Subject: kern/2953: scsi sense problem
To: None <gnats-bugs@gnats.netbsd.org>
From: None <dms@celtech.com>
List: netbsd-bugs
Date: 11/19/1996 17:22:41
>Number:         2953
>Category:       kern
>Synopsis:       scsi sense error
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Nov 19 17:35:00 1996
>Last-Modified:
>Originator:     David M. Stanhope Dms
>Organization:
CTS
>Release:        1.2_ALPHA
>Environment:
i386, netbsd
System: NetBSD bsdpc.celtech.com 1.2 NetBSD 1.2 (NETBSD) #2: Wed Nov 13 01:11:36 PST 1996 dms@netbsd.celtech.com:/usr2/src/sys/arch/i386/compile/NETBSD i386


>Description:
	on older scsi drives the mode-sense-4 won't reurn an error,
        and the scsi probe code will use bogus data in the scsi-sense buffer.
	there used to be a bzero of the sense buffer prior to doing the
	mode-sense command but this was remove in later versions of the
	driver (sd.c).  A test for 'bogus' values of zero now fails since the
	buffer can hold arbitrary data.
>How-To-Repeat:
	try to use a fairly old scsi drive
>Fix:

diff -c sd.c sd.C

*** sd.c        Wed Jun  5 04:41:26 1996
--- sd.C        Fri Jun 21 21:08:50 1996
***************
*** 842,847 ****
--- 842,848 ----
        /*
         * do a "mode sense page 4"
         */
+       bzero(&scsi_sense, sizeof(scsi_sense)); /* DMS */
        bzero(&scsi_cmd, sizeof(scsi_cmd));
        scsi_cmd.opcode = MODE_SENSE;
        scsi_cmd.page = 4;


>Audit-Trail:
>Unformatted: