diff mbox series

ACPI: x86: Adjust Microsoft LPS0 _DSM handling sequence

Message ID 20230601233953.1332-1-mario.limonciello@amd.com
State Accepted
Commit f198478cfdc8105a1c8d8945918904f0498d19be
Headers show
Series ACPI: x86: Adjust Microsoft LPS0 _DSM handling sequence | expand

Commit Message

Mario Limonciello June 1, 2023, 11:39 p.m. UTC
In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
and 8->6->4 for resume like Linux currently does.

Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
Align this calling order for Linux as well.

Link: https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
---
 drivers/acpi/x86/s2idle.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Comments

Mario Limonciello June 2, 2023, 1:46 a.m. UTC | #1
On 6/1/2023 8:31 PM, David E. Box wrote:
> On Thu, 2023-06-01 at 18:39 -0500, Mario Limonciello wrote:
>> In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
>> and 8->6->4 for resume like Linux currently does.
>>
>> Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
>> Align this calling order for Linux as well.
>>
>> Link:
>> https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
> I didn't catch the ordering in the link.

Yeah it's tough to interpret from the link, because the picture at the 
bottom
is missing annotations.

Basically if you look at the picture the blue part is the screen on/off.

The green part is "modern standby" and then the little "humps" are LPS0 
enter/exit.

> Was there any issue that prompted this
> change?


We were debugging an unrelated problem and noticed the difference 
comparing the

BIOS debugging log from Windows and Linux.

If an OEM depends on this call order in that code used in LPS0 phase 
requires
changes from MS phase I could hypothesize this fixes it.


> David

BTW - is there interest in supporting the Microsoft _DSM GUID for Intel 
side too?

It's an incongruity today that we run both AMD GUID and Microsoft GUID 
for AMD systems
but only run Intel GUID for Intel systems.

>
>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>> ---
>>   drivers/acpi/x86/s2idle.c | 14 +++++++-------
>>   1 file changed, 7 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c
>> index e499c60c4579..7214197c15a0 100644
>> --- a/drivers/acpi/x86/s2idle.c
>> +++ b/drivers/acpi/x86/s2idle.c
>> @@ -485,11 +485,11 @@ int acpi_s2idle_prepare_late(void)
>>                                          ACPI_LPS0_ENTRY,
>>                                          lps0_dsm_func_mask, lps0_dsm_guid);
>>          if (lps0_dsm_func_mask_microsoft > 0) {
>> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
>> -                               lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>>                  /* modern standby entry */
>>                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY,
>>                                  lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
>> +                               lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>>          }
>>   
>>          list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) {
>> @@ -524,11 +524,6 @@ void acpi_s2idle_restore_early(void)
>>                  if (handler->restore)
>>                          handler->restore();
>>   
>> -       /* Modern standby exit */
>> -       if (lps0_dsm_func_mask_microsoft > 0)
>> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
>> -                               lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>> -
>>          /* LPS0 exit */
>>          if (lps0_dsm_func_mask > 0)
>>                  acpi_sleep_run_lps0_dsm(acpi_s2idle_vendor_amd() ?
>> @@ -539,6 +534,11 @@ void acpi_s2idle_restore_early(void)
>>                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT,
>>                                  lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>>   
>> +       /* Modern standby exit */
>> +       if (lps0_dsm_func_mask_microsoft > 0)
>> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
>> +                               lps0_dsm_func_mask_microsoft,
>> lps0_dsm_guid_microsoft);
>> +
>>          /* Screen on */
>>          if (lps0_dsm_func_mask_microsoft > 0)
>>                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON,
David E. Box June 2, 2023, 3:06 a.m. UTC | #2
On Thu, 2023-06-01 at 20:46 -0500, Limonciello, Mario wrote:
> 
> On 6/1/2023 8:31 PM, David E. Box wrote:
> > On Thu, 2023-06-01 at 18:39 -0500, Mario Limonciello wrote:
> > > In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
> > > and 8->6->4 for resume like Linux currently does.
> > > 
> > > Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
> > > Align this calling order for Linux as well.
> > > 
> > > Link:
> > > https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
> > I didn't catch the ordering in the link.
> 
> Yeah it's tough to interpret from the link, because the picture at the 
> bottom
> is missing annotations.
> 
> Basically if you look at the picture the blue part is the screen on/off.
> 
> The green part is "modern standby" and then the little "humps" are LPS0 
> enter/exit.
> 
> > Was there any issue that prompted this
> > change?
> 
> 
> We were debugging an unrelated problem and noticed the difference 
> comparing the
> 
> BIOS debugging log from Windows and Linux.
> 
> If an OEM depends on this call order in that code used in LPS0 phase 
> requires
> changes from MS phase I could hypothesize this fixes it.
> 
> 
> > David
> 
> BTW - is there interest in supporting the Microsoft _DSM GUID for Intel 
> side too?
> 
> It's an incongruity today that we run both AMD GUID and Microsoft GUID 
> for AMD systems
> but only run Intel GUID for Intel systems.

There hasn't been a need yet. Rafael have you look at it?

David

> 
> > 
> > > Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> > > ---
> > >   drivers/acpi/x86/s2idle.c | 14 +++++++-------
> > >   1 file changed, 7 insertions(+), 7 deletions(-)
> > > 
> > > diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c
> > > index e499c60c4579..7214197c15a0 100644
> > > --- a/drivers/acpi/x86/s2idle.c
> > > +++ b/drivers/acpi/x86/s2idle.c
> > > @@ -485,11 +485,11 @@ int acpi_s2idle_prepare_late(void)
> > >                                          ACPI_LPS0_ENTRY,
> > >                                          lps0_dsm_func_mask,
> > > lps0_dsm_guid);
> > >          if (lps0_dsm_func_mask_microsoft > 0) {
> > > -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
> > > -                               lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > >                  /* modern standby entry */
> > >                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY,
> > >                                  lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > > +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
> > > +                               lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > >          }
> > >   
> > >          list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node)
> > > {
> > > @@ -524,11 +524,6 @@ void acpi_s2idle_restore_early(void)
> > >                  if (handler->restore)
> > >                          handler->restore();
> > >   
> > > -       /* Modern standby exit */
> > > -       if (lps0_dsm_func_mask_microsoft > 0)
> > > -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
> > > -                               lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > > -
> > >          /* LPS0 exit */
> > >          if (lps0_dsm_func_mask > 0)
> > >                  acpi_sleep_run_lps0_dsm(acpi_s2idle_vendor_amd() ?
> > > @@ -539,6 +534,11 @@ void acpi_s2idle_restore_early(void)
> > >                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT,
> > >                                  lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > >   
> > > +       /* Modern standby exit */
> > > +       if (lps0_dsm_func_mask_microsoft > 0)
> > > +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
> > > +                               lps0_dsm_func_mask_microsoft,
> > > lps0_dsm_guid_microsoft);
> > > +
> > >          /* Screen on */
> > >          if (lps0_dsm_func_mask_microsoft > 0)
> > >                  acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON,
Mario Limonciello June 2, 2023, 4:17 a.m. UTC | #3
On 6/1/2023 10:06 PM, David E. Box wrote:
> On Thu, 2023-06-01 at 20:46 -0500, Limonciello, Mario wrote:
>> On 6/1/2023 8:31 PM, David E. Box wrote:
>>> On Thu, 2023-06-01 at 18:39 -0500, Mario Limonciello wrote:
>>>> In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
>>>> and 8->6->4 for resume like Linux currently does.
>>>>
>>>> Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
>>>> Align this calling order for Linux as well.
>>>>
>>>> Link:
>>>> https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
>>> I didn't catch the ordering in the link.
>> Yeah it's tough to interpret from the link, because the picture at the
>> bottom
>> is missing annotations.
>>
>> Basically if you look at the picture the blue part is the screen on/off.
>>
>> The green part is "modern standby" and then the little "humps" are LPS0
>> enter/exit.
>>
>>> Was there any issue that prompted this
>>> change?
>>
>> We were debugging an unrelated problem and noticed the difference
>> comparing the
>>
>> BIOS debugging log from Windows and Linux.
>>
>> If an OEM depends on this call order in that code used in LPS0 phase
>> requires
>> changes from MS phase I could hypothesize this fixes it.
>>
>>
>>> David
>> BTW - is there interest in supporting the Microsoft _DSM GUID for Intel
>> side too?
>>
>> It's an incongruity today that we run both AMD GUID and Microsoft GUID
>> for AMD systems
>> but only run Intel GUID for Intel systems.
> There hasn't been a need yet. Rafael have you look at it?
>
> David

AFAICT from Linux side it should be a one line patch to drop:

         lps0_dsm_func_mask_microsoft = -EINVAL;

>>>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>>>> ---
>>>>    drivers/acpi/x86/s2idle.c | 14 +++++++-------
>>>>    1 file changed, 7 insertions(+), 7 deletions(-)
>>>>
>>>> diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c
>>>> index e499c60c4579..7214197c15a0 100644
>>>> --- a/drivers/acpi/x86/s2idle.c
>>>> +++ b/drivers/acpi/x86/s2idle.c
>>>> @@ -485,11 +485,11 @@ int acpi_s2idle_prepare_late(void)
>>>>                                           ACPI_LPS0_ENTRY,
>>>>                                           lps0_dsm_func_mask,
>>>> lps0_dsm_guid);
>>>>           if (lps0_dsm_func_mask_microsoft > 0) {
>>>> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
>>>> -                               lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>>                   /* modern standby entry */
>>>>                   acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY,
>>>>                                   lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
>>>> +                               lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>>           }
>>>>    
>>>>           list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node)
>>>> {
>>>> @@ -524,11 +524,6 @@ void acpi_s2idle_restore_early(void)
>>>>                   if (handler->restore)
>>>>                           handler->restore();
>>>>    
>>>> -       /* Modern standby exit */
>>>> -       if (lps0_dsm_func_mask_microsoft > 0)
>>>> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
>>>> -                               lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>> -
>>>>           /* LPS0 exit */
>>>>           if (lps0_dsm_func_mask > 0)
>>>>                   acpi_sleep_run_lps0_dsm(acpi_s2idle_vendor_amd() ?
>>>> @@ -539,6 +534,11 @@ void acpi_s2idle_restore_early(void)
>>>>                   acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT,
>>>>                                   lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>>    
>>>> +       /* Modern standby exit */
>>>> +       if (lps0_dsm_func_mask_microsoft > 0)
>>>> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
>>>> +                               lps0_dsm_func_mask_microsoft,
>>>> lps0_dsm_guid_microsoft);
>>>> +
>>>>           /* Screen on */
>>>>           if (lps0_dsm_func_mask_microsoft > 0)
>>>>                   acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON,
Rafael J. Wysocki June 2, 2023, 2:51 p.m. UTC | #4
On Fri, Jun 2, 2023 at 5:06 AM David E. Box <david.e.box@linux.intel.com> wrote:
>
> On Thu, 2023-06-01 at 20:46 -0500, Limonciello, Mario wrote:
> >
> > On 6/1/2023 8:31 PM, David E. Box wrote:
> > > On Thu, 2023-06-01 at 18:39 -0500, Mario Limonciello wrote:
> > > > In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
> > > > and 8->6->4 for resume like Linux currently does.
> > > >
> > > > Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
> > > > Align this calling order for Linux as well.
> > > >
> > > > Link:
> > > > https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
> > > I didn't catch the ordering in the link.
> >
> > Yeah it's tough to interpret from the link, because the picture at the
> > bottom
> > is missing annotations.
> >
> > Basically if you look at the picture the blue part is the screen on/off.
> >
> > The green part is "modern standby" and then the little "humps" are LPS0
> > enter/exit.
> >
> > > Was there any issue that prompted this
> > > change?
> >
> >
> > We were debugging an unrelated problem and noticed the difference
> > comparing the
> >
> > BIOS debugging log from Windows and Linux.
> >
> > If an OEM depends on this call order in that code used in LPS0 phase
> > requires
> > changes from MS phase I could hypothesize this fixes it.
> >
> >
> > > David
> >
> > BTW - is there interest in supporting the Microsoft _DSM GUID for Intel
> > side too?
> >
> > It's an incongruity today that we run both AMD GUID and Microsoft GUID
> > for AMD systems
> > but only run Intel GUID for Intel systems.
>
> There hasn't been a need yet. Rafael have you look at it?

Nothing official ATM AFAICS.

But I guess it'll need to be used on Intel at one point too.
Rafael J. Wysocki June 5, 2023, 5:24 p.m. UTC | #5
On Fri, Jun 2, 2023 at 1:40 AM Mario Limonciello
<mario.limonciello@amd.com> wrote:
>
> In Windows the Microsoft _DSM doesn't call functions 3->5->7 for suspend
> and 8->6->4 for resume like Linux currently does.
>
> Rather it calls 3->7->5 for suspend and 6->8->4 for resume.
> Align this calling order for Linux as well.
>
> Link: https://learn.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-states
> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> ---
>  drivers/acpi/x86/s2idle.c | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c
> index e499c60c4579..7214197c15a0 100644
> --- a/drivers/acpi/x86/s2idle.c
> +++ b/drivers/acpi/x86/s2idle.c
> @@ -485,11 +485,11 @@ int acpi_s2idle_prepare_late(void)
>                                         ACPI_LPS0_ENTRY,
>                                         lps0_dsm_func_mask, lps0_dsm_guid);
>         if (lps0_dsm_func_mask_microsoft > 0) {
> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
> -                               lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
>                 /* modern standby entry */
>                 acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY,
>                                 lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
> +                               lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
>         }
>
>         list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) {
> @@ -524,11 +524,6 @@ void acpi_s2idle_restore_early(void)
>                 if (handler->restore)
>                         handler->restore();
>
> -       /* Modern standby exit */
> -       if (lps0_dsm_func_mask_microsoft > 0)
> -               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
> -                               lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
> -
>         /* LPS0 exit */
>         if (lps0_dsm_func_mask > 0)
>                 acpi_sleep_run_lps0_dsm(acpi_s2idle_vendor_amd() ?
> @@ -539,6 +534,11 @@ void acpi_s2idle_restore_early(void)
>                 acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT,
>                                 lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
>
> +       /* Modern standby exit */
> +       if (lps0_dsm_func_mask_microsoft > 0)
> +               acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
> +                               lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
> +
>         /* Screen on */
>         if (lps0_dsm_func_mask_microsoft > 0)
>                 acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON,
> --

Applied as 6.5 material, thanks!
diff mbox series

Patch

diff --git a/drivers/acpi/x86/s2idle.c b/drivers/acpi/x86/s2idle.c
index e499c60c4579..7214197c15a0 100644
--- a/drivers/acpi/x86/s2idle.c
+++ b/drivers/acpi/x86/s2idle.c
@@ -485,11 +485,11 @@  int acpi_s2idle_prepare_late(void)
 					ACPI_LPS0_ENTRY,
 					lps0_dsm_func_mask, lps0_dsm_guid);
 	if (lps0_dsm_func_mask_microsoft > 0) {
-		acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
-				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
 		/* modern standby entry */
 		acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_ENTRY,
 				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
+		acpi_sleep_run_lps0_dsm(ACPI_LPS0_ENTRY,
+				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
 	}
 
 	list_for_each_entry(handler, &lps0_s2idle_devops_head, list_node) {
@@ -524,11 +524,6 @@  void acpi_s2idle_restore_early(void)
 		if (handler->restore)
 			handler->restore();
 
-	/* Modern standby exit */
-	if (lps0_dsm_func_mask_microsoft > 0)
-		acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
-				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
-
 	/* LPS0 exit */
 	if (lps0_dsm_func_mask > 0)
 		acpi_sleep_run_lps0_dsm(acpi_s2idle_vendor_amd() ?
@@ -539,6 +534,11 @@  void acpi_s2idle_restore_early(void)
 		acpi_sleep_run_lps0_dsm(ACPI_LPS0_EXIT,
 				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
 
+	/* Modern standby exit */
+	if (lps0_dsm_func_mask_microsoft > 0)
+		acpi_sleep_run_lps0_dsm(ACPI_LPS0_MS_EXIT,
+				lps0_dsm_func_mask_microsoft, lps0_dsm_guid_microsoft);
+
 	/* Screen on */
 	if (lps0_dsm_func_mask_microsoft > 0)
 		acpi_sleep_run_lps0_dsm(ACPI_LPS0_SCREEN_ON,