NetBSD Problem Report #22160

Received: (qmail 27478 invoked by uid 605); 16 Jul 2003 17:45:02 -0000
Message-Id: <20030716174457.1258A11152@narn.netbsd.org>
Date: Wed, 16 Jul 2003 17:44:57 +0000 (UTC)
From: dundas@caltech.edu
Sender: gnats-bugs-owner@NetBSD.org
Reply-To: dundas@caltech.edu
To: gnats-bugs@gnats.netbsd.org
Subject: panic using SCSI TK50 in NetBSD 1.6.1
X-Send-Pr-Version: www-1.0

>Number:         22160
>Category:       port-alpha
>Synopsis:       panic using SCSI TK50 in NetBSD 1.6.1
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-alpha-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Jul 16 17:46:01 +0000 2003
>Closed-Date:    
>Last-Modified:  Sat Dec 31 23:12:42 +0000 2022
>Originator:     John Dundas
>Release:        1.6.1
>Organization:
>Environment:
TBD
>Description:
I have an AS200 (4/233) with a TK50Z-GA attached.  The boot messages from dmesg can be viewed here:

<http://www.its.caltech.edu/~dundas/alpha/dmesg.txt>

First, it appears that the TK50 is not recognized other than as a generic SCSI tape drive (st0).  I suppose that's fine, though I wonder if it is the cause of the crash I am experiencing (to be described).  In googling
through some of the old netbsd archives, I see some patches specifically
aimed at this drive (though they were in the VAX port).  However none of
this code is in the current st driver.  Should/could it be there?

Well, what I'm trying to accomplish is to make .TAP archives of TK50 tapes. I have written a small program to read the tape and write an appropriately formatted disk file.  The program may be viewed here:

<http://www.its.caltech.edu/~dundas/alpha/tapin.c>

The program has generally worked just fine.  However I have encountered a
system crash (and thus the motivation for the email) with, I believe, one
type of tape.  [Specifically, the last or only tape in a VMS backup save
set archive.  Seems consistent with these.]  The most recent crash,
yesterday, is listed below.  I apologize in advance for the
hand-transcribed crash data as I don't have a way to get this
electronically at this time.  If there are any errors it is due to my
transcription; the transcript has been edited slightly.

st0 (siop0:0:0:0): Check condition on CDB: 0x08 00 00 80 00 00
    SENSE KEY: Recovered Error
               Incorrect Length Indicator Set
   INFO FIELD: 16867328
 COMMAND INFO: 671088743 (0x28000067)
     ASC/ASCQ: No Index/Sector Signal

panic: Kernel diagnostic assertion "done >= 0" failed
       file ...kern/kern_physio.c, line 222

Stopped in pid 236 (tapin) at cpu_Debugger+0x4: ret zero,(ra)
db>

The tapin program was indeed running at the time.  Partial output from the program can be viewed from the script output at:

<http://www.its.caltech.edu/~dundas/alpha/crash.txt>

Note that this is only a partial transcript of stdout as the complete file was not recovered after the system crashed.

>How-To-Repeat:
See description.
>Fix:
I am unable to fix.
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->feedback
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sun, 30 Nov 2008 07:04:44 +0000
State-Changed-Why:
Have you seen this problem on a recent kernel?


From: "John A. Dundas III" <dundas@caltech.edu>
To: gnats-bugs@NetBSD.org
Cc: "John A. Dundas III" <dundas@caltech.edu>
Subject: Re: port-alpha/22160 (panic using SCSI TK50 in NetBSD 1.6.1)
Date: Mon, 1 Dec 2008 07:47:12 -0800

 >Synopsis: panic using SCSI TK50 in NetBSD 1.6.1
 >
 >State-Changed-From-To: open->feedback
 >State-Changed-By: dholland@NetBSD.org
 >State-Changed-When: Sun, 30 Nov 2008 07:04:44 +0000
 >State-Changed-Why:
 >Have you seen this problem on a recent kernel?

 Gee, I filed that ticket several years ago :-)

 I haven't tried recently.  Is there any reason to expect a change in 
 behavior with a more recent kernel than the one it was filed under?

 Thanks,

 John
 -- 
 John A. Dundas III
 Director, IMSS Voice & Data Networks, Caltech
 Mail Code: 1-10, Pasadena, CA  91125
 Phone: 626.395.3392 FAX: 626.449.6973

From: David Holland <dholland-bugs@netbsd.org>
To: "John A. Dundas III" <dundas@caltech.edu>
Cc: port-alpha-maintainer@netbsd.org, gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org, gnats-bugs@netbsd.org
Subject: Re: port-alpha/22160 (panic using SCSI TK50 in NetBSD 1.6.1)
Date: Tue, 2 Dec 2008 16:38:02 +0000

 On Mon, Dec 01, 2008 at 05:10:05PM +0000, John A. Dundas III wrote:
  >  >Synopsis: panic using SCSI TK50 in NetBSD 1.6.1
  >  >
  >  >State-Changed-From-To: open->feedback
  >  >State-Changed-By: dholland@NetBSD.org
  >  >State-Changed-When: Sun, 30 Nov 2008 07:04:44 +0000
  >  >State-Changed-Why:
  >  >Have you seen this problem on a recent kernel?
  >  
  >  Gee, I filed that ticket several years ago :-)

 And it's sat there ever since :(  I've been trying to get ones like
 that some attention.

  >  I haven't tried recently.  Is there any reason to expect a change in 
  >  behavior with a more recent kernel than the one it was filed under?

 There have been a number of changes to physio (and st) since 1.6,
 so... maybe. I think in -current or 5.0 beta it probably at least
 won't crash the system any more just because you get an I/O error from
 the system. (Although on the other hand, beware of PR 38643.)

 I don't have any idea if anything's been done with the vax-specific
 patches, but if you can find them it should be possible to figure out.

 The urls in the original report are all stale, I'm afraid.

 -- 
 David A. Holland
 dholland@netbsd.org

From: "John A. Dundas III" <dundas@caltech.edu>
To: David Holland <dholland-bugs@netbsd.org>
Cc: "John A. Dundas III" <dundas@caltech.edu>,
 port-alpha-maintainer@netbsd.org, gnats-admin@netbsd.org,
 netbsd-bugs@netbsd.org, gnats-bugs@netbsd.org
Subject: Re: port-alpha/22160 (panic using SCSI TK50 in NetBSD 1.6.1)
Date: Tue, 2 Dec 2008 08:46:50 -0800

 At 4:38 PM +0000 12/2/08, David Holland wrote:
 >On Mon, Dec 01, 2008 at 05:10:05PM +0000, John A. Dundas III wrote:
 >  >  I haven't tried recently.  Is there any reason to expect a change in
 >  >  behavior with a more recent kernel than the one it was filed under?
 >
 >There have been a number of changes to physio (and st) since 1.6,
 >so... maybe. I think in -current or 5.0 beta it probably at least
 >won't crash the system any more just because you get an I/O error from
 >the system. (Although on the other hand, beware of PR 38643.)

 OK, thanks for letting me know.  I'll look up PR 38643.  If I have 
 some time this month, I'll try the current version and see what 
 happens.

 >I don't have any idea if anything's been done with the vax-specific
 >patches, but if you can find them it should be possible to figure out.
 >
 >The urls in the original report are all stale, I'm afraid.

 I'll attempt to update the ticket with anything new I learn.  I'll 
 also see if I can find the old URL contents.

 Thanks,

 John
 -- 
 John A. Dundas III
 Director, IMSS Voice & Data Networks, Caltech
 Mail Code: 1-10, Pasadena, CA  91125
 Phone: 626.395.3392 FAX: 626.449.6973

State-Changed-From-To: feedback->open
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sat, 31 Dec 2022 23:12:42 +0000
State-Changed-Why:
14-year feedback timeout :-|

This really needs to be tested one way or the other, no real reason to
suppose it's fixed. So, if anyone has an alpha with a tape drive...


>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.47 2022/09/11 19:34:41 kim Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2022 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.