Project

General

Profile

Actions

Bug #1270

closed

System hangs after /sbin/init on SMP+APIC enabled kernel

Added by russiane39 about 15 years ago. Updated over 14 years ago.

Status:
Closed
Priority:
High
Assignee:
-
Category:
-
Target version:
-
Start date:
Due date:
% Done:

0%

Estimated time:

Description

System will hang if I compile kernel with SMP+APIC options in kernel config.
System hw specs:
Intel S3210SHLX motherboard
Intel Core2 Quad Q6600 CPU
In attached file you could see output from serial console. Unfortunately but I'm
unable to get a coredump from system.


Files

s3210shlx.txt (25.4 KB) s3210shlx.txt russiane39, 02/06/2009 03:53 AM
s3210shlxps.txt (6 KB) s3210shlxps.txt russiane39, 02/06/2009 10:50 AM
Actions #1

Updated by dillon about 15 years ago

:System will hang if I compile kernel with SMP+APIC options in kernel config=
:.
:System hw specs:
:Intel S3210SHLX motherboard
:Intel Core2 Quad Q6600 CPU
:In attached file you could see output from serial console. Unfortunately bu=
:t I'm
:unable to get a coredump from system.

Well, its booting, but stops after configuring lo0.  Try hitting
^T there instead of escaping into the debugger and see if it tells
you what command it is trying to run.
-Matt
Actions #2

Updated by russiane39 about 15 years ago

I've tried patches from http://bugs.dragonflybsd.org/issue1267 with no luck.
Also, system not responds to ^T.

Actions #3

Updated by sepherosa about 15 years ago

On Fri, Feb 6, 2009 at 1:02 PM, Michael Kosarev (via DragonFly issue
tracker) <> wrote:

Michael Kosarev <> added the comment:

I've tried patches from http://bugs.dragonflybsd.org/issue1267 with no luck.
Also, system not responds to ^T.

From the dmesg, it is not related.

May be you should try 'ps' in ddb?

Best Regards,
sephe

Actions #4

Updated by russiane39 about 15 years ago

Results of ps in enclosed file.

Actions #5

Updated by dillon about 15 years ago

:Michael Kosarev <> added the comment:
:
:Results of ps in enclosed file.

Very odd.  It doesn't look like a hang to me, it looks like fsck might
be running on ad6, which is a 1TB drive. It should be possible to ^C
it. A fsck on a huge UFS partition can take hours.
If you have physical access to the box, can you see if the drive light
continues to show activity after the machine has hung (and not dropping
into ddb)?
If it can boot that far you should be able to boot -s to get into single
user mode and mess around from there... maybe run the fsck's manually
to determine if an actual lockup is occuring or whether it is simply
a long-running fsck.
-Matt
Actions #6

Updated by russiane39 about 15 years ago

Well, HDD LED is not lights, 1Tb hard drive is empty (I mean no partitions), I
can boot system with non-ACPI kernel. It is definitely not a long-running fsck.

Actions #7

Updated by dillon about 15 years ago

:
:Michael Kosarev <> added the comment:
:
:Well, HDD LED is not lights, 1Tb hard drive is empty (I mean no partitions)=
:, I
:can boot system with non-ACPI kernel. It is definitely not a long-running f=
:sck.

Ah, ok.  Well, if you can run with SMP and without ACPI then that is 
how you are going to have to do it. We have always had issues with ACPI
on certain BIOSes.
-Matt
Matthew Dillon
&lt;&gt;
Actions #8

Updated by russiane39 over 14 years ago

Most issues with newer Intel motherboard is resolved with importing of newpci and
polishing MADT code.

Actions

Also available in: Atom PDF