[PATCH] ARM: fix exynos randconfig build error

[PATCH] ARM: fix exynos randconfig build error

[Russell King]

The following error was observed with SMP=n in v4.0-rc1:

arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]

As the code unconditionally calls a function only available with SMP=y,
make the Exynos PM support depend on SMP.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
---
Alternatively, the code needs fixing not to call this function on non-SMP
platforms.

 arch/arm/mach-exynos/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/mach-exynos/Kconfig b/arch/arm/mach-exynos/Kconfig
index 603820e5aba7..55a1ada28350 100644
--- a/arch/arm/mach-exynos/Kconfig
+++ b/arch/arm/mach-exynos/Kconfig
@@ -131,6 +131,7 @@ config EXYNOS5420_MCPM
 
 config EXYNOS_CPU_SUSPEND
 	bool
+	depends on SMP
 	select ARM_CPU_SUSPEND
 	default PM_SLEEP || ARM_EXYNOS_CPUIDLE
 
-- 
1.8.3.1

[PATCH] ARM: fix exynos randconfig build error

[Krzysztof Kozlowski]

2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
> The following error was observed with SMP=n in v4.0-rc1:
>
> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
>
> As the code unconditionally calls a function only available with SMP=y,
> make the Exynos PM support depend on SMP.
>
> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

Hi,

Thanks for the patch but this already waits for Kukjin top be picked
up. The first patch was similar to yours (adds dependency on SMP),
sent on 4th of February:
https://patchwork.ozlabs.org/patch/436231/

But later Bartlomiej fixed this in other way (allowing to use cpuidle
on non-SMP):
https://patchwork.ozlabs.org/patch/436445/

Unfortunately none of them were picked up.

Best regards,
Krzysztof

[PATCH] ARM: fix exynos randconfig build error

[Kukjin Kim]

On 02/25/15 20:46, Krzysztof Kozlowski wrote:
> 2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
>> The following error was observed with SMP=n in v4.0-rc1:
>>
>> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
>> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
>>
>> As the code unconditionally calls a function only available with SMP=y,
>> make the Exynos PM support depend on SMP.
>>
>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> 
> Hi,
> 
> Thanks for the patch but this already waits for Kukjin top be picked
> up. The first patch was similar to yours (adds dependency on SMP),
> sent on 4th of February:
> https://patchwork.ozlabs.org/patch/436231/
> 
> But later Bartlomiej fixed this in other way (allowing to use cpuidle
> on non-SMP):
> https://patchwork.ozlabs.org/patch/436445/
> 
> Unfortunately none of them were picked up.
> 
I've missed the fix, sorry.

BTW, as you know, all of exynos SoCs are based on SMP so generally (in
normal case) there is no reason to use non-SMP on exynos
platforms...even though I understand the build error should be fixed...

Anyway, I'll have a look Bart's patch and Russell's fix in this weekend.

Thanks for your kindly pointing out.
Kukjin

[PATCH] ARM: fix exynos randconfig build error

[Kukjin Kim]

On 02/27/15 06:30, Kukjin Kim wrote:
> On 02/25/15 20:46, Krzysztof Kozlowski wrote:
>> 2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
>>> The following error was observed with SMP=n in v4.0-rc1:
>>>
>>> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
>>> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
>>>
>>> As the code unconditionally calls a function only available with SMP=y,
>>> make the Exynos PM support depend on SMP.
>>>
>>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
>>
>> Hi,
>>
>> Thanks for the patch but this already waits for Kukjin top be picked
>> up. The first patch was similar to yours (adds dependency on SMP),
>> sent on 4th of February:
>> https://patchwork.ozlabs.org/patch/436231/
>>
>> But later Bartlomiej fixed this in other way (allowing to use cpuidle
>> on non-SMP):
>> https://patchwork.ozlabs.org/patch/436445/
>>
>> Unfortunately none of them were picked up.
>>
> I've missed the fix, sorry.
> 
> BTW, as you know, all of exynos SoCs are based on SMP so generally (in
> normal case) there is no reason to use non-SMP on exynos
> platforms...even though I understand the build error should be fixed...
> 
> Anyway, I'll have a look Bart's patch and Russell's fix in this weekend.
> 

Firstly, let me take rmk's patch for the randconfig build error...BTW
I'm still wondering exynos stuff needs to support non-SMP and need to
think more about its usefulness?...

- Kukjin

[PATCH] ARM: fix exynos randconfig build error

[Bartlomiej Zolnierkiewicz]

Hi,

On Tuesday, March 03, 2015 04:40:02 AM Kukjin Kim wrote:
> On 02/27/15 06:30, Kukjin Kim wrote:
> > On 02/25/15 20:46, Krzysztof Kozlowski wrote:
> >> 2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
> >>> The following error was observed with SMP=n in v4.0-rc1:
> >>>
> >>> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
> >>> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
> >>>
> >>> As the code unconditionally calls a function only available with SMP=y,
> >>> make the Exynos PM support depend on SMP.
> >>>
> >>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> >>
> >> Hi,
> >>
> >> Thanks for the patch but this already waits for Kukjin top be picked
> >> up. The first patch was similar to yours (adds dependency on SMP),
> >> sent on 4th of February:
> >> https://patchwork.ozlabs.org/patch/436231/
> >>
> >> But later Bartlomiej fixed this in other way (allowing to use cpuidle
> >> on non-SMP):
> >> https://patchwork.ozlabs.org/patch/436445/
> >>
> >> Unfortunately none of them were picked up.
> >>
> > I've missed the fix, sorry.
> > 
> > BTW, as you know, all of exynos SoCs are based on SMP so generally (in
> > normal case) there is no reason to use non-SMP on exynos
> > platforms...even though I understand the build error should be fixed...
> > 
> > Anyway, I'll have a look Bart's patch and Russell's fix in this weekend.
> > 
> 
> Firstly, let me take rmk's patch for the randconfig build error...BTW

What is wrong with picking my patch instead?  It is non-invasive and fixes
cpuidle support on UP (which is a regression from previous kernels)?

https://lkml.org/lkml/2015/2/4/521

> I'm still wondering exynos stuff needs to support non-SMP and need to
> think more about its usefulness?...

Currently UP is supported and at least I find it useful for testing/debug
purposes.  If you want to to make Exynos SMP only thats OK but it should
be done globally for Exynos arch support not just for cpuidle support.

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics

[PATCH] ARM: fix exynos randconfig build error

[Russell King - ARM Linux]

On Tue, Mar 03, 2015 at 01:44:34PM +0100, Bartlomiej Zolnierkiewicz wrote:
> 
> Hi,
> 
> On Tuesday, March 03, 2015 04:40:02 AM Kukjin Kim wrote:
> > On 02/27/15 06:30, Kukjin Kim wrote:
> > > On 02/25/15 20:46, Krzysztof Kozlowski wrote:
> > >> 2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
> > >>> The following error was observed with SMP=n in v4.0-rc1:
> > >>>
> > >>> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
> > >>> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
> > >>>
> > >>> As the code unconditionally calls a function only available with SMP=y,
> > >>> make the Exynos PM support depend on SMP.
> > >>>
> > >>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> > >>
> > >> Hi,
> > >>
> > >> Thanks for the patch but this already waits for Kukjin top be picked
> > >> up. The first patch was similar to yours (adds dependency on SMP),
> > >> sent on 4th of February:
> > >> https://patchwork.ozlabs.org/patch/436231/
> > >>
> > >> But later Bartlomiej fixed this in other way (allowing to use cpuidle
> > >> on non-SMP):
> > >> https://patchwork.ozlabs.org/patch/436445/
> > >>
> > >> Unfortunately none of them were picked up.
> > >>
> > > I've missed the fix, sorry.
> > > 
> > > BTW, as you know, all of exynos SoCs are based on SMP so generally (in
> > > normal case) there is no reason to use non-SMP on exynos
> > > platforms...even though I understand the build error should be fixed...
> > > 
> > > Anyway, I'll have a look Bart's patch and Russell's fix in this weekend.
> > > 
> > 
> > Firstly, let me take rmk's patch for the randconfig build error...BTW
> 
> What is wrong with picking my patch instead?  It is non-invasive and fixes
> cpuidle support on UP (which is a regression from previous kernels)?
> 
> https://lkml.org/lkml/2015/2/4/521
> 
> > I'm still wondering exynos stuff needs to support non-SMP and need to
> > think more about its usefulness?...
> 
> Currently UP is supported and at least I find it useful for testing/debug
> purposes.  If you want to to make Exynos SMP only thats OK but it should
> be done globally for Exynos arch support not just for cpuidle support.

Has either of these happened yet?  I don't see either in arm-soc.

(lkml.org seems to be - as seems typical - having problems.  The above
URL now returns a blank message.)

-- 
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.

[PATCH] ARM: fix exynos randconfig build error

[Bartlomiej Zolnierkiewicz]

Hi,

On Monday, March 30, 2015 10:37:42 PM Russell King - ARM Linux wrote:
> On Tue, Mar 03, 2015 at 01:44:34PM +0100, Bartlomiej Zolnierkiewicz wrote:
> > 
> > Hi,
> > 
> > On Tuesday, March 03, 2015 04:40:02 AM Kukjin Kim wrote:
> > > On 02/27/15 06:30, Kukjin Kim wrote:
> > > > On 02/25/15 20:46, Krzysztof Kozlowski wrote:
> > > >> 2015-02-25 12:26 GMT+01:00 Russell King <rmk+kernel@arm.linux.org.uk>:
> > > >>> The following error was observed with SMP=n in v4.0-rc1:
> > > >>>
> > > >>> arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
> > > >>> arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
> > > >>>
> > > >>> As the code unconditionally calls a function only available with SMP=y,
> > > >>> make the Exynos PM support depend on SMP.
> > > >>>
> > > >>> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> > > >>
> > > >> Hi,
> > > >>
> > > >> Thanks for the patch but this already waits for Kukjin top be picked
> > > >> up. The first patch was similar to yours (adds dependency on SMP),
> > > >> sent on 4th of February:
> > > >> https://patchwork.ozlabs.org/patch/436231/
> > > >>
> > > >> But later Bartlomiej fixed this in other way (allowing to use cpuidle
> > > >> on non-SMP):
> > > >> https://patchwork.ozlabs.org/patch/436445/
> > > >>
> > > >> Unfortunately none of them were picked up.
> > > >>
> > > > I've missed the fix, sorry.
> > > > 
> > > > BTW, as you know, all of exynos SoCs are based on SMP so generally (in
> > > > normal case) there is no reason to use non-SMP on exynos
> > > > platforms...even though I understand the build error should be fixed...
> > > > 
> > > > Anyway, I'll have a look Bart's patch and Russell's fix in this weekend.
> > > > 
> > > 
> > > Firstly, let me take rmk's patch for the randconfig build error...BTW
> > 
> > What is wrong with picking my patch instead?  It is non-invasive and fixes
> > cpuidle support on UP (which is a regression from previous kernels)?
> > 
> > https://lkml.org/lkml/2015/2/4/521
> > 
> > > I'm still wondering exynos stuff needs to support non-SMP and need to
> > > think more about its usefulness?...
> > 
> > Currently UP is supported and at least I find it useful for testing/debug
> > purposes.  If you want to to make Exynos SMP only thats OK but it should
> > be done globally for Exynos arch support not just for cpuidle support.
> 
> Has either of these happened yet?  I don't see either in arm-soc.
> 
> (lkml.org seems to be - as seems typical - having problems.  The above
> URL now returns a blank message.)

My fix has been in Kukjin's tree (and thus linux-next) for the last two weeks:
http://git.kernel.org/cgit/linux/kernel/git/kgene/linux-samsung.git/commit/?h=v4.0-samsung-fixes-2&id=cfdda3535f87e752780ca18a57d13db58f6a6913

Unfortunately it seems that Samsung fixes branch has not been pushed to
arm-soc tree yet.

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics