_TAGS) I'll look into this, but I need more information: 1) Complete verbose boot message log, if possible (I.e. Boot with "-v" option, send "/sbin/dmesg" output). 2) When does this happen ? At boot time, or when the system had been heavily loaded ? Which devices are connected ? active ? There is a timeout function in the driver, which should reset the NCR chip after some 10 seconds of no progress on any outstanding command. This ought to clear the error condition ... } Also, can someone tell me what the maxusers option in the kernel } config file does exactly? Thanks! Look for uses of MAXUSERS in "/sys/conf/param.c". 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