Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Rafael J. Wysocki]

On Tuesday, January 04, 2011, Jiri Slaby wrote:
> On 01/04/2011 02:40 PM, Jiri Slaby wrote:
> > On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:
> >> The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to
> > 
> > Hi, this kernel regresses with respect to suspend to ram in comparison
> > with mmotm 2010-12-16-14-56.
> > 
> > This is OK:
> > echo devices > /sys/power/pm_test
> > pm-suspend
> > This hangs at suspend phase:
> > echo platform > /sys/power/pm_test
> > pm-suspend

Hmm.  So it looks like _PTS hangs?

> > Note that this kernel is based on next-20101221. Should I try newer (and
> > clean) -next?
> 
> Ok, bisected down to:
> 16dc39c98a6ca56a27f22f7ac6731d8223237a2e is first bad commit
> commit 16dc39c98a6ca56a27f22f7ac6731d8223237a2e
> Author: Len Brown <len.brown@intel.com>
> Date:   Thu Dec 16 23:12:23 2010 -0500
> 
>     ACPI: use ioremap_cache()
> 
>     Although the temporary boot-time ACPI table mappings
>     were set up with CPU caching enabled, the permanent table
>     mappings and AML run-time region memory accesses were
>     set up with ioremap(), which on x86 is a synonym for
>     ioremap_nocache().
> 
>     Changing this to ioremap_cache() improves performance as
>     seen when accessing the tables via acpidump,
>     or /sys/firmware/acpi/tables.  It should also improve
>     AML run-time performance.
> 
>     No change on ia64.
> 
>     Reported-by: Jack Steiner <steiner@sgi.com>
>     Signed-off-by: Len Brown <len.brown@intel.com>
> 
> :040000 040000 be35c5e8f214f10f94688c1a27f33ecfb8505220
> 52581222d0edf190f160f3e5aa5d2c1af8e76988 M      arch
> :040000 040000 ccdca0d41938b8312e946cde3c01c59b32d1c17c
> 96ccf2357f2ac4a31d19cc41f5728d9f87b6cac0 M      drivers
> 
> Revert of that patch fixes the problem.

Can you send a dmesg output and acpidump output from the failing system, please?

Rafael

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Jiri Slaby]

On 01/04/2011 11:57 PM, Rafael J. Wysocki wrote:
> On Tuesday, January 04, 2011, Jiri Slaby wrote:
>> On 01/04/2011 02:40 PM, Jiri Slaby wrote:
>>> On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:
>>>> The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to
>>>
>>> Hi, this kernel regresses with respect to suspend to ram in comparison
>>> with mmotm 2010-12-16-14-56.
>>>
>>> This is OK:
>>> echo devices > /sys/power/pm_test
>>> pm-suspend
>>> This hangs at suspend phase:
>>> echo platform > /sys/power/pm_test
>>> pm-suspend
> 
> Hmm.  So it looks like _PTS hangs?

No _PST here :).

>>> Note that this kernel is based on next-20101221. Should I try newer (and
>>> clean) -next?
>>
>> Ok, bisected down to:
>> 16dc39c98a6ca56a27f22f7ac6731d8223237a2e is first bad commit
>> commit 16dc39c98a6ca56a27f22f7ac6731d8223237a2e
...
>> Revert of that patch fixes the problem.
> 
> Can you send a dmesg output and acpidump output from the failing system, please?

Here they are:
http://www.fi.muni.cz/~xslaby/sklad/panics/dmesg-no-susp.txt
http://www.fi.muni.cz/~xslaby/sklad/adump.txt

They are captured from the mmotm minus 16dc39c98a6ca56 if that's OK?
Ignore the BUG there, it's caused by qemu and fixed in later -next.

regards,
-- 
js

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Rafael J. Wysocki]

On Wednesday, January 05, 2011, Jiri Slaby wrote:
> On 01/04/2011 11:57 PM, Rafael J. Wysocki wrote:
> > On Tuesday, January 04, 2011, Jiri Slaby wrote:
> >> On 01/04/2011 02:40 PM, Jiri Slaby wrote:
> >>> On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:
> >>>> The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to
> >>>
> >>> Hi, this kernel regresses with respect to suspend to ram in comparison
> >>> with mmotm 2010-12-16-14-56.
> >>>
> >>> This is OK:
> >>> echo devices > /sys/power/pm_test
> >>> pm-suspend
> >>> This hangs at suspend phase:
> >>> echo platform > /sys/power/pm_test
> >>> pm-suspend
> > 
> > Hmm.  So it looks like _PTS hangs?
> 
> No _PST here :).

I really meant _PTS and yes, there is one and it does quite a lot of
interesting stuff.

Are you able to collect serial console (or equivalent) logs from that machine?

Rafael

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Jiri Slaby]

On 01/05/2011 11:39 PM, Rafael J. Wysocki wrote:
> On Wednesday, January 05, 2011, Jiri Slaby wrote:
>> On 01/04/2011 11:57 PM, Rafael J. Wysocki wrote:
>>> On Tuesday, January 04, 2011, Jiri Slaby wrote:
>>>> On 01/04/2011 02:40 PM, Jiri Slaby wrote:
>>>>> On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:
>>>>>> The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to
>>>>>
>>>>> Hi, this kernel regresses with respect to suspend to ram in comparison
>>>>> with mmotm 2010-12-16-14-56.
>>>>>
>>>>> This is OK:
>>>>> echo devices > /sys/power/pm_test
>>>>> pm-suspend
>>>>> This hangs at suspend phase:
>>>>> echo platform > /sys/power/pm_test
>>>>> pm-suspend
>>>
>>> Hmm.  So it looks like _PTS hangs?
>>
>> No _PST here :).
> 
> I really meant _PTS and yes, there is one and it does quite a lot of
> interesting stuff.

Yeah, I grepped for _PTS, but in a wrong file. My bad.

> Are you able to collect serial console (or equivalent) logs from that machine?

Unfortunately I don't have a second machine with serial port. What
exactly you want to dig out of the logs? Will netconsole work at that phase?

regards,
-- 
js

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Rafael J. Wysocki]

On Wednesday, January 05, 2011, Jiri Slaby wrote:
> On 01/05/2011 11:39 PM, Rafael J. Wysocki wrote:
> > On Wednesday, January 05, 2011, Jiri Slaby wrote:
> >> On 01/04/2011 11:57 PM, Rafael J. Wysocki wrote:
> >>> On Tuesday, January 04, 2011, Jiri Slaby wrote:
> >>>> On 01/04/2011 02:40 PM, Jiri Slaby wrote:
> >>>>> On 12/24/2010 01:58 AM, akpm@linux-foundation.org wrote:
> >>>>>> The mm-of-the-moment snapshot 2010-12-23-16-58 has been uploaded to
> >>>>>
> >>>>> Hi, this kernel regresses with respect to suspend to ram in comparison
> >>>>> with mmotm 2010-12-16-14-56.
> >>>>>
> >>>>> This is OK:
> >>>>> echo devices > /sys/power/pm_test
> >>>>> pm-suspend
> >>>>> This hangs at suspend phase:
> >>>>> echo platform > /sys/power/pm_test
> >>>>> pm-suspend
> >>>
> >>> Hmm.  So it looks like _PTS hangs?
> >>
> >> No _PST here :).
> > 
> > I really meant _PTS and yes, there is one and it does quite a lot of
> > interesting stuff.
> 
> Yeah, I grepped for _PTS, but in a wrong file. My bad.
> 
> > Are you able to collect serial console (or equivalent) logs from that machine?
> 
> Unfortunately I don't have a second machine with serial port. What
> exactly you want to dig out of the logs? Will netconsole work at that phase?

I don't think so, the adapter will have been already suspended.

Hmm.  I guess you can boot with no_console_suspend in the kernel command line
and use the video console to get some debug info.  At this point we would like
to know where exactly it hangs, so you'd need to enable AML tracing in the
kernel and basically see what's printed last. :-)

Two more questions:
(1) Does the failing machine power off correctly with the kernel that fails to
    suspend?
(2) Is the failing suspend behavior reproducible if you run
    "echo disk > /sys/power/state" instead of "pm-suspend" in the platform test?

Rafael

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Jiri Slaby]

On 01/06/2011 12:28 AM, Rafael J. Wysocki wrote:
> Two more questions:
> (1) Does the failing machine power off correctly with the kernel that fails to
>     suspend?

Poweroff is OK.

> (2) Is the failing suspend behavior reproducible if you run
>     "echo disk > /sys/power/state" instead of "pm-suspend" in the platform test?

Thanks for the hint no_console_suspend with hibernate shows somwthing
(opposing to suspend). I've sent a patch for the oops which occured due
to ioremap change. The ioremap change still needs to be reverted, obviously.

The oops in question:
http://www.fi.muni.cz/~xslaby/sklad/panics/suspend-oops.png

/proc/mtrr:
reg00: base=0x0b0000000 ( 2816MB), size=  256MB, count=1: uncachable
reg01: base=0x0c0000000 ( 3072MB), size= 1024MB, count=1: uncachable
reg02: base=0x000000000 (    0MB), size= 4096MB, count=1: write-back
reg03: base=0x100000000 ( 4096MB), size= 2048MB, count=1: write-back
reg04: base=0x180000000 ( 6144MB), size= 1024MB, count=1: write-back
reg05: base=0x1c0000000 ( 7168MB), size=  128MB, count=1: write-back
reg06: base=0x1c8000000 ( 7296MB), size=   64MB, count=1: write-back
reg07: base=0x0af600000 ( 2806MB), size=    2MB, count=1: uncachable

/debug/x86/pat_memtype_list:
PAT memtype list:
uncached-minus @ 0xaf5b0000-0xaf5b7000
uncached-minus @ 0xaf5be000-0xaf5c1000
uncached-minus @ 0xaf5be000-0xaf5c1000
uncached-minus @ 0xaf5be000-0xaf5bf000
uncached-minus @ 0xb0200000-0xb0201000
write-combining @ 0xd0000000-0xe0000000
write-combining @ 0xd0001000-0xd0021000
write-combining @ 0xd0030000-0xd0530000
uncached-minus @ 0xe0000000-0xf0000000
uncached-minus @ 0xfe6ff000-0xfe700000
uncached-minus @ 0xfe7e0000-0xfe800000
uncached-minus @ 0xfea00000-0xfea80000
uncached-minus @ 0xfeb40000-0xfeb60000
uncached-minus @ 0xfeb78000-0xfeb7c000
uncached-minus @ 0xfeb7c000-0xfeb7d000
uncached-minus @ 0xfeb7d000-0xfeb7e000
uncached-minus @ 0xfeb7f000-0xfeb80000
uncached-minus @ 0xfeb7f000-0xfeb80000
uncached-minus @ 0xfeb80000-0xfec00000
uncached-minus @ 0xfeb80000-0xfec00000
uncached-minus @ 0xfed00000-0xfed01000
uncached-minus @ 0xfed1f000-0xfed20000
uncached-minus @ 0xffff0000-0xffff1000

regards,
-- 
js

Re: suspend hangs at platform phase [was: mmotm 2010-12-23-16-58 uploaded]

[Rafael J. Wysocki]

On Thursday, January 06, 2011, Jiri Slaby wrote:
> On 01/06/2011 12:28 AM, Rafael J. Wysocki wrote:
> > Two more questions:
> > (1) Does the failing machine power off correctly with the kernel that fails to
> >     suspend?
> 
> Poweroff is OK.
> 
> > (2) Is the failing suspend behavior reproducible if you run
> >     "echo disk > /sys/power/state" instead of "pm-suspend" in the platform test?
> 
> Thanks for the hint no_console_suspend with hibernate shows somwthing
> (opposing to suspend). I've sent a patch for the oops which occured due
> to ioremap change.

Can you send a link to the patch, please?

> The ioremap change still needs to be reverted, obviously.

Does it need to be reverted with your patch applied too?

Rafael

Re: [PATCH 1/1] PM: fix oops in suspend/hibernate code

[Rafael J. Wysocki]

On Thursday, January 06, 2011, Jiri Slaby wrote:
> When ioremap fails (which might happen for some reason),

If it happens, something is seriously wrong (see below).

BTW, to keep things in context, please post fixes like this in the same thread
in which you reported the problem.  At lease please retain the CC list from
there.

> we nicely oops in suspend_nvs_save due to NULL dereference by memcpy in
> there.  Fail gracefully instead.
> 
> Signed-off-by: Jiri Slaby <jslaby@suse.cz>
> Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
> ---
>  drivers/acpi/sleep.c    |    5 ++---
>  include/linux/suspend.h |    4 ++--
>  kernel/power/nvs.c      |    8 +++++++-
>  3 files changed, 11 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
> index c423231..f94c9a9 100644
> --- a/drivers/acpi/sleep.c
> +++ b/drivers/acpi/sleep.c
> @@ -124,8 +124,7 @@ static int acpi_pm_freeze(void)
>  static int acpi_pm_pre_suspend(void)
>  {
>  	acpi_pm_freeze();
> -	suspend_nvs_save();
> -	return 0;
> +	return suspend_nvs_save();
>  }
>  
>  /**
> @@ -151,7 +150,7 @@ static int acpi_pm_prepare(void)
>  {
>  	int error = __acpi_pm_prepare();
>  	if (!error)
> -		acpi_pm_pre_suspend();
> +		error = acpi_pm_pre_suspend();
>  
>  	return error;
>  }
> diff --git a/include/linux/suspend.h b/include/linux/suspend.h
> index c1f4998..3ac2551 100644
> --- a/include/linux/suspend.h
> +++ b/include/linux/suspend.h
> @@ -262,7 +262,7 @@ static inline bool system_entering_hibernation(void) { return false; }
>  extern int suspend_nvs_register(unsigned long start, unsigned long size);
>  extern int suspend_nvs_alloc(void);
>  extern void suspend_nvs_free(void);
> -extern void suspend_nvs_save(void);
> +extern int suspend_nvs_save(void);
>  extern void suspend_nvs_restore(void);
>  #else /* CONFIG_SUSPEND_NVS */
>  static inline int suspend_nvs_register(unsigned long a, unsigned long b)
> @@ -271,7 +271,7 @@ static inline int suspend_nvs_register(unsigned long a, unsigned long b)
>  }
>  static inline int suspend_nvs_alloc(void) { return 0; }
>  static inline void suspend_nvs_free(void) {}
> -static inline void suspend_nvs_save(void) {}
> +static inline int suspend_nvs_save(void) {}
>  static inline void suspend_nvs_restore(void) {}
>  #endif /* CONFIG_SUSPEND_NVS */
>  
> diff --git a/kernel/power/nvs.c b/kernel/power/nvs.c
> index 1836db6..57c6fab 100644
> --- a/kernel/power/nvs.c
> +++ b/kernel/power/nvs.c
> @@ -105,7 +105,7 @@ int suspend_nvs_alloc(void)
>  /**
>   *	suspend_nvs_save - save NVS memory regions
>   */
> -void suspend_nvs_save(void)
> +int suspend_nvs_save(void)
>  {
>  	struct nvs_page *entry;
>  
> @@ -114,8 +114,14 @@ void suspend_nvs_save(void)
>  	list_for_each_entry(entry, &nvs_list, node)
>  		if (entry->data) {
>  			entry->kaddr = ioremap(entry->phys_start, entry->size);

I wonder what happens if you simply change the ioremap() here to
ioremap_nocache() without any other modifications?

It _really_ shouldn't fail here, because the NVS pages are known to be present.

> +			if (!entry->kaddr) {
> +				suspend_nvs_free();
> +				return -ENOMEM;
> +			}
>  			memcpy(entry->data, entry->kaddr, entry->size);
>  		}
> +
> +	return 0;
>  }
>  
>  /**
> 

Rafael

Re: [PATCH 1/1] PM: fix oops in suspend/hibernate code

[Jiri Slaby]

On 01/06/2011 04:57 PM, Rafael J. Wysocki wrote:
> On Thursday, January 06, 2011, Jiri Slaby wrote:
>> When ioremap fails (which might happen for some reason),
> 
> If it happens, something is seriously wrong (see below).

I agree that something is broken, however ioremap may fail for dozen of
reasons. Ignoring the retval is a *bad* idea and it took me a while to
sort out what is wrong. Especially if one has no console like throughout
suspend. If it was handled properly, I would know immediately. (There
should be a message printed out which I forgot to add.)

> BTW, to keep things in context, please post fixes like this in the same thread
> in which you reported the problem.  At lease please retain the CC list from
> there.

I actually did, there is:
In-Reply-To: <201101060028.43342.rjw@sisk.pl>
and it successfully threaded to the conversation for me in TB.

>> we nicely oops in suspend_nvs_save due to NULL dereference by memcpy in
>> there.  Fail gracefully instead.
>>
>> Signed-off-by: Jiri Slaby <jslaby@suse.cz>
>> Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
>> ---
>>  drivers/acpi/sleep.c    |    5 ++---
>>  include/linux/suspend.h |    4 ++--
>>  kernel/power/nvs.c      |    8 +++++++-
>>  3 files changed, 11 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
>> index c423231..f94c9a9 100644
>> --- a/drivers/acpi/sleep.c
>> +++ b/drivers/acpi/sleep.c
>> @@ -124,8 +124,7 @@ static int acpi_pm_freeze(void)
>>  static int acpi_pm_pre_suspend(void)
>>  {
>>  	acpi_pm_freeze();
>> -	suspend_nvs_save();
>> -	return 0;
>> +	return suspend_nvs_save();
>>  }
>>  
>>  /**
>> @@ -151,7 +150,7 @@ static int acpi_pm_prepare(void)
>>  {
>>  	int error = __acpi_pm_prepare();
>>  	if (!error)
>> -		acpi_pm_pre_suspend();
>> +		error = acpi_pm_pre_suspend();
>>  
>>  	return error;
>>  }
>> diff --git a/include/linux/suspend.h b/include/linux/suspend.h
>> index c1f4998..3ac2551 100644
>> --- a/include/linux/suspend.h
>> +++ b/include/linux/suspend.h
>> @@ -262,7 +262,7 @@ static inline bool system_entering_hibernation(void) { return false; }
>>  extern int suspend_nvs_register(unsigned long start, unsigned long size);
>>  extern int suspend_nvs_alloc(void);
>>  extern void suspend_nvs_free(void);
>> -extern void suspend_nvs_save(void);
>> +extern int suspend_nvs_save(void);
>>  extern void suspend_nvs_restore(void);
>>  #else /* CONFIG_SUSPEND_NVS */
>>  static inline int suspend_nvs_register(unsigned long a, unsigned long b)
>> @@ -271,7 +271,7 @@ static inline int suspend_nvs_register(unsigned long a, unsigned long b)
>>  }
>>  static inline int suspend_nvs_alloc(void) { return 0; }
>>  static inline void suspend_nvs_free(void) {}
>> -static inline void suspend_nvs_save(void) {}
>> +static inline int suspend_nvs_save(void) {}
>>  static inline void suspend_nvs_restore(void) {}
>>  #endif /* CONFIG_SUSPEND_NVS */
>>  
>> diff --git a/kernel/power/nvs.c b/kernel/power/nvs.c
>> index 1836db6..57c6fab 100644
>> --- a/kernel/power/nvs.c
>> +++ b/kernel/power/nvs.c
>> @@ -105,7 +105,7 @@ int suspend_nvs_alloc(void)
>>  /**
>>   *	suspend_nvs_save - save NVS memory regions
>>   */
>> -void suspend_nvs_save(void)
>> +int suspend_nvs_save(void)
>>  {
>>  	struct nvs_page *entry;
>>  
>> @@ -114,8 +114,14 @@ void suspend_nvs_save(void)
>>  	list_for_each_entry(entry, &nvs_list, node)
>>  		if (entry->data) {
>>  			entry->kaddr = ioremap(entry->phys_start, entry->size);
> 
> I wonder what happens if you simply change the ioremap() here to
> ioremap_nocache() without any other modifications?

ioremap *is* ioremap_nocache on x86. And that's the conflict it
complains about I guess? Don't you mean ioremap_cache?

> It _really_ shouldn't fail here, because the NVS pages are known to be present.

It fails because of conflicting maps as can be seen in the photo. At
least I think so.

regards,
-- 
js

Re: [PATCH 1/1] PM: fix oops in suspend/hibernate code

[Jiri Slaby]

On 01/06/2011 05:09 PM, Jiri Slaby wrote:
>>> --- a/kernel/power/nvs.c
>>> +++ b/kernel/power/nvs.c
>>> @@ -105,7 +105,7 @@ int suspend_nvs_alloc(void)
>>>  /**
>>>   *	suspend_nvs_save - save NVS memory regions
>>>   */
>>> -void suspend_nvs_save(void)
>>> +int suspend_nvs_save(void)
>>>  {
>>>  	struct nvs_page *entry;
>>>  
>>> @@ -114,8 +114,14 @@ void suspend_nvs_save(void)
>>>  	list_for_each_entry(entry, &nvs_list, node)
>>>  		if (entry->data) {
>>>  			entry->kaddr = ioremap(entry->phys_start, entry->size);
>>
>> I wonder what happens if you simply change the ioremap() here to
>> ioremap_nocache() without any other modifications?
> 
> ioremap *is* ioremap_nocache on x86. And that's the conflict it
> complains about I guess? Don't you mean ioremap_cache?

Using ioremap_cache indeed fixes the problem...

thanks,
-- 
js

Re: [PATCH 1/1] PM: fix oops in suspend/hibernate code

[Rafael J. Wysocki]

On Thursday, January 06, 2011, Jiri Slaby wrote:
> On 01/06/2011 04:57 PM, Rafael J. Wysocki wrote:
> > On Thursday, January 06, 2011, Jiri Slaby wrote:
> >> When ioremap fails (which might happen for some reason),
> > 
> > If it happens, something is seriously wrong (see below).
> 
> I agree that something is broken, however ioremap may fail for dozen of
> reasons. Ignoring the retval is a *bad* idea and it took me a while to
> sort out what is wrong. Especially if one has no console like throughout
> suspend. If it was handled properly, I would know immediately. (There
> should be a message printed out which I forgot to add.)

It wasn't handled, because it _never_ failed previously.  The ACPI mapping
change apparently revealed a deeper problem.

I'm not saying the patch isn't useful, though, and I'm going to take it
for 2.6.38 (perhaps with minor modifications).

> > BTW, to keep things in context, please post fixes like this in the same thread
> > in which you reported the problem.  At lease please retain the CC list from
> > there.
> 
> I actually did, there is:
> In-Reply-To: <201101060028.43342.rjw@sisk.pl>
> and it successfully threaded to the conversation for me in TB.

But you trimmed the CC line, didn't you?  Which caused my filter to put the
patch into a different folder. :-)

> >> we nicely oops in suspend_nvs_save due to NULL dereference by memcpy in
> >> there.  Fail gracefully instead.
> >>
> >> Signed-off-by: Jiri Slaby <jslaby@suse.cz>
> >> Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
> >> ---
> >>  drivers/acpi/sleep.c    |    5 ++---
> >>  include/linux/suspend.h |    4 ++--
> >>  kernel/power/nvs.c      |    8 +++++++-
> >>  3 files changed, 11 insertions(+), 6 deletions(-)
> >>
> >> diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
> >> index c423231..f94c9a9 100644
> >> --- a/drivers/acpi/sleep.c
> >> +++ b/drivers/acpi/sleep.c
> >> @@ -124,8 +124,7 @@ static int acpi_pm_freeze(void)
> >>  static int acpi_pm_pre_suspend(void)
> >>  {
> >>  	acpi_pm_freeze();
> >> -	suspend_nvs_save();
> >> -	return 0;
> >> +	return suspend_nvs_save();
> >>  }
> >>  
> >>  /**
> >> @@ -151,7 +150,7 @@ static int acpi_pm_prepare(void)
> >>  {
> >>  	int error = __acpi_pm_prepare();
> >>  	if (!error)
> >> -		acpi_pm_pre_suspend();
> >> +		error = acpi_pm_pre_suspend();
> >>  
> >>  	return error;
> >>  }
> >> diff --git a/include/linux/suspend.h b/include/linux/suspend.h
> >> index c1f4998..3ac2551 100644
> >> --- a/include/linux/suspend.h
> >> +++ b/include/linux/suspend.h
> >> @@ -262,7 +262,7 @@ static inline bool system_entering_hibernation(void) { return false; }
> >>  extern int suspend_nvs_register(unsigned long start, unsigned long size);
> >>  extern int suspend_nvs_alloc(void);
> >>  extern void suspend_nvs_free(void);
> >> -extern void suspend_nvs_save(void);
> >> +extern int suspend_nvs_save(void);
> >>  extern void suspend_nvs_restore(void);
> >>  #else /* CONFIG_SUSPEND_NVS */
> >>  static inline int suspend_nvs_register(unsigned long a, unsigned long b)
> >> @@ -271,7 +271,7 @@ static inline int suspend_nvs_register(unsigned long a, unsigned long b)
> >>  }
> >>  static inline int suspend_nvs_alloc(void) { return 0; }
> >>  static inline void suspend_nvs_free(void) {}
> >> -static inline void suspend_nvs_save(void) {}
> >> +static inline int suspend_nvs_save(void) {}
> >>  static inline void suspend_nvs_restore(void) {}
> >>  #endif /* CONFIG_SUSPEND_NVS */
> >>  
> >> diff --git a/kernel/power/nvs.c b/kernel/power/nvs.c
> >> index 1836db6..57c6fab 100644
> >> --- a/kernel/power/nvs.c
> >> +++ b/kernel/power/nvs.c
> >> @@ -105,7 +105,7 @@ int suspend_nvs_alloc(void)
> >>  /**
> >>   *	suspend_nvs_save - save NVS memory regions
> >>   */
> >> -void suspend_nvs_save(void)
> >> +int suspend_nvs_save(void)
> >>  {
> >>  	struct nvs_page *entry;
> >>  
> >> @@ -114,8 +114,14 @@ void suspend_nvs_save(void)
> >>  	list_for_each_entry(entry, &nvs_list, node)
> >>  		if (entry->data) {
> >>  			entry->kaddr = ioremap(entry->phys_start, entry->size);
> > 
> > I wonder what happens if you simply change the ioremap() here to
> > ioremap_nocache() without any other modifications?
> 
> ioremap *is* ioremap_nocache on x86. And that's the conflict it
> complains about I guess? Don't you mean ioremap_cache?

Yes, I meant ioremap_cache(), sorry.  Using ioremap_cache() here fixes the
problem for Len (he's seeing the same issue on his test machine).

The question is why it helps, though.  My theory is that we have mapped the
same area already using ioremap_cache() and now we're trying to map it again
using ioremap_nocache(), hence the conflict.  I need to confirm this.

> > It _really_ shouldn't fail here, because the NVS pages are known to be present.
> 
> It fails because of conflicting maps as can be seen in the photo. At
> least I think so.

Yes, I think so too.  Which is _suspicious_.

Thanks,
Rafael

Re: [PATCH 1/1] PM: fix oops in suspend/hibernate code

[Len Brown]

> ... My theory is that we have mapped the
> same area already using ioremap_cache() and now we're trying to map it again
> using ioremap_nocache(), hence the conflict.  I need to confirm this.

On my test box.

BIOS-e820: 0000000077626000 - 0000000077632000 (ACPI NVS)

yet there are at least two tables (FACS and SSDT)
that live in that region, and there are several
run-time AML memory opregions residing in that range too.
So the region has already been mapped by acpi_os_map_memory()
(now using ioremap_cache()) before suspend_nvs_save() runs.

Is there a reason that suspend_nvs_save()
requires a non-cached mapping?

cheers,
Len Brown, Intel Open Source Technology Center