Skip to contentSkip to conversation

zaurus needs generic pxa suspend/resume routines

PATCH Pavel Machek 4 participants 7 messages Jun 6, 2009 – Jun 9, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

For suspend/resume to work, spitz needs pxa_pm_suspend/resume to be called. Otherwise PSPR is not set properly, and system will die during resume.

PATCH FOLLOWS
KernelVersion: 2.6.30-rc6

arch/arm/common/sharpsl_pm.c renamed from arch/arm.ofic/common/sharpsl_pm.c +3 -1
@@ -553,7 +556,7 @@
 			corgi_goto_sleep(alarm_time, alarm_enable, state);
 			return 1;
 		}
-		if(sharpsl_off_charge_battery()) {
+		if (sharpsl_off_charge_battery()) {
 			dev_dbg(sharpsl_pm.dev, "Charging. Suspend...\n");
 			corgi_goto_sleep(alarm_time, alarm_enable, state);
 			return 1;
@@ -774,7 +777,9 @@
 }
 
 #ifdef CONFIG_PM
 static struct platform_suspend_ops sharpsl_pm_ops = {
+	.prepare	= pxa_pm_prepare,
+	.finish		= pxa_pm_finish,
 	.enter		= corgi_pxa_pm_enter,
 	.valid		= suspend_valid_only_mem,
 };
arch/arm/mach-pxa/include/mach/pm.h renamed from arch/arm.ofic/mach-pxa/include/mach/pm.h +2 -0
diff -ur linux-rc/arch/arm.ofic/mach-pxa/include/mach/pm.h linux-rc/arch/arm/mach-pxa/include/mach/pm.h
@@ -27,6 +27,8 @@
 extern void pxa_cpu_resume(void);
 
 extern int pxa_pm_enter(suspend_state_t state);
+extern int pxa_pm_prepare(void);
+extern void pxa_pm_finish(void);
 
 /* NOTE: this is for PM debugging on Lubbock,  it's really a big
  * ugly, but let's keep the crap minimum here, instead of direct
arch/arm/mach-pxa/pm.c renamed from arch/arm.ofic/mach-pxa/pm.c +2 -2
diff -ur linux-rc/arch/arm.ofic/mach-pxa/pm.c linux-rc/arch/arm/mach-pxa/pm.c
@@ -79,7 +84,7 @@
 	return -EINVAL;
 }
 
-static int pxa_pm_prepare(void)
+int pxa_pm_prepare(void)
 {
 	int ret = 0;
 
@@ -89,7 +94,7 @@
 	return ret;
 }
 
-static void pxa_pm_finish(void)
+void pxa_pm_finish(void)
 {
 	if (pxa_cpu_pm_fns && pxa_cpu_pm_fns->finish)
 		pxa_cpu_pm_fns->finish();
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#1 lore raw
Quoted 3 lines · Pavel Machek

For suspend/resume to work, spitz needs pxa_pm_suspend/resume to be called. Otherwise PSPR is not set properly, and system will die during resume.

100 points for copying lots of people. 0 points for omitting the right person. Add Eric.

Eric, ack?

Quoted 68 lines

PATCH FOLLOWS
KernelVersion: 2.6.30-rc6

arch/arm/common/sharpsl_pm.c renamed from arch/arm.ofic/common/sharpsl_pm.c +3 -1
@@ -553,7 +556,7 @@
 			corgi_goto_sleep(alarm_time, alarm_enable, state);
 			return 1;
 		}
-		if(sharpsl_off_charge_battery()) {
+		if (sharpsl_off_charge_battery()) {
 			dev_dbg(sharpsl_pm.dev, "Charging. Suspend...\n");
 			corgi_goto_sleep(alarm_time, alarm_enable, state);
 			return 1;
@@ -774,7 +777,9 @@
 }
 
 #ifdef CONFIG_PM
 static struct platform_suspend_ops sharpsl_pm_ops = {
+	.prepare	= pxa_pm_prepare,
+	.finish		= pxa_pm_finish,
 	.enter		= corgi_pxa_pm_enter,
 	.valid		= suspend_valid_only_mem,
 };

diff -ur linux-rc/arch/arm.ofic/mach-pxa/include/mach/pm.h linux-rc/arch/arm/mach-pxa/include/mach/pm.h

arch/arm/mach-pxa/include/mach/pm.h renamed from arch/arm.ofic/mach-pxa/include/mach/pm.h +2 -0
@@ -27,6 +27,8 @@
 extern void pxa_cpu_resume(void);
 
 extern int pxa_pm_enter(suspend_state_t state);
+extern int pxa_pm_prepare(void);
+extern void pxa_pm_finish(void);
 
 /* NOTE: this is for PM debugging on Lubbock,  it's really a big
  * ugly, but let's keep the crap minimum here, instead of direct

diff -ur linux-rc/arch/arm.ofic/mach-pxa/pm.c linux-rc/arch/arm/mach-pxa/pm.c

arch/arm/mach-pxa/pm.c renamed from arch/arm.ofic/mach-pxa/pm.c +2 -2
@@ -79,7 +84,7 @@
 	return -EINVAL;
 }
 
-static int pxa_pm_prepare(void)
+int pxa_pm_prepare(void)
 {
 	int ret = 0;
 
@@ -89,7 +94,7 @@
 	return ret;
 }
 
-static void pxa_pm_finish(void)
+void pxa_pm_finish(void)
 {
 	if (pxa_cpu_pm_fns && pxa_cpu_pm_fns->finish)
 		pxa_cpu_pm_fns->finish();
Signature · 8 lines
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php
Acked-by
#2 lore raw
Quoted 10 lines · Russell King - ARM Linux
Quoted 3 lines · Pavel Machek

For suspend/resume to work, spitz needs pxa_pm_suspend/resume to be called. Otherwise PSPR is not set properly, and system will die during resume.

100 points for copying lots of people.  0 points for omitting the right person.  Add Eric.

Eric, ack?

Yes, I think this is the best trade-off for the moment.

And Russell you may want to queue this for this -rc phase, looks it's still -rc8-git5.

#3 lore raw
Quoted 13 lines · Eric Miao
Quoted 10 lines · Russell King - ARM Linux
Quoted 3 lines · Pavel Machek

For suspend/resume to work, spitz needs pxa_pm_suspend/resume to be called. Otherwise PSPR is not set properly, and system will die during resume.

100 points for copying lots of people.  0 points for omitting the right person.  Add Eric.

Eric, ack?

Quoted 6 lines

Yes, I think this is the best trade-off for the moment.

And Russell you may want to queue this for this -rc phase, looks it's still -rc8-git5.

I think it's missed .30, unless Linus does a -rc9. I sent what was probably my final pull request last night.

#4 lore raw
Quoted 16 lines · Russell King - ARM Linux
Quoted 15 lines · Eric Miao
Quoted 10 lines · Russell King - ARM Linux
Quoted 3 lines · Pavel Machek

For suspend/resume to work, spitz needs pxa_pm_suspend/resume to be called. Otherwise PSPR is not set properly, and system will die during resume.

100 points for copying lots of people.  0 points for omitting the right person.  Add Eric.

Eric, ack?

Yes, I think this is the best trade-off for the moment.

Quoted 8 lines
Quoted 4 lines

And Russell you may want to queue this for this -rc phase, looks it's still -rc8-git5.

I think it's missed .30, unless Linus does a -rc9.  I sent what was probably my final pull request last night.

That should be fine. There are some other changes in my 'sharpsl_pm' branch, so I'd queue this patch for the next merge window and make some adjustment.

Signature · 2 lines
Cheers
- eric
#5 lore raw
Quoted 4 lines · Eric Miao
Quoted 4 lines
Quoted 4 lines

Yes, I think this is the best trade-off for the moment.

However resume still does not work with this patch on top of yesterday's Linus tree snapshot (SL-C3200, ccc0d38). It seems to go to resume for ~300 milliseconds (orange LED lights for a moment, power consumption increases) and then runs back to the sleep. It seems that it disables offline charging, too. When battery cover switch is open it goes to resume just for ~50 milliseconds (orange LED does not light).

The last Pavel's hack from last week at least fixed resume at cost of breaking offline charging (and broken serial after resume).

Signature · 9 lines
Best Regards / S pozdravem,

Stanislav Brabec
software developer
---------------------------------------------------------------------
SUSE LINUX, s. r. o.                          e-mail: [email protected]
Lihovarská 1060/12           tel: +420 284 028 966, +49 911 740538747
190 00 Praha 9                                  fax: +420 284 028 951
Czech Republic                                    http://www.suse.cz/
#6 lore raw
Quoted 7 lines · Stanislav Brabec
Quoted 4 lines · Eric Miao
Quoted 4 lines
Quoted 4 lines

Yes, I think this is the best trade-off for the moment.

Quoted 6 lines

However resume still does not work with this patch on top of yesterday's Linus tree snapshot (SL-C3200, ccc0d38). It seems to go to resume for ~300 milliseconds (orange LED lights for a moment, power consumption increases) and then runs back to the sleep. It seems that it disables offline charging, too. When battery cover switch is open it goes to resume just for ~50 milliseconds (orange LED does not light).

Yes, the patch fixes just the "cpu registers not set correctly" problem. Problem with charging remains. Unfortunately, I do have patch to fix that, yet. (Yes, it can be worked around by disabling offline charge, but that is not a fix).

Quoted 2 lines

The last Pavel's hack from last week at least fixed resume at cost of breaking offline charging (and broken serial after resume).

Hmm, do you have a serial cable? :-).
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html