[RFC] change suspend_finish order slightly

[RFC] change suspend_finish order slightly

[Shaohua Li]

[-- Attachment #1: Type: text/plain, Size: 967 bytes --]

The BIOS guys here require ACPI _WAK method which is called in
pm_ops->finish is invoked after all CPUs are resumed. Detail info can be
found at http://bugzilla.kernel.org/show_bug.cgi?id=5651. To do this,
this patch tries to change the order slightly. Does this break other
platforms?

Thanks,
Shaohua
---

 linux-2.6.14-root/kernel/power/main.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff -puN kernel/power/main.c~smp_s3_order kernel/power/main.c
--- linux-2.6.14/kernel/power/main.c~smp_s3_order	2005-11-03 14:18:29.000000000 +0800
+++ linux-2.6.14-root/kernel/power/main.c	2005-11-03 14:21:05.000000000 +0800
@@ -133,10 +133,10 @@ static int suspend_enter(suspend_state_t
 static void suspend_finish(suspend_state_t state)
 {
 	device_resume();
-	if (pm_ops && pm_ops->finish)
-		pm_ops->finish(state);
 	thaw_processes();
 	enable_nonboot_cpus();
+	if (pm_ops && pm_ops->finish)
+		pm_ops->finish(state);
 	pm_restore_console();
 }
 
_



[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: [RFC] change suspend_finish order slightly

[Patrick Mochel]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1081 bytes --]


On Fri, 2 Dec 2005, Shaohua Li wrote:

> The BIOS guys here require ACPI _WAK method which is called in
> pm_ops->finish is invoked after all CPUs are resumed. Detail info can be
> found at http://bugzilla.kernel.org/show_bug.cgi?id=5651. To do this,
> this patch tries to change the order slightly. Does this break other
> platforms?

>  linux-2.6.14-root/kernel/power/main.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff -puN kernel/power/main.c~smp_s3_order kernel/power/main.c
> --- linux-2.6.14/kernel/power/main.c~smp_s3_order	2005-11-03 14:18:29.000000000 +0800
> +++ linux-2.6.14-root/kernel/power/main.c	2005-11-03 14:21:05.000000000 +0800
> @@ -133,10 +133,10 @@ static int suspend_enter(suspend_state_t
>  static void suspend_finish(suspend_state_t state)
>  {
>  	device_resume();
> -	if (pm_ops && pm_ops->finish)
> -		pm_ops->finish(state);
>  	thaw_processes();
>  	enable_nonboot_cpus();
> +	if (pm_ops && pm_ops->finish)
> +		pm_ops->finish(state);
>  	pm_restore_console();
>  }
>
> _

ACKed-by Patrick Mochel <mochel@linux.intel.com>

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: [RFC] change suspend_finish order slightly

[Nigel Cunningham]

[-- Attachment #1: Type: text/plain, Size: 1448 bytes --]

Hi.

On Sat, 2005-12-03 at 13:15, Shaohua Li wrote:
> The BIOS guys here require ACPI _WAK method which is called in
> pm_ops->finish is invoked after all CPUs are resumed. Detail info can be
> found at http://bugzilla.kernel.org/show_bug.cgi?id=5651. To do this,
> this patch tries to change the order slightly. Does this break other
> platforms?

Is it needed even if you call the prepare after hot unplugging other
cpus? (If so, I need to modify suspend2, too).

Regards,

Nigel

> Thanks,
> Shaohua
> ---
> 
>  linux-2.6.14-root/kernel/power/main.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff -puN kernel/power/main.c~smp_s3_order kernel/power/main.c
> --- linux-2.6.14/kernel/power/main.c~smp_s3_order	2005-11-03 14:18:29.000000000 +0800
> +++ linux-2.6.14-root/kernel/power/main.c	2005-11-03 14:21:05.000000000 +0800
> @@ -133,10 +133,10 @@ static int suspend_enter(suspend_state_t
>  static void suspend_finish(suspend_state_t state)
>  {
>  	device_resume();
> -	if (pm_ops && pm_ops->finish)
> -		pm_ops->finish(state);
>  	thaw_processes();
>  	enable_nonboot_cpus();
> +	if (pm_ops && pm_ops->finish)
> +		pm_ops->finish(state);
>  	pm_restore_console();
>  }
>  
> _
> 
> 
> 
> ______________________________________________________________________
> _______________________________________________
> linux-pm mailing list
> linux-pm@lists.osdl.org
> https://lists.osdl.org/mailman/listinfo/linux-pm
-- 



[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

RE: [RFC] change suspend_finish order slightly

[Li, Shaohua]

[-- Attachment #1: Type: text/plain, Size: 566 bytes --]

Hi,
>
>On Sat, 2005-12-03 at 13:15, Shaohua Li wrote:
>> The BIOS guys here require ACPI _WAK method which is called in
>> pm_ops->finish is invoked after all CPUs are resumed. Detail info can
be
>> found at http://bugzilla.kernel.org/show_bug.cgi?id=5651. To do this,
>> this patch tries to change the order slightly. Does this break other
>> platforms?
>
>Is it needed even if you call the prepare after hot unplugging other
>cpus? (If so, I need to modify suspend2, too).
Yes. The ACPI _WAK method requires all CPU are booted per the BIOS guys.

Thanks,
Shaohua


[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

RE: [RFC] change suspend_finish order slightly

[Nigel Cunningham]

[-- Attachment #1: Type: text/plain, Size: 648 bytes --]

Hi.

On Tue, 2005-12-06 at 09:09, Li, Shaohua wrote:
> Hi,
> >
> >On Sat, 2005-12-03 at 13:15, Shaohua Li wrote:
> >> The BIOS guys here require ACPI _WAK method which is called in
> >> pm_ops->finish is invoked after all CPUs are resumed. Detail info can
> be
> >> found at http://bugzilla.kernel.org/show_bug.cgi?id=5651. To do this,
> >> this patch tries to change the order slightly. Does this break other
> >> platforms?
> >
> >Is it needed even if you call the prepare after hot unplugging other
> >cpus? (If so, I need to modify suspend2, too).
> Yes. The ACPI _WAK method requires all CPU are booted per the BIOS guys.

Ok. Thanks!

Nigel


[-- Attachment #2: Type: text/plain, Size: 0 bytes --]