Skip to contentSkip to conversation

"reboot" swsusp mode leaves moon icon blinking

PATCH Pavel Machek 3 participants 9 messages Mar 20, 2007

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi!

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

Unfortunately I see no easy solution, and this may/will cause other problems -- in case of broken bios and user telling us not to call that bios, we'll call it anyway.

(Ouch and I think this is regression after 2.6.20?)

Pavel

kernel/power/disk.c +0 -8
index 873cdf8..dee0ff4 100644
@@ -241,18 +241,11 @@ static int software_resume(void)
 		goto Done;
 	}
 
-	error = platform_prepare();
-	if (error) {
-		swsusp_free();
-		goto Thaw;
-	}
-
 	pr_debug("PM: Reading swsusp image.\n");
 
 	error = swsusp_read();
 	if (error) {
 		swsusp_free();
-		platform_finish();
 		goto Thaw;
 	}
 
@@ -270,7 +263,6 @@ static int software_resume(void)
 	enable_nonboot_cpus();
  Free:
 	swsusp_free();
-	platform_finish();
 	device_resume();
 	resume_console();
  Thaw:
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

Hi,

Quoted 7 lines · Pavel Machek

Date: Tue, 20 Mar 2007 13:32:45 +0100
From: Pavel Machek <[email protected]>
To: "Rafael J. Wysocki" <[email protected]>,
kernel list <[email protected]>
Cc: Andrew Morton <[email protected]>
Subject: "reboot" swsusp mode leaves moon icon blinking

Quoted 6 lines

Hi!

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

What does the "reboot" swsusp mean?

I am having this (or a similar problem):

echo shutdown > /sys/power/disk; echo disk > /sys/power/state

suspends normally. But after a resume, the moon keeps blinking _and_ I don't get ACPI events anymore. At least all ibm-acpi features refuse to work (hotkeys, lcd-brightness, lid, ...).

I saved some /proc files (interrupts, acpi/processor/power, ...) and compared them before the suspend and after the resume but I could not find any differences.

Quoted 5 lines
kernel/power/disk.c +0 -0
index 873cdf8..dee0ff4 100644

kernel/power/disk.c static int software_resume(void) · line 249

Quoted 9 lines
kernel/power/disk.c +0 -6
@@ -241,18 +241,11 @@ static int software_resume(void)
 		goto Done;
 	}
 
-	error = platform_prepare();
-	if (error) {
-		swsusp_free();
-		goto Thaw;
-	}
-

What is platform_prepare() actually good for when _resuming_? I tried to follow the function calls and it seems to disable a lot of thing instead of re-enabling them...?

=Hannes

#2 lore raw
Quoted 39 lines · Johannes Weiner

Hi,

Quoted 7 lines · Pavel Machek

Date: Tue, 20 Mar 2007 13:32:45 +0100
From: Pavel Machek <[email protected]>
To: "Rafael J. Wysocki" <[email protected]>,
kernel list <[email protected]>
Cc: Andrew Morton <[email protected]>
Subject: "reboot" swsusp mode leaves moon icon blinking

Quoted 6 lines

Hi!

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

What does the "reboot" swsusp mean?

I am having this (or a similar problem):

echo shutdown > /sys/power/disk; echo disk > /sys/power/state

suspends normally. But after a resume, the moon keeps blinking _and_ I don't get ACPI events anymore. At least all ibm-acpi features refuse to work (hotkeys, lcd-brightness, lid, ...).

I saved some /proc files (interrupts, acpi/processor/power, ...) and compared them before the suspend and after the resume but I could not find any differences.

Quoted 8 lines
kernel/power/disk.c +0 -0
index 873cdf8..dee0ff4 100644
@@ -241,18 +241,11 @@ static int software_resume(void)
 		goto Done;
 	}
 
Quoted 10 lines

kernel/power/disk.c static int software_resume(void) · line 249

Quoted 6 lines
kernel/power/disk.c +0 -6
-	error = platform_prepare();
-	if (error) {
-		swsusp_free();
-		goto Thaw;
-	}
-

What is platform_prepare() actually good for when _resuming_? I tried to follow the function calls and it seems to disable a lot of thing instead of re-enabling them...?

Well, if you suspend using the platform mode, platform_finish() will be called after the image has been restored (ie. by the "old" kernel).

Greetings,
Rafael

Signature · 3 lines
If you don't have the time to read,
you don't have the time or the tools to write.
		- Stephen King
#3 lore raw
In reply to Johannes Weiner

Hi!

Quoted 6 lines
Quoted 4 lines

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

What does the "reboot" swsusp mean?

echo reboot > ...disk.

Quoted 3 lines

I am having this (or a similar problem):

echo shutdown > /sys/power/disk; echo disk > /sys/power/state

Yep, that's duplicate. Try patch below.
Pavel

Quoted 29 lines

I saved some /proc files (interrupts, acpi/processor/power, ...) and compared them before the suspend and after the resume but I could not find any differences.

Quoted 5 lines
kernel/power/disk.c +0 -0
index 873cdf8..dee0ff4 100644

kernel/power/disk.c static int software_resume(void) · line 249

Quoted 9 lines
kernel/power/disk.c +0 -6
@@ -241,18 +241,11 @@ static int software_resume(void)
 		goto Done;
 	}
 
-	error = platform_prepare();
-	if (error) {
-		swsusp_free();
-		goto Thaw;
-	}
-

What is platform_prepare() actually good for when _resuming_? I tried to follow the function calls and it seems to disable a lot of thing instead of re-enabling them...?

=Hannes
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

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

Hi,

Quoted 3 lines · Pavel Machek
Quoted 1 line

What does the "reboot" swsusp mean?

echo reboot > ...disk.

Thank you.

Quoted 5 lines
Quoted 3 lines

I am having this (or a similar problem):

echo shutdown > /sys/power/disk; echo disk > /sys/power/state

Yep, that's duplicate. Try patch below.

I applied it and the box is suspending/resuming properly again.

=Hannes

#5 lore raw
In reply to Pavel Machek
Quoted 4 lines · Pavel Machek

Hi!

...and cause is really simple.

Quoted 8 lines

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

Unfortunately I see no easy solution, and this may/will cause other problems -- in case of broken bios and user telling us not to call that bios, we'll call it anyway.

(Ouch and I think this is regression after 2.6.20?)

I'm not sure, will check that later.

The solution would be to pass the information about the suspend mode used in the image header, I think (that's what uswsusp does after my recent patch).

Greetings,
Rafael

#6 lore raw

Hi!

Quoted 9 lines
Quoted 9 lines

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

Unfortunately I see no easy solution, and this may/will cause other problems -- in case of broken bios and user telling us not to call that bios, we'll call it anyway.

Quoted 6 lines
Quoted 1 line

(Ouch and I think this is regression after 2.6.20?)

I'm not sure, will check that later.

The solution would be to pass the information about the suspend mode used in the image header, I think (that's what uswsusp does after my recent patch).

Yes, we could do that.

OTOH, we could simply avoid calling the platform code in resume path. It worked ok for a long while, and it seems to have no downsides...

Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#7 lore raw
Quoted 15 lines · Pavel Machek

Hi!

Quoted 13 lines
Quoted 10 lines

...and cause is really simple.

During resume, we do not know that "reboot" method was used, so we assume plaform and make the led blink...

Unfortunately I see no easy solution, and this may/will cause other problems -- in case of broken bios and user telling us not to call that bios, we'll call it anyway.

(Ouch and I think this is regression after 2.6.20?)

I'm not sure, will check that later.

Quoted 8 lines
Quoted 2 lines

The solution would be to pass the information about the suspend mode used in the image header, I think (that's what uswsusp does after my recent patch).

Yes, we could do that.

OTOH, we could simply avoid calling the platform code in resume path. It worked ok for a long while, and it seems to have no downsides...

No, we have always done it, actually, but not in this particular place. ;-)

In theory, avoiding it could be problematic, because of the pm_ops->finish() that gets called after the image has been restored (if the platform mode was used for suspending). Still, I think we can try.

Greetings,
Rafael

#8 lore raw

Hi!

Quoted 11 lines
Quoted 5 lines

Yes, we could do that.

OTOH, we could simply avoid calling the platform code in resume path. It worked ok for a long while, and it seems to have no downsides...

No, we have always done it, actually, but not in this particular place. ;-)

In theory, avoiding it could be problematic, because of the pm_ops->finish() that gets called after the image has been restored (if the platform mode was used for suspending). Still, I think we can try.

If we can return to the method "how we did it in 2.6.20", that would indeed be preffered ;-).
Pavel

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