DaemonForums  

Go Back   DaemonForums > OpenBSD > OpenBSD Installation and Upgrading

OpenBSD Installation and Upgrading Installing and upgrading OpenBSD.

Reply
 
Thread Tools Display Modes
  #1   (View Single Post)  
Old 9th September 2011
marcv marcv is offline
New User
 
Join Date: Sep 2011
Posts: 5
Default OpenBSD ACER aspire 9300 laptop install panic

Hello all,

Sorry for the lengthy post but I want to explain my problem as good as possible!

Since about a week I've been working on installing OpenBSD 4.9 (i386) from CD on this laptop and had some problems - very similar to the ones described in by user Sepuku in June. I've tried the steps outlined in the reply #9 by jggimi. Sometimes it works, usually it doesn't.

What I've noticed/tried:
- upon booting from CD and disabling ACPI, sometimes I can run the full install. After booting from the HD OpenBSD crashes again with a panic screen. I've successfully installed OpenBSD 2 times now but cannot get it to boot reliably
- if I don't disable ACPI I always get a panic screen during install
- if I boot from HD, I've been able to get OpenBSD running without disabling ACPI once.
- the liveCD of jggmi doesn't boot - I get a similar panic screen
- the panic screen is shortly after the boot starts - I don't see any boot messages on a blue background. It takes less then a second to crash.

I don't think it is due to a hardware problem:
- the laptop has been running Windows XP without problems for years
- I've tried booting and installing several distro's of linux (ubuntu, mint, clearos, ...) and it all works - no problem whatsoever.
- I've tried installing OpenBSD 4.0 from the official CD's and that failed in the same way.
- I've done an extensive memory test (11 hours) using memtest
- I've done a thorough disktest and that came out good (apart from 2 sector reallocations)
- looks like ACPI has no influence ...

So I'm beginning to suspect a bug in OpenBSD.
I've added 2 attachments:
- a zip file containing evidence of the panic screens
- dmesg file

Some help to nail this further down is greatly appreciated.

Regards,

Marc
Attached Files
File Type: zip Acerboot.zip (891.4 KB, 127 views)
File Type: txt dmesg_openbsd.txt (5.8 KB, 90 views)
Reply With Quote
  #2   (View Single Post)  
Old 9th September 2011
jggimi's Avatar
jggimi jggimi is offline
More noise than signal
 
Join Date: May 2008
Location: USA
Posts: 7,975
Default

Hello and welcome. Thank you for describing your problem so clearly, and including your dmesg and what you could capture with your camera.

If I've interpreted it all properly, when you boot the ramdisk kernel (the one used to install the OS), you must always disable ACPI otherwise you often get an invalid opcode fault. When you boot the full kernel (after installation, booting the hard drive) you often get a protection fault trap.

Unfortunately, the kernel stack traces on screen aren't helpful to me; at the least they require a diagnostic build of the matching -release kernel to map addresses to kernel source code components. And while I could build one, I would not necessarily interpret the resulting analysis correctly.

I should ask -- the next time you successfully get the OS running -- check to see if you have any files stored in /var/crash; this directory will store information from a panic on restart, if possible.

I am going to make a set of recommendations for you:

1) Try an i386-current snapshot. The -current environment is already a couple of months of development beyond 5.0-release, which is expected November 1. 4.9-release was built in February, so it is possible whatever this is has been fixed between then and now. See your nearest mirror, typically in /pub/OpenBSD/snapshots/i386/.

2) Try an amd64 version -- it's the same OS, just a different architecture, and if I read your dmesg correctly, your processor is 64-bit capable. You may get different results. Look in /pub/OpenBSD/4.9/amd64/ and /pub/OpenBSD/snapshots/amd64/

3) Post a detailed report to the misc@ mailing list. About half of the developers follow it, as well as many technically astute users. You are likely to get some good advice if you approach the list requesting recommendations for assistance. Be sure to post your dmesg in-line with your message (misc@ does not accept attachments), and you may provide a direct link to your .zip you've attached here, as long as you let the list know the link points to photos of the console.

Posts to misc@ (and to all the other mailing lists) need to be plain text, and wrap at around 72 columns. See http://www.openbsd.org/mail.html for further information on posting to the lists.
Reply With Quote
  #3   (View Single Post)  
Old 9th September 2011
marcv marcv is offline
New User
 
Join Date: Sep 2011
Posts: 5
Default

Quote:
Originally Posted by jggimi View Post
Hello and welcome. Thank you for describing your problem so clearly, and including your dmesg and what you could capture with your camera.
Thanks for the welcome !

Quote:
If I've interpreted it all properly, when you boot the ramdisk kernel (the one used to install the OS), you must always disable ACPI otherwise you often get an invalid opcode fault. When you boot the full kernel (after installation, booting the hard drive) you often get a protection fault trap.
Let me put it another way: without disabling ACPI on the ramdisk kernel I've not been able to boot it properly. However, I get the feeling that it is possible to boot from the ramdisk kernel when ACPI is enabled but that the chances are slim.

Quote:
...
I should ask -- the next time you successfully get the OS running -- check to see if you have any files stored in /var/crash; this directory will store information from a panic on restart, if possible.
Did that - nothing there besides a file called minfree that has the number 4096 in it. I guess it's too early in the boot to write something there ...

Quote:
I am going to make a set of recommendations for you:

1) Try an i386-current snapshot. The -current environment is already a couple of months of development beyond 5.0-release, which is expected November 1. 4.9-release was built in February, so it is possible whatever this is has been fixed between then and now. See your nearest mirror, typically in /pub/OpenBSD/snapshots/i386/.
I'll give it a go.

Quote:
2) Try an amd64 version -- it's the same OS, just a different architecture, and if I read your dmesg correctly, your processor is 64-bit capable. You may get different results. Look in /pub/OpenBSD/4.9/amd64/ and /pub/OpenBSD/snapshots/amd64/
I'll try that too.

Quote:
3) Post a detailed report to the misc@ mailing list. About half of the developers follow it, as well as many technically astute users. You are likely to get some good advice if you approach the list requesting recommendations for assistance. Be sure to post your dmesg in-line with your message (misc@ does not accept attachments), and you may provide a direct link to your .zip you've attached here, as long as you let the list know the link points to photos of the console.

Posts to misc@ (and to all the other mailing lists) need to be plain text, and wrap at around 72 columns. See removed_url_because_of_limitation_as_a_new_user for further information on posting to the lists.
Thanks for the help and recommendations. I'll try the -current snapshot first to see what that gives. If I still get a failure I'll post to the misc@ list.

Regards,

Marc
Reply With Quote
  #4   (View Single Post)  
Old 9th September 2011
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

Indeed, you should always try a snapshot if you suspect the problem may be with OpenBSD. They represent the latest code available.

This system is 64-bit compatible, OpenBSD/amd64 may work where OpenBSD/i386 fails, as they are treated as completely separate systems and do many things differently.

It should also be noted that Acer has a BIOS update for the Aspire 9300, you can find it below.. it is possible that the problem was on their end.

http://support.acer.com/us/en/produc...=5&modelId=135
or
http://global-download.acer.com/Step...&Category=BIOS
Reply With Quote
  #5   (View Single Post)  
Old 10th September 2011
marcv marcv is offline
New User
 
Join Date: Sep 2011
Posts: 5
Default

OK - update:

I've taken the current snapshot (8th of Spetember) of i386-current and that indeed had another behaviour. It did not crash but neither did it go further beyond the 'entry point at 0x200120' line - - the install simply stops. No panic screen - it just hangs.

The AMD release gets to the 'boot>' prompt. After a short wait it shows a spinning cursor and 1 number (in the other releases it seems like calculating and displays something like '5965416+946088 [61+227984+215947]=0x703e08'). Then it crashes without a panic screen and the system reboots automatically.

Then I've upgraded the BIOS to V1.20 - the latest available. I did have to wipe out OpenBSD and reinstall XP in order to get that done

I've tried out all the different versions of OpenBSD again and I still get the same results. Upgrading the BIOS did not solve it.

Guess it's time to write an email to misc@openbsd ...

Regards,

Marc
Reply With Quote
  #6   (View Single Post)  
Old 12th September 2011
marcv marcv is offline
New User
 
Join Date: Sep 2011
Posts: 5
Default

Quote:
Originally Posted by marcv View Post
OK - update:

I've taken the current snapshot (8th of September) of i386-current and that indeed had another behaviour. It did not crash but neither did it go further beyond the 'entry point at 0x200120' line - - the install simply stops. No panic screen - it just hangs.

....

Guess it's time to write an email to misc@openbsd ...

Regards,

Marc
I've posted my problem on the misc@ list. No takers their either.

The good news is that after a lot of investigation and trail/error I found a workaround that avoids the panic screen.
If I press F12 - the button that invokes the bootdevice selection - and select the hard disk then the laptop boots as expected. If I don't do that, the second stage boot crashes with a panic screen (trap: 0(0): invalid opcode fault).

Boot order of the devices in the BIOS doesn't matter. Pressing F12 does the trick.

Anyone that can help figure this out?

This workaraound has been tested on i386-current.

Regards,

Marc
Reply With Quote
  #7   (View Single Post)  
Old 12th September 2011
jggimi's Avatar
jggimi jggimi is offline
More noise than signal
 
Join Date: May 2008
Location: USA
Posts: 7,975
Default

I've followed the thread on misc@, Marc. Glad you found a workaround!
Reply With Quote
  #8   (View Single Post)  
Old 12th September 2011
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

x86 bootloaders make a lot of assumptions about the state the system is in when the BIOS passes control over to it.. legacy structures in memory, register contents, and how other legacy hardware is configured.

I guess when this system goes to the F12 menu it takes some extra precautions to ensure the legacy environment is more compatible.. it's really hard to say why.

A lot of newer x86 firmware is actually UEFI/EFI with legacy BIOS emulation as a module.

It's definitely frustrating, and it may indeed be a bug in OpenBSD's bootloader, however it does boot on a large amount of systems.. the problem is likely pretty obscure and will be difficult to hunt down, contacting a developer who has recently committed in that area may be able to help isolate it.
Reply With Quote
  #9   (View Single Post)  
Old 13th September 2011
marcv marcv is offline
New User
 
Join Date: Sep 2011
Posts: 5
Default

Quote:
Originally Posted by BSDfan666 View Post
x86 bootloaders make a lot of assumptions about the state the system is in when the BIOS passes control over to it.. legacy structures in memory, register contents, and how other legacy hardware is configured.

I guess when this system goes to the F12 menu it takes some extra precautions to ensure the legacy environment is more compatible.. it's really hard to say why.
Yes - I can understand that a bootloader is complex.

Quote:
A lot of newer x86 firmware is actually UEFI/EFI with legacy BIOS emulation as a module.

It's definitely frustrating, and it may indeed be a bug in OpenBSD's bootloader, however it does boot on a large amount of systems.. the problem is likely pretty obscure and will be difficult to hunt down, contacting a developer who has recently committed in that area may be able to help isolate it.
Since all the linux distros I used do work would it be a viable option to compare their assumptions of the BIOS state against OpenBSD maybe?

Or for that matter 'dump' the BIOS state(s), registers, ... in say /var/crash and then kick off the second stage boot. If that crashes, we still have the contents saved there and can compare it against a 'good' run where I used F12.

But then I definitvely need guidance from a developer to tell me how to save the BIOS states ...

Regards,

Marc
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Obsd won't install to acer laptop sepuku OpenBSD Installation and Upgrading 16 7th July 2011 03:26 PM
OpenBSD's X11 on an Acer Aspire netbook bruguiea OpenBSD General 6 20th April 2010 06:10 AM
any possible way to make install46.iso boot from usb to install to a old laptop vinnie_vinodh OpenBSD Installation and Upgrading 10 21st February 2010 06:37 PM
audio and ethernet problem on on acer aspire 4740G Stellar FreeBSD Installation and Upgrading 3 21st February 2010 07:27 AM
Anyone tried FBSD on Acer Aspire One? michaelrmgreen FreeBSD General 1 5th September 2008 12:39 PM


All times are GMT. The time now is 04:56 AM.


Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Content copyright © 2007-2010, the authors
Daemon image copyright ©1988, Marshall Kirk McKusick