help needed with invalid method in GPE block

help needed with invalid method in GPE block

[Charl P. Botha]

Dear list,

I'm still trying to get my stupid sleep/wake-button to wake my laptop.  This
is quite interesting:

ACPI Namespace successfully loaded at root c02cdcfc
evxfevnt-0092 [04] acpi_enable           : Transition to ACPI mode successful
   evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at 0000000000001028
   evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE0 to GPE15
   evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L1D is not valid
   evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at 000000000000102C
   evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE16 to GPE31
   evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L05 is not valid
   evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L1D is not valid

Coincidentally, in my dsdt _L1D (the one with invalid number above) is a
control method for my sleep button (control method type, not hard-wired).
This _could_ explain why the button is not doing anything when pressed
during sleep:

        Method (_L1D, 0, NotSerialized) 
        {
            Notify (\_SB.SLPB, 0x02)
        }

This is in the \_GPE scope.  I've tried hacking the control method sleep
button example on page 66 of ACPIspec-2-0a.pdf into my dsdt as method _L01,
but to no avail.  Then ACPI just claims that _L01 is invalid for the
*second* GPE block:

evxfevnt-0092 [04] acpi_enable           : Transition to ACPI mode successful
evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at 0000000000001028
evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE0 to GPE15
evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L1D is not valid
evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at 000000000000102C
evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE16 to GPE31
evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L05 is not valid
evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L1D is not valid
evgpe-0138 [08] ev_save_method_info   : GPE number associated with method _L01 is not valid

In both cases, my fixed DSDT compiled with iasl 20030228 with no errors and
no warnings.

Can anyone explain to me why I have two GPE blocks?  This means (according
to the code logic in the acpi driver) that a control method is ALWAYS going
to have an invalid GPE number in at least one of the two blocks.

If anyone is interested, my original dsdt is here:
http://cpbotha.net/thingies/dsdt.asl.original
and the hacked one is here:
http://cpbotha.net/thingies/dsdt.asl

The object of the game is to get my wake button to wake. :)

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open! 
Get cracking and register here for some mind boggling fun and 
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en

RE: help needed with invalid method in GPE block

[Moore, Robert]

You can ignore the GPE error messages, they are leftover cruft and have no
meaning.


> -----Original Message-----
> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org]
> Sent: Monday, March 17, 2003 1:35 PM
> To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> Subject: [ACPI] help needed with invalid method in GPE block
> 
> Dear list,
> 
> I'm still trying to get my stupid sleep/wake-button to wake my laptop.
> This
> is quite interesting:
> 
> ACPI Namespace successfully loaded at root c02cdcfc
> evxfevnt-0092 [04] acpi_enable           : Transition to ACPI mode
> successful
>    evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at
> 0000000000001028
>    evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE0 to
> GPE15
>    evgpe-0138 [08] ev_save_method_info   : GPE number associated with
> method _L1D is not valid
>    evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at
> 000000000000102C
>    evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE16 to
> GPE31
>    evgpe-0138 [08] ev_save_method_info   : GPE number associated with
> method _L05 is not valid
>    evgpe-0138 [08] ev_save_method_info   : GPE number associated with
> method _L1D is not valid
> 
> Coincidentally, in my dsdt _L1D (the one with invalid number above) is a
> control method for my sleep button (control method type, not hard-wired).
> This _could_ explain why the button is not doing anything when pressed
> during sleep:
> 
>         Method (_L1D, 0, NotSerialized)
>         {
>             Notify (\_SB.SLPB, 0x02)
>         }
> 
> This is in the \_GPE scope.  I've tried hacking the control method sleep
> button example on page 66 of ACPIspec-2-0a.pdf into my dsdt as method
> _L01,
> but to no avail.  Then ACPI just claims that _L01 is invalid for the
> *second* GPE block:
> 
> evxfevnt-0092 [04] acpi_enable           : Transition to ACPI mode
> successful
> evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at
> 0000000000001028
> evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE0 to GPE15
> evgpe-0138 [08] ev_save_method_info   : GPE number associated with method
> _L1D is not valid
> evgpe-0416 [06] ev_create_gpe_block   : GPE Block: 2 registers at
> 000000000000102C
> evgpe-0421 [06] ev_create_gpe_block   : GPE Block defined as GPE16 to
> GPE31
> evgpe-0138 [08] ev_save_method_info   : GPE number associated with method
> _L05 is not valid
> evgpe-0138 [08] ev_save_method_info   : GPE number associated with method
> _L1D is not valid
> evgpe-0138 [08] ev_save_method_info   : GPE number associated with method
> _L01 is not valid
> 
> In both cases, my fixed DSDT compiled with iasl 20030228 with no errors
> and
> no warnings.
> 
> Can anyone explain to me why I have two GPE blocks?  This means (according
> to the code logic in the acpi driver) that a control method is ALWAYS
> going
> to have an invalid GPE number in at least one of the two blocks.
> 
> If anyone is interested, my original dsdt is here:
> http://cpbotha.net/thingies/dsdt.asl.original
> and the hacked one is here:
> http://cpbotha.net/thingies/dsdt.asl
> 
> The object of the game is to get my wake button to wake. :)
> 
> Thanks,
> Charl
> 
> --
> charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by:Crypto Challenge is now open!
> Get cracking and register here for some mind boggling fun and
> the chance of winning an Apple iPod:
> http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel


-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open! 
Get cracking and register here for some mind boggling fun and 
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Mon, Mar 17, 2003 at 03:07:31PM -0800, Moore, Robert wrote:
> You can ignore the GPE error messages, they are leftover cruft and have no
> meaning.

In my kernel (2.5.64-bk12), if one sees that error message, it means that
that event is not stored.  acpi_ev_save_method_info() in evgpeblk.c returns
without adding the method to the passed gpe_event_info struct.  Is this
really not a problem, i.e. is this code as well as the gpe_event_info
structure not relevant anymore?

Do you perhaps have any other ideas as to why my sleep key just continues
sleeping?  :)

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open! 
Get cracking and register here for some mind boggling fun and 
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Mon, Mar 17, 2003 at 03:57:00PM -0800, Moore, Robert wrote:
> Here's what the code should look like
> 
>     /* Ensure that we have a valid GPE number for this GPE block */
> 
>     if ((GpeNumber < GpeBlock->BlockBaseNumber) ||
>         (GpeNumber >= (GpeBlock->BlockBaseNumber + (GpeBlock->RegisterCount
> * 8))))
>     {
>         /* 
>          * Not valid for this GPE block, just ignore it
>          * However, it may be valid for a different GPE block, since GPE0
> and GPE1
>          * methods both appear under \_GPE.
>          */
>         return (AE_OK);
>     }

Aaaah, the "if" is now clearly corrected.  The way it was did constitute a
bug, right?  It resulted in control methods that were valid for the second
GPE block not getting added.

My frikking sleep button is still not waking my laptop up though.  I would
appreciate any other tips for debugging that problem.

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open! 
Get cracking and register here for some mind boggling fun and 
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en

RE: help needed with invalid method in GPE block

[Grover, Andrew]

> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> My frikking sleep button is still not waking my laptop up 
> though.  I would
> appreciate any other tips for debugging that problem.

I would guess that we are disabling your sleep button's GPE when we are
sleeping.

To verify this, you might try calling _PRW from the button driver (or
just look at the ASL), and then using that GPE number call
acpi_enable_gpe(gpe_num, NULL, ACPI_EVENT_WAKE_ENABLE).

This is not the correct long-term solution but it should at least verify
the root cause of the problem...

Regards -- Andy


-------------------------------------------------------
This SF.net email is sponsored by:Crypto Challenge is now open! 
Get cracking and register here for some mind boggling fun and 
the chance of winning an Apple iPod:
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en

RE: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, 2003-03-18 at 01:34, Grover, Andrew wrote:
> > From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > My frikking sleep button is still not waking my laptop up 
> > though.  I would
> > appreciate any other tips for debugging that problem.
> 
> I would guess that we are disabling your sleep button's GPE when we are
> sleeping.

Yes!  After going to bed last night, I remembered reading about the
various wake-up capable devices being enabled or disabled.

> To verify this, you might try calling _PRW from the button driver (or
> just look at the ASL), and then using that GPE number call
> acpi_enable_gpe(gpe_num, NULL, ACPI_EVENT_WAKE_ENABLE).

Hmmm, there are several other wake-up devices writing to GPE bit 0x1D,
e.g.:

Device (LID)
{
   Name (_HID, EisaId ("PNP0C0D"))
   Name (_PRW, Package (0x02)
   {
      0x1D, 
      0x03
   })
...

but the corresponding method in the _GPE namespace, _L1D(), Notify()s
the OS about wakeup (event 0x02) and sends only the button object along.

If I do acpe_enable_gpe(0x1D, NULL, ACPI_EVENT_WAKE_ENABLE), will it
mean that all these other devices will also be able to wake the laptop
up from S3?  My DSDT is at http://cpbotha.net/thingies/dsdt.asl.original

> This is not the correct long-term solution but it should at least verify
> the root cause of the problem...

Great.  I will try this tonight, thank you very much.  What would be the
Right Way(tm) of fixing this?

Thanks again,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, 2003-03-18 at 01:19, Charl P. Botha wrote:
> On Mon, Mar 17, 2003 at 03:57:00PM -0800, Moore, Robert wrote:
> > Here's what the code should look like
> > 
> >     /* Ensure that we have a valid GPE number for this GPE block */
> > 
> >     if ((GpeNumber < GpeBlock->BlockBaseNumber) ||
> >         (GpeNumber >= (GpeBlock->BlockBaseNumber + (GpeBlock->RegisterCount
> > * 8))))
> >     {
> >         /* 
> >          * Not valid for this GPE block, just ignore it
> >          * However, it may be valid for a different GPE block, since GPE0
> > and GPE1
> >          * methods both appear under \_GPE.
> >          */
> >         return (AE_OK);
> >     }
> 
> Aaaah, the "if" is now clearly corrected.  The way it was did constitute a
> bug, right?  It resulted in control methods that were valid for the second
> GPE block not getting added.

BTW, before making this change, my laptop would go INSTANTLY to sleep if
I pressed the sleep button.  After applying it, messages flash by
indicating that the kernel is preparing devices for the suspend.

This would mean that, in my case, the handler for method _L1D() was not
being added (in either GPE block), so the Notify() by that method was
not getting handled by the kernel, which was a bad thing.

Will this fix go into the next BK push, or whatever it's called?

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

RE: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, 2003-03-18 at 01:34, Grover, Andrew wrote:
> > From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > My frikking sleep button is still not waking my laptop up 
> > though.  I would
> > appreciate any other tips for debugging that problem.
> 
> I would guess that we are disabling your sleep button's GPE when we are
> sleeping.
> 
> To verify this, you might try calling _PRW from the button driver (or
> just look at the ASL), and then using that GPE number call
> acpi_enable_gpe(gpe_num, NULL, ACPI_EVENT_WAKE_ENABLE).

Can I assume that you mean:
acpe_enable_event(gpe_num, ACPI_EVENT_GPE, ACPI_EVENT_WAKE_ENABLE);
where gpe_num is in my case 0x1D (the method in the asl is _L1D)?

I was planning on adding this just after the
status = acpi_hw_enable_gpe (gpe_event_info);
call in the acpi_ev_save_method_info() function of evgpeblk.c IF the
gpe_num is 0x1D.  Is this okay for verifying?

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

RE: help needed with invalid method in GPE block

[Grover, Andrew]

> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > To verify this, you might try calling _PRW from the button 
> driver (or
> > just look at the ASL), and then using that GPE number call
> > acpi_enable_gpe(gpe_num, NULL, ACPI_EVENT_WAKE_ENABLE).
> 
> Can I assume that you mean:
> acpe_enable_event(gpe_num, ACPI_EVENT_GPE, ACPI_EVENT_WAKE_ENABLE);
> where gpe_num is in my case 0x1D (the method in the asl is _L1D)?

Hehe yes. Did I mention the next version will change that interface? :)

> I was planning on adding this just after the
> status = acpi_hw_enable_gpe (gpe_event_info);
> call in the acpi_ev_save_method_info() function of evgpeblk.c IF the
> gpe_num is 0x1D.  Is this okay for verifying?

I would think so, but there's only one way to find out...

Regards -- Andy


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, Mar 18, 2003 at 09:21:15AM -0800, Grover, Andrew wrote:
> > From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > Can I assume that you mean:
> > acpe_enable_event(gpe_num, ACPI_EVENT_GPE, ACPI_EVENT_WAKE_ENABLE);
> > where gpe_num is in my case 0x1D (the method in the asl is _L1D)?
> 
> Hehe yes. Did I mention the next version will change that interface? :)
> 
> > I was planning on adding this just after the
> > status = acpi_hw_enable_gpe (gpe_event_info);
> > call in the acpi_ev_save_method_info() function of evgpeblk.c IF the
> > gpe_num is 0x1D.  Is this okay for verifying?
> 
> I would think so, but there's only one way to find out...

Yay, my frikking sleep button is now waking the laptop from sleep! Nevermind
the fact that everything else is still dead (these problems seem a tad more
tractable), the BUTTON IS WORKING.

Through printk()s all over, it seems that the acpi code never explicitly
enables the 0x1D event on my machine at all, nevermind enabling it for
waking.  Is this a problem with my firmware/dsdt or with the acpi kernel
code?

I ran into some *cough* Microsoft documents *cough* online that claimed that
*cough* Windows *cough* enabled/disabled certain devices for waking.
According to these documents, the sleep button was always enabled.  Should
Linux ACPI be doing the same?

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

RE: help needed with invalid method in GPE block

[Grover, Andrew]

> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > > I was planning on adding this just after the
> > > status = acpi_hw_enable_gpe (gpe_event_info);
> > > call in the acpi_ev_save_method_info() function of 
> evgpeblk.c IF the
> > > gpe_num is 0x1D.  Is this okay for verifying?
> > 
> > I would think so, but there's only one way to find out...
> 
> Yay, my frikking sleep button is now waking the laptop from 
> sleep! Nevermind
> the fact that everything else is still dead (these problems 
> seem a tad more
> tractable), the BUTTON IS WORKING.

Great.

> Through printk()s all over, it seems that the acpi code never 
> explicitly
> enables the 0x1D event on my machine at all, nevermind enabling it for
> waking.  Is this a problem with my firmware/dsdt or with the 
> acpi kernel
> code?

Are you sure? We should be calling acpi_ev_save_method_info for each
_Lxx or _Exx entry under _GPE scope, and that function enables the GPE.
Please let me know if this is not the case.

> I ran into some *cough* Microsoft documents *cough* online 
> that claimed that
> *cough* Windows *cough* enabled/disabled certain devices for waking.
> According to these documents, the sleep button was always 
> enabled.  Should
> Linux ACPI be doing the same?

Yes. The button driver should be enabling its devices for wakeup. I
think the ACPI bus code should  export a function that makes this easy
to do.

I am going to stick this in kernel.bugzilla.org so we don't forget about
it, and I will add you to the CC list.

Regards -- Andy


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, Mar 18, 2003 at 01:54:11PM -0800, Grover, Andrew wrote:
> > Through printk()s all over, it seems that the acpi code never 
> > explicitly
> > enables the 0x1D event on my machine at all, nevermind enabling it for
> > waking.  Is this a problem with my firmware/dsdt or with the 
> > acpi kernel
> > code?
> 
> Are you sure? We should be calling acpi_ev_save_method_info for each
> _Lxx or _Exx entry under _GPE scope, and that function enables the GPE.
> Please let me know if this is not the case.

I'm quite sure.  Yes, you do call acpi_ev_save_method_info() for each _Lxx
and _Exx method, but in acpi_ev_save_method_info() you only call
acpi_hw_enable_gpe() and NEVER acpi_hw_enable_gpe_for_wakeup().  As a matter
of fact, shouldn't that call rather be acpi_enable_event() (INSTEAD of
acpi_hw_enable_gpe()) with the WAKEUP parameter according to some mask?

Fact is, with my ACPI-untrained eyes, I couldn't find any occurrence where
acpi_hw_enable_gpe_for_wakeup() is directly or indirectly called for the
sleep button method (_L1D in my case).

Your thoughts?

Thanks,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

RE: help needed with invalid method in GPE block

[Grover, Andrew]

> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > > Through printk()s all over, it seems that the acpi code never 
> > > explicitly
> > > enables the 0x1D event on my machine at all, nevermind 
> enabling it for
> > > waking.  Is this a problem with my firmware/dsdt or with the 
> > > acpi kernel
> > > code?
> > 
> > Are you sure? We should be calling acpi_ev_save_method_info for each
> > _Lxx or _Exx entry under _GPE scope, and that function 
> enables the GPE.
> > Please let me know if this is not the case.
> 
> I'm quite sure.  Yes, you do call acpi_ev_save_method_info() 
> for each _Lxx
> and _Exx method, but in acpi_ev_save_method_info() you only call
> acpi_hw_enable_gpe() and NEVER 
> acpi_hw_enable_gpe_for_wakeup().  As a matter
> of fact, shouldn't that call rather be acpi_enable_event() (INSTEAD of
> acpi_hw_enable_gpe()) with the WAKEUP parameter according to 
> some mask?

I agree that it is currently never enabled for wakeup, but you said we
never enable it at all, the opposite of what your last email said. I
guess I'll assume the second statement is what you meant.

BTW here is the bug:
http://bugme.osdl.org/show_bug.cgi?id=470

Apparently it only allows CCs to registered people, so I couldn't add
you, but if you have any extra comments, feel free to register and add
them there.

Regards -- Andy



-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

RE: help needed with invalid method in GPE block

[Moore, Robert]

So,
We are calling acpi_hw_enable_gpe() on 0x1D or not?



> -----Original Message-----
> From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org]
> Sent: Tuesday, March 18, 2003 2:09 PM
> To: Grover, Andrew
> Cc: Moore, Robert; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> Subject: Re: [ACPI] help needed with invalid method in GPE block
> 
> On Tue, Mar 18, 2003 at 01:54:11PM -0800, Grover, Andrew wrote:
> > > Through printk()s all over, it seems that the acpi code never
> > > explicitly
> > > enables the 0x1D event on my machine at all, nevermind enabling it for
> > > waking.  Is this a problem with my firmware/dsdt or with the
> > > acpi kernel
> > > code?
> >
> > Are you sure? We should be calling acpi_ev_save_method_info for each
> > _Lxx or _Exx entry under _GPE scope, and that function enables the GPE.
> > Please let me know if this is not the case.
> 
> I'm quite sure.  Yes, you do call acpi_ev_save_method_info() for each _Lxx
> and _Exx method, but in acpi_ev_save_method_info() you only call
> acpi_hw_enable_gpe() and NEVER acpi_hw_enable_gpe_for_wakeup().  As a
> matter
> of fact, shouldn't that call rather be acpi_enable_event() (INSTEAD of
> acpi_hw_enable_gpe()) with the WAKEUP parameter according to some mask?
> 
> Fact is, with my ACPI-untrained eyes, I couldn't find any occurrence where
> acpi_hw_enable_gpe_for_wakeup() is directly or indirectly called for the
> sleep button method (_L1D in my case).
> 
> Your thoughts?
> 
> Thanks,
> Charl
> 
> --
> charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: Does your code think in ink?
> You could win a Tablet PC. Get a free Tablet PC hat just for playing.
> What are you waiting for?
> http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, Mar 18, 2003 at 02:25:11PM -0800, Moore, Robert wrote:
> We are calling acpi_hw_enable_gpe() on 0x1D or not?

Sorry, my answer was ambiguous.  Yes, you are calling acpi_hw_enable_gpe()
on 0x1D.

However, this doesn't enable it for wakeup, which I guess is the problem.

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en

Re: help needed with invalid method in GPE block

[Charl P. Botha]

On Tue, Mar 18, 2003 at 02:24:31PM -0800, Grover, Andrew wrote:
> > From: Charl P. Botha [mailto:c.p.botha-ra0OqMccq8edIhEUFHOBvg@public.gmane.org] 
> > I'm quite sure.  Yes, you do call acpi_ev_save_method_info() 
> > for each _Lxx
> > and _Exx method, but in acpi_ev_save_method_info() you only call
> > acpi_hw_enable_gpe() and NEVER 
> > acpi_hw_enable_gpe_for_wakeup().  As a matter
> > of fact, shouldn't that call rather be acpi_enable_event() (INSTEAD of
> > acpi_hw_enable_gpe()) with the WAKEUP parameter according to 
> > some mask?
> 
> I agree that it is currently never enabled for wakeup, but you said we
> never enable it at all, the opposite of what your last email said. I
> guess I'll assume the second statement is what you meant.

Your assumption is correct. :)  For all clarity, the event is enabled, but
not enabled for wakeup.

> BTW here is the bug:
> http://bugme.osdl.org/show_bug.cgi?id=470

Excellent, thank you very much.  So, if I understand correctly, this would
mean that the button driver would call this function that you're going to
add in order to indicate that it, the button, should be able to wake the
system?

Thanks again,
Charl

-- 
charl p. botha http://cpbotha.net/ http://visualisation.tudelft.nl/


-------------------------------------------------------
This SF.net email is sponsored by: Does your code think in ink? 
You could win a Tablet PC. Get a free Tablet PC hat just for playing. 
What are you waiting for?
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr5043en