Skip to contentSkip to conversation

2.6.25-rc1 does not boot on Alpha

REGRESSION Bob Tracy 4 participants 10 messages Feb 13, 2008 – Feb 14, 2008

lore
Share & tools
lore

Bob Tracy

author
#0 lore raw

This isn't going to be terribly useful other than giving someone a heads-up there's a problem with something in 2.6.25-rc1 on the Alpha PWS 433au. I get the usual messages out of "aboot", including

aboot: zero-filling 210392 bytes at 0xfffffc0000776740
aboot: starting kernel vmlinux.new.gz with argument ro root=/dev/sda3

and then the screen background switches to red and the machine locks up solid. No further console output. Completely reproducible. The 2.6.24 kernel is fine.

Unless someone has a good idea what's causing this, I see a massive "git bisect" project in my future. Won't be able to spend any time with this until at least the weekend :-(.

Signature · 4 lines
------------------------------------------------------------------------
Bob Tracy          |  "I was a beta tester for dirt.  They never did
[email protected]       |   get all the bugs out." - Steve McGrew on /.
------------------------------------------------------------------------
#1 lore raw
Quoted 7 lines · Bob Tracy

This isn't going to be terribly useful other than giving someone a heads-up there's a problem with something in 2.6.25-rc1 on the Alpha PWS 433au. I get the usual messages out of "aboot", including

aboot: zero-filling 210392 bytes at 0xfffffc0000776740 aboot: starting kernel vmlinux.new.gz with argument ro root=/dev/sda3

Quoted 7 lines

and then the screen background switches to red and the machine locks up solid. No further console output. Completely reproducible. The 2.6.24 kernel is fine.

Unless someone has a good idea what's causing this, I see a massive "git bisect" project in my future. Won't be able to spend any time with this until at least the weekend :-(.

hm, is this SMP? Alpha uses asm-generic/percpu.h. One "early bootup" effect on Alpha that i can think of could have been this family of commits:

commit acdac87202a408133ee8f7985076de9d2e0dc5ab
Author: [email protected] <[email protected]>
Date: Wed Jan 30 13:32:52 2008 +0100

percpu: make the asm-generic/percpu.h more "generic"

you wont be able to revert this patch cleanly, perhaps try to compile commit cf8fa920cb4271f17e0265c863d64bea1b31941a and see whether it boots. (that tree does not have the percpu changes yet)

if that doesnt help then i guess it's the bisection of 9000+ commits :-/

Ingo

#2 lore raw
Quoted 7 lines · Ingo Molnar
Quoted 3 lines

Unless someone has a good idea what's causing this, I see a massive "git bisect" project in my future. Won't be able to spend any time with this until at least the weekend :-(.

hm, is this SMP? Alpha uses asm-generic/percpu.h. One "early bootup" effect on Alpha that i can think of could have been this family of commits:

another possibility would be to turn off PRINTK_TIMESTAMPS [if they are on in your .config].

Ingo

#3 lore raw
Quoted 10 lines · Ingo Molnar
Quoted 7 lines
Quoted 3 lines

Unless someone has a good idea what's causing this, I see a massive "git bisect" project in my future. Won't be able to spend any time with this until at least the weekend :-(.

hm, is this SMP? Alpha uses asm-generic/percpu.h. One "early bootup" effect on Alpha that i can think of could have been this family of commits:

another possibility would be to turn off PRINTK_TIMESTAMPS [if they are on in your .config].

No SMP, no PRINTK_TIMESTAMPS in my case. Looks like it dies trying to to switch to vga console, but I had no time to debug this yet...

Ivan.

author
#4 lore raw
Quoted 2 lines · Ivan Kokshaysky

No SMP, no PRINTK_TIMESTAMPS in my case. Looks like it dies trying to to switch to vga console, but I had no time to debug this yet...

Same basic configuration as Ivan. Concur with the observation that death seems to happen during the console switch. Using the radeonfb driver (built-in, default graphics mode 80x30).

Signature · 4 lines
------------------------------------------------------------------------
Bob Tracy          |  "I was a beta tester for dirt.  They never did
[email protected]       |   get all the bugs out." - Steve McGrew on /.
------------------------------------------------------------------------
#5 lore raw
Quoted 7 lines · Bob Tracy
Quoted 2 lines · Ivan Kokshaysky

No SMP, no PRINTK_TIMESTAMPS in my case. Looks like it dies trying to to switch to vga console, but I had no time to debug this yet...

Same basic configuration as Ivan. Concur with the observation that death seems to happen during the console switch. Using the radeonfb driver (built-in, default graphics mode 80x30).

As Peter pointed out in other mail, this is fixed by

http://lkml.org/lkml/2008/1/28/100

Ivan.

author
#6 lore raw
Quoted 5 lines · Ivan Kokshaysky
Quoted 4 lines · Bob Tracy
Quoted 2 lines · Ivan Kokshaysky

No SMP, no PRINTK_TIMESTAMPS in my case. Looks like it dies trying to to switch to vga console, but I had no time to debug this yet...

Quoted 7 lines
Quoted 3 lines

Same basic configuration as Ivan. Concur with the observation that death seems to happen during the console switch. Using the radeonfb driver (built-in, default graphics mode 80x30).

As Peter pointed out in other mail, this is fixed by

http://lkml.org/lkml/2008/1/28/100

ACK. I'll confirm it works for me this evening. Good catch, all.

Signature · 4 lines
------------------------------------------------------------------------
Bob Tracy          |  "I was a beta tester for dirt.  They never did
[email protected]       |   get all the bugs out." - Steve McGrew on /.
------------------------------------------------------------------------
author
#7 lore raw
In reply to Ivan Kokshaysky
Quoted 5 lines · Ivan Kokshaysky
Quoted 4 lines · Bob Tracy
Quoted 2 lines · Ivan Kokshaysky

No SMP, no PRINTK_TIMESTAMPS in my case. Looks like it dies trying to to switch to vga console, but I had no time to debug this yet...

Quoted 7 lines
Quoted 3 lines

Same basic configuration as Ivan. Concur with the observation that death seems to happen during the console switch. Using the radeonfb driver (built-in, default graphics mode 80x30).

As Peter pointed out in other mail, this is fixed by

http://lkml.org/lkml/2008/1/28/100

Confirmed. This fixes 2.6.25-rc1 for me. I'm running on it as I type this.

Signature · 4 lines
------------------------------------------------------------------------
Bob Tracy          |  "I was a beta tester for dirt.  They never did
[email protected]       |   get all the bugs out." - Steve McGrew on /.
------------------------------------------------------------------------
#8 lore raw
In reply to Bob Tracy
Quoted 7 lines · Bob Tracy

This isn't going to be terribly useful other than giving someone a heads-up there's a problem with something in 2.6.25-rc1 on the Alpha PWS 433au. I get the usual messages out of "aboot", including

aboot: zero-filling 210392 bytes at 0xfffffc0000776740
aboot: starting kernel vmlinux.new.gz with argument ro root=/dev/sda3

Quoted 7 lines

and then the screen background switches to red and the machine locks up solid. No further console output. Completely reproducible. The 2.6.24 kernel is fine.

Unless someone has a good idea what's causing this, I see a massive "git bisect" project in my future. Won't be able to spend any time with this until at least the weekend :-(.

Does: http://lkml.org/lkml/2008/1/28/100

help?

author
#9 lore raw
Quoted 3 lines · Peter Zijlstra

Does: http://lkml.org/lkml/2008/1/28/100

help?

I'll give that a try this evening or tomorrow and let you know.

Signature · 4 lines
------------------------------------------------------------------------
Bob Tracy          |  "I was a beta tester for dirt.  They never did
[email protected]       |   get all the bugs out." - Steve McGrew on /.
------------------------------------------------------------------------