FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Kernel

 
 
LinkBack Thread Tools
 
Old 09-02-2012, 08:47 PM
"Hans-J. Ullrich"
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

Am Sonntag, 2. September 2012 schrieb Jonathan Nieder:
> tags 649253 + moreinfo
> quit
>
> Hi Hans,
>
Hi Jonathan,

I will be pleased to answer.

> Hans-J. Ullrich wrote:
> > [Subject: Bug can be closed]
>
> Please keep in mind that these appear as emails in a crowded inbox, so
> the subject line can be a good place to put valuable context.
>
> [...]
>
> > as there is no hardwaresupport on this old device, I suggest, to close
> > this bug. I am running the old kernel module with the workaround
> > "acpi="Linux", which is working like a charm.
>
> I'd rather see such bugs fixed so others do not have to suffer from
> them.
>


Just to explain: eeepc-wmi will not be fixed, as the developers told me (almost
2 years ago). The problem of this, is that the old hardware specs of my EEEPC
1005 HGO are not made available by the vendor. So it is of course impossible,
to write a kernel module.

All newer hardware, which is in most newer processors will be supported, of
course! But as I said, my EEEPC 1005 HGO with the old single core cpu (Intel
N280), still got the old and undocumentd API.
> What is the newest kernel you've experienced this bug with? What
> other kernels have you tested and what happened with each --- is it a
> regression?
>
I am running now linux-image-3.2.0-3-i386 (stock debian kernel of testing).

> What steps exactly do you use to work around the bug? What do you
> mean by "running the old kernel module"? Are you using an older
> kernel?
>

No, I can choose either to load eeepc-wmi or use the option "acpi_osi="Linux"
which let me load eeepc-laptop at startup. The second one I call the "old"
module, as this is the old fancontroller module developed for my old cpu.

If eeepc-laptop is used, then eeepc-wmi will not be loaded.

> I imagine upstream would want to know about this, so please do test a
> 3.5.y kernel from experimental and let us know the result if you have
> time. Output from "reportbug --template linux-image-$(uname -r)" and
> "acpidump" and full "dmesg" output from an affected boot (as
> attachments) would be useful as well.
>

Of course, I will be pleased to test, but I doubt, that it will work. The
developers of the modiule "eeepc-wmi" wrote me a mail, where they explained,
that they will not continue further developing of eeepc-wmi. Not because they
do not want to, but because they can't - because missing informations of the
hardware API!

So I changed to eeepc-laptop, which is working fine for me. And for these
reasons I claimed this bugreport closable.

However, I will test a newer kernel (3.5-X) in the next days, maybe not
tomorrow, but during the coming week. Response and report will then be sent.


> Hope that helps,
> Jonathan

Thank you for your help!

Best regards

Hans


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201209022247.02111.hans.ullrich@loop.de">http://lists.debian.org/201209022247.02111.hans.ullrich@loop.de
 
Old 09-03-2012, 02:11 AM
Ben Hutchings
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

On Sun, 2012-09-02 at 22:47 +0200, Hans-J. Ullrich wrote:
[...]
> Just to explain: eeepc-wmi will not be fixed, as the developers told me (almost
> 2 years ago). The problem of this, is that the old hardware specs of my EEEPC
> 1005 HGO are not made available by the vendor. So it is of course impossible,
> to write a kernel module.
>
> All newer hardware, which is in most newer processors will be supported, of
> course! But as I said, my EEEPC 1005 HGO with the old single core cpu (Intel
> N280), still got the old and undocumentd API.
>
> > What steps exactly do you use to work around the bug? What do you
> > mean by "running the old kernel module"? Are you using an older
> > kernel?
> >
>
> No, I can choose either to load eeepc-wmi or use the option "acpi_osi="Linux"
> which let me load eeepc-laptop at startup. The second one I call the "old"
> module, as this is the old fancontroller module developed for my old cpu.
>
> If eeepc-laptop is used, then eeepc-wmi will not be loaded.
[...]

Corentin, I assume you understand what's going on with these different
drivers. Shouldn't we add a dmi_enable_osi_linux quirk for the affected
models in drivers/acpi/blacklist.c?

Ben.

--
Ben Hutchings
In a hierarchy, every employee tends to rise to his level of incompetence.
 
Old 09-03-2012, 02:19 AM
Jonathan Nieder
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

reassign 649253 src:linux 3.2.23-1
found 649253 linux-2.6/3.0.0-6
quit

Hans-J. Ullrich wrote:

> Just to explain: eeepc-wmi will not be fixed, as the developers told me (almost
> 2 years ago). The problem of this, is that the old hardware specs of my EEEPC
> 1005 HGO are not made available by the vendor. So it is of course impossible,
> to write a kernel module.

Thanks. Do you have a pointer to that discussion, so we can look it
over in case it has additional useful information?

If I understand correctly, you have said that

(1) the eeepc-wmi module is loaded automatically
(2) loading that module on this machine makes the kernel behave
worse

If that is correct, then it is something that must be fixed. No
special specs are needed in order for the kernel not to automatically
load a certain module.

[...]
> I am running now linux-image-3.2.0-3-i386 (stock debian kernel of testing).

I am guessing you mean 3.2.23-1 (and marking so). You can get the
currently running kernel version by running

cat /proc/version

and looking for a version in parentheses or get the installed kernel
version by running

dpkg-query -W linux-image-$(uname -r)

See also [1].

[...]
> No, I can choose either to load eeepc-wmi or use the option "acpi_osi="Linux"
> which let me load eeepc-laptop at startup. The second one I call the "old"
> module, as this is the old fancontroller module developed for my old cpu.

Does blacklisting the eeepc-wmi module to prevent it from loading
allow eeepc-laptop to load and work well?

[...]
> However, I will test a newer kernel (3.5-X) in the next days, maybe not
> tomorrow, but during the coming week. Response and report will then be sent.

Thanks, nice to hear. I look forward to hearing how it goes.

Sincerely,
Jonathan

[1] http://kernel-handbook.alioth.debian.org/ch-versions.html#s-version-types
or the corresponding page in the debian-kernel-handbook package


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: http://lists.debian.org/20120903021923.GA2769@mannheim-rule.local
 
Old 09-03-2012, 06:51 AM
Corentin Chary
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

On Mon, Sep 3, 2012 at 3:11 AM, Ben Hutchings <ben@decadent.org.uk> wrote:
> On Sun, 2012-09-02 at 22:47 +0200, Hans-J. Ullrich wrote:
> [...]
>> Just to explain: eeepc-wmi will not be fixed, as the developers told me (almost
>> 2 years ago). The problem of this, is that the old hardware specs of my EEEPC
>> 1005 HGO are not made available by the vendor. So it is of course impossible,
>> to write a kernel module.
>>
>> All newer hardware, which is in most newer processors will be supported, of
>> course! But as I said, my EEEPC 1005 HGO with the old single core cpu (Intel
>> N280), still got the old and undocumentd API.
>>
>> > What steps exactly do you use to work around the bug? What do you
>> > mean by "running the old kernel module"? Are you using an older
>> > kernel?
>> >
>>
>> No, I can choose either to load eeepc-wmi or use the option "acpi_osi="Linux"
>> which let me load eeepc-laptop at startup. The second one I call the "old"
>> module, as this is the old fancontroller module developed for my old cpu.
>>
>> If eeepc-laptop is used, then eeepc-wmi will not be loaded.
> [...]
>
> Corentin, I assume you understand what's going on with these different
> drivers. Shouldn't we add a dmi_enable_osi_linux quirk for the affected
> models in drivers/acpi/blacklist.c?
>
> Ben.

I think the wmi interface should be used when available, because it's
the one asus is using on Windows. Fan control was kind of working
using a methods that were probably not covered by Asus ACPI "ABI". So
my point of view is "not supported by vendor on windows, hard to do
with wmi: won't support it". Individuals can always boot with
acpi_os=Linux if they want to, but I don't want to make that the
default.

Thanks



--
Corentin Chary
http://xf.iksaif.net


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: CAHR064i6kdDfpO72oSqivyk3WW5d1Eq+v-Opx4zAK9ubcgKydw@mail.gmail.com">http://lists.debian.org/CAHR064i6kdDfpO72oSqivyk3WW5d1Eq+v-Opx4zAK9ubcgKydw@mail.gmail.com
 
Old 09-05-2012, 06:02 PM
"Hans-J. Ullrich"
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

Hi Jonathan,

I tried the latest kernel from experimental, but as expected, there is no
difference. Same as with the other kernels.

For those, who are interested, I advise not to use eeepc-wmi, as doing so, the
fan will NOT speed up, even when the cpu is getting hot. To say correct: It is
speeding up, but only a very little bit! You may overheat your cpu.

I reverted back to the "acpi-osi="Linux"-workaround, which is working very
well on this computer.

Ben, I think, you understand the situation, as we talked about it some time
ago. I might remember, that you also told me, that the hardware of this
computer is not documented and so developing might be impossible.

I doubt, that many people are using an EEEPC 1005HGO with linux, most people
are having a better machine (with N280 cpu or better) and the newer API.

In my case I can use the workaround, and doing development for a single person
is really(!) too much of a good thing!

So it is ok for me, when this bug may be closed.

Otherwise I will be pleased, when I can help with some testing.

Thank you very much for your time and help!


Best regards

Hans


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201209052002.03421.hans.ullrich@loop.de">http://lists.debian.org/201209052002.03421.hans.ullrich@loop.de
 
Old 09-06-2012, 07:35 AM
Jonathan Nieder
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

found 649253 linux/3.5.2-1~experimental.1
quit

Hi,

Hans-J. Ullrich wrote:

> I tried the latest kernel from experimental, but as expected, there is no
> difference.

Marking so. Thanks for the quick feedback.

[...]
> For those, who are interested, I advise not to use eeepc-wmi, as doing so, the
> fan will NOT speed up, even when the cpu is getting hot. To say correct: It is
> speeding up, but only a very little bit! You may overheat your cpu.

To spell this out for me: if you blacklist eeepc-wmi in
/etc/modprobe.d, does the fan start to behave correctly, or do you
also need to tell the kernel to respond TRUE to OSI(Linux) queries
to mimic the pre-2.6.23 situation? Please do attach "acpidump"
output.

[...]
> In my case I can use the workaround, and doing development for a single person
> is really(!) too much of a good thing!

I doubt you are the only one using this machine, and more importantly,
adding a quirk to work around it would help kernel developers to see
what has been discovered about this family of machines.

> Otherwise I will be pleased, when I can help with some testing.

I'll be happy to write a patch to test that adds a quirk, but first,
please provide the information listed before ("acpidump" output, and
behavior without acpi_osi=Linux when eeepc-wmi is not loaded).

Thanks again for your help,
Jonathan


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: http://lists.debian.org/20120906073502.GA490@mannheim-rule.local
 
Old 09-06-2012, 04:36 PM
"Hans-J. Ullrich"
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

Hi Jonathan,


> To spell this out for me: if you blacklist eeepc-wmi in
> /etc/modprobe.d, does the fan start to behave correctly, or do you
> also need to tell the kernel to respond TRUE to OSI(Linux) queries
> to mimic the pre-2.6.23 situation? Please do attach "acpidump"
> output.

Well, with acpi-osi=Linux eeepci-wmi is NOT loaded (no need to blacklist it),
anbd eeepc-laptop is loaded instead. In this case, pwmcapable kernel-modules
are loaded, so that pwmconfig will start and is working.

If eeepc-wmi is loaded, and IO start the command pwmconfig, then I get the
message no pwmcapable hardware found.

I suppose, I already send the exact output before, if needed, I will sed it
again.

> I'll be happy to write a patch to test that adds a quirk, but first,
> please provide the information listed before ("acpidump" output, and
> behavior without acpi_osi=Linux when eeepc-wmi is not loaded).
>
Great, I will likely test it. Can you build a kernel-package for it? So it is
easy for me to install this kernel and deinstall after testing, without
getting in danger to destroy my system. I need it for work!

> Thanks again for your help,
> Jonathan

I am happy if I can help with my little knowledge.

Hans


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 201209061836.40611.hans.ullrich@loop.de">http://lists.debian.org/201209061836.40611.hans.ullrich@loop.de
 
Old 09-06-2012, 07:47 PM
Jonathan Nieder
 
Default Bug#649253: eeepc-wmi does not seem to adjust fan speed, but prevents use of fancontrol

Hans-J. Ullrich wrote:

>> I'll be happy to write a patch to test that adds a quirk, but first,
>> please provide the information listed before ("acpidump" output, and
>> behavior without acpi_osi=Linux when eeepc-wmi is not loaded).
>
> Great, I will likely test it. Can you build a kernel-package for it? So it is
> easy for me to install this kernel and deinstall after testing, without
> getting in danger to destroy my system. I need it for work!

Sorry for the lack of clarity. I meant the following:

1. Please attach "acpidump" output.

2. Please blacklist the eeepc-wmi module and test without the
acpi_osi=Linux parameter.

3. Using data from the above two steps, I will be ready to prepare
a patch and instructions for testing (which will involve
building and installing a Debian package).

Thanks again,
Jonathan


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: http://lists.debian.org/20120906194709.GB159@mannheim-rule.local
 

Thread Tools




All times are GMT. The time now is 07:14 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org