Discussion:
Fréquence du processeur
(trop ancien pour répondre)
JKB
2010-07-31 18:37:35 UTC
Permalink
Bonjour à tous,

Je progresse, je progresse... J'arrive maintenant à bricoler des
choses avec autoconf/automake et tout le toutim sur une machine qui
tourne avec eCS 2.0. Les deux coeurs de processeur sont reconnus
(Pentium E5200 de mémoire) mais j'ai l'affreuse impression que le
processeur en question tourne toujours à fond. Sa température en
fonctionnement est de 56 °C à vide (alors qu'avec un Linux sur la
même machine, il tourne aux alentours de 35 °C). J'ai essayé un
certain nombres d'utilitaires sans trop de succès. Par ailleurs
setgetstate.exe me renvoie une erreur.

Est-ce grave, docteurs ?

Merci de vos lumières,

JKB
--
Le cerveau, c'est un véritable scandale écologique. Il représente 2% de notre
masse corporelle, mais disperse à lui seul 25% de l'énergie que nous
consommons tous les jours.
=> http://grincheux.de-charybde-en-scylla.fr
Guillaume Gay
2010-07-31 20:00:19 UTC
Permalink
Bonsoir;
Post by JKB
Je progresse, je progresse... J'arrive maintenant à bricoler des
choses avec autoconf/automake et tout le toutim
Ça a l'air d'être encourageant !
Post by JKB
(Pentium E5200 de mémoire) mais j'ai l'affreuse impression que le
processeur en question tourne toujours à fond.
Cela ne m'étonnerait pas trop.
Post by JKB
Sa température en fonctionnement est de 56 °C à vide (alors
qu'avec un Linux sur la même machine, il tourne aux alentours
de 35 °C).
Je me demande si le "pilote" ACPI permet véritablement de contrôler
correctement le fonctionnement du processeur. Il semblerait qu'il ait été
plus testé sur des Intel que des AMD. Il semblerait qu'il y ait des
problèmes différents en fonction des versions. Je n'ai malheureusement pas
encore installé eCS 2.0 sur une machine multi-processeurs.
Post by JKB
J'ai essayé un certain nombres d'utilitaires sans trop de succès.
Par ailleurs setgetstate.exe me renvoie une erreur.
Tout ce que j'ai pu trouvé à propos de processeur qui s'emballe (c'est pas
cher...), c'est ça (et c'est pour l'AMD) et peut-être même que tu l'as déjà
lu :

======================================================================
From: "Doug Bissett" <*********@***********>
Newsgroups: ecomstation.support.hardware
Subject: Re: eCS 2.0, ACPI 3.18 & CPU throttling.
Date: Sat, 26 Jun 2010 16:57:18 GMT
Post by JKB
Does anyone know whether CPU throttling is supposed to work
with the AMD64 X4 9650?
Yes it is supposed to work (but may not because the ACPI support has,
so far, totally ignored AMD processors). <see below>
Post by JKB
Running "SetGetThrtl" shows:-
...snip...
Post by JKB
Trying "SetGetThrtl 1 0"/"SetGetThrtl 1 100" does nothing, i.e.
running "SetGetThrtl" again still shows the above.
Wrong commands...
Post by JKB
Running "SetGetThrtl 1 50" stops the CPU requiring removal of power
to to get the PC working again.
That command works for me. <see below>.
Post by JKB
All obvious settings in the BIOS seem to be set & it does work when
booted to Windows or Linux.
The motherboard is an ASUS M3A78-EM.
Same as mine, and mine is also a quad core 9650 processor.
Post by JKB
Thanks in advance for any info.
Dave
First, what version of ACPI are you using? I am using what comes with
eCS 2.0 GA (3.18). I remember having troubles like what you describe,
some time ago, but I don't remember which version of ACPI I was using.
With the current version throttling seems to be working, but precious
little else works with ACPI.

FWIW, I am working on a program to do some power control, using the
ACPI controls. It will be using the programs:
SETGETTHRTL.EXE
SETGETSTATE.EXE
ACPIBATTERY.EXE
along with a few other things to attempt to use a low(er) power
setting, to reduce heat and power usage. It is really aimed at
laptops, but should work, in a more limited way, with desktops too.

The specifics of what I have found out about the M3A78-EM motherboard,
so far, are these:
SETGETSTATE.EXE doesn't work (always rc=1).
ACPIBATTERY.EXE is, of course not useful, since there is no battery.
SETGETTHRTL.EXE does seem to work, with the ACPI from eCS 2.0 GA.

You have a quad core system, and some multi CPU systems will throtle
individual CPUs, but this setup supports throttling only on CPU 1,
and, apparently, that throttles all of them.

From the CPU.EXE program (from the CPUSPEED1.4 package at HOBBES - not
to be confused with the CPU.EXE program that is in the \eCS\BIN
directory), I get this information:
======================
Using CPU 1 for Throttling.

Minimum Scanned Speed is 13 %
SPEED SHIFT detected for 25%, 37%, 50%, 62%, 75%, 87%, 0%,

Speed Steps to be used are: 13 %
Speed Steps to be used are: 25 %
Speed Steps to be used are: 38 %
Speed Steps to be used are: 50 %
Speed Steps to be used are: 63 %
Speed Steps to be used are: 75 %
Speed Steps to be used are: 88 %
Speed Steps to be used are: 100 %


Throttling steps initialized.
This CPU supports 8 stpes for throttling.

minimum: 13 %, maximum: 100 %

Currently running at 87 %
=======================
The important part here is the list of Speed Steps. Those are really
the only valid numbers to use. When you set it to 100, throttling is
disabled. I have never seen that program step down below 87%. by
itself.

After doing some experimenting, I have found that the most useful
setting is at 75%. It keeps the heat down, and the performance doesn't
seem to be affected enough to notice. Here is the text shown:
====================
[C:\ecs\acpi\util]setgetthrtl 1 75
Searching processors
------------------------------------
Processor [P001] id:1 Possible for throttle setting
------------------------------------
Processor [P002] id:2
------------------------------------
Processor [P003] id:3
------------------------------------
Processor [P004] id:4
------------------------------------
Processor [P005] id:5
------------------------------------
Processor [P006] id:6
------------------------------------
Trottling to 1 set succesfull

[C:\ecs\acpi\util]setgetthrtl
Searching processors
------------------------------------
Processor [P001] id:1 Possible for throttle setting
------------------------------------
Processor [P002] id:2
------------------------------------
Processor [P003] id:3
------------------------------------
Processor [P004] id:4
------------------------------------
Processor [P005] id:5
------------------------------------
Processor [P006] id:6
------------------------------------
CPU number 1
ProcId: 0x1
P_BLK: 0x810
P_BLK len: 0x6
Min set is 13
CPU1 Perfomance=75 %
------------------------------------
==========================

You will notice, that it lists 6 processors. I have no idea why, and
the last two seem to be dummy entries, and can be ignored.

The bottom line is that it will work, but you do need the very latest
"official" ACPI version.

From my research, it seems that there are many different ways that
power control is/can be done. No wonder the ACPI development team is
having so much trouble.

Hope this helps...
--
From the eComStation of Doug Bissett
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)
======================================================================
======================================================================
From: "Doug Bissett" <*********@***********>
Newsgroups: ecomstation.support.hardware
Subject: Re: eCS 2.0, ACPI 3.18 & CPU throttling.
Date: Sun, 27 Jun 2010 16:13:51 GMT
Post by JKB
Hi Doug,
Thanks for the info, I didn't know about the cpuspeed package.
My results are unfortunately different from yours.
First, CPU.EXE without any arguments runs as far as stating:-
Minimum Scanned Speed is 13%
and then halts the CPU requiring a power off for about a minute
to get it to restart.
Running "CPU 63 100" displays correctly but doesn't appear to do
anything. Running SetGetThrtl in another session states that
throttling is disabled.
That is not the right command. You are telling CPU number 63 to
throttle to 100%. You need to tell CPU1 to throttle to something
valid. Try this command:
SETGETTHRTL 1 75
and it should throttle all of your processors to 75%. Of course, if
you don't have the proper version of ACPI, properly installed, it may
not do what it is supposed to do. You will see "throttling is
disabled" when it is set to 100%.
Post by JKB
Using SetGetThrtl with the figures you show gives the following:-
SetGetThrtl 1 [63|75|88], the program runs ok but it does not appear
to do anything. Running SetGetThrtl now without any arguments shows
CPU1 Performance as expected however the fan speed does not change &
\ecs\bin\cpuspeed still shows
No. of CPUs: 4
CPU 1: 2310 MHz
CPU 2: 2310 MHz
CPU 3: 2310 MHz
CPU 4: 2310 MHz
i.e. no change.
Throttling does not change the CPU speed. It also does not really make
a lot of difference to the temperature, compared to changing the CPU
speed. Unfortunately, SETGETSTATE does not work on the M3A78-EM
motherboard. SETGETSTATE is what changes the CPU speed. Your fans may
not change speed anyway, unless you have the BIOS set to use the
Smart-Q Function, and have all of the fans connected to controled
power connectors (mine has one 120 mm fan that is connected directly
to the 12 volts, using an adapter that connects to one of the disk/CD
power connectrors). in the HW Monitor section. On my system, the only
real indication that throttling is actually doing anything, is thet
running SETGETTHRTL.EXE with no parameters tells me what it is set to,
and I notice that the hot air coming from the exhaust fan is not
nearly as hot as it is when it runs at 100%. The temperature drop can
take a while, because the internal components take a while to cool
off. Unfortunately, the THERMAL program also doesn't work on these
motherboards, so you really can't monitor the internal temperature
while running.
Post by JKB
Using SetGetThrtl or CPU.EXE with any other percentages halts the
CPU.
Which version of ACPI do you have installed? How did you install it?
Do you have any traces of older versions on your system? Older
versions did freeze the system. The version that comes with eCS 2.0 GA
does not (at least for me). I did a clean install of eCS 2.0.
Post by JKB
So, since we have the same board & processor what is the difference?
I have "PSD=ACPI.PSD /SMP /PIC" in config.sys. /APIC won't boot.
Yes, that is what I found too.
Post by JKB
What else could affect it?
See above...
Post by JKB
BIOS date is 08-28-08.
Hmmm. That may have something to do with it. I am using the 2003
version of BIOS, dated Oct. 29 2009 (2003 is the version number, not
the date). It was the latest version, last time I looked, a couple of
months ago.
--
From the eComStation of Doug Bissett
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)
======================================================================
Sinon, à part taper dans les options du "pilote" ACPI (dans le CONFIG.SYS),
ou du moins en testant eCS en mode mono-processeur (pour voir si c'est de
là que ça vient) je ne vois pas trop où chercher. Désolé.

En hopant que ça helpera.
Bonne fin de ouikenne !
--
Guillaume Gay [à] bigfoot . com [ Anti-SPAM : Paf_Le_Bot! ]
=========> Marche à l'OS/2 ! [ « Ich bin ein Merliner »]
=(IRC)=> "GG" sur #eCSfr, #eCS, ... [ irc://irc.ecomstation.nl/ ]
=(IRC)=> "GG_*" sur #netlabs [ irc://irc.freenode.net/ ]
JKB
2010-07-31 20:39:46 UTC
Permalink
Le Sat, 31 Jul 2010 20:00:19 +0000 (UTC),
Post by Guillaume Gay
Bonsoir;
Post by JKB
Je progresse, je progresse... J'arrive maintenant à bricoler des
choses avec autoconf/automake et tout le toutim
Ça a l'air d'être encourageant !
Ouaips, sauf que je n'arrive toujours pas à compiler ni OpenSSL
1.0.0a, ni GSL 1.14. Autant je vais réussir à bricoler OpenSSL,
autant l'erreur de gsl est une erreur franchement étrange de
libtool. D'ailleurs, si quelqu'un avait un tuyau pour GSL, je suis
preneur...
Post by Guillaume Gay
Post by JKB
(Pentium E5200 de mémoire) mais j'ai l'affreuse impression que le
processeur en question tourne toujours à fond.
Cela ne m'étonnerait pas trop.
Lorsque je dis "tourne à fond", c'est 1% d'occupation pour chacun
des deux coeurs et 2,5 GHz de fréquence d'horloge. J'ai récupéré ce
processeur et pour avoir fait des tests, il correspond à au T7500
que j'ai sur un portable. Intel l'a appelé Pentium parce qu'il n'a
que 2 Mo de cache par coeur contrairement au core2duo de même
fréquence.
Post by Guillaume Gay
Post by JKB
Sa température en fonctionnement est de 56 °C à vide (alors
qu'avec un Linux sur la même machine, il tourne aux alentours
de 35 °C).
Je me demande si le "pilote" ACPI permet véritablement de contrôler
correctement le fonctionnement du processeur. Il semblerait qu'il ait été
plus testé sur des Intel que des AMD. Il semblerait qu'il y ait des
problèmes différents en fonction des versions. Je n'ai malheureusement pas
encore installé eCS 2.0 sur une machine multi-processeurs.
Post by JKB
J'ai essayé un certain nombres d'utilitaires sans trop de succès.
Par ailleurs setgetstate.exe me renvoie une erreur.
Tout ce que j'ai pu trouvé à propos de processeur qui s'emballe (c'est pas
cher...), c'est ça (et c'est pour l'AMD) et peut-être même que tu l'as déjà
Je vair relire ça avec attention.

Merci pour le tuyau.

JKB
Post by Guillaume Gay
======================================================================
Newsgroups: ecomstation.support.hardware
Subject: Re: eCS 2.0, ACPI 3.18 & CPU throttling.
Date: Sat, 26 Jun 2010 16:57:18 GMT
Post by JKB
Does anyone know whether CPU throttling is supposed to work
with the AMD64 X4 9650?
Yes it is supposed to work (but may not because the ACPI support has,
so far, totally ignored AMD processors). <see below>
Post by JKB
Running "SetGetThrtl" shows:-
...snip...
Post by JKB
Trying "SetGetThrtl 1 0"/"SetGetThrtl 1 100" does nothing, i.e.
running "SetGetThrtl" again still shows the above.
Wrong commands...
Post by JKB
Running "SetGetThrtl 1 50" stops the CPU requiring removal of power
to to get the PC working again.
That command works for me. <see below>.
Post by JKB
All obvious settings in the BIOS seem to be set & it does work when
booted to Windows or Linux.
The motherboard is an ASUS M3A78-EM.
Same as mine, and mine is also a quad core 9650 processor.
Post by JKB
Thanks in advance for any info.
Dave
First, what version of ACPI are you using? I am using what comes with
eCS 2.0 GA (3.18). I remember having troubles like what you describe,
some time ago, but I don't remember which version of ACPI I was using.
With the current version throttling seems to be working, but precious
little else works with ACPI.
FWIW, I am working on a program to do some power control, using the
SETGETTHRTL.EXE
SETGETSTATE.EXE
ACPIBATTERY.EXE
along with a few other things to attempt to use a low(er) power
setting, to reduce heat and power usage. It is really aimed at
laptops, but should work, in a more limited way, with desktops too.
The specifics of what I have found out about the M3A78-EM motherboard,
SETGETSTATE.EXE doesn't work (always rc=1).
ACPIBATTERY.EXE is, of course not useful, since there is no battery.
SETGETTHRTL.EXE does seem to work, with the ACPI from eCS 2.0 GA.
You have a quad core system, and some multi CPU systems will throtle
individual CPUs, but this setup supports throttling only on CPU 1,
and, apparently, that throttles all of them.
From the CPU.EXE program (from the CPUSPEED1.4 package at HOBBES - not
to be confused with the CPU.EXE program that is in the \eCS\BIN
======================
Using CPU 1 for Throttling.
Minimum Scanned Speed is 13 %
SPEED SHIFT detected for 25%, 37%, 50%, 62%, 75%, 87%, 0%,
Speed Steps to be used are: 13 %
Speed Steps to be used are: 25 %
Speed Steps to be used are: 38 %
Speed Steps to be used are: 50 %
Speed Steps to be used are: 63 %
Speed Steps to be used are: 75 %
Speed Steps to be used are: 88 %
Speed Steps to be used are: 100 %
Throttling steps initialized.
This CPU supports 8 stpes for throttling.
minimum: 13 %, maximum: 100 %
Currently running at 87 %
=======================
The important part here is the list of Speed Steps. Those are really
the only valid numbers to use. When you set it to 100, throttling is
disabled. I have never seen that program step down below 87%. by
itself.
After doing some experimenting, I have found that the most useful
setting is at 75%. It keeps the heat down, and the performance doesn't
====================
[C:\ecs\acpi\util]setgetthrtl 1 75
Searching processors
------------------------------------
Processor [P001] id:1 Possible for throttle setting
------------------------------------
Processor [P002] id:2
------------------------------------
Processor [P003] id:3
------------------------------------
Processor [P004] id:4
------------------------------------
Processor [P005] id:5
------------------------------------
Processor [P006] id:6
------------------------------------
Trottling to 1 set succesfull
[C:\ecs\acpi\util]setgetthrtl
Searching processors
------------------------------------
Processor [P001] id:1 Possible for throttle setting
------------------------------------
Processor [P002] id:2
------------------------------------
Processor [P003] id:3
------------------------------------
Processor [P004] id:4
------------------------------------
Processor [P005] id:5
------------------------------------
Processor [P006] id:6
------------------------------------
CPU number 1
ProcId: 0x1
P_BLK: 0x810
P_BLK len: 0x6
Min set is 13
CPU1 Perfomance=75 %
------------------------------------
==========================
You will notice, that it lists 6 processors. I have no idea why, and
the last two seem to be dummy entries, and can be ignored.
The bottom line is that it will work, but you do need the very latest
"official" ACPI version.
From my research, it seems that there are many different ways that
power control is/can be done. No wonder the ACPI development team is
having so much trouble.
Hope this helps...
--
From the eComStation of Doug Bissett
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)
======================================================================
======================================================================
Newsgroups: ecomstation.support.hardware
Subject: Re: eCS 2.0, ACPI 3.18 & CPU throttling.
Date: Sun, 27 Jun 2010 16:13:51 GMT
Post by JKB
Hi Doug,
Thanks for the info, I didn't know about the cpuspeed package.
My results are unfortunately different from yours.
First, CPU.EXE without any arguments runs as far as stating:-
Minimum Scanned Speed is 13%
and then halts the CPU requiring a power off for about a minute
to get it to restart.
Running "CPU 63 100" displays correctly but doesn't appear to do
anything. Running SetGetThrtl in another session states that
throttling is disabled.
That is not the right command. You are telling CPU number 63 to
throttle to 100%. You need to tell CPU1 to throttle to something
SETGETTHRTL 1 75
and it should throttle all of your processors to 75%. Of course, if
you don't have the proper version of ACPI, properly installed, it may
not do what it is supposed to do. You will see "throttling is
disabled" when it is set to 100%.
Post by JKB
Using SetGetThrtl with the figures you show gives the following:-
SetGetThrtl 1 [63|75|88], the program runs ok but it does not appear
to do anything. Running SetGetThrtl now without any arguments shows
CPU1 Performance as expected however the fan speed does not change &
\ecs\bin\cpuspeed still shows
No. of CPUs: 4
CPU 1: 2310 MHz
CPU 2: 2310 MHz
CPU 3: 2310 MHz
CPU 4: 2310 MHz
i.e. no change.
Throttling does not change the CPU speed. It also does not really make
a lot of difference to the temperature, compared to changing the CPU
speed. Unfortunately, SETGETSTATE does not work on the M3A78-EM
motherboard. SETGETSTATE is what changes the CPU speed. Your fans may
not change speed anyway, unless you have the BIOS set to use the
Smart-Q Function, and have all of the fans connected to controled
power connectors (mine has one 120 mm fan that is connected directly
to the 12 volts, using an adapter that connects to one of the disk/CD
power connectrors). in the HW Monitor section. On my system, the only
real indication that throttling is actually doing anything, is thet
running SETGETTHRTL.EXE with no parameters tells me what it is set to,
and I notice that the hot air coming from the exhaust fan is not
nearly as hot as it is when it runs at 100%. The temperature drop can
take a while, because the internal components take a while to cool
off. Unfortunately, the THERMAL program also doesn't work on these
motherboards, so you really can't monitor the internal temperature
while running.
Post by JKB
Using SetGetThrtl or CPU.EXE with any other percentages halts the
CPU.
Which version of ACPI do you have installed? How did you install it?
Do you have any traces of older versions on your system? Older
versions did freeze the system. The version that comes with eCS 2.0 GA
does not (at least for me). I did a clean install of eCS 2.0.
Post by JKB
So, since we have the same board & processor what is the difference?
I have "PSD=ACPI.PSD /SMP /PIC" in config.sys. /APIC won't boot.
Yes, that is what I found too.
Post by JKB
What else could affect it?
See above...
Post by JKB
BIOS date is 08-28-08.
Hmmm. That may have something to do with it. I am using the 2003
version of BIOS, dated Oct. 29 2009 (2003 is the version number, not
the date). It was the latest version, last time I looked, a couple of
months ago.
--
From the eComStation of Doug Bissett
dougb007 at telus dot net
(Please make the obvious changes, to e-mail me)
======================================================================
Sinon, à part taper dans les options du "pilote" ACPI (dans le CONFIG.SYS),
ou du moins en testant eCS en mode mono-processeur (pour voir si c'est de
là que ça vient) je ne vois pas trop où chercher. Désolé.
En hopant que ça helpera.
Bonne fin de ouikenne !
--
Le cerveau, c'est un véritable scandale écologique. Il représente 2% de notre
masse corporelle, mais disperse à lui seul 25% de l'énergie que nous
consommons tous les jours.
=> http://grincheux.de-charybde-en-scylla.fr
Guillaume Gay
2010-08-01 07:55:09 UTC
Permalink
Bonjour,
Post by JKB
Ouaips, sauf que je n'arrive toujours pas à compiler ni OpenSSL
1.0.0a, ni GSL 1.14. Autant je vais réussir à bricoler OpenSSL,
autant l'erreur de gsl est une erreur franchement étrange de
libtool. D'ailleurs, si quelqu'un avait un tuyau pour GSL, je suis
preneur...
Il y a eu une discussion à propos de OpenSSH sur ecomstation.support.misc.
Le fil de discussion est plutôt long et il s'agit au départ d'un problème
de plantage sous eCS 2.0 en mode SMP.

======================================================================
From: "Paul Smedley" <***********************>
Newsgroups: ecomstation.support.misc
Subject: Re: SMP crash ECS2.0 - Got one
Date: Tue, 20 Jul 2010 09:36:24 GMT
Message-ID: <POn1Xm9ddZOp-pn2-***@smedley.info>
References: <fV45K0OBJxbE-pn2-***@paddington.bear.den>
<c1.2b8.3LlHbJ$***@news.ecomstation.com>
<fV45K0OBJxbE-pn2-***@paddington.bear.den>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<c1.2b8.3Lwy28$***@news.ecomstation.com>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slat42-1.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<POn1Xm9ddZOp-pn2-***@smedley.info>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>
<fV45K0OBJxbE-pn2-***@localhost>
<11p86vVJT4Oe-pn2-***@slamain.slainc.com>

Hi Steven,

On Mon, 19 Jul 2010 20:13:56 UTC, "Steven Levine"
Post by JKB
On Mon, 19 Jul 2010 13:36:15 UTC, "Dave Saville"
One thing I have established is that Paul's published environment is
no longer the same as he uses so some of the problems I see he doesn't
- different compilers/headers etc.
Yes. Since Paul is effectively the gcc maintainer, he is usually
running some form of 4.3.x. The zipped up environment is pretty much
3.3.5.
Not in the case of the build environment available from
http://os2ports.smedley.info/index.php?page=BuildEnvDVD

This includes GCC 3.3.5, 3.4.6, 4.3.x and 4.4.x

Of course literally no two discs are the same, as it's a fresh zip of
what's on my u: drive ;)
Post by JKB
I suspect the message might be comming from the OS/2 shell. I say
this because the messages I see in the openssh code are much more
verbose.
I'd say this is coming from the implementation of openpty() in
bsd-openpty.c - that tries to open '/dev/ptyxx' which of course
fails...
--
Cheers,

Paul.
======================================================================
D'autres informations sont disponibles dans ce fil de discussion.

Cela n'a peut-être rien à voir avec ton problème, mais il semblerait tout
de même que libc ne retourne pas toutes les fonctions attendue par ce
projet. Aussi te conseillerais-je de te mettre en relation avec Paul
Smedley ou Dave Saville (ou les deux) afin d'éviter de réinventer la roue.
Paul a toujours une session IRC ("smedley") ouverte sur #netlabs
(irc://irc.freenode.net/). David ("dcs"), lui, n'est là que très rarement,
mais il peut se rendre disponible pour une discussion, je pense ?
Post by JKB
processeur et pour avoir fait des tests, il correspond à au T7500
que j'ai sur un portable. Intel l'a appelé Pentium parce qu'il n'a
que 2 Mo de cache par coeur contrairement au core2duo de même
fréquence.
Heu ! Je dois avouer que j'ai complètement arrêté d'essayer de comprendre
les dénominations délirantes des gammes de processeurs Intel.

Sur ce, je vous souhaite un "bon dimanche, sous vos applaudissements" !

Oui, bon... Vous pouvez laisser tomber les applaudissements.
Topette !
--
Guillaume Gay [à] bigfoot . com [ Anti-SPAM : Paf_Le_Bot! ]
=========> Marche à l'OS/2 ! [ « Ich bin ein Merliner »]
=(IRC)=> "GG" sur #eCSfr, #eCS, ... [ irc://irc.ecomstation.nl/ ]
=(IRC)=> "GG_*" sur #netlabs [ irc://irc.freenode.net/ ]
JKB
2010-08-01 08:17:46 UTC
Permalink
Le Sun, 1 Aug 2010 07:55:09 +0000 (UTC),
Post by Guillaume Gay
Bonjour,
Post by JKB
Ouaips, sauf que je n'arrive toujours pas à compiler ni OpenSSL
1.0.0a, ni GSL 1.14. Autant je vais réussir à bricoler OpenSSL,
autant l'erreur de gsl est une erreur franchement étrange de
libtool. D'ailleurs, si quelqu'un avait un tuyau pour GSL, je suis
preneur...
Il y a eu une discussion à propos de OpenSSH sur ecomstation.support.misc.
Le fil de discussion est plutôt long et il s'agit au départ d'un problème
de plantage sous eCS 2.0 en mode SMP.
Ce n'est pas une histoire de plantage mais de compilation qui
merdoie. Je m'occuperai de ça après avoir réussi à compiler GSL
1.14. De toute façon, dès que j'aurai trouvé une solution, je ne
manquerai pas d'en faire part.
Post by Guillaume Gay
======================================================================
From: "Paul Smedley" <***********************>
Newsgroups: ecomstation.support.misc
Subject: Re: SMP crash ECS2.0 - Got one
Date: Tue, 20 Jul 2010 09:36:24 GMT
Hi Steven,
On Mon, 19 Jul 2010 20:13:56 UTC, "Steven Levine"
Post by JKB
On Mon, 19 Jul 2010 13:36:15 UTC, "Dave Saville"
One thing I have established is that Paul's published environment is
no longer the same as he uses so some of the problems I see he doesn't
- different compilers/headers etc.
Yes. Since Paul is effectively the gcc maintainer, he is usually
running some form of 4.3.x. The zipped up environment is pretty much
3.3.5.
Not in the case of the build environment available from
http://os2ports.smedley.info/index.php?page=BuildEnvDVD
This includes GCC 3.3.5, 3.4.6, 4.3.x and 4.4.x
Of course literally no two discs are the same, as it's a fresh zip of
what's on my u: drive ;)
Post by JKB
I suspect the message might be comming from the OS/2 shell. I say
this because the messages I see in the openssh code are much more
verbose.
I'd say this is coming from the implementation of openpty() in
bsd-openpty.c - that tries to open '/dev/ptyxx' which of course
fails...
--
Cheers,
Paul.
======================================================================
D'autres informations sont disponibles dans ce fil de discussion.
Cela n'a peut-être rien à voir avec ton problème, mais il semblerait tout
de même que libc ne retourne pas toutes les fonctions attendue par ce
projet. Aussi te conseillerais-je de te mettre en relation avec Paul
Smedley ou Dave Saville (ou les deux) afin d'éviter de réinventer la roue.
Paul a toujours une session IRC ("smedley") ouverte sur #netlabs
(irc://irc.freenode.net/). David ("dcs"), lui, n'est là que très rarement,
mais il peut se rendre disponible pour une discussion, je pense ?
Post by JKB
processeur et pour avoir fait des tests, il correspond à au T7500
que j'ai sur un portable. Intel l'a appelé Pentium parce qu'il n'a
que 2 Mo de cache par coeur contrairement au core2duo de même
fréquence.
Heu ! Je dois avouer que j'ai complètement arrêté d'essayer de comprendre
les dénominations délirantes des gammes de processeurs Intel.
Ah, toi aussi ? J'ai récupéré cette machine chez un client, elle
avait tout juste un an et avait été massacrée. Ce qu'Intel appelle
Pentium Exxxx, c'est juste un core2duo avec un poil moins de cache
de niveau 2. En fait, c'est tout à fait acceptable pour une
utilisation courante.
Post by Guillaume Gay
Sur ce, je vous souhaite un "bon dimanche, sous vos applaudissements" !
Oui, bon... Vous pouvez laisser tomber les applaudissements.
Topette !
Cordialement,

JKB
--
Le cerveau, c'est un véritable scandale écologique. Il représente 2% de notre
masse corporelle, mais disperse à lui seul 25% de l'énergie que nous
consommons tous les jours.
=> http://grincheux.de-charybde-en-scylla.fr
JKB
2010-08-03 14:07:17 UTC
Permalink
Le Sun, 1 Aug 2010 08:17:46 +0000 (UTC),
Post by JKB
Le Sun, 1 Aug 2010 07:55:09 +0000 (UTC),
Post by Guillaume Gay
Bonjour,
Post by JKB
Ouaips, sauf que je n'arrive toujours pas à compiler ni OpenSSL
1.0.0a, ni GSL 1.14. Autant je vais réussir à bricoler OpenSSL,
autant l'erreur de gsl est une erreur franchement étrange de
libtool. D'ailleurs, si quelqu'un avait un tuyau pour GSL, je suis
preneur...
Il y a eu une discussion à propos de OpenSSH sur ecomstation.support.misc.
Le fil de discussion est plutôt long et il s'agit au départ d'un problème
de plantage sous eCS 2.0 en mode SMP.
Ce n'est pas une histoire de plantage mais de compilation qui
merdoie. Je m'occuperai de ça après avoir réussi à compiler GSL
1.14. De toute façon, dès que j'aurai trouvé une solution, je ne
manquerai pas d'en faire part.
A y'est... Ne reste plus qu'à émuler les sémaphores POSIX par
l'horrible semop(). D'ailleurs si quelqu'un avait un bout de code
sous la main pour sem_open, sem_close, sem_post, sem_wait et
sem_getvalue et les quelques autres fonctions, ce serait gentil de
penser à moi ;-)

Les quick and dirty patches sont trop longs à expliquer. Je mettrai
le résultat en ligne dès qu'il compilera jusqu'au bout, ce qui ne
saurait tarder.

Cordialement,

JKB
--
Le cerveau, c'est un véritable scandale écologique. Il représente 2% de notre
masse corporelle, mais disperse à lui seul 25% de l'énergie que nous
consommons tous les jours.
=> http://grincheux.de-charybde-en-scylla.fr
Guillaume Gay
2010-08-03 20:16:11 UTC
Permalink
Bonjour,
Post by JKB
D'ailleurs si quelqu'un avait un bout de code
sous la main pour sem_open, sem_close, sem_post, sem_wait et
sem_getvalue et les quelques autres fonctions, ce serait gentil de
penser à moi ;-)
Far beyond of my knowledge.

Sois le bienvenu sur irc://irc.freenode.net/ canal #netlabs (il existe une
interface web).

En hopant que ça pourra helper.

Sur ce, topette les gars !
--
Guillaume Gay [à] bigfoot . com [ Anti-SPAM : Paf_Le_Bot! ]
=========> Marche à l'OS/2 ! [ « Ich bin ein Merliner »]
=(IRC)=> "GG" sur #eCSfr, #eCS, ... [ irc://irc.ecomstation.nl/ ]
=(IRC)=> "GG_*" sur #netlabs [ irc://irc.freenode.net/ ]
Loading...