Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Pavel Machek]

On Thu 2008-04-17 22:21:25, Andrew Morton wrote:
> This is the Vaio.
> 
> This:
> 
> <cold boot>
> <suspend-to-ram>
> <resume-from-ram>
> <suspend-to-ram>
> <resume-from-ram>
> 
> works OK.
> 
> 
> This:
> 
> <cold boot>
> <suspend-to-disk>
> <resume-from-disk>
> 
> works OK.
> 
> 
> But this:
> 
> <cold boot>
> <suspend-to-ram>
> <resume-from-ram>
> <suspend-to-disk>
> <resume-from-disk>
> 
> hangs at the end of resume-from-disk.

You could try echo disk > /sys/power/state to see if it is ACPI acting
funny or some driver

> During the suspend-to-disk I saw:
> 
> ipw2200: Radio Frequency Kill Switch is On:
> Kill switch must be turned off for wireless networking to work.
> PM: Device usb1 failed to (unknown PM event): error -113
> PM: Device usb2 failed to (unknown PM event): error -113
> PM: Device usb3 failed to (unknown PM event): error -113
> PM: Device usb4 failed to (unknown PM event): error -113
> Suspending console(s)
> 
> But it suspends OK.



> Resume _almost_ works.  It gets to:
> 
> calling  hpet_insert_resource+0x0/0x1e()
> initcall hpet_insert_resource+0x0/0x1e() returned 1 after 0 msecs
> initcall hpet_insert_resource+0x0/0x1e() returned with error code 1
> calling  init_lapic_nmi_sysfs+0x0/0x33()
> initcall init_lapic_nmi_sysfs+0x0/0x33() returned 0 after 0 msecs
> calling  io_apic_bug_finalize+0x0/0x1c()
> initcall io_apic_bug_finalize+0x0/0x1c() returned 0 after 0 msecs
> calling  check_early_ioremap_leak+0x0/0x3d()
> initcall check_early_ioremap_leak+0x0/0x3d() returned 0 after 0 msecs
> calling  print_ipi_mode+0x0/0x26()
> Using IPI Shortcut mode
> initcall print_ipi_mode+0x0/0x26() returned 0 after 2 msecs
> calling  init_oops_id+0x0/0x23()
> initcall init_oops_id+0x0/0x23() returned 0 after 0 msecs
> calling  disable_boot_consoles+0x0/0x36()
> initcall disable_boot_consoles+0x0/0x36() returned 0 after 0 msecs
> calling  pm_qos_power_init+0x0/0x54()
> initcall pm_qos_power_init+0x0/0x54() returned 0 after 0 msecs
> calling  software_resume+0x0/0x179()
> Freezing user space processes ... (elapsed 0.00 seconds) done.
> Freezing remaining freezable tasks ... <6>input: PS/2 Mouse as /class/input/input3
> input: AlpsPS/2 ALPS GlidePoint as /class/input/input4
> (elapsed 0.14 seconds) done.
> PM: Loading image data pages (88965 pages) ... done
> PM: Read 355860 kbytes in 16.36 seconds (21.75 MB/s)
> Suspending console(s)

no_console_suspend should get you a bit more output...

> I guess I get to bisect this.

Either that, or you can try if it works with minimum config (no
wireless, no usb, no fancy drivers).
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Pavel Machek]

Hi!

> This is the Vaio.
> 
> This:
> 
> <cold boot>
> <suspend-to-ram>
> <resume-from-ram>
> <suspend-to-ram>
> <resume-from-ram>
> 
> works OK.
> 
> 
> This:
> 
> <cold boot>
> <suspend-to-disk>
> <resume-from-disk>
> 
> works OK.
> 
> 
> But this:
> 
> <cold boot>
> <suspend-to-ram>
> <resume-from-ram>
> <suspend-to-disk>
> <resume-from-disk>

Works for me on thinkpad x60.

Problems I'm aware in this area: MCE driver does not like suspend
(compile it out), and iommu has problems (iommu=soft solves it).
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Andrew Morton]

On Fri, 18 Apr 2008 13:08:13 +0200 Pavel Machek <pavel@ucw.cz> wrote:

> Hi!
> 
> > This is the Vaio.
> > 
> > This:
> > 
> > <cold boot>
> > <suspend-to-ram>
> > <resume-from-ram>
> > <suspend-to-ram>
> > <resume-from-ram>
> > 
> > works OK.
> > 
> > 
> > This:
> > 
> > <cold boot>
> > <suspend-to-disk>
> > <resume-from-disk>
> > 
> > works OK.
> > 
> > 
> > But this:
> > 
> > <cold boot>
> > <suspend-to-ram>
> > <resume-from-ram>
> > <suspend-to-disk>
> > <resume-from-disk>
> 
> Works for me on thinkpad x60.
> 
> Problems I'm aware in this area: MCE driver does not like suspend
> (compile it out), and iommu has problems (iommu=soft solves it).

Yeah.  I got down to bisect the Vaio failure and it went away.  But it was
there.

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Alan Stern]

On Thu, 17 Apr 2008, Andrew Morton wrote:

> But this:
> 
> <cold boot>
> <suspend-to-ram>
> <resume-from-ram>
> <suspend-to-disk>
> <resume-from-disk>
> 
> hangs at the end of resume-from-disk.
> 
> 
> During the suspend-to-disk I saw:
> 
> ipw2200: Radio Frequency Kill Switch is On:
> Kill switch must be turned off for wireless networking to work.
> PM: Device usb1 failed to (unknown PM event): error -113
> PM: Device usb2 failed to (unknown PM event): error -113
> PM: Device usb3 failed to (unknown PM event): error -113
> PM: Device usb4 failed to (unknown PM event): error -113

These -113 errors can be ignored.  They are fallout from the recent 
conversion of the driver core's suspend/resume API.  Matching changes 
have not yet been added to the USB core.

The "(unknown PM event)" strings look like a bug, though.

Alan Stern

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Greg KH]

On Fri, Apr 18, 2008 at 11:19:51AM -0400, Alan Stern wrote:
> On Thu, 17 Apr 2008, Andrew Morton wrote:
> 
> > But this:
> > 
> > <cold boot>
> > <suspend-to-ram>
> > <resume-from-ram>
> > <suspend-to-disk>
> > <resume-from-disk>
> > 
> > hangs at the end of resume-from-disk.
> > 
> > 
> > During the suspend-to-disk I saw:
> > 
> > ipw2200: Radio Frequency Kill Switch is On:
> > Kill switch must be turned off for wireless networking to work.
> > PM: Device usb1 failed to (unknown PM event): error -113
> > PM: Device usb2 failed to (unknown PM event): error -113
> > PM: Device usb3 failed to (unknown PM event): error -113
> > PM: Device usb4 failed to (unknown PM event): error -113
> 
> These -113 errors can be ignored.  They are fallout from the recent 
> conversion of the driver core's suspend/resume API.  Matching changes 
> have not yet been added to the USB core.

Ick, that's not good, is anyone going to at least send me a patch to
make them go away?  :)

thanks,

greg k-h

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Alan Stern]

On Fri, 18 Apr 2008, Greg KH wrote:

> > These -113 errors can be ignored.  They are fallout from the recent 
> > conversion of the driver core's suspend/resume API.  Matching changes 
> > have not yet been added to the USB core.
> 
> Ick, that's not good, is anyone going to at least send me a patch to
> make them go away?  :)

It's included in a 9-part series I'm debugging right now.  If you want, 
I can send just that portion separately this afternoon.

Alan Stern

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Greg KH]

On Fri, Apr 18, 2008 at 12:45:21PM -0400, Alan Stern wrote:
> On Fri, 18 Apr 2008, Greg KH wrote:
> 
> > > These -113 errors can be ignored.  They are fallout from the recent 
> > > conversion of the driver core's suspend/resume API.  Matching changes 
> > > have not yet been added to the USB core.
> > 
> > Ick, that's not good, is anyone going to at least send me a patch to
> > make them go away?  :)
> 
> It's included in a 9-part series I'm debugging right now.  If you want, 
> I can send just that portion separately this afternoon.

No, that's fine, I can wait, just want to make sure that it is taken
care of.

thanks,

greg k-h

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Friday, 18 of April 2008, Alan Stern wrote:
> On Thu, 17 Apr 2008, Andrew Morton wrote:
> 
> > But this:
> > 
> > <cold boot>
> > <suspend-to-ram>
> > <resume-from-ram>
> > <suspend-to-disk>
> > <resume-from-disk>
> > 
> > hangs at the end of resume-from-disk.
> > 
> > 
> > During the suspend-to-disk I saw:
> > 
> > ipw2200: Radio Frequency Kill Switch is On:
> > Kill switch must be turned off for wireless networking to work.
> > PM: Device usb1 failed to (unknown PM event): error -113
> > PM: Device usb2 failed to (unknown PM event): error -113
> > PM: Device usb3 failed to (unknown PM event): error -113
> > PM: Device usb4 failed to (unknown PM event): error -113
> 
> These -113 errors can be ignored.  They are fallout from the recent 
> conversion of the driver core's suspend/resume API.  Matching changes 
> have not yet been added to the USB core.
> 
> The "(unknown PM event)" strings look like a bug, though.

Yes, it does.

I'm trying to figure this out.

Thanks,
Rafael

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> On Friday, 18 of April 2008, Alan Stern wrote:
> > On Thu, 17 Apr 2008, Andrew Morton wrote:
> > 
> > > But this:
> > > 
> > > <cold boot>
> > > <suspend-to-ram>
> > > <resume-from-ram>
> > > <suspend-to-disk>
> > > <resume-from-disk>
> > > 
> > > hangs at the end of resume-from-disk.
> > > 
> > > 
> > > During the suspend-to-disk I saw:
> > > 
> > > ipw2200: Radio Frequency Kill Switch is On:
> > > Kill switch must be turned off for wireless networking to work.
> > > PM: Device usb1 failed to (unknown PM event): error -113
> > > PM: Device usb2 failed to (unknown PM event): error -113
> > > PM: Device usb3 failed to (unknown PM event): error -113
> > > PM: Device usb4 failed to (unknown PM event): error -113
> > 
> > These -113 errors can be ignored.  They are fallout from the recent 
> > conversion of the driver core's suspend/resume API.  Matching changes 
> > have not yet been added to the USB core.
> > 
> > The "(unknown PM event)" strings look like a bug, though.
> 
> Yes, it does.
> 
> I'm trying to figure this out.

This particular problem should be fixed by the appended patch.

Andrew, can you please check if you see "recover" instead of
"(unknown PM event)" in those messages with the patch applied?

Thanks,
Rafael

---
From: Rafael J. Wysocki <rjw@sisk.pl>

If PMSG_RECOVER is passed to it, the pm_verb() function will return
a pointer to "(unknown PM event)", which is wrong.  Fix that.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 drivers/base/power/main.c |    2 ++
 1 file changed, 2 insertions(+)

Index: linux-2.6/drivers/base/power/main.c
===================================================================
--- linux-2.6.orig/drivers/base/power/main.c
+++ linux-2.6/drivers/base/power/main.c
@@ -259,6 +259,8 @@ static char *pm_verb(int event)
 		return "thaw";
 	case PM_EVENT_RESTORE:
 		return "restore";
+	case PM_EVENT_RECOVER:
+		return "recover";
 	default:
 		return "(unknown PM event)";
 	}

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> > On Friday, 18 of April 2008, Alan Stern wrote:
> > > On Thu, 17 Apr 2008, Andrew Morton wrote:
> > > 
> > > > But this:
> > > > 
> > > > <cold boot>
> > > > <suspend-to-ram>
> > > > <resume-from-ram>
> > > > <suspend-to-disk>
> > > > <resume-from-disk>
> > > > 
> > > > hangs at the end of resume-from-disk.
> > > > 
> > > > 
> > > > During the suspend-to-disk I saw:
> > > > 
> > > > ipw2200: Radio Frequency Kill Switch is On:
> > > > Kill switch must be turned off for wireless networking to work.
> > > > PM: Device usb1 failed to (unknown PM event): error -113
> > > > PM: Device usb2 failed to (unknown PM event): error -113
> > > > PM: Device usb3 failed to (unknown PM event): error -113
> > > > PM: Device usb4 failed to (unknown PM event): error -113
> > > 
> > > These -113 errors can be ignored.  They are fallout from the recent 
> > > conversion of the driver core's suspend/resume API.  Matching changes 
> > > have not yet been added to the USB core.
> > > 
> > > The "(unknown PM event)" strings look like a bug, though.
> > 
> > Yes, it does.
> > 
> > I'm trying to figure this out.
> 
> This particular problem should be fixed by the appended patch.

The following patch is also necessary to make hibernation work as documented
in pm.h.

Thanks,
Rafael

---
From: Rafael J. Wysocki <rjw@sisk.pl>

Make hibernation work as documented in include/linux/pm.h .

In pm.h it is documented that device_power_up() and device_resume()
will be called with PMSG_RECOVER if the creation of a hibernation
image fails or the restoration of memory contents from a
hibernation image fails.  Otherwise, it should be called either
with PMSG_THAW or with PMSG_RESTORE, depending on the context.
Make that happen.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
---
 kernel/power/disk.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

Index: linux-2.6/kernel/power/disk.c
===================================================================
--- linux-2.6.orig/kernel/power/disk.c
+++ linux-2.6/kernel/power/disk.c
@@ -225,7 +225,8 @@ static int create_image(int platform_mod
 	/* NOTE:  device_power_up() is just a resume() for devices
 	 * that suspended with irqs off ... no overall powerup.
 	 */
-	device_power_up(in_suspend ? PMSG_RECOVER : PMSG_RESTORE);
+	device_power_up(in_suspend ?
+		(error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE);
  Enable_irqs:
 	local_irq_enable();
 	device_pm_unlock();
@@ -282,7 +283,8 @@ int hibernation_snapshot(int platform_mo
  Finish:
 	platform_finish(platform_mode);
  Resume_devices:
-	device_resume(in_suspend ? PMSG_RECOVER : PMSG_RESTORE);
+	device_resume(in_suspend ?
+		(error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE);
  Resume_console:
 	resume_console();
  Close:
@@ -332,7 +334,7 @@ static int resume_target_kernel(void)
 	swsusp_free();
 	restore_processor_state();
 	touch_softlockup_watchdog();
-	device_power_up(PMSG_THAW);
+	device_power_up(PMSG_RECOVER);
  Enable_irqs:
 	local_irq_enable();
 	device_pm_unlock();

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Greg KH]

On Fri, Apr 18, 2008 at 10:38:44PM +0200, Rafael J. Wysocki wrote:
> On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> > On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> > > On Friday, 18 of April 2008, Alan Stern wrote:
> > > > On Thu, 17 Apr 2008, Andrew Morton wrote:
> > > > 
> > > > > But this:
> > > > > 
> > > > > <cold boot>
> > > > > <suspend-to-ram>
> > > > > <resume-from-ram>
> > > > > <suspend-to-disk>
> > > > > <resume-from-disk>
> > > > > 
> > > > > hangs at the end of resume-from-disk.
> > > > > 
> > > > > 
> > > > > During the suspend-to-disk I saw:
> > > > > 
> > > > > ipw2200: Radio Frequency Kill Switch is On:
> > > > > Kill switch must be turned off for wireless networking to work.
> > > > > PM: Device usb1 failed to (unknown PM event): error -113
> > > > > PM: Device usb2 failed to (unknown PM event): error -113
> > > > > PM: Device usb3 failed to (unknown PM event): error -113
> > > > > PM: Device usb4 failed to (unknown PM event): error -113
> > > > 
> > > > These -113 errors can be ignored.  They are fallout from the recent 
> > > > conversion of the driver core's suspend/resume API.  Matching changes 
> > > > have not yet been added to the USB core.
> > > > 
> > > > The "(unknown PM event)" strings look like a bug, though.
> > > 
> > > Yes, it does.
> > > 
> > > I'm trying to figure this out.
> > 
> > This particular problem should be fixed by the appended patch.
> 
> The following patch is also necessary to make hibernation work as documented
> in pm.h.
> 
> Thanks,
> Rafael
> 
> ---
> From: Rafael J. Wysocki <rjw@sisk.pl>
> 
> Make hibernation work as documented in include/linux/pm.h .
> 
> In pm.h it is documented that device_power_up() and device_resume()
> will be called with PMSG_RECOVER if the creation of a hibernation
> image fails or the restoration of memory contents from a
> hibernation image fails.  Otherwise, it should be called either
> with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> Make that happen.

Should either of these patches be merged / added to the ones you sent me
before?

thanks,

greg k-h

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Friday, 18 of April 2008, Greg KH wrote:
> On Fri, Apr 18, 2008 at 10:38:44PM +0200, Rafael J. Wysocki wrote:
> > On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> > > On Friday, 18 of April 2008, Rafael J. Wysocki wrote:
> > > > On Friday, 18 of April 2008, Alan Stern wrote:
> > > > > On Thu, 17 Apr 2008, Andrew Morton wrote:
> > > > > 
> > > > > > But this:
> > > > > > 
> > > > > > <cold boot>
> > > > > > <suspend-to-ram>
> > > > > > <resume-from-ram>
> > > > > > <suspend-to-disk>
> > > > > > <resume-from-disk>
> > > > > > 
> > > > > > hangs at the end of resume-from-disk.
> > > > > > 
> > > > > > 
> > > > > > During the suspend-to-disk I saw:
> > > > > > 
> > > > > > ipw2200: Radio Frequency Kill Switch is On:
> > > > > > Kill switch must be turned off for wireless networking to work.
> > > > > > PM: Device usb1 failed to (unknown PM event): error -113
> > > > > > PM: Device usb2 failed to (unknown PM event): error -113
> > > > > > PM: Device usb3 failed to (unknown PM event): error -113
> > > > > > PM: Device usb4 failed to (unknown PM event): error -113
> > > > > 
> > > > > These -113 errors can be ignored.  They are fallout from the recent 
> > > > > conversion of the driver core's suspend/resume API.  Matching changes 
> > > > > have not yet been added to the USB core.
> > > > > 
> > > > > The "(unknown PM event)" strings look like a bug, though.
> > > > 
> > > > Yes, it does.
> > > > 
> > > > I'm trying to figure this out.
> > > 
> > > This particular problem should be fixed by the appended patch.
> > 
> > The following patch is also necessary to make hibernation work as documented
> > in pm.h.
> > 
> > Thanks,
> > Rafael
> > 
> > ---
> > From: Rafael J. Wysocki <rjw@sisk.pl>
> > 
> > Make hibernation work as documented in include/linux/pm.h .
> > 
> > In pm.h it is documented that device_power_up() and device_resume()
> > will be called with PMSG_RECOVER if the creation of a hibernation
> > image fails or the restoration of memory contents from a
> > hibernation image fails.  Otherwise, it should be called either
> > with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> > Make that happen.
> 
> Should either of these patches be merged / added to the ones you sent me
> before?

Yes, both of them, please.

Thanks,
Rafael

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Greg KH]

On Fri, Apr 18, 2008 at 11:30:39PM +0200, Rafael J. Wysocki wrote:
> On Friday, 18 of April 2008, Greg KH wrote:
> > > ---
> > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > > 
> > > Make hibernation work as documented in include/linux/pm.h .
> > > 
> > > In pm.h it is documented that device_power_up() and device_resume()
> > > will be called with PMSG_RECOVER if the creation of a hibernation
> > > image fails or the restoration of memory contents from a
> > > hibernation image fails.  Otherwise, it should be called either
> > > with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> > > Make that happen.
> > 
> > Should either of these patches be merged / added to the ones you sent me
> > before?
> 
> Yes, both of them, please.

Ok, which is it?  Added at the end of the series, or merged in to
prevent git bisection problems?  If merged, which patch should I merge
it to?

thanks,

greg k-h

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Saturday, 19 of April 2008, Greg KH wrote:
> On Fri, Apr 18, 2008 at 11:30:39PM +0200, Rafael J. Wysocki wrote:
> > On Friday, 18 of April 2008, Greg KH wrote:
> > > > ---
> > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > > > 
> > > > Make hibernation work as documented in include/linux/pm.h .
> > > > 
> > > > In pm.h it is documented that device_power_up() and device_resume()
> > > > will be called with PMSG_RECOVER if the creation of a hibernation
> > > > image fails or the restoration of memory contents from a
> > > > hibernation image fails.  Otherwise, it should be called either
> > > > with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> > > > Make that happen.
> > > 
> > > Should either of these patches be merged / added to the ones you sent me
> > > before?
> > 
> > Yes, both of them, please.
> 
> Ok, which is it?  Added at the end of the series, or merged in to
> prevent git bisection problems?  If merged, which patch should I merge
> it to?

Please merge both into
pm-introduce-new-top-level-suspend-and-hibernation-callbacks.patch ,
if possible, the changes belong in there.

Thanks,
Rafael

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Greg KH]

On Sat, Apr 19, 2008 at 01:16:16AM +0200, Rafael J. Wysocki wrote:
> On Saturday, 19 of April 2008, Greg KH wrote:
> > On Fri, Apr 18, 2008 at 11:30:39PM +0200, Rafael J. Wysocki wrote:
> > > On Friday, 18 of April 2008, Greg KH wrote:
> > > > > ---
> > > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > > > > 
> > > > > Make hibernation work as documented in include/linux/pm.h .
> > > > > 
> > > > > In pm.h it is documented that device_power_up() and device_resume()
> > > > > will be called with PMSG_RECOVER if the creation of a hibernation
> > > > > image fails or the restoration of memory contents from a
> > > > > hibernation image fails.  Otherwise, it should be called either
> > > > > with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> > > > > Make that happen.
> > > > 
> > > > Should either of these patches be merged / added to the ones you sent me
> > > > before?
> > > 
> > > Yes, both of them, please.
> > 
> > Ok, which is it?  Added at the end of the series, or merged in to
> > prevent git bisection problems?  If merged, which patch should I merge
> > it to?
> 
> Please merge both into
> pm-introduce-new-top-level-suspend-and-hibernation-callbacks.patch ,
> if possible, the changes belong in there.

Hm, I just tried this (sorry for the delay) and I think it's no longer
needed with your latest version of the patches, right?

thanks,

greg k-h

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Rafael J. Wysocki]

On Tuesday, 29 of April 2008, Greg KH wrote:
> On Sat, Apr 19, 2008 at 01:16:16AM +0200, Rafael J. Wysocki wrote:
> > On Saturday, 19 of April 2008, Greg KH wrote:
> > > On Fri, Apr 18, 2008 at 11:30:39PM +0200, Rafael J. Wysocki wrote:
> > > > On Friday, 18 of April 2008, Greg KH wrote:
> > > > > > ---
> > > > > > From: Rafael J. Wysocki <rjw@sisk.pl>
> > > > > > 
> > > > > > Make hibernation work as documented in include/linux/pm.h .
> > > > > > 
> > > > > > In pm.h it is documented that device_power_up() and device_resume()
> > > > > > will be called with PMSG_RECOVER if the creation of a hibernation
> > > > > > image fails or the restoration of memory contents from a
> > > > > > hibernation image fails.  Otherwise, it should be called either
> > > > > > with PMSG_THAW or with PMSG_RESTORE, depending on the context.
> > > > > > Make that happen.
> > > > > 
> > > > > Should either of these patches be merged / added to the ones you sent me
> > > > > before?
> > > > 
> > > > Yes, both of them, please.
> > > 
> > > Ok, which is it?  Added at the end of the series, or merged in to
> > > prevent git bisection problems?  If merged, which patch should I merge
> > > it to?
> > 
> > Please merge both into
> > pm-introduce-new-top-level-suspend-and-hibernation-callbacks.patch ,
> > if possible, the changes belong in there.
> 
> Hm, I just tried this (sorry for the delay) and I think it's no longer
> needed with your latest version of the patches, right?

Right.  The last version already contains it.

Thanks,
Rafael

Re: 2.6.25-mm1: suspend-to-disk warning, resume failure

[Pavel Machek]

> > > > > ipw2200: Radio Frequency Kill Switch is On:
> > > > > Kill switch must be turned off for wireless networking to work.
> > > > > PM: Device usb1 failed to (unknown PM event): error -113
> > > > > PM: Device usb2 failed to (unknown PM event): error -113
> > > > > PM: Device usb3 failed to (unknown PM event): error -113
> > > > > PM: Device usb4 failed to (unknown PM event): error -113
> > > > 
> > > > These -113 errors can be ignored.  They are fallout from the recent 
> > > > conversion of the driver core's suspend/resume API.  Matching changes 
> > > > have not yet been added to the USB core.
> > > > 
> > > > The "(unknown PM event)" strings look like a bug, though.
> > > 
> > > Yes, it does.
> > > 
> > > I'm trying to figure this out.
> > 
> > This particular problem should be fixed by the appended patch.
> 
> The following patch is also necessary to make hibernation work as documented
> in pm.h.

ACK.

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

2.6.25-mm1: suspend-to-disk warning, resume failure

[Andrew Morton]

This is the Vaio.

This:

<cold boot>
<suspend-to-ram>
<resume-from-ram>
<suspend-to-ram>
<resume-from-ram>

works OK.


This:

<cold boot>
<suspend-to-disk>
<resume-from-disk>

works OK.


But this:

<cold boot>
<suspend-to-ram>
<resume-from-ram>
<suspend-to-disk>
<resume-from-disk>

hangs at the end of resume-from-disk.


During the suspend-to-disk I saw:

ipw2200: Radio Frequency Kill Switch is On:
Kill switch must be turned off for wireless networking to work.
PM: Device usb1 failed to (unknown PM event): error -113
PM: Device usb2 failed to (unknown PM event): error -113
PM: Device usb3 failed to (unknown PM event): error -113
PM: Device usb4 failed to (unknown PM event): error -113
Suspending console(s)

But it suspends OK.

Resume _almost_ works.  It gets to:

calling  hpet_insert_resource+0x0/0x1e()
initcall hpet_insert_resource+0x0/0x1e() returned 1 after 0 msecs
initcall hpet_insert_resource+0x0/0x1e() returned with error code 1
calling  init_lapic_nmi_sysfs+0x0/0x33()
initcall init_lapic_nmi_sysfs+0x0/0x33() returned 0 after 0 msecs
calling  io_apic_bug_finalize+0x0/0x1c()
initcall io_apic_bug_finalize+0x0/0x1c() returned 0 after 0 msecs
calling  check_early_ioremap_leak+0x0/0x3d()
initcall check_early_ioremap_leak+0x0/0x3d() returned 0 after 0 msecs
calling  print_ipi_mode+0x0/0x26()
Using IPI Shortcut mode
initcall print_ipi_mode+0x0/0x26() returned 0 after 2 msecs
calling  init_oops_id+0x0/0x23()
initcall init_oops_id+0x0/0x23() returned 0 after 0 msecs
calling  disable_boot_consoles+0x0/0x36()
initcall disable_boot_consoles+0x0/0x36() returned 0 after 0 msecs
calling  pm_qos_power_init+0x0/0x54()
initcall pm_qos_power_init+0x0/0x54() returned 0 after 0 msecs
calling  software_resume+0x0/0x179()
Freezing user space processes ... (elapsed 0.00 seconds) done.
Freezing remaining freezable tasks ... <6>input: PS/2 Mouse as /class/input/input3
input: AlpsPS/2 ALPS GlidePoint as /class/input/input4
(elapsed 0.14 seconds) done.
PM: Loading image data pages (88965 pages) ... done
PM: Read 355860 kbytes in 16.36 seconds (21.75 MB/s)
Suspending console(s)

then stops with fans whirring.

http://userweb.kernel.org/~akpm/config-sony.txt
http://userweb.kernel.org/~akpm/dmesg-sony.txt


I guess I get to bisect this.