NetBSD Problem Report #11726

Received: (qmail 2611 invoked from network); 13 Dec 2000 15:31:52 -0000
Message-Id: <200012131531.HAA25463@nbwww.isc.org>
Date: Wed, 13 Dec 2000 07:31:51 -0800 (PST)
From: angelo.morando@libero.it
Reply-To: angelo.morando@libero.it
To: gnats-bugs@gnats.netbsd.org
Subject: driver for TMC-950 based SCSI host adapters hangs the system
X-Send-Pr-Version: www-1.0

>Number:         11726
>Category:       kern
>Synopsis:       driver for TMC-950 based SCSI host adapters hangs the system
>Confidential:   no
>Severity:       critical
>Priority:       low
>Responsible:    kern-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 13 15:32:00 +0000 2000
>Closed-Date:    Fri Jun 15 06:54:37 +0000 2007
>Last-Modified:  Fri Jun 15 06:54:37 +0000 2007
>Originator:     Gilberto Morando
>Release:        1.4.1
>Organization:
>Environment:
i386, 16Mb.
>Description:
the SCSI driver correctly finds the adapter, does the search for peripherals
but hangs ("timeout") when it finds one because IDENTIFY command fails
>How-To-Repeat:
Insert this driver into the kernel and correctly configure it.
Connect a SCSI cdrom. (probably the same with a disk)
>Fix:

>Release-Note:
>Audit-Trail:

From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: angelo.morando@libero.it
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/11726: driver for TMC-950 based SCSI host adapters hangs the system
Date: Wed, 13 Dec 2000 17:01:01 +0100

 On Wed, Dec 13, 2000 at 07:31:51AM -0800, angelo.morando@libero.it wrote:
 > 
 > >Number:         11726
 > >Category:       kern
 > >Synopsis:       driver for TMC-950 based SCSI host adapters hangs the system
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       low
 > >Responsible:    kern-bug-people
 > >State:          open
 > >Class:          sw-bug
 > >Submitter-Id:   net
 > >Arrival-Date:   Wed Dec 13 07:32:00 PST 2000
 > >Closed-Date:
 > >Last-Modified:
 > >Originator:     Gilberto Morando
 > >Release:        1.4.1
 > >Organization:
 > >Environment:
 > i386, 16Mb.
 > >Description:
 > the SCSI driver correctly finds the adapter, does the search for peripherals
 > but hangs ("timeout") when it finds one because IDENTIFY command fails

 If IDENTIFY fail, either this device is severely broken, or you SCSI bus
 is broken. Check cables and terminations.

 --
 Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
 --

From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: "angelo.morando@libero.it" <angelo.morando@libero.it>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: kern/11726: driver for TMC-950 based SCSI host adapters hangs the system
Date: Sat, 16 Dec 2000 19:03:41 +0100

 --envbJBWh7q8WU6mo
 Content-Type: text/plain; charset=us-ascii

 [ Please keep <gnats-bugs@gnats.netbsd.org> in Cc: so that the messages
   show up in the bug tracking system. It's important for developers ]

 On Fri, Dec 15, 2000 at 11:40:48AM +0100, angelo.morando@libero.it wrote:
 > Dear Mr. Bouyer,
 > 
 > Yesterday I had trouble recompiling the kernel (config, & depend go OK,
 > but make complains about some missing PCMCIA reference in a module....
 > I actually stripped all PCMCIA reference from GENERIC, but this wasn't
 > expected to cause problems, since the compile directory was void...
 > Besides, the day before the compilation went well...
 > But that was another day, wasn't it ? :-)))))))

 Are you using 1.5 release or -current ? -current can, on occasion, fail to
 build.
 If you removed all pcmcia devices, did you also remove all references to
 tcic and pcic ?

 > 
 > I therefore still can't send you the /var/log/messages you requested me,
 > since the right kernel isn't running. I'll try this evening.
 > 
 > However i remember fairly well the messages i got.
 > It was:
 > 
 > "found controller sea0 at de000 irq 5"
 > "testing 8 targets, 8 id per target"
 > "testing (0,0,0) not found"
 > "testing (0,0,0) not found AGAIN"
 > "testing (0,1,0) not found"
 > "testing (0,1,0) not found AGAIN"
 > 
 > (note that these complaints are correct: no peripheral exist on IDs 0
 > and 1).
 > 
 > "testing (0,2,0)"
 > "expecting IDENTIFY, got 0x80)"
 > "timeout error"
 > "timeout error"
 > 
 > and so on.... (it proceeds ad infinitum like that)
 > 
 > It isn't an absolutely precise report, I know, but the relevant messages
 > are all there (I remember them well because I noted yesterday that

 Doesn't like NetBSD boot messages :)

 However I think I found the problem, it's just a reverted test logic.
 Can you try the attached patch ?

 --
 Manuel Bouyer <bouyer@antioche.eu.org>
 --

 --envbJBWh7q8WU6mo
 Content-Type: text/plain; charset=us-ascii
 Content-Disposition: attachment; filename="seagate.diff"

 --- seagate.c.old	Sat Dec 16 18:58:35 2000
 +++ seagate.c	Sat Dec 16 19:00:22 2000
 @@ -896,7 +896,7 @@
  	phase = PH_MSGIN;
  	sea_transfer_pio(sea, &phase, &len, &data); 

 -	if (MSG_ISIDENTIFY(msg[0])) {
 +	if (!MSG_ISIDENTIFY(msg[0])) {
  		printf("%s: expecting IDENTIFY message, got 0x%x\n",
  		    sea->sc_dev.dv_xname, msg[0]);
  		abort = 1;

 --envbJBWh7q8WU6mo--

From: "angelo.morando@libero.it"<angelo.morando@libero.it>
To: bouyer@antioche.lip6.fr
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: kern/11726: driver for TMC-950 based SCSI host adapters hangs the 
	system
Date: Sun, 17 Dec 2000 00:38:51 +0100

 I tried the patch you sent me.
 There's good news and bad news
 The good news: now the system recognizes the SCSI adapter and the 
 attached peripherals
 The bad news: after mounting one device (i tried cdrom, id 3, there's 
 also a hd in id 2) the kernel panics.

 How to repeat:
 1) Mount the device
 2) Do whatever you want, or, more simply, let the kernel do its thing

 The kernel is guaranteed to panic in less than 1 minute.

 Now the question ?

 1) Do I have to turn the DEBUG flag on ? (Ok, ok, I'll do. I asked just 
 to avoid >1h of recompilation :-))
 2) Is there a convenient way to send you the output you need other than
 copying the screen output ? (I've got a dialup connection, so remote 
 debugging is not an option. I've got another UNIX box, if this can be 
 useful).
 3) Do I have to open anither bug in the list ?
 I didn't so because you're already working on the driver, but I don't 
 know the practice that NetBSD Foundation follows, because it's the 
 first bug I report.

 Thank you for the work you've done
                    Best regards
                           Gilberto


From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: "angelo.morando@libero.it" <angelo.morando@libero.it>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: kern/11726: driver for TMC-950 based SCSI host adapters hangs the system
Date: Sun, 17 Dec 2000 20:42:59 +0100

 On Sun, Dec 17, 2000 at 12:38:51AM +0100, angelo.morando@libero.it wrote:
 > I tried the patch you sent me.
 > There's good news and bad news
 > The good news: now the system recognizes the SCSI adapter and the 
 > attached peripherals

 Good, I'll commit the fix.

 > The bad news: after mounting one device (i tried cdrom, id 3, there's 
 > also a hd in id 2) the kernel panics.
 > 
 > How to repeat:
 > 1) Mount the device
 > 2) Do whatever you want, or, more simply, let the kernel do its thing
 > 
 > The kernel is guaranteed to panic in less than 1 minute.

 It doesn't panic immediadly ? I'll be harder to debug :(

 > 
 > Now the question ?
 > 
 > 1) Do I have to turn the DEBUG flag on ? (Ok, ok, I'll do. I asked just 
 > to avoid >1h of recompilation :-))

 No, but ddb would be nice (it's in the generic kernel).

 > 2) Is there a convenient way to send you the output you need other than
 > copying the screen output ? (I've got a dialup connection, so remote 
 > debugging is not an option. I've got another UNIX box, if this can be 
 > useful).

 A serial console: you need a crossover cable connected between a serial
 port of your PC and the other Unix machine. Then use kermit, or tip or
 whatever is available to connect to the serial port on the other unix
 machine. On your PC, at the boot prompt, type 'consdev com0' (assuming
 you've connected  your cable to com0). Now the boot prompt should show
 up over the serial connection.

 Now there's another way: when the kernel panic, type 'call cpu_reboot(0x104)'
 the kernel should dump a core. Before this you can type 'tr' to get a stack
 trace.
 At reboot, it will save this core dump in
 /var/crash  if there's enouth place (if not, make /var/crash a symlink to
 a directory with enouth space), the core dump will be as big as your RAM.

 Then it's possible to use gdb on the core dump to get infos. If you have some
 way to make the kernel and core dump available for ftp it would be nice :)

 > 3) Do I have to open anither bug in the list ?
 > I didn't so because you're already working on the driver, but I don't 
 > know the practice that NetBSD Foundation follows, because it's the 
 > first bug I report.

 We can follow up on this one.

 --
 Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
 --
State-Changed-From-To: open->closed
State-Changed-By: dsl@netbsd.org
State-Changed-When: Fri, 15 Jun 2007 06:54:37 +0000
State-Changed-Why:
1.4.1 is not longer supported
The code has almost certainly changed since then!


>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.