Linux Archive

Linux Archive (http://www.linux-archive.org/)
-   Fedora Development (http://www.linux-archive.org/fedora-development/)
-   -   Custom DSDT Table in the fedora kernel at boot time (http://www.linux-archive.org/fedora-development/112738-custom-dsdt-table-fedora-kernel-boot-time.html)

"Joshua C." 06-24-2008 04:05 PM

Custom DSDT Table in the fedora kernel at boot time
 
Hello guys,

I had a serios problems with my bios and at the end it was the dsdt
table that was broken. this gave me alot of kernel panics untill i
figured it out. Now I have to recompile the kernel every time a new
one goes out in order to put in my custom dsdt table and enable it.
this takes about 30 min and I cannot use any of the devel kernels.

There is, however, a kernel patch at gaugusch.at/kernel.shtml which
enables the kernel to load a custom dsdt table from the initrd. and
this is alot faster than the previous method with the kernel
recompilation.

Will it be possible for this patch to go in the mainstream fedora
kernel?

Some other distros are using it and I don't see any reason why
fedora shouldn't use it either. I know this kind of tainting the
kernel is "dangerous" but it is the only way for user (like me)
with broken bios/dsdt table to run their kernels. And because lots
of bioses are tested only on windows there are no other way than
inserting a custom dsdt for linux.

Any sugestions or downsides of this method that can prove the
exclusion of this patch from the mainstream fedora kernel?

Cheers,
Joshua

--
Internet Security Software - Click here.
http://tagline.hushmail.com/fc/Ioyw6h4dJxHDn2puyZ3Jrv4LFpXqrJ3oKIWGfQomgsD4N9EJi3 jVn1/

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Dominik 'Rathann' Mierzejewski" 06-24-2008 05:20 PM

Custom DSDT Table in the fedora kernel at boot time
 
Hi,

On Tuesday, 24 June 2008 at 18:05, Joshua C. wrote:
> I had a serios problems with my bios and at the end it was the dsdt
> table that was broken. this gave me alot of kernel panics untill i
> figured it out. Now I have to recompile the kernel every time a new
> one goes out in order to put in my custom dsdt table and enable it.
> this takes about 30 min and I cannot use any of the devel kernels.
>
> There is, however, a kernel patch at gaugusch.at/kernel.shtml which
> enables the kernel to load a custom dsdt table from the initrd. and
> this is alot faster than the previous method with the kernel
> recompilation.
>
> Will it be possible for this patch to go in the mainstream fedora
> kernel?

This has already been discussed and the answer was no:
https://bugzilla.redhat.com/show_bug.cgi?id=110511
https://bugzilla.redhat.com/show_bug.cgi?id=140215
https://bugzilla.redhat.com/show_bug.cgi?id=169014

Actually these bug reports should be marked as duplicates.

> Some other distros are using it and I don't see any reason why
> fedora shouldn't use it either. I know this kind of tainting the
> kernel is "dangerous" but it is the only way for user (like me)
> with broken bios/dsdt table to run their kernels. And because lots
> of bioses are tested only on windows there are no other way than
> inserting a custom dsdt for linux.
>
> Any sugestions or downsides of this method that can prove the
> exclusion of this patch from the mainstream fedora kernel?

Quoting Dave Jones (https://bugzilla.redhat.com/show_bug.cgi?id=169014#c2):

| I've already explained this several times already.
| The only way this is going into the fedora kernel is via upstream.

| If users start manipulating their DSDTs, and getting panics etc, they file bugs
| here, and would 'neglect' to say how they hacked their dsdt's

| It's a support nightmare waiting to happen.

So there you go.

Regards,
R.

--
Fedora http://fedoraproject.org/wiki/User:Rathann
Livna http://rpm.livna.org | MPlayer http://mplayerhq.hu
"Faith manages."
-- Delenn to Lennier in Babylon 5:"Confessions and Lamentations"

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Joshua C." 06-24-2008 08:24 PM

Custom DSDT Table in the fedora kernel at boot time
 
I read these comments but they are from year 2004 and 2005. And the
propoced patch was in some of the 2.6.25-rc versions.

This is not an extra functionallity but the only option (for some
of us) to run linux on our laptops. The ACPI code is quite
difficult to play with, so I don't see any "nightmare waiting to
happen" here. If at the end there are so many bug reports connected
with DSDT, then just drop the patch again.

I don't see that much bug reports on opensuse, ubuntu and mandriva
complaining about a kernel panic based on broken dsdt. Maybe Dave
Jones should reconsider his opinion about it. Or at least try to
explain to the notebook producers that they should test their bios
to work with linux, too.

Fedora is supposed to be "on the bleeding edge". so i really don't
see any reasons against (at least trying to) inserting this patch
in some RCs. Lets see what'll happen. it is always possible to
revert the patch.

--
Click to get a free auto insurance quotes from top companies.
http://tagline.hushmail.com/fc/Ioyw6h4d8EIXAvwILUXMaz5GGa1ZtyyaAXI851HuaA9ykDdFoQ LNoH/

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Jeff Spaleta" 06-24-2008 08:31 PM

Custom DSDT Table in the fedora kernel at boot time
 
On Tue, Jun 24, 2008 at 12:24 PM, Joshua C. <jjj123jjj@hushmail.com> wrote:
> Fedora is supposed to be "on the bleeding edge". so i really don't
> see any reasons against (at least trying to) inserting this patch
> in some RCs. Lets see what'll happen. it is always possible to
> revert the patch.


You missed a very important point. Its not just bleeding edge... but
sustainable bleed-edge.
We work closely with upstream. And in the case of this specific kernel
patch. Doesn't everyone benefit by getting that patch into the
mainline kernel?

-jef

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

Arjan van de Ven 06-24-2008 09:40 PM

Custom DSDT Table in the fedora kernel at boot time
 
On Tue, 24 Jun 2008 12:31:50 -0800
"Jeff Spaleta" <jspaleta@gmail.com> wrote:

> On Tue, Jun 24, 2008 at 12:24 PM, Joshua C. <jjj123jjj@hushmail.com>
> wrote:
> > Fedora is supposed to be "on the bleeding edge". so i really don't
> > see any reasons against (at least trying to) inserting this patch
> > in some RCs. Lets see what'll happen. it is always possible to
> > revert the patch.
>
>
> You missed a very important point. Its not just bleeding edge... but
> sustainable bleed-edge.
> We work closely with upstream. And in the case of this specific kernel
> patch. Doesn't everyone benefit by getting that patch into the
> mainline kernel?
>

said upstream patch was merged by Linus for a short while and then
reverted for quality reasons.... so that would be a strong indication
of "maybe not a good idea"

--
If you want to reach me at my work email, use arjan@linux.intel.com
For development, discussion and tips for power savings,
visit http://www.lesswatts.org

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Jeff Spaleta" 06-24-2008 10:03 PM

Custom DSDT Table in the fedora kernel at boot time
 
On Tue, Jun 24, 2008 at 1:40 PM, Arjan van de Ven <arjan@infradead.org> wrote:
> said upstream patch was merged by Linus for a short while and then
> reverted for quality reasons.... so that would be a strong indication
> of "maybe not a good idea"

Even better... everyone benefits by not being exposed to a poor quality patch.

The question, which I can't personally answer, becomes what is it
going to take to clean this patch up so it can be upstreamed.

-jef

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

Eric Sandeen 06-24-2008 10:05 PM

Custom DSDT Table in the fedora kernel at boot time
 
Jeff Spaleta wrote:
> On Tue, Jun 24, 2008 at 1:40 PM, Arjan van de Ven <arjan@infradead.org> wrote:
>> said upstream patch was merged by Linus for a short while and then
>> reverted for quality reasons.... so that would be a strong indication
>> of "maybe not a good idea"
>
> Even better... everyone benefits by not being exposed to a poor quality patch.
>
> The question, which I can't personally answer, becomes what is it
> going to take to clean this patch up so it can be upstreamed.

It looks like it's already underway, at least as far as the author is
concerned. From http://gaugusch.at/kernel.shtml :

Update as of 2008-03-15: Several maintainers in the Linux kernel have
started to reconsider their position on the feature. The patch as even
made a brief appearance in some development versions of the kernel
2.6.25 (it has been reverted because deemed too likely to bring bugs,
but a new approach, patch v0.9, is already available for next
development cycle).

-Eric

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Joshua C." 06-24-2008 11:18 PM

Custom DSDT Table in the fedora kernel at boot time
 
Eric Sandeen wrote:
>It looks like it's already underway, at least as far as the author
is
>concerned. From http://gaugusch.at/kernel.shtml :
>
>Update as of 2008-03-15: Several maintainers in the Linux kernel
have
>started to reconsider their position on the feature. The patch as
even
>made a brief appearance in some development versions of the kernel
>2.6.25 (it has been reverted because deemed too likely to bring
bugs,
>but a new approach, patch v0.9, is already available for next
>development cycle).
>
>-Eric
>

ok guys,

then it is up to whether the patch is "clean" enough to go in the
mainstream kernel.

I think the decision about a custom dsdt table in the kernel has
been taken long time ago. and the answer is the option that enables
the kernel to load a custom table. what the patch does is that it
pospones the loading of the table until the initram is loaded. so
the custom dsdt can be read from it.

If someone thinks the patch isn't good enough, then he should
suggest his own patch. but for now this is the only one available.
Either in the kernel or in the initram, in both cases we can load
our custom dsdt table. the only difference is that the kernel
recompilation takes about 30 mins (compared to the 2mins for the
repacking of the initram). I really don't see any objections to
simplifying this process.

Please, prove me wrong, but I'm just trying to make it easier to
run linux on my machine. And the patch simlifies this process.


PS: There is one more option:
Email the acpi-dev-team my custom dsdt so that they can include it
in the mainline kernel :). Since this is the "missing" code that
breaks my machine, it should be send as a patch for inclussion. And
this is the only way to make my hardware work. So maybe I should
email the kernel-acpi-dev-team :). As you can see, this is a lot
"dirtier" than the above solution. Any other ideas how to make my
hardware work? :)

Joshua

--
Shop &amp; save on the supplements you want. Click now!
http://tagline.hushmail.com/fc/Ioyw6h4fK5SEhRB76ZjQFrf3fIDbozD9Jz2WUnmOvXQBcaKi9I oiwL/

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

"Naheem Zaffar" 06-24-2008 11:38 PM

Custom DSDT Table in the fedora kernel at boot time
 
If it fixes the problem for good instead of just working around it, I
doubt many would see that as a dirtier solution.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list

Arjan van de Ven 06-25-2008 02:38 AM

Custom DSDT Table in the fedora kernel at boot time
 
On Wed, 25 Jun 2008 01:18:56 +0200
"Joshua C." <jjj123jjj@hushmail.com> wrote:
>
> PS: There is one more option:
> Email the acpi-dev-team my custom dsdt so that they can include it
> in the mainline kernel :). Since this is the "missing" code that
> breaks my machine, it should be send as a patch for inclussion. And
> this is the only way to make my hardware work. So maybe I should
> email the kernel-acpi-dev-team :). As you can see, this is a lot
> "dirtier" than the above solution. Any other ideas how to make my
> hardware work? :)

as someone who works for Intel near the ACPI dev team (yes one cube
away)... lets put it this way:
If Windows works with the BIOS, and Linux does not, that is a bug that
needs fixing in Linux. The ACPI team takes this and other bugs VERY
seriously; especially if the impact is severe functional impairment and
Windows does work. I would strongly suggest you file a bug in
bugzilla.kernel.org attaching your dsdt....

--
If you want to reach me at my work email, use arjan@linux.intel.com
For development, discussion and tips for power savings,
visit http://www.lesswatts.org

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list


All times are GMT. The time now is 07:12 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.