the role of the dstd

the role of the dstd

[Joeri Hendrickx]

Hi

Since my asus laptop's acpi won't work decently in Linux (lid close 
won't generate event and such) I've been reading up and I've come to the 
conclusion that the problem must be a messed up DSTD.

I do have some questions about it, though.

For one, if the DSTD is part of the BIOS, how come windows does work, 
while it must rely on the same data (I haven't installed any drivers for 
power mgt yet lid close does work, and so do the volume butons)  If 
windows knows something we don't, can't we get the correct data from there?

Kind regards,
J. Hendrickx, hoping to see my suspend to ram working someday

Re: the role of the dstd

[Sanjoy Mahajan]

> For one, if the DSTD is part of the BIOS, how come windows does
> work, while it must rely on the same data

I'm not an ACPI expert but here is my understanding.  Hopefully it's
close enough to reality that the experts need to make only minor
corrections and can keep hacking.

Often, or rather always, the manufacturer tests the DSDT only with
Windows.  If Windows makes assumptions about how the DSDT code should
be interpreted that are different from what Linux assumes, even if
Linux's assumptions are compatible with the ACPI spec, then the DSDT
might fail when run with Linux.

For example, Windows serializes (runs in one thread) procedures
automatically, even if a procedure is labelled NotSerialized.  But
Linux ACPI, unless you pass a boot option to tell it to act like
Windows, will not serialize those procedures, and each one can run in
its own thread.

Therefore, race conditions from a sloppy DSDT, which might be hidden
in Windows, are often exposed in Linux.  And so S3 suspend might hang
because of one of those races or locking problems.  My old Thinkpad
600X improved its S3 suspend behavior somewhat when booting with the
"Act like Windows" option.

ACPI development on Linux is often: "My laptop doesn't suspend in
Linux" (for a modern laptop it's almost certain that it suspends in
Windows) and then people finding what bug needs to be introduced into
the Linux DSDT interpreter so that Linux can act just like Windows.

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

Re: the role of the dstd

[Ruslán Ledesma Garza]

Is this option "act like windows" really available? Is it a boot parameter?

Thanks in advance.

On 8/24/06, Sanjoy Mahajan <sanjoy@mrao.cam.ac.uk> wrote:
> > For one, if the DSTD is part of the BIOS, how come windows does
> > work, while it must rely on the same data
>
> I'm not an ACPI expert but here is my understanding.  Hopefully it's
> close enough to reality that the experts need to make only minor
> corrections and can keep hacking.
>
> Often, or rather always, the manufacturer tests the DSDT only with
> Windows.  If Windows makes assumptions about how the DSDT code should
> be interpreted that are different from what Linux assumes, even if
> Linux's assumptions are compatible with the ACPI spec, then the DSDT
> might fail when run with Linux.
>
> For example, Windows serializes (runs in one thread) procedures
> automatically, even if a procedure is labelled NotSerialized.  But
> Linux ACPI, unless you pass a boot option to tell it to act like
> Windows, will not serialize those procedures, and each one can run in
> its own thread.
>
> Therefore, race conditions from a sloppy DSDT, which might be hidden
> in Windows, are often exposed in Linux.  And so S3 suspend might hang
> because of one of those races or locking problems.  My old Thinkpad
> 600X improved its S3 suspend behavior somewhat when booting with the
> "Act like Windows" option.
>
> ACPI development on Linux is often: "My laptop doesn't suspend in
> Linux" (for a modern laptop it's almost certain that it suspends in
> Windows) and then people finding what bug needs to be introduced into
> the Linux DSDT interpreter so that Linux can act just like Windows.
>
> -Sanjoy
>
> `Never underestimate the evil of which men of power are capable.'
>          --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.
> -
> 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: the role of the dstd

[Sanjoy Mahajan]

> Is this option "act like windows" really available? Is it a boot
> parameter?

It is.  Add these options to your boot parameters:

  acpi_os_name="Microsoft Windows" acpi_serialize

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

RE: the role of the dstd

[Moore, Robert]

> For example, Windows serializes (runs in one thread) procedures 
> automatically, even if a procedure is labelled NotSerialized.  But 
> Linux ACPI, unless you pass a boot option to tell it to act like 
> Windows, will not serialize those procedures, and each one can run in 
> its own thread.

Actually, we think that Windows does NOT serialize everything within the AML interpreter.


> -----Original Message-----
> From: linux-acpi-owner@vger.kernel.org [mailto:linux-acpi-
> owner@vger.kernel.org] On Behalf Of Sanjoy Mahajan
> Sent: Wednesday, September 13, 2006 11:11 AM
> To: Ruslán Ledesma Garza
> Cc: linux-acpi@vger.kernel.org
> Subject: Re: the role of the dstd
> 
> > Is this option "act like windows" really available? Is it a boot
> > parameter?
> 
> It is.  Add these options to your boot parameters:
> 
>   acpi_os_name="Microsoft Windows" acpi_serialize
> 
> -Sanjoy
> 
> `Never underestimate the evil of which men of power are capable.'
>          --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.
> -
> 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
-
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: the role of the dstd

[Sanjoy Mahajan]

Interesting; thanks for the correction.  So acpi_serialize isn't
necessarily the right option to use?

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

RE: the role of the dstd

[Moore, Robert]

We have a "slack mode" for the interpreter which enables the "MS compatibility" features. For Linux, this mode is enabled by default.

Acpi_serialize is in fact used to serialize all methods, for those systems that have trouble with multiple threads. 

We do see systems that apparently do not fail on Windows, but fail on ACPICA based systems (fail with multithread issues.) We are still investigating these, perhaps timing related.

Bob


> -----Original Message-----
> From: linux-acpi-owner@vger.kernel.org [mailto:linux-acpi-
> owner@vger.kernel.org] On Behalf Of Sanjoy Mahajan
> Sent: Wednesday, September 13, 2006 12:05 PM
> To: Moore, Robert
> Cc: Ruslán Ledesma Garza; linux-acpi@vger.kernel.org
> Subject: Re: the role of the dstd
> 
> Interesting; thanks for the correction.  So acpi_serialize isn't
> necessarily the right option to use?
> 
> -Sanjoy
> 
> `Never underestimate the evil of which men of power are capable.'
>          --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.
> -
> 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
-
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: the role of the dstd

[Ruslán Ledesma Garza]

I've tested this boot option but seems to have no effect on my dell
inspiron 1150 bios version A07. Someone told me to file a bug on
bugzilla, so I expect to do it soon.

On 9/13/06, Moore, Robert <robert.moore@intel.com> wrote:
> We have a "slack mode" for the interpreter which enables the "MS compatibility" features. For Linux, this mode is enabled by default.
>
> Acpi_serialize is in fact used to serialize all methods, for those systems that have trouble with multiple threads.
>
> We do see systems that apparently do not fail on Windows, but fail on ACPICA based systems (fail with multithread issues.) We are still investigating these, perhaps timing related.
>
> Bob
>
>
> > -----Original Message-----
> > From: linux-acpi-owner@vger.kernel.org [mailto:linux-acpi-
> > owner@vger.kernel.org] On Behalf Of Sanjoy Mahajan
> > Sent: Wednesday, September 13, 2006 12:05 PM
> > To: Moore, Robert
> > Cc: Ruslán Ledesma Garza; linux-acpi@vger.kernel.org
> > Subject: Re: the role of the dstd
> >
> > Interesting; thanks for the correction.  So acpi_serialize isn't
> > necessarily the right option to use?
> >
> > -Sanjoy
> >
> > `Never underestimate the evil of which men of power are capable.'
> >          --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.
> > -
> > 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
>
-
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