From nobody@FreeBSD.org  Sat Mar 30 13:46:44 2002
Return-Path: <nobody@FreeBSD.org>
Received: from freefall.freebsd.org (freefall.FreeBSD.org [216.136.204.21])
	by hub.freebsd.org (Postfix) with ESMTP id 609ED37B416
	for <freebsd-gnats-submit@FreeBSD.org>; Sat, 30 Mar 2002 13:46:44 -0800 (PST)
Received: (from nobody@localhost)
	by freefall.freebsd.org (8.11.6/8.11.6) id g2ULkiL64453;
	Sat, 30 Mar 2002 13:46:44 -0800 (PST)
	(envelope-from nobody)
Message-Id: <200203302146.g2ULkiL64453@freefall.freebsd.org>
Date: Sat, 30 Mar 2002 13:46:44 -0800 (PST)
From: Matthias Andree <matthias.andree@web.de>
To: freebsd-gnats-submit@FreeBSD.org
Subject: sym driver fails on Tekram DC-390U in 486 machine
X-Send-Pr-Version: www-1.0

>Number:         36549
>Category:       kern
>Synopsis:       sym driver fails on Tekram DC-390U in 486 machine
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sat Mar 30 13:50:01 PST 2002
>Closed-Date:    Sun Nov 02 14:14:26 PST 2003
>Last-Modified:  Sun Nov 02 14:14:26 PST 2003
>Originator:     Matthias Andree
>Release:        4.5-RELEASE
>Organization:
>Environment:
(cannot be obtained) Cyrix 5x86/100 with UMC8881/8886 PCI chip set. 24 MB RAM.
>Description:
At Kernel boot time, the sym driver hangs the kernel when probing devices. The last lines I see are (not verbatim):

1. sending SCSI reset
2. waiting 15 seconds for SCSI devices to settle
3. sym0:0 ERROR (81:0) (8-0-0) (0/5) @ (mem 8:f000f7c8)
4. register dump (not copied here)

It's a DC-390U card, which has a Symbios 53C875 manufactured 9752R. The DC-390U has two 50-pin (narrow SCSI) connectors.

The only attached device is a Plextor PX-20 TSi CD-ROM at unit #2 (the adaptor has unit #7). PCI IRQ is unshared. The Plextor is terminated and is detected at boot time or by NetBSD 1.5.

The card is fine with NetBSD 1.5 or Linux 2.2/2.4. The card is also fine in a newer main board (Gigabyte 7ZX-R, with Duron/700 and VIA KT133 chip set).
>How-To-Repeat:
      
>Fix:
      
>Release-Note:
>Audit-Trail:

From: =?ISO-8859-1?Q?G=E9rard_Roudier?= <groudier@free.fr>
To: Matthias Andree <matthias.andree@web.de>
Cc: freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: kern/36549: sym driver fails on Tekram DC-390U in 486 machine
Date: Mon, 1 Apr 2002 21:23:44 +0200 (CEST)

 On Sat, 30 Mar 2002, Matthias Andree wrote:
 
 >
 > >Number:         36549
 > >Category:       kern
 > >Synopsis:       sym driver fails on Tekram DC-390U in 486 machine
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       high
 > >Responsible:    freebsd-bugs
 > >State:          open
 > >Quarter:
 > >Keywords:
 > >Date-Required:
 > >Class:          sw-bug
 > >Submitter-Id:   current-users
 > >Arrival-Date:   Sat Mar 30 13:50:01 PST 2002
 > >Closed-Date:
 > >Last-Modified:
 > >Originator:     Matthias Andree
 > >Release:        4.5-RELEASE
 > >Organization:
 > >Environment:
 > (cannot be obtained) Cyrix 5x86/100 with UMC8881/8886 PCI chip set. 24 MB=
  RAM.
 > >Description:
 > At Kernel boot time, the sym driver hangs the kernel when probing devices=
 =2E The last lines I see are (not verbatim):
 >
 > 1. sending SCSI reset
 > 2. waiting 15 seconds for SCSI devices to settle
 > 3. sym0:0 ERROR (81:0) (8-0-0) (0/5) @ (mem 8:f000f7c8)
 
 Means the SCRIPTS processor did try to execute instruction at bus
 physical address 8 (assumed memory).
 
 > 4. register dump (not copied here)
 >
 > It's a DC-390U card, which has a Symbios 53C875 manufactured 9752R. The D=
 C-390U has two 50-pin (narrow SCSI) connectors.
 >
 > The only attached device is a Plextor PX-20 TSi CD-ROM at unit #2 (the ad=
 aptor has unit #7). PCI IRQ is unshared. The Plextor is terminated and is d=
 etected at boot time or by NetBSD 1.5.
 >
 > The card is fine with NetBSD 1.5 or Linux 2.2/2.4.
 > The card is also fine in a newer main board (Gigabyte 7ZX-R, with Duron/7=
 00 and VIA KT133 chip set).
 
 Do you mean that the DC-390U with the Plextor attached doesn't encounter
 any problem with FreeBSD-4.5 on the Gigabyte 7ZX-R Mainboard?
 
 I got a couple of so weird SCRIPTS execution breakages under FreeBSD 4.5.
 For one of them 4.3 booted ok, but 4.4 nd 4.5 failed. Could this be tried?
 
 Btw, the driver changes between 4.3 and 4.4 should have no effect, given
 that they apply to the new CAM transfer code that is #ifed out for FreeBSD
 4.x (at least for now). There is a single endian related change between
 4.4 and 4.5 that has no effect on little endian machines.
 
 
 Regards,
   G=E9rard.
 

From: Matthias Andree <matthias.andree@stud.uni-dortmund.de>
To: =?iso-8859-15?Q?G=E9rard?= Roudier <groudier@free.fr>
Cc: Matthias Andree <matthias.andree@web.de>,
	freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: kern/36549: sym driver fails on Tekram DC-390U in 486 machine
Date: Tue, 2 Apr 2002 11:04:07 +0200

 On Mon, 01 Apr 2002, G=E9rard Roudier wrote:
 
 > > The card is fine with NetBSD 1.5 or Linux 2.2/2.4.
 > > The card is also fine in a newer main board (Gigabyte 7ZX-R, with Duron=
 /700 and VIA KT133 chip set).
 >=20
 > Do you mean that the DC-390U with the Plextor attached doesn't encounter
 > any problem with FreeBSD-4.5 on the Gigabyte 7ZX-R Mainboard?
 
 Yes, but with a different number of devices: one Plextor PX-32TS
 (CD-ROM drive), one Tandberg TDC 4222 (SLR 4 DC tape streamer), one
 Yamaha CRW4416 (CD writer).
 
 > I got a couple of so weird SCRIPTS execution breakages under FreeBSD 4.5.
 > For one of them 4.3 booted ok, but 4.4 nd 4.5 failed. Could this be tried?
 
 I'll try to boot 4.4 and see how it does. More later.

From: Matthias Andree <matthias.andree@stud.uni-dortmund.de>
To: =?iso-8859-15?Q?G=E9rard?= Roudier <groudier@free.fr>
Cc: Matthias Andree <matthias.andree@web.de>,
	freebsd-gnats-submit@FreeBSD.ORG
Subject: Re: kern/36549: sym driver fails on Tekram DC-390U in 486 machine
Date: Mon, 15 Apr 2002 15:23:22 +0200

 The sad thing is, I can no longer reproduce this.
 
 I'm getting this instead:
 
 sym0: <875> port 0xfc00-0xfcff mem 0xffbea000-0xffbeafff,0xffbebf00-0xffbebfff irq 11 at device 14.0 on pci0
 sym0: Tekram NVRAM, ID 7, Fast-20, SE, parity checking
 sym0: open drain IRQ line driver, using on-chip SRAM
 sym0: using LOAD/STORE-based firmware.
 pci0: <Matrox MGA Millennium II 2164W graphics accelerator> (vendor=0x102b, dev= 0x051b) at 15.0 irq 9
 pcib1: <UMC UM8881 HB4 486 PCI Chipset> at device 16.0 on pci0
 ...
 ad0: 9671MB <IBM-DTTA-351010> [19650/16/63] at ata0-master BIOSPIO
 Waiting 2 seconds for SCSI devices to settle
 (noperiph:sym0:0:-1:-1): SCSI BUS reset delivered.
 (probe2:sym0:0:2:0): INQUIRY. CDB: 12 1 80 0 ff 0 
 (probe2:sym0:0:2:0): ILLEGAL REQUEST asc:24,0
 (probe2:sym0:0:2:0): Invalid field in CDB
 Creating DISK cd0
 pass0 at sym0 bus 0 target 2 lun 0
 pass0: <PLEXTOR CD-ROM PX-20TS 1.00> Removable CD-ROM SCSI-2 device 
 pass0: 10.000MB/s transfers (10.000MHz, offset 15)
 
 I did upgrade the kernel to 4.5-RELEASE-p2 and change its configuration,
 I can send the diffs of the kernels if so desired. Most important
 changes: * drop cpu I686_CPU and cpu I386_CPU * drop device ncr (keeping sym)
 * change SCSI_DELAY=2000 (maybe the Plextor is not ready at the time it
 is queried, Plextor drives don't boot from CD with Tekram unless the
 reset->inquiry delay is >= 5 sec -- default 3 sec) * drop tons of other
 devices

From: =?ISO-8859-1?Q?G=E9rard_Roudier?= <groudier@free.fr>
To: Matthias Andree <matthias.andree@stud.uni-dortmund.de>
Cc: Matthias Andree <matthias.andree@web.de>,
	<freebsd-gnats-submit@FreeBSD.ORG>
Subject: Re: kern/36549: sym driver fails on Tekram DC-390U in 486 machine
Date: Mon, 15 Apr 2002 23:31:43 +0200 (CEST)

 On Mon, 15 Apr 2002, Matthias Andree wrote:
 
 > The sad thing is, I can no longer reproduce this.
 
 What's sad?
 A system that failed and is now working is a great news. :-)
 
 > I'm getting this instead:
 >
 > sym0: <875> port 0xfc00-0xfcff mem 0xffbea000-0xffbeafff,0xffbebf00-0xffb=
 ebfff irq 11 at device 14.0 on pci0
 > sym0: Tekram NVRAM, ID 7, Fast-20, SE, parity checking
 > sym0: open drain IRQ line driver, using on-chip SRAM
 > sym0: using LOAD/STORE-based firmware.
 > pci0: <Matrox MGA Millennium II 2164W graphics accelerator> (vendor=3D0x1=
 02b, dev=3D 0x051b) at 15.0 irq 9
 > pcib1: <UMC UM8881 HB4 486 PCI Chipset> at device 16.0 on pci0
 > ...
 > ad0: 9671MB <IBM-DTTA-351010> [19650/16/63] at ata0-master BIOSPIO
 > Waiting 2 seconds for SCSI devices to settle
 > (noperiph:sym0:0:-1:-1): SCSI BUS reset delivered.
 > (probe2:sym0:0:2:0): INQUIRY. CDB: 12 1 80 0 ff 0
 > (probe2:sym0:0:2:0): ILLEGAL REQUEST asc:24,0
 > (probe2:sym0:0:2:0): Invalid field in CDB
 > Creating DISK cd0
 > pass0 at sym0 bus 0 target 2 lun 0
 > pass0: <PLEXTOR CD-ROM PX-20TS 1.00> Removable CD-ROM SCSI-2 device
 > pass0: 10.000MB/s transfers (10.000MHz, offset 15)
 >
 > I did upgrade the kernel to 4.5-RELEASE-p2 and change its configuration,
 > I can send the diffs of the kernels if so desired. Most important
 > changes: * drop cpu I686_CPU and cpu I386_CPU * drop device ncr (keeping =
 sym)
 > * change SCSI_DELAY=3D2000 (maybe the Plextor is not ready at the time it
 > is queried, Plextor drives don't boot from CD with Tekram unless the
 > reset->inquiry delay is >=3D 5 sec -- default 3 sec) * drop tons of other
 > devices
 
 I just diffed 4.5 againt 4.5-release and didn't find anything that can
 explain the new behaviour of your system. Only a small change in
 tcp_syncache.c seems significant, but I donnot know enough about this code
 to actually see if it may explain something.
 
 The drop of the CPU types does affect slightly the machine dependant code,
 but the additional code should have no effect for a 486 CPU.
 
 Btw, there are so few changes, it seems, between the system that failed
 and the one that almost work that it could be possible to decrementaly (or
 incrementaly) add (or remove) the changes that seem related to the
 difference and see which one makes difference.
 
 You may for example try:
 
 1) Add again 686 and/or 386 CPU
 2) Reverse the patch from tcp_syncache.c
 3) Configure again ncr
 4) Restore previous SCSI_DELAY
 =2E..) All the changes I missed
 
 Until your system fails again.
 
 Regards,
   G=E9rard.
 

From: Matthias Andree <matthias.andree@gmx.de>
To: freebsd-gnats-submit@FreeBSD.org, matthias.andree@web.de
Cc:  
Subject: Re: kern/36549: sym driver fails on Tekram DC-390U in 486 machine
Date: Sun, 02 Nov 2003 23:07:23 +0100

 I think this PR can be closed.
 
 My Tekram DC-390U in a Tyan S1590 ("Trinity 100AT")
 Super Socket 7  (VIA MVP3 chip set) board works,
 and I no longer have i486 stuff to check the report on.
 
 -- 
 Matthias Andree
State-Changed-From-To: open->closed 
State-Changed-By: linimon 
State-Changed-When: Sun Nov 2 14:12:21 PST 2003 
State-Changed-Why:  
Submitter reports that "my Tekram DC-390U in a Tyan S1590 
(Trinity 100AT) Super Socket 7 (VIA MVP3 chip set) board 
works, and I no longer have i486 stuff to check the report on." 

http://www.freebsd.org/cgi/query-pr.cgi?pr=36549 
>Unformatted:
