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 15th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default OpenBSD 4.7 issue?

I just installed OpenBSD 4.7 AMD64 on a desktop using a CD I bought. When I boot up it asks for commands like this:

ddb{0}>

I had used 4.6 briefly and don't remember it being like that. There should have just been a $ or something indicating you were in shell. I can enter basic commands after it but not anything like package installation or reading man pages. It says something about a kernel page fault trap two lines above it, so that might be the issue. I'm wondering if the install went wrong, if there's a hardware issue (I can list more details if needed) or maybe there's nothing wrong at all and I just need to type something to get into a normal shell. If I had to I could just leave the CD in and type 's' when it asks for install, upgrade or shell, but I don't see why it would be necessary. I appreciate any help.
Reply With Quote
  #2   (View Single Post)  
Old 15th May 2010
jggimi's Avatar
jggimi jggimi is offline
More noise than signal
 
Join Date: May 2008
Location: USA
Posts: 7,975
Default

1. ddb is the kernel debugger. Your kernel entered a panic. More information should be right -above- the ddb> prompt as for the reason for the panic. See this man page for general guidance on panics and crashes: http://www.openbsd.org/cgi-bin/man.c...86&format=html

2. 4.7 is not -yet- supported. It will become supported the day that 4.7 is officially released by the project. Right now, only 4.5, 4.6, and -current are supported.
Reply With Quote
  #3   (View Single Post)  
Old 15th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

As stated, the kernel has encountered a problem and panicked.. you can now issue commands to obtain further information about the crash and report the issue to developers.

4.7 isn't officially supported until May 19th, but if you could provide a dmesg in [code][/code] blocks.. along with the message indicated above the debugger console, and a backtrace, the developers might be able to help you out further (..so send an email).

The {0} next to ddb indicates you're running the GENERIC.MP(bsd.mp) kernel, can you try running GENERIC? it may be named bsd.sp in a default installation on an dual processor/core.
Reply With Quote
  #4   (View Single Post)  
Old 15th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

Thanks. From reading that man page it looks like it might be hardware-related. We could wait until May 19 (official 4.7 release) to diagnose it, but in any case here is my specs:

Gateway GT5012 PC
Pentium D 2.8 Ghz dual-core processor
3 GB memory (2 512 MB stock RAM, 2 1 GB sticks added in after)
700W power supply

I can't post URLs yet, but if you Google "Gateway GT5012" and click the second result that will give you detailed info.

I doubt the extra RAM or upgraded power supply would cause it, but I could try anyway. Let me know if more info on that or the error message is needed. To reiterate, the problem was a kernel page fault trap.
Reply With Quote
  #5   (View Single Post)  
Old 15th May 2010
jggimi's Avatar
jggimi jggimi is offline
More noise than signal
 
Join Date: May 2008
Location: USA
Posts: 7,975
Default

One possibility is an ACPI related issue. If you conducted an installation, you did so from the ramdisk kernel, which does not use ACPI.

Easy enough to test: disable ACPI in the kernel during bootup, and see what happens.

At the "boot>" prompt, type "-c" to enter the User Kernel Configurator.

The kernel will load, and before doing much else, will give you a "UKC>" prompt. Type "disable acpi", then type, "quit" -- the kernel will continue onward, with its ACPI code disabled. If the OS comes up, you'll know your problem is ACPI-related.

The ACPI standard is very loosely written, and every hardware manufacturer's implementation is a little bit different. Some are downright broken. If it works with Windows, it's usually good enough for them.
Reply With Quote
  #6   (View Single Post)  
Old 15th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

Quote:
Originally Posted by jggimi View Post
At the "boot>" prompt, type "-c" to enter the User Kernel Configurator.

The kernel will load, and before doing much else, will give you a "UKC>" prompt. Type "disable acpi", then type, "quit" -- the kernel will continue onward, with its ACPI code disabled. If the OS comes up, you'll know your problem is ACPI-related.
So how do I get it from ddb{0}> to boot>? Typing '-c' the way it is now just returns a question mark.
Reply With Quote
  #7   (View Single Post)  
Old 15th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

The boot> prompt appears before the kernel is executed, you'll only see a ddb> prompt after the kernel has loaded.. you can reboot the machine from ddb using "boot reboot" as documented in the ddb(4) man page.

You haven't included output, simple stating "a kernel page fault" is not adequate.. and you'll need to be far more verbose when reporting the problem to the developers.

As was said, you can try disabling acpi (..although on newer systems, this isn't always a good idea), or try booting the single-processor kernel instead.

Reading on screen instructions is also a good thing to do, as they are displayed when the kernel panics.
Reply With Quote
  #8   (View Single Post)  
Old 15th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

Also, Pentium 4/D are members of the "NetBurst" family of Intel processors.. even if the bug you're experiencing is discovered, the performance of OpenBSD isn't all that great on them.

Intel abandoned this "ground up" redesign of their x86 family due to the performance issues (..cache,soft interrupts,certain instructions), and went back to the P6 architecture (..Pro/PII/PIII) with their new "Core" processors.

If you do wish to use OpenBSD on this system, you should be aware of this.

Also of note, Pentium D does support i386.. if you were unaware of this, you may wish to try OpenBSD/i386, this is on CD1.
Reply With Quote
  #9   (View Single Post)  
Old 16th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

Quote:
Originally Posted by BSDfan666 View Post
The boot> prompt appears before the kernel is executed, you'll only see a ddb> prompt after the kernel has loaded.. you can reboot the machine from ddb using "boot reboot" as documented in the ddb(4) man page.
Okay thanks. I just had to type -c right at the start before it loaded the kernel. I followed jggimi's instructions to disable acpi and now get a protection fault trap, detailed below.

Without ACPI:
Code:
inteldrm0 at vga1kernel: protection fault trap, code=0
Stopped at intr_find_mpmapping+0x2f
Originally, with ACPI:
Code:
uvm_fault(0xffffffff80c9e980, 0x0, 0, 1) -> e
kernel: page fault trap, code=0
Stopped at azalia_set_command+0x25: movq 0x68(%rdi),%rsi
If you need much more info than that, I could just film myself on the desktop, as it would be easier for me (and possibly all of you helping me) than me writing out everything.

Also, there doesn't seem to be multiquote on this forum, but I may try that i386 idea if you think it has a high likelihood of working. I just don't want to uninstall what I have now.
Reply With Quote
Old 16th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

It is possible, AMD was the first to come up with the 64-bit extension to x86.. and Intel cloned it, I know that some early Intel x86-64 processors weren't entirely compatible with the AMD64 implementation.

So perhaps OpenBSD is having issues because of that, trying out OpenBSD/i386 would at least exclude that as a possibility.

Taking pictures is apparently one of the only options on modern systems (..without a serial console), so consider doing that.. none of us here are the developers, so try sending an email.

One thing that hasn't came up is bad memory, as a last ditch effort.. try obtaining & running something called "memtest86+", I don't believe this to be the issue but it surely is possible.

http://www.openbsd.org/report.html
http://www.openbsd.org/mail.html
Reply With Quote
Old 16th May 2010
jggimi's Avatar
jggimi jggimi is offline
More noise than signal
 
Join Date: May 2008
Location: USA
Posts: 7,975
Default

Since the basic installation completed, but the OS failed to function, you don't have much installed yet. Nothing too much to a reinstall, but five minutes of your time. the 32-bit OS will probably outperform the 64-bit for most applications on that platform anyway.
Reply With Quote
Old 16th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

Quote:
Originally Posted by jggimi View Post
Since the basic installation completed, but the OS failed to function, you don't have much installed yet. Nothing too much to a reinstall, but five minutes of your time. the 32-bit OS will probably outperform the 64-bit for most applications on that platform anyway.
I guess it depends, as mentioned, the NetBurst family has issues with software interrupts.. but OpenBSD/amd64 implements support for the SYSCALL/SYSRET fast system call instructions which would alleviate that.

Didn't really consider that he did successfully install it using a 64-bit RAMDISK kernel, so the crashes may not be due to some early Intel bug.

Testing the i386 kernels would help confirm if it's some sort of ACPI error, if that's the case.. perhaps he can check and see if any BIOS updates are available.
Reply With Quote
Old 18th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

I made a video of me installing the i386 over the existing amd64 installation and then booting both with and then without ACPI enabled. I have a description next to the video that you can read as well. Hopefully this will give you a much better idea of the issue.

http://www.youtube.com/watch?v=IcGeFykSiLQ
Reply With Quote
Old 18th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

Not everyone has access to the content on YouTube, perhaps uploading this to some user space you have available elsewhere.. or simply taking photos of the boot messages, panic, and the output of trace/ps as previously suggested.

Sending such a link to the mailing lists is also unwise, pictures would be far more acceptable (..not as attachments, hosted externally, or even here on the forums).

Have you attempted to boot a single-processor kernel yet?
Reply With Quote
Old 18th May 2010
BSDfan666 BSDfan666 is offline
Real Name: N/A, this is the interweb.
Banned
 
Join Date: Apr 2008
Location: Ontario, Canada
Posts: 2,223
Default

I managed to grab a copy of the video using an extension for the Mozilla Firefox browser, had you have reduced the size of your home partition the duration of the video would have been shorter.

The installation proceeds properly, but as I stated, after you type reboot and the BIOS post "splash" finishes, it boots the 'bsd' kernel.

Since at 4.7, the installer automatically uses 'bsd.mp' and renames the single-processor kernel to 'bsd.sp'.

While the boot> prompt is visible, type boot /bsd.sp and see if it continues to panic.

The first panic does seem to be occurring sometime while initializing the audio hardware, azalia(4), If the single-processor kernel fails at the same point, try booting into UKC and disabling 'azalia'.

The second panic is related to there being no alternative legacy MP/multiprocessor mappings for pre-ACPI OS's (..hence my suggestion to boot without utilizing the other cores).

Try the single processor kernel:
boot> boot /bsd.sp
...
Or enter kernel config:
boot> boot -c
...
UKC> disable azalia

If this works, you won't have sound.. but at least you'll have a working system, and form there can help the developers (..not us) provide additional clues.

You might also want to try burning a snapshot release to a CD and seeing if this has been fixed in -CURRENT yet.

Good luck.

Last edited by BSDfan666; 18th May 2010 at 04:40 AM.
Reply With Quote
Old 18th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

Quote:
Originally Posted by BSDfan666 View Post
While the boot> prompt is visible, type boot /bsd.sp and see if it continues to panic.
It works! Thanks. I'll send a formal report to the developers tomorrow, as 4.7 will be officially released and they will be eager to read it.

Thanks again for all your quick replies BSDfan666 and jggimi. I appreciate your time.
Reply With Quote
Old 20th May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

By the way, does multi-processor (bsd.mp) mean multi-core as well or just actual multi-processor? I have a dual-core, single-processor machine and am thinking maybe the error came from me only having one processor as the bsd.sp worked fine. I'm guessing bsd.mp is for dual-quad core servers that have actual multiple processors and not just cores.
Reply With Quote
Old 20th May 2010
ocicat ocicat is offline
Administrator
 
Join Date: Apr 2008
Posts: 3,318
Default

Quote:
Originally Posted by rpindy View Post
...does multi-processor (bsd.mp) mean multi-core as well or just actual multi-processor?
As far as the kernel is concerned, there is no difference between multiple cores & multiple processors. They are treated similarly in OpenBSD.

If your question then becomes, "Why did the single-processor kernel work & the multiprocessor kernel panic?", the answer is not clear at this moment. Debugging via the ddb(0)> prompt by someone experienced with the kernel's internals may reveal the root cause.
Reply With Quote
Old 21st May 2010
rpindy rpindy is offline
Fdisk Soldier
 
Join Date: May 2010
Posts: 59
Default

Thanks ocicat. I'll send a report to the developers. By the way, when I type "show panic" at the ddb{0}> prompt, it says "the kernel did not panic" despite there being a kernel page fault trap. Maybe it's a more generic error, but in any case the developers should be able to help me from here.
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
Downloading free OpenBSD issue karolina OpenBSD General 1 19th June 2009 10:36 AM
WindowMaker 0.92.0p7 (OpenBSD 4.4/i386 Packages) configuration issue. xixobrax OpenBSD General 1 3rd May 2009 04:04 PM
Possible SMP Issue? MetalHead OpenBSD General 1 25th November 2008 03:52 AM
How to Fix Security Issue In OpenBSD 4.1 Stable ? openbsdspirit OpenBSD General 4 21st June 2008 11:33 AM
RAM issue nikkon FreeBSD General 5 7th May 2008 04:26 AM


All times are GMT. The time now is 02:46 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