From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from cthulhu.engr.sgi.com (cthulhu.engr.sgi.com [192.26.80.2]) by neteng.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id PAA78371 for ; Fri, 26 Mar 1999 15:50:54 -0800 (PST) Return-Path: Received: (from majordomo-owner@localhost) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id PAA02003 for linux-list; Fri, 26 Mar 1999 15:49:36 -0800 (PST) mail_from (owner-linux@relay.engr.sgi.com) Received: from fir.engr.sgi.com (fir.engr.sgi.com [150.166.40.90]) by cthulhu.engr.sgi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) via SMTP id PAA52501; Fri, 26 Mar 1999 15:49:33 -0800 (PST) mail_from (wje@fir.engr.sgi.com) Received: (from wje@localhost) by fir.engr.sgi.com (950413.SGI.8.6.12/950213.SGI.AUTOCF) id PAA26663; Fri, 26 Mar 1999 15:49:32 -0800 Date: Fri, 26 Mar 1999 15:49:32 -0800 Message-Id: <199903262349.PAA26663@fir.engr.sgi.com> From: "William J. Earl" To: Thomas Bogendoerfer Cc: linux@cthulhu.engr.sgi.com Subject: Re: Help needed to solve SCSI problem In-Reply-To: <19990327002321.A3539@alpha.franken.de> References: <19990327002321.A3539@alpha.franken.de> Sender: owner-linux@cthulhu.engr.sgi.com Precedence: bulk Thomas Bogendoerfer writes: > I'm now total stuck with the remaining SCSI problem, and need some help. > > I have a really broken cdrom drive. With this cdrom I'm able produce a > scsi timeout, which will cause a hang of the complete machine. I've > traced the problem a little bit further, and detected, that I get > interrupts on interrupt vector 0 (it's named VECTOR_GIO0 in the Irix > header files). I always see 17 irq 0 then the machine hangs completly. > > Please, could someone tell what this interrupt means, so I can either > solve it by servicing the interrupt the right way or avoid it by doing > something else ? I believe that is the two bits: #define EXTIO_HPC3_BUSERR 0x0010 #define EXTIO_SG_STAT_0 0x0001 HPC3_BUSERR means that the HPC3 got a bus error reading from memory. The driver is probably broken and giving bad commands to the HPC3. You should be able to find the error address in the GIO_ERR_ADDR register: #define GIO_ERR_ADDR 0x1fa000f4 /* GIO error address (w, ro) */ (0xbfa000f4 as a K1SEG address).