Skip to contentSkip to conversation

patch for timer.c - two dmesgs

DISCUSSION Uwe Bugla 3 participants 6 messages Jul 12, 2006 – Jul 13, 2006

lore
Share & tools
lore

Uwe Bugla

author
#0 lore raw

Hi Roman,
Trying to boot kernel 2.6.18-rc1-mm1 on a Siemens-Fujitsu mainbord with an Intel Pentium 3 processor running at 600 Mhz the last message I can read is the following:
Starting PostgreSQL 7.4 database server: main_
(The _ behind main stands for the hanging cursors).

Then the boot process does not take any break at all (like in kernel 2.6.18-rc1 and in kernels 2.6.17-mm*), but simply stops completely. About 7 message lines are missing before X starts for presenting the graphical login prompt (proftpd, xprint etc.). Perhaps two dmesgs help: one for a functionable 2.6.17.4 kernel (dmesg17), another for the kernel in question (dmesg18). This does not happen on two Pentium 4 machines with Asus mainboards and Pentium 4 processors running at 1800 Mhz.

Please see two dmesg attachments.

Regards

Uwe

P. S.: Please a humble request: I'd deeply appreciate that Valdis stops telling me things like: „Odd.. there wasn't anything P4-specific in that patch..“ or „Are you *positive* that your Pentium3 box is hanging up for the same reason? It could be it's something else entirely.“

Simply my intuition tells me that a system timer performs very different on two very different machines with two very different CPU frequencies and two very different main processors. For example absolutely NO Pentium 3 or Pentium 2 processor contains hyperthreading technology, far apart from the totally different CPU frequencies!
In so far it not only could be but simply MUST be a timer.c issue!!

Signature · 2 lines
Echte DSL-Flatrate dauerhaft für 0,- Euro*!
"Feel free" mit GMX DSL! http://www.gmx.net/de/go/dsl
#1 lore raw

Hi,

Quoted 3 lines · Uwe Bugla

Then the boot process does not take any break at all (like in kernel 2.6.18-rc1 and in kernels 2.6.17-mm*), but simply stops completely. About 7 message lines are missing before X starts for presenting the graphical login prompt (proftpd, xprint etc.). Perhaps two dmesgs help: one for a functionable 2.6.17.4 kernel (dmesg17), another for the kernel in question (dmesg18).

A lot has changed since then...
Did you try using SysRq+P or Alt+ScrollLock? (A SysRq+T might be useful too).

Quoted 3 lines

Simply my intuition tells me that a system timer performs very different on two very different machines with two very different CPU frequencies and two very different main processors.

No offense, it's no intuition, but a shot in the dark.
If you at least tried a few more kernels (at least 2.6.18-mm1 and 2.6.18-mm2), it would be a much better indication whether it's the timer patch.

bye, Roman

author
#2 lore raw

-------- Original-Nachricht --------
Datum: Wed, 12 Jul 2006 14:17:04 +0200 (CEST)
Von: Roman Zippel <[email protected]>
An: Uwe Bugla <[email protected]>
Betreff: Re: patch for timer.c - two dmesgs

Quoted 6 lines

Hi,

Quoted 1 line · Uwe Bugla

Then the boot process does not take any break at all (like in kernel

2.6.18-rc1 and in kernels 2.6.17-mm*), but simply stops completely.

Quoted 8 lines
Quoted 1 line

About 7 message lines are missing before X starts for presenting the

graphical login prompt (proftpd, xprint etc.).

Quoted 1 line

Perhaps two dmesgs help: one for a functionable 2.6.17.4 kernel

(dmesg17), another for the kernel in question (dmesg18).

A lot has changed since then...
Did you try using SysRq+P or Alt+ScrollLock? (A SysRq+T might be useful too).

Sorry for this stupid sounding question:
At what point of the boot process do I have to use those keyboard combinations please? And what is the output / product of them please? Sorry if I simply lack experience in those questions.

Quoted 11 lines
Quoted 4 lines

Simply my intuition tells me that a system timer performs very different on two very different machines with two very different CPU frequencies and two very different main processors.

No offense, it's no intuition, but a shot in the dark.
If you at least tried a few more kernels (at least 2.6.18-mm1 and 2.6.18-mm2), it would be a much better indication whether it's the timer patch.

bye, Roman

Signature · 2 lines
Echte DSL-Flatrate dauerhaft für 0,- Euro*!
"Feel free" mit GMX DSL! http://www.gmx.net/de/go/dsl
#3 lore raw

Hi,

Quoted 5 lines · Uwe Bugla
Quoted 3 lines

A lot has changed since then...
Did you try using SysRq+P or Alt+ScrollLock? (A SysRq+T might be useful too).

Sorry for this stupid sounding question:
At what point of the boot process do I have to use those keyboard combinations please? And what is the output / product of them please?

Just press them the boot stops and send the kernel log or wite it down if it's not in the log.

bye, Roman

#5 lore raw
In reply to Uwe Bugla
Quoted 16 lines · Uwe Bugla

-------- Original-Nachricht --------
Datum: Wed, 12 Jul 2006 14:17:04 +0200 (CEST)
Von: Roman Zippel <[email protected]>
An: Uwe Bugla <[email protected]>
Betreff: Re: patch for timer.c - two dmesgs

Quoted 9 lines

Hi,

Quoted 1 line · Uwe Bugla

Then the boot process does not take any break at all (like in kernel

2.6.18-rc1 and in kernels 2.6.17-mm*), but simply stops completely.

Quoted 1 line

About 7 message lines are missing before X starts for presenting the

graphical login prompt (proftpd, xprint etc.).

Quoted 1 line

Perhaps two dmesgs help: one for a functionable 2.6.17.4 kernel

Quoted 9 lines
Quoted 5 lines

(dmesg17), another for the kernel in question (dmesg18).

A lot has changed since then...
Did you try using SysRq+P or Alt+ScrollLock? (A SysRq+T might be useful too).

Sorry for this stupid sounding question:
At what point of the boot process do I have to use those keyboard combinations please? And what is the output / product of them please? Sorry if I simply lack experience in those questions.

Quoted 1 line

This might help : http://sosdg.org/~coywolf/lxr/source/Documentation/sysrq.txt
I can't answer at what point you should get the data though.

Signature · 3 lines
Jesper Juhl <[email protected]>
Don't top-post  http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please      http://www.expita.com/nomime.html
author
#4 lore raw
In reply to Roman Zippel

-------- Original-Nachricht --------
Datum: Wed, 12 Jul 2006 20:41:40 +0200 (CEST)
Von: Roman Zippel <[email protected]>
An: Uwe Bugla <[email protected]>
Betreff: Re: Re: patch for timer.c - two dmesgs

Quoted 8 lines

Hi,

Quoted 2 lines · Uwe Bugla
Quoted 2 lines

A lot has changed since then...
Did you try using SysRq+P or Alt+ScrollLock? (A SysRq+T might be

useful

Quoted 1 line
Quoted 1 line

too).

Quoted 8 lines
Quoted 2 lines

Sorry for this stupid sounding question:
At what point of the boot process do I have to use those keyboard

combinations please? And what is the output / product of them please?

Just press them the boot stops and send the kernel log or wite it down if it's not in the log.

bye, Roman

Thanks, Roman.
To avoid misunderstandings:
I in fact invest more time and energy in testing mm- or other kernels than any comparable user or admin. I in fact tested 3 or 4 different ones somewhere between 2.6.17 and 2.6.18-rc1-mm1, but got exhausted and demotivated as the only output was inability to boot. I informed Andrew as quick as possible. And I never in my life stumbled over kernels behaving like this. In so far as a consequence I am missing tools or tricks how to produce system output for constructive feedback.
Question: What difference is there between the output of dmesg when kernel stops out of inconvenience or just out of containing buggy code on the one hand and the output of SysRq+P or Alt+ScrollLock on the other hand?

Regards
Uwe

Signature · 2 lines
Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
Ideal für Modem und ISDN: http://www.gmx.net/de/go/smartsurfer