From nobody@FreeBSD.org  Fri Apr 20 13:04:30 2001
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 A518537B423
	for <freebsd-gnats-submit@FreeBSD.org>; Fri, 20 Apr 2001 13:04:28 -0700 (PDT)
	(envelope-from nobody@FreeBSD.org)
Received: (from nobody@localhost)
	by freefall.freebsd.org (8.11.1/8.11.1) id f3KK4Si80109;
	Fri, 20 Apr 2001 13:04:28 -0700 (PDT)
	(envelope-from nobody)
Message-Id: <200104202004.f3KK4Si80109@freefall.freebsd.org>
Date: Fri, 20 Apr 2001 13:04:28 -0700 (PDT)
From: lazaro@online.no, salem@statoil.com
To: freebsd-gnats-submit@FreeBSD.org
Subject: System freeze booting from (i386) 4.3 floppies on HP Vectra VL 6/233 Series 7 
X-Send-Pr-Version: www-1.0

>Number:         26736
>Category:       i386
>Synopsis:       System freeze booting from (i386) 4.3 floppies on HP Vectra VL 6/233 Series 7
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Apr 20 13:10:01 PDT 2001
>Closed-Date:    Fri Aug 09 12:27:48 PDT 2002
>Last-Modified:  Mon Aug 12 08:30:03 PDT 2002
>Originator:     Lazaro D. Salem
>Release:        4.3RC4
>Organization:
none
>Environment:
Sorry, was not possible to complete the boot to run 'uname -a'.
The system is a HP-Vectra VL 6/233 Series 7 (i386)
with a Phoenix BIOS 4.0 Release 6.0.P  (1997)
>Description:
I am using the 4.3RC4 floppies. 
After inserting the mfsroot.flp and configuring drivers w/Kernel 
config Menu in visual mode, I left (the minimal) following 
Active Drivers:
          Dev---IRQ--Port 
Storage: 
          ata0   14  0x1f0
          ata1   15  0x170
          fdc0   6   0x3f0
Communications:
          ppc0   7
          sio0   4   0x3f8
Input:
         atkbd0  1
         psm0    12
         sc0
Miscellaneous:
         npx0    13   0xf0
Them quit and saved config, and let the kernel autoprobe devices:

--- begin transcript of last lines seen on the screen: 
md0: Preloaded image </mfsroot> 2949120 bytes at 0xc03a0628
md1: Malloc disk
npx0: <math processor> on motherboard
npx0: INT 16 interface
pcbi0: <Intel 82443LX (440 LX) host to PCI bridge> on motherboard
pci0: <PCI bus> on pcib0
pcbi1: <Intel 82443LX (440 LX) PCI-PCI (AGP) bridge> at device 1.0 on pci0
pci1: <PCI bus> on pcib1
pci1: <Cirrus Logic GD5465 SVGA controller> at 0.0 irq 9
isab0: <Intel 82371AB PCI to ISA bridge> at device 4.0 on pci0
isa0: <ISA bus> on isab0
atapci0: <Intel PIIX4 ATA33 controller> port 0xfcd0-0xfcdf at device 4.1 on pci0

ata0: at 0x1f0 irq 14 on atapci0
ata1: at 0x170 irq 15 on atapci0
uhci0: <Intel 82731AB/EB (PIIX4) USB controller> port 0xfce0-0xfcff irq 9 at device 4.2 on pci0
usb0: USB revision 1.0
uhub0: 2 ports with 2 removable, self powered
chip1: <Intel 82371AB Power management controchip1: <Intel 82371AB Power management controller> port 0x8800-0x880f at device 4.3 on pci0
pci0: <Number Nine Imagine 128 graphics accelerator at 12.0 irq 9
--- End of transcript

After which the system completely hangs.
The only way to reboot/turn-off the system is by pulling
the cable from the wall. No keyboard or on/off switch reaction.

--

Notice the IRQ 9 is used by the two graphic cards on this system, and
by the USB, although at different ports.

I then tried disabling IRQ 9 on the  BIOS setup as following:
From the PhoenixBIOS Setup Utility 
-> Advanced -> ISA Resource Exclusion (*)
-> I then changed the value of the IRQ 9 item from "Available" to 
"Reserved" (**). 
All the other IRQ items (3,4,5,7,10,11 and 15) were left "Available".

Trying to boot again after this BIOS setup, does not help: 
The system freezes at the same point with the same messages as above 
except for the following: every instance of 'irq 9' in the text above 
is replaced by 'irq 11'. 

If in addition to disabling the IRQ 9, I also disable the IRQ 11 
and retry, the machine still hangs ('irq 11' is now replaced 
by 'irq 10').

Notes:
(*) For this option, from the 'Item Specific Help' panel in the 
BIOS setup, I read: 
Reserves interrupts for legacy ISA devices to prevent conflict with 
PCI/PnP devices 
(**) The option Available allows to allocate IRQ for PCI or ISA PnP devices.

>How-To-Repeat:

>Fix:
Workaround: 

    Maybe disabling probing of USB _and_ Power management in GENERIC?

Fix: 

    I do not really know. 

First, notice the system has presently two graphic cards: 
one integrated in the motherboard (AGP Cirrus Logic Laguna GD5465)
and an external one (Number 9 Imagine 128) and both are 
on the PCI bus.
The USB and the two graphic cards presently are reported to 
use the same IRQ, although at different ports. 
Should that happen at all? 

I am also not really sure that this is an IRQ asignement 
problem, as a first look suggests.
It might have to do with the power management controller...

It is worth to mention I had this problem every time I tried
booting from releases floppies in the 4.x branch.
In all those attempts (specifically using 4.0-R,4.1.1-R and 
4.2-R floppies) I had only _one_ graphic card. 
I still observed the same problem, which was reported 
in PR misc/18201.
The USB messages were then slightly different. I guess that is 
due to the work done on the USB code since 4.0-R. 
But still the IRQ 9 was shared by the USB and the single graphic 
card (Cirrus Logic GD5465). In those cases the system stopped 
responding just after printing:

[snip previous lines]
chip1: <Intel 82371AB Power management controller> port 0x8800-0x880f at device 4.3 on pci0
--
After that I had always to pull the chord to reboot.

Puzzle: with 4.3RC4 the system hangs after recognizing the 
2nd graphic card, printed after probing the Power management 
controller. If the 82371AB was the culprit, shouldn't it hang 
before the last line when trying 4.3RC floppies?.

Another failed attempt was as follows:
When playing with the 4.1.1 CD, I managed to install the 
src/sys tree, edited GENERIC, commented out all USB support 
in it (at least I thought so) and finally compiled a 4.x kernel
which I copied to the kernel.flp diskette. No changes to 
mfsroot.flp, though.
I tried in this way to eliminate the sharing of one IRQ by two devices. 
The system still hanged at boot time with these floppies.

----
The box presently runs 3.5-STABLE without no other inconvenience 
than be "forced" to do so.
I am willing to test any floppy pair you send me to overcome 
this problem. Thank you for reading this long post.
>Release-Note:
>Audit-Trail:

From: "Lazaro D. Salem" <salem@statoil.com>
To: freebsd-gnats-submit@FreeBSD.org, lazaro@online.no,
	salem@statoil.com
Cc: Sheldon Hearn <sheldonh@uunet.co.za>
Subject: Re: i386/26736: System freeze booting from (i386) 4.3 floppies on HP 
 Vectra VL 6/233 Series 7
Date: Mon, 21 May 2001 16:18:46 +0200

 A followup to a related PR has been submitted with a work around this
 problem
 See:
 http://www.FreeBSD.org/cgi/query-pr.cgi?pr=26686
 also related is
 http://www.FreeBSD.org/cgi/query-pr.cgi?pr=18201
 
 The IRQ peculiarity mentioned here has nothing to do with the freeze.
 Neither does the USB or Power management intrinsic to the HP/Vectra.
 It has been suggested the culprit is the ex driver probing an installed
 the NE2000 NIC. Disabling the NE2000 did not not help.
 Physically removing the NE2000 did help and the boot completed
 without problems. If the ex driver is the culprit, why cannot we
 disabled
 at the kernel configuration stage? This was possible to do in the
 3.x releases,  but is not possible to do with the 4.x releases.
 
 Cheers
 

From: Hiten Pandya <hitmaster2k@yahoo.com>
To: freebsd-gnats-submit@FreeBSD.org, lazaro@online.no,
	salem@statoil.com
Cc:  
Subject: Re: i386/26736: System freeze booting from (i386) 4.3 floppies on HP 
 Vectra VL 6/233 Series 7
Date: Wed, 02 Jan 2002 16:22:25 +0000

 hi,
 
 According to the Active drivers in your Visual Kernel Config:
 
 ==================
 Active Drivers:
                Dev---IRQ--Port 
      Storage: 
                ata0   14  0x1f0
                ata1   15  0x170
                fdc0   6   0x3f0
      Communications:
                ppc0   7
                sio0   4   0x3f8
      Input:
               atkbd0  1
               psm0    12
               sc0
      Miscellaneous:
               npx0    13   0xf0
 
 =================
 
 The 'pci' controller has been removed from the 'Miscelleneous' category,
 and
 according to my doubts, it hangs right where a PCI scan starts, this
 means that
 you will have to leave the 'pci0' controller in, otherwise the system
 will hang.
 
 This means, doing nothing to the Miscelleneous Category of the Kernel
 Configuration
 step, when booting of the installation.
 
 regards,
  - Hiten
  - <hiten@uk.FreeBSD.org>
 
 -- 
 1024 45:a5:9c:f2:fb:07:da:70:18:02:0b:f3:63:f1:7a:a6 hitenp@hpdi.ath.cx
State-Changed-From-To: open->feedback 
State-Changed-By: sheldonh 
State-Changed-When: Mon Jan 14 03:47:28 PST 2002 
State-Changed-Why:  
We need to know whether the problem Mr Pandya guesses is the  
cause is indeed the problem affecting the originator. 

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=26736 
State-Changed-From-To: feedback->closed 
State-Changed-By: schweikh 
State-Changed-When: Fri Aug 9 12:27:16 PDT 2002 
State-Changed-Why:  
Feedback timeout. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=26736 

From: Jens Schweikhardt <schweikh@schweikhardt.net>
To: Lazaro Daniel Salem <salem@statoil.com>
Cc: bug-followup@freebsd.org
Subject: Re: i386/26736: System freeze booting from (i386) 4.3 floppies on HP Vectra VL 6/233 Series 7
Date: Mon, 12 Aug 2002 12:17:51 +0200

 Dear Lazaro,
 
 [please Cc: answers to bug-followup@freebsd.org so GNATS knows about them.]
 
 On Mon, Aug 12, 2002 at 09:29:02AM +0200, Lazaro Daniel Salem wrote:
 # 
 # Dear Jens,
 # I cannot recall receiving a request for feedback I have not answered.
 
 I was referring to this state change:
 
 	State-Changed-From-To: open->feedback 
 	State-Changed-By: sheldonh 
 	State-Changed-When: Mon Jan 14 03:47:28 PST 2002 
 	State-Changed-Why:  
 	We need to know whether the problem Mr Pandya guesses is the  
 	cause is indeed the problem affecting the originator. 
 
 Is the answer to this in a related PR?
 
 # I still observe the same problems with 4.6R. Please read the related PRs
 # mentioned in this one.
 # 
 # The 3 PRs revolve on the same same thing, something I did not known at
 # first, but when
 # I got the clue I did mention in all of them by following-up.
 # 
 # I had recently had a crash on th HD where 4.4 stable was installed so I am
 # quite sure
 # 4.5R and 4.6R did not solve the problem as I tried to install from the CDs.
 # In both cases I observed the same problem already reported.
 
 Do you want me to reopen this PR or is this problem adequately handled
 by one of the mentioned PRs? We should try to avoid duplicates if we
 can.
 
 Regards,
 
 	Jens
 -- 
 Jens Schweikhardt http://www.schweikhardt.net/
 SIGSIG -- signature too long (core dumped)

From: "Lazaro Daniel Salem" <salem@statoil.com>
To: bug-followup@freebsd.org
Cc:  
Subject: Re: i386/26736: System freeze booting from (i386) 4.3 floppies on HP Vectra
 VL 6/233 Series 7
Date: Mon, 12 Aug 2002 17:20:59 +0200

 > Is the answer to this in a related PR?
 
 No. the answer to Mr Pandya is not in this or other PR as I could not make
 sense of
 his suggestion (I should have answered anyway).
 Notice the ne2000 (ISA card) was/is not using the IRQ assigned to the pci.
 When I remove the ISA NE200 card from the box the machine boots without
 any problem (which confirms Mr Pandya observation misses the real problem).
 
 Reading the followups in the other PRs
     http://www.FreeBSD.org/cgi/query-pr.cgi?pr=26686
     http://www.FreeBSD.org/cgi/query-pr.cgi?pr=18201
 sheds lights on the problem, and also suggests a solution originally
 proposed  by G.W.Roberts <mas039"@bangor.ac.uk> on 27-Apr-2001.
 
 There is certainly a duplication of the PR, but at the time I wrote those
 I did not know. May I suggest you collect the relevant info in the latest
 PR's
 with references to the other (probably to-be-closed ones)?
 
 Thanks for such a timely and good "user support" work!
 
 Lazaro D. Salem
 
 
 -------------------------------------------------------------------
 The information contained in this message may be CONFIDENTIAL and is
 intended for the addressee only. Any unauthorised use, dissemination of the
 information or copying of this message is prohibited. If you are not the
 addressee, please notify the sender immediately by return e-mail and delete
 this message.
 Thank you.
 
>Unformatted:
