Date: Tue, 9 Jan 1996 19:51:09 +0100 From: se@zpr.uni-koeln.de (Stefan Esser) To: kowa@sunbayer45.informatik.tu-muenchen.de (Wolfgang Kowarschick) Cc: freebsd-scsi@freebsd.org Subject: Re: scsi problems (Repost) Message-ID: <199601091851.AA13775@Sysiphos> In-Reply-To: kowa@sunbayer45.informatik.tu-muenchen.de (Wolfgang Kowarschick) "Re: scsi problems (Repost)" (Jan 8, 14:12)
next in thread | raw e-mail | index | archive | help
On Jan 8, 14:12, Wolfgang Kowarschick wrote: } Subject: Re: scsi problems (Repost) } I'm running FreeBSD 2.1.0 on an Asus P55TP4XE Board with } an Asus PCI-SC200 controller (NCR 53c810). I've three internal } devices (Quantum Atlas 2GB, NEC CD-Rom 512, SQ 555 (44MB)) } and one external device (HP Dat Streamer). } The termination of the scsi-bus seems to be correct } (I'm not quite sure with the disk as I've no description } of the jumper settings, but I think it's ok). } } kernel output (copied by hand, so there may be some typos): } } /kernel: st0(ncr0:3:0): NOT READY asc:4,1 } logical unit is in process of becomming ready } phase change 2-7 10@004fc558 resid=9 } ncr0:0: ERROR(81:0) (f-aa-0) (8/13) @ (eq8:48000000) } script cond = c0000008 The 81 dstat indicates an illegal instruction. There are two possible reasons: 1) The current instruction is not defined at all. 2) The current instruction can't be executed in the current context (e.g. state of the SCSI bus). In your case 1) seems to apply. There is no instruction starting with 48. This indicates a memory problem, IMHO. Guess that the DRAM speed is marginally sufficient, but sometimes just a little to low ... If this can be reproduced, then please send more such logs. (You mentioned the possibility of typos: Please make sure those numbers are excactly what was written in the error log. At least ALL numbers in the two lines quoted above MUST match reality, or I'm at a loss :-) } vnode_papger_input: I/O read error } vm_fault: pager input (problably hardware) error, PID 310 failes } Note that the vm_fault error does also occur sometimes when the dat-streamer } is switched off. Well, then just DON't do it ... SCSI is not meant to be hot-pluggable! And switching off devices may cause spikes on the bus lines. 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/~se <se@ZPR.Uni-Koeln.DE>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601091851.AA13775>