[RFC] poweroff via pm_power_off if set

[RFC] poweroff via pm_power_off if set

[Christian Krafft]

[-- Attachment #1: Type: text/plain, Size: 1627 bytes --]

This patch is needed to make ipmi_poweroff working on powerpc.
It straightens the poweroff procedure to match with other architectures.
At the moment powerpc plattforms can define their poweroff method using
ppc_md.power_off. The only way for plattform independent driver (ipmi_poweroff)
to register it's poweroff function is to use pm_power_off. So machine_power_off
should check whether pm_power_off has been set and if so it should be used.
If not, plattform dependent ppc_md.power_off should be called.

Signed-off-by: Christian Krafft <krafft@de.ibm.com>

Index: linux.git/arch/powerpc/kernel/setup-common.c
===================================================================
--- linux.git.orig/arch/powerpc/kernel/setup-common.c
+++ linux.git/arch/powerpc/kernel/setup-common.c
@@ -121,6 +121,8 @@ void machine_restart(char *cmd)
 void machine_power_off(void)
 {
 	machine_shutdown();
+	if (pm_power_off)
+		pm_power_off();
 	if (ppc_md.power_off)
 		ppc_md.power_off();
 #ifdef CONFIG_SMP
@@ -133,7 +135,7 @@ void machine_power_off(void)
 /* Used by the G5 thermal driver */
 EXPORT_SYMBOL_GPL(machine_power_off);
 
-void (*pm_power_off)(void) = machine_power_off;
+void (*pm_power_off)(void) = NULL;
 EXPORT_SYMBOL_GPL(pm_power_off);
 
 void machine_halt(void)


-- 
Mit freundlichen Gruessen,
kind regards,

Christian Krafft
IBM Systems & Technology Group,
Linux Kernel Development
IT Specialist


Vorsitzender des Aufsichtsrats:	Martin Jetter
Geschaeftsfuehrung:		Herbert Kircher
Sitz der Gesellschaft:		Boeblingen
Registriergericht:		Amtsgericht Stuttgart, HRB 243294

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

Re: [RFC] poweroff via pm_power_off if set

[Kumar Gala]

On Jun 13, 2008, at 10:36 AM, Christian Krafft wrote:

> This patch is needed to make ipmi_poweroff working on powerpc.
> It straightens the poweroff procedure to match with other  
> architectures.
> At the moment powerpc plattforms can define their poweroff method  
> using
> ppc_md.power_off. The only way for plattform independent driver  
> (ipmi_poweroff)
> to register it's poweroff function is to use pm_power_off. So  
> machine_power_off
> should check whether pm_power_off has been set and if so it should  
> be used.
> If not, plattform dependent ppc_md.power_off should be called.

It seems like this should be the other way around.  Meaning  
ppc_md.power_off should be called first than pm_power_off.  If a  
platform provides a power off mechanism that should presuming take  
precedence over the ipmi_poweroff.

- k

Re: [RFC] poweroff via pm_power_off if set

[Christian Krafft]

[-- Attachment #1: Type: text/plain, Size: 1593 bytes --]

On Fri, 13 Jun 2008 11:00:33 -0500
Kumar Gala <galak@kernel.crashing.org> wrote:

> 
> On Jun 13, 2008, at 10:36 AM, Christian Krafft wrote:
> 
> > This patch is needed to make ipmi_poweroff working on powerpc.
> > It straightens the poweroff procedure to match with other  
> > architectures.
> > At the moment powerpc plattforms can define their poweroff method  
> > using
> > ppc_md.power_off. The only way for plattform independent driver  
> > (ipmi_poweroff)
> > to register it's poweroff function is to use pm_power_off. So  
> > machine_power_off
> > should check whether pm_power_off has been set and if so it should  
> > be used.
> > If not, plattform dependent ppc_md.power_off should be called.
> 
> It seems like this should be the other way around.  Meaning  
> ppc_md.power_off should be called first than pm_power_off.  If a  
> platform provides a power off mechanism that should presuming take  
> precedence over the ipmi_poweroff.
> 
> - k

I'd see the pm_power_off as an option to override platform specific poweroff
function. If you want to use ppc_md.power_off you simply don't need to load
ipmi_poweroff module. Also it would not be possible to use ipmi_poweroff
on plattforms that provide a ppc_md.power_off. 


-- 
Mit freundlichen Gruessen,
kind regards,

Christian Krafft
IBM Systems & Technology Group,
Linux Kernel Development
IT Specialist


Vorsitzender des Aufsichtsrats:	Martin Jetter
Geschaeftsfuehrung:		Herbert Kircher
Sitz der Gesellschaft:		Boeblingen
Registriergericht:		Amtsgericht Stuttgart, HRB 243294

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