bonjour à tous,

J'ai un petit souci, j'ai depuis qlq temps une nouvelle machine: Asus
Barbone, carte Asus, PentiumIV 3.2Gh Hyper-Threading.
Lorsque j'active l'Hyper-Threading (via SMP), ça fonctionne parfaitement
jusqu'à ce que je mette ma carte wireless en route, c'est une E-Tech WGPI03
avec chipset Ralink. Dés que le module rt2500 est chargé et que la carte
commence ses échanges, le clavier se fige puis la souris. Et là, c'est la
fin, la seule solution est le reset.
Avec un kernel non SMP, je n'ai pas de souci. J'ai essayé la Slack
(différents kernels: 2.4.29 et 31, 2.6.11, 12 et 13), la mandriva2006 et la
Suse9.3. Toutes réagissent de la même manière: clavier et ensuite souris
figés.

Si quelqu'un a une idée sur ce problème, il est le bienvenu.
Et est-ce que ce bug doit être rapporté aux dev du kernel?

Benoit

Le Sun, 18 Sep 2005 12:40:39 +0200, Cornet B a écrit*:

> bonjour à tous,

Hello Benoit,

Here is an international-pidgin-english group, please avoid
to post in french.

Though I guess your post should be posted in other groups
to get a good chance of some vital hint, I included a translation
of your post at the bottom.
I set a crosspost to alt.fr.os.slackware too, though the
group is far from a vivid one :-)

I don't know your stuff but I'd say that your problem
is probably more in a hardware syzygy than with the software
and or distro. If you coul test your Wireless card in other
computers, and also test your computer with othere Wireless cards,
that would give a clue. And also it could be a problem of overheating
triggered by the HT ...

Now, others may give some real help :-)

> J'ai un petit souci, j'ai depuis qlq temps une nouvelle machine: Asus
> Barbone, carte Asus, PentiumIV 3.2Gh Hyper-Threading.
> Lorsque j'active l'Hyper-Threading (via SMP), ça fonctionne parfaitement
> jusqu'à ce que je mette ma carte wireless en route, c'est une E-Tech WGPI03
> avec chipset Ralink. Dés que le module rt2500 est chargé et que la carte
> commence ses échanges, le clavier se fige puis la souris. Et là, c'est la
> fin, la seule solution est le reset.
> Avec un kernel non SMP, je n'ai pas de souci. J'ai essayé la Slack
> (différents kernels: 2.4.29 et 31, 2.6.11, 12 et 13), la mandriva2006 et la
> Suse9.3. Toutes réagissent de la même manière: clavier et ensuite souris
> figés.
>
> Si quelqu'un a une idée sur ce problème, il est le bienvenu.
> Et est-ce que ce bug doit être rapporté aux dev du kernel?
>
> Benoit

Hello,

I have a little problem with this new machine: Asus Barebone
Asus card, P4 3.2GHz HT.

When I activate, thru SMP, the Hyper Threading, all is well
up to the point when I launch my Wireless card. (an E-Tech WGPI03
- Ralink chipset).
Then, as soon as the rt2500 module gets up and the Wireless starts
its I/Os, the keyboard freezes, then the mouse. Then the only
exit is to reset.

It doesn't happen with a non-SMP kernel, no beef. I tried with
Slackware (different kernels: 2.4.29 and .31, 2.6.11, .12 and .13),
the mandriva2006 and the Suse9.3.
They all have the save big end : KB frozen, then mouse stuck.

If anyone has some hint on this snag, he's welcome.
Should this bug be reported to the kernel developpers ?

Benoit.

Loki Harfagr <loki@DarkDesign.free.fr> wrote:
> When I activate, thru SMP, the Hyper Threading, all is well up to the
> point when I launch my Wireless card. (an E-Tech WGPI03 - Ralink chipset).
> Then, as soon as the rt2500 module gets up and the Wireless starts
> its I/Os, the keyboard freezes, then the mouse. Then the only exit is to
> reset.
>
> It doesn't happen with a non-SMP kernel,

This sounds like the classic mistake to use an SMP kernel together with
non-SMP modules. Was the module compiled with SMP support? Recompiling a
kernel with SMP and/or highmem support also means that all modules has to
be recompiled. All modules includes all modules which belongs to the
kernel as well as other modules which might belong to other packages.

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc7(at)uthyres.com Examples of addresses which go to spammers:
root@variousus.net info@k-soft.se info@k-software.biz root@localhost

Le Mon, 19 Sep 2005 20:06:08 +0200, Henrik Carlqvist a écrit*:

> Loki Harfagr <loki@DarkDesign.free.fr> wrote:

Nope, that was an adaptation from the OP:
From: Cornet B <cornetb@pt.lu>
Message-ID: <432d4427@news.vo.lu>


I fu2 this new adaptation to alt.fr.os.slackware,
maybe next post we'll know if it was not necessary for the OP :-)
(and yes, I know, I Xposted because I fsck'd the previous FU2 ...-|}

>> When I activate, thru SMP, the Hyper Threading, all is well up to the
>> point when I launch my Wireless card. (an E-Tech WGPI03 - Ralink chipset).
>> Then, as soon as the rt2500 module gets up and the Wireless starts
>> its I/Os, the keyboard freezes, then the mouse. Then the only exit is to
>> reset.
>>
>> It doesn't happen with a non-SMP kernel,
>
> This sounds like the classic mistake to use an SMP kernel together with
> non-SMP modules. Was the module compiled with SMP support? Recompiling a
> kernel with SMP and/or highmem support also means that all modules has to
> be recompiled. All modules includes all modules which belongs to the
> kernel as well as other modules which might belong to other packages.
>
> regards Henrik


On dirait l'erreur classique d'utilisation d'un noyau SMP avec des
modules compilés pour mono.

Le module a-t-il été compilé pour SMP ? Recompiler un noyau avec
SMP et/ou HiMem implique de recompiler également tous les modules.
("Tous" veut dire, non seulement ceux inclus avec le noyau
mais *aussi* ceux provenant de packages additionels)

Loki Harfagr wrote:

> Le Mon, 19 Sep 2005 20:06:08 +0200, Henrik Carlqvist a écrit*:
>
>> Loki Harfagr <loki@DarkDesign.free.fr> wrote:
>
> Nope, that was an adaptation from the OP:
> From: Cornet B <cornetb@pt.lu>
> Message-ID: <432d4427@news.vo.lu>
>
>
> I fu2 this new adaptation to alt.fr.os.slackware,
> maybe next post we'll know if it was not necessary for the OP :-)
> (and yes, I know, I Xposted because I fsck'd the previous FU2 ...-|}
>
>>> When I activate, thru SMP, the Hyper Threading, all is well up to the
>>> point when I launch my Wireless card. (an E-Tech WGPI03 - Ralink
>>> chipset).
>>> Then, as soon as the rt2500 module gets up and the Wireless starts
>>> its I/Os, the keyboard freezes, then the mouse. Then the only exit is to
>>> reset.
>>>
>>> It doesn't happen with a non-SMP kernel,
>>
>> This sounds like the classic mistake to use an SMP kernel together with
>> non-SMP modules. Was the module compiled with SMP support? Recompiling a
>> kernel with SMP and/or highmem support also means that all modules has to
>> be recompiled. All modules includes all modules which belongs to the
>> kernel as well as other modules which might belong to other packages.
>>
>> regards Henrik
>
>
> On dirait l'erreur classique d'utilisation d'un noyau SMP avec des
> modules compilés pour mono.
>
> Le module a-t-il été compilé pour SMP ? Recompiler un noyau avec
> SMP et/ou HiMem implique de recompiler également tous les modules.
> ("Tous" veut dire, non seulement ceux inclus avec le noyau
> mais *aussi* ceux provenant de packages additionels)


Excuse me for the french post, it's an error.

For my problem of wireless, I noted that the driver does not support the smp
yet (http://rt2x00.serialmonkey.com/phpBB...pic.php?t=380). That
belonged to the priorities of the developers.
Patience!

And the next time, I will read documentation before putting questions in
French!

Benoit

Le Wed, 21 Sep 2005 20:12:09 +0200, Cornet B a écrit*:

> Loki Harfagr wrote:
>
>> Le Mon, 19 Sep 2005 20:06:08 +0200, Henrik Carlqvist a écrit*:
>>
>>> Loki Harfagr <loki@DarkDesign.free.fr> wrote:
>>
>> Nope, that was an adaptation from the OP:
>> From: Cornet B <cornetb@pt.lu>
>> Message-ID: <432d4427@news.vo.lu>
>>
>>
>> I fu2 this new adaptation to alt.fr.os.slackware,
>> maybe next post we'll know if it was not necessary for the OP :-)
>> (and yes, I know, I Xposted because I fsck'd the previous FU2 ...-|}
>>
>>>> When I activate, thru SMP, the Hyper Threading, all is well up to the
>>>> point when I launch my Wireless card. (an E-Tech WGPI03 - Ralink
>>>> chipset).
>>>> Then, as soon as the rt2500 module gets up and the Wireless starts
>>>> its I/Os, the keyboard freezes, then the mouse. Then the only exit is to
>>>> reset.
>>>>
>>>> It doesn't happen with a non-SMP kernel,
>>>
>>> This sounds like the classic mistake to use an SMP kernel together with
>>> non-SMP modules. Was the module compiled with SMP support? Recompiling a
>>> kernel with SMP and/or highmem support also means that all modules has to
>>> be recompiled. All modules includes all modules which belongs to the
>>> kernel as well as other modules which might belong to other packages.
>>>
>>> regards Henrik
>>
>>
>> On dirait l'erreur classique d'utilisation d'un noyau SMP avec des
>> modules compilés pour mono.
>>
>> Le module a-t-il été compilé pour SMP ? Recompiler un noyau avec
>> SMP et/ou HiMem implique de recompiler également tous les modules.
>> ("Tous" veut dire, non seulement ceux inclus avec le noyau
>> mais *aussi* ceux provenant de packages additionels)
>
>
> Excuse me for the french post, it's an error.

That's allright now :-)

> For my problem of wireless, I noted that the driver does not support the
> smp yet (http://rt2x00.serialmonkey.com/phpBB...pic.php?t=380).
> That belonged to the priorities of the developers. Patience!

Then Henrik spot it first shot !
And yet, some folks still pretend[brawl] this group never helps any
poster...

> And the next time, I will read documentation before putting questions in
> French!

Good move, just hope it's not a "click and a promise" ;-)