Why is it so hard to (get) help on ACPI?

Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 683 bytes --]

Hi,

what do I have to do to upload my DSDT patch and find help on how to make it 
better?

I've already tried registering on http://acpi.sourceforge.net/dsdt/ , but my 
account just doesn't show up on the list.
I've already sent my DSDT patch to this mailing list (after joining it, of 
course) with connected questions, but it didn't get through?

Why is it so hard to help or get help on ACPI? What am I doing wrong?

regards,
Matija Šuklje

-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Why is it so hard to (get) help on ACPI?

[Alexey Starikovskiy]

Hi,

First of all, why do you want/need to change your DSDT?
Second, patching DSDT is meant to be a last resort if any other
method, like finding error in our implementation of ACPI or drivers.

So "better" DSDT patch is patch of zero length -- don't touch it, fix
ACPI or drivers.

Regards,
Alex.

On 5/17/07, Matija Šuklje <matija.suklje@rutka.net> wrote:
> Hi,
>
> what do I have to do to upload my DSDT patch and find help on how to make it
> better?
>
> I've already tried registering on http://acpi.sourceforge.net/dsdt/ , but my
> account just doesn't show up on the list.
> I've already sent my DSDT patch to this mailing list (after joining it, of
> course) with connected questions, but it didn't get through?
>
> Why is it so hard to help or get help on ACPI? What am I doing wrong?
>
> regards,
> Matija Šuklje
>
> --
> gsm: +386 41 849 552
> e-mail: matija.suklje@rutka.net
> www: http://matija.suklje.name
>
> aim: hookofsilver
> icq: 110183360
> jabber/g-talk: matija.suklje@jabber.com
> msn: matija.suklje@rutka.net
> yahoo: matija_suklje
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Why is it so hard to (get) help on ACPI?

[Norbert Preining]

On Don, 17 Mai 2007, Alexey Starikovskiy wrote:
> First of all, why do you want/need to change your DSDT?
> Second, patching DSDT is meant to be a last resort if any other
> method, like finding error in our implementation of ACPI or drivers.

Well, of course, but frustration level is sometimes getting high, and
one tries to fix as much as possible to get it working. I for example
removed my modified DSDT some time ago to test the `official' ACPI
implementation. Then I tried to get s2ram to working, to no avail.
Restarting was freezing the kernel beyond repair (there is already a bug
reported, but no real action).

Now of course I tried also to fix the compile errors in the DSDT etc.

> So "better" DSDT patch is patch of zero length -- don't touch it, fix
> ACPI or drivers.

Well, tell me how?

Best wishes

Norbert

-------------------------------------------------------------------------------
Dr. Norbert Preining <preining@logic.at>                    Università di Siena
Debian Developer <preining@debian.org>                         Debian TeX Group
gpg DSA: 0x09C5B094      fp: 14DF 2E6C 0307 BE6D AD76  A9C0 D2BF 4AA3 09C5 B094
-------------------------------------------------------------------------------
DEEPING ST NICHOLAS (n.)
			--- Douglas Adams, The Meaning of Liff
What street-wise kids do at Christmas. They hide on the rooftops
waiting for Santa Claus so that if he arrives and goes down the
chimney, they can rip stuff off from his sleigh.
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 2067 bytes --]

Dne četrtek 17 maj 2007 20:58 ste napisali:
> First of all, why do you want/need to change your DSDT?
> Second, patching DSDT is meant to be a last resort if any other
> method, like finding error in our implementation of ACPI or drivers.

Thank you for the reply.

I *think* I need it because I didn't find any better solution. On my Acer 
Aspire 5022 for some time I've been getting APIC errors (most commonly "40
(40)") - which seem to cause system hangups, high cpu io-wait for (to me) no 
apparent reasons, high temperature and disability to resume from a suspend. 
This has been happening for quite a few kernel versions (gentoo kernel 
flavour). I was told the APIC errors should go away with a BIOS upgrade - so 
I did ...with no luck.

Then someone suggested trying the Linux-ready Firmware Developer Kit ( 
http://www.linuxfirmwarekit.org ), on which my laptop('s BIOS) failed on 9 
tests out of 21 (some of which even from DSDT). I can attach the results if 
you find them useful.

Currently I run my system with the "noapic" kernel flag and with my DSDT 
patch. It seems to be happier then before I did this, but I feel it's still 
far from optimal.

> So "better" DSDT patch is patch of zero length -- don't touch it, fix
> ACPI or drivers.

I agree that a solution by someone who knows what they're doing is a lot 
better then me (being a non-coder) trying to fiddle with ASL code. But it 
just drove me nuts...

Btw, the patch I wrote eliminates 9 errors which seem to be connected with 
memory.

But this still doesn't explain why it seems like I cannot make a user account 
on the acpi.sf.net/dsdt page and why my previous e-mail did not arrive to the 
list. Is the attachment size quota set to less then 1 MiB?

best regards and kudos for all your hard work on ACPI!

Matija Šuklje

-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 921 bytes --]

Dne četrtek 17 maj 2007 21:56 ste napisali:
> mail lists do not allow for size of attachment greater than 64k.

Makes sense, it would be nice if Majordomo would notify one of such 
limitations though.

> bugzilla is the place to put such things.

Kernel bugzilla?

> acpi.sf.net was near dead for several years already, since the time
> then there was  a decision that if something works under _other_ OS,
> it should work under Linux too. This made hacks to DSDT, etc obsolete.

Meaning ugly hacking should be done in the kernel itself?

So, what do you propose I should do? File a bug to kernel.org with the results 
I got from the firmware dev kit?

regards,
Matija Šuklje

-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Why is it so hard to (get) help on ACPI?

[Matthew Garrett]

On Thu, May 17, 2007 at 10:10:11PM +0200, Matija Šuklje wrote:
> Dne četrtek 17 maj 2007 21:56 ste napisali:
> > acpi.sf.net was near dead for several years already, since the time
> > then there was  a decision that if something works under _other_ OS,
> > it should work under Linux too. This made hacks to DSDT, etc obsolete.
> 
> Meaning ugly hacking should be done in the kernel itself?

The kernel's job is to make hardware work. If the hardware works under 
another OS, then it failing to work in Linux is a bug in Linux. So, yes, 
the correct place to fix this sort of thing is in the kernel rather than 
your DSDT.

However, please note that the DSDT is almost entirely irrelevant when it 
comes to things like suspend/resume. Modifying it is hugely unlikely to 
make things work there.

-- 
Matthew Garrett | mjg59@srcf.ucam.org
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Why is it so hard to (get) help on ACPI?

[Johannes Engel]

Matthew Garrett schrieb:
> The kernel's job is to make hardware work. If the hardware works under 
> another OS, then it failing to work in Linux is a bug in Linux. So, yes, 
> the correct place to fix this sort of thing is in the kernel rather than 
> your DSDT.
>
> However, please note that the DSDT is almost entirely irrelevant when it 
> comes to things like suspend/resume. Modifying it is hugely unlikely to 
> make things work there.
Hm, I do not totally agree. Since in some cases you know that a BIOS
update caused the trouble it seems to be likely in my eyes that these
problems could be fixed by DSDT changes. Please correct me if I'm wrong.

Greetings, Johannes

Re: Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 630 bytes --]

Dne petek 18 maj 2007 00:16 je Johannes Engel napisal(a):
> Hm, I do not totally agree. Since in some cases you know that a BIOS
> update caused the trouble it seems to be likely in my eyes that these
> problems could be fixed by DSDT changes. Please correct me if I'm wrong.

No, no ...the problems existed already before the BIOS update. They just 
persisted after it as well.

regards,
Matija

-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Why is it so hard to (get) help on ACPI?

[Matthew Garrett]

On Fri, May 18, 2007 at 12:16:30AM +0200, Johannes Engel wrote:

> Hm, I do not totally agree. Since in some cases you know that a BIOS
> update caused the trouble it seems to be likely in my eyes that these
> problems could be fixed by DSDT changes. Please correct me if I'm wrong.

The DSDT is far from the only bit of the BIOS used during 
suspend/resume. When you wake up, the system jumps back into the BIOS. 
An arbitrarily large amount of code can be executed there before control 
is passed back to the kernel. That's likely to be much more significant.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 2067 bytes --]

Dne četrtek 17 maj 2007 20:58 ste napisali:
> First of all, why do you want/need to change your DSDT?
> Second, patching DSDT is meant to be a last resort if any other
> method, like finding error in our implementation of ACPI or drivers.

Thank you for the reply.

I *think* I need it because I didn't find any better solution. On my Acer 
Aspire 5022 for some time I've been getting APIC errors (most commonly "40
(40)") - which seem to cause system hangups, high cpu io-wait for (to me) no 
apparent reasons, high temperature and disability to resume from a suspend. 
This has been happening for quite a few kernel versions (gentoo kernel 
flavour). I was told the APIC errors should go away with a BIOS upgrade - so 
I did ...with no luck.

Then someone suggested trying the Linux-ready Firmware Developer Kit ( 
http://www.linuxfirmwarekit.org ), on which my laptop('s BIOS) failed on 9 
tests out of 21 (some of which even from DSDT). I can attach the results if 
you find them useful.

Currently I run my system with the "noapic" kernel flag and with my DSDT 
patch. It seems to be happier then before I did this, but I feel it's still 
far from optimal.

> So "better" DSDT patch is patch of zero length -- don't touch it, fix
> ACPI or drivers.

I agree that a solution by someone who knows what they're doing is a lot 
better then me (being a non-coder) trying to fiddle with ASL code. But it 
just drove me nuts...

Btw, the patch I wrote eliminates 9 errors which seem to be connected with 
memory.

But this still doesn't explain why it seems like I cannot make a user account 
on the acpi.sf.net/dsdt page and why my previous e-mail did not arrive to the 
list. Is the attachment size quota set to less then 1 MiB?

best regards and kudos for all your hard work on ACPI!

Matija Šuklje

-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: Why is it so hard to (get) help on ACPI?

[Len Brown]

> I *think* I need it because I didn't find any better solution. On my Acer 
> Aspire 5022 for some time I've been getting APIC errors (most commonly "40
> (40)")

APIC errors have nothing to do with ACPI.
(other than the fact that on lots of systems, ACPI is required
if you want to configure the APIC).

If you don't like the APIC errors, then run with "noapic".
very few laptop uses would notice a difference.

> But this still doesn't explain why it seems like I cannot make a user account 
> on the acpi.sf.net/dsdt page and why my previous e-mail did not arrive to the 
> list. Is the attachment size quota set to less then 1 MiB?
> 
> best regards and kudos for all your hard work on ACPI!

acpi.sf.net is used as a project home page, but we otherwise
don't use the services of sourceforge.

we use bugzilla.kernel.org for Linux/ACPI bugs
we use this list on vger.kernel.org -- which has a 100K limit.
If you want to file a bug and attach a dsdt, use the bugzilla above.

cheers,
-Len

Re: Why is it so hard to (get) help on ACPI?

[Matija Šuklje]

[-- Attachment #1: Type: text/plain, Size: 380 bytes --]

Thanks for the reply and sorry about the late previous e-mail - the 
server "was full" and it took half an hour before it came through.

regards,
matija
-- 
gsm: +386 41 849 552
e-mail: matija.suklje@rutka.net
www: http://matija.suklje.name

aim: hookofsilver
icq: 110183360
jabber/g-talk: matija.suklje@jabber.com
msn: matija.suklje@rutka.net
yahoo: matija_suklje

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]