Adds PMU and S2R support for exynos5420

From: Abhilash Kesavan <[email protected]>

Adds intial PMU settings for exynos5420. This is required for future S2R and Switching support.

Signed-off-by Thomas Abraham <[email protected]>

Signed-off-by Abhilash Kesavan <[email protected]>

Signed-off-by Vikas Sajjan <[email protected]>

---
 arch/arm/mach-exynos/pmu.c      |  287 +++++++++++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/regs-pmu.h |  227 +++++++++++++++++++++++++++++++
 2 files changed, 514 insertions(+)

Hellos Vikas,

Quoted 5 lines · Vikas Sajjan

From: Abhilash Kesavan <[email protected]>

Adds intial PMU settings for exynos5420. This is required for future S2R and Switching support.

Quoted 7 lines

Signed-off-by Thomas Abraham <[email protected]>

Signed-off-by Abhilash Kesavan <[email protected]>

Signed-off-by Vikas Sajjan <[email protected]>

---
 arch/arm/mach-exynos/pmu.c      |  287 +++++++++++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/regs-pmu.h |  227 +++++++++++++++++++++++++++++++
 2 files changed, 514 insertions(+)

S2R is working on my Exynos5420 Peach Pit Chromebook with your series:

Tested-by Javier Martinez Canillas <[email protected]>

Best regards,
Javier

Adds Suspend-to-RAM support for EXYNOS5420

Signed-off-by Vikas Sajjan <[email protected]>

Signed-off-by Abhilash Kesavan <[email protected]>

---
 arch/arm/mach-exynos/suspend.c |  151 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 149 insertions(+), 2 deletions(-)

Hello Vikas,

Quoted 2 lines · Vikas Sajjan

Adds Suspend-to-RAM support for EXYNOS5420

Quoted 6 lines

Signed-off-by Vikas Sajjan <[email protected]>

Signed-off-by Abhilash Kesavan <[email protected]>

---
 arch/arm/mach-exynos/suspend.c |  151 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 149 insertions(+), 2 deletions(-)

Tested-by Javier Martinez Canillas <[email protected]>

Best regards,
Javier

Rebased on
1] Kukjin Kim's tree, for-next branch
https://git.kernel.org/cgit/linux/kernel/git/kgene/linux-samsung.git/log/?h=for-next
2] Pankaj Dubey's v9 PMU patchset
http://www.spinics.net/lists/arm-kernel/msg367939.html

changes since v8:
- addressed abhilash's comments to remove the restoring of the CPU0 low power state register, since it will taken care in mcpm s2r patch from abhilash.

changes since v7:
- rebased on pankaj's latest patchset.

changes since v6:
- rebased on 3.17.rc1.

changes since v5:
- Refactored pm.c to use DT based lookup as suggested by Tomasz Figa.

changes since v4:
- Adressed comments from Tomasz figa and rebased on Pankaj Dubey's v5 PMU patchset

changes since v3:
Addressed the following comments from Pankaj Dubey, Bartlomiej Zolnierkiewicz, Tomasz Figa and Alim Akhtar:
- Moved EXYNOS5420_USE_STANDBY_WFI_ALL define to regs-pmu.h.
- Merged exynos5420_set_core_flag function into powerdown_conf.
- Removed XXTI_DURATION3 register setting.
- Updated the commit message and ordered the clock registers in clock patch.
- Removed the code for SYS_DISP1_BLK_CFG handling.
- Modified SoC checks to A9 specific checks in PM code.
- Updated some comments in the code and added macros for register offsets.
- Fixed code which was changing pad retention code for older SoCs.

changes since v2:
- Addressed comments from Tomasz figa
- rebased on Pankaj's V3 patchset https://lkml.org/lkml/2014/5/2/612
- dropped patch "ARM: dts: Add node for GPIO keys on SMDK5420", will be sent separately.

changes since v1:
- Addressed comments from Tomasz figa.
- restructured/consolidated as per Tomasz figa's PM consolidations for exynos

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

Abhilash Kesavan (1):
ARM: exynos5: Add PMU support for 5420

Vikas Sajjan (1):
ARM: exynos5: Add Suspend-to-RAM support for 5420

 arch/arm/mach-exynos/pmu.c      |  287 +++++++++++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/regs-pmu.h |  227 +++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/suspend.c  |  151 +++++++++++++++++++-
 3 files changed, 663 insertions(+), 2 deletions(-)

--
1.7.9.5

Hi Kukjin,

Quoted 69 lines

Rebased on
1] Kukjin Kim's tree, for-next branch
https://git.kernel.org/cgit/linux/kernel/git/kgene/linux-samsung.git/log/?h=for-next
2] Pankaj Dubey's v9 PMU patchset
http://www.spinics.net/lists/arm-kernel/msg367939.html

changes since v8:
- addressed abhilash's comments to remove the restoring of the CPU0 low power state register, since it will taken care in mcpm s2r patch from abhilash.

changes since v7:
- rebased on pankaj's latest patchset.

changes since v6:
- rebased on 3.17.rc1.

changes since v5:
- Refactored pm.c to use DT based lookup as suggested by Tomasz Figa.

changes since v4:
- Adressed comments from Tomasz figa and rebased on Pankaj Dubey's v5 PMU patchset

changes since v3:
Addressed the following comments from Pankaj Dubey, Bartlomiej Zolnierkiewicz, Tomasz Figa and Alim Akhtar:
- Moved EXYNOS5420_USE_STANDBY_WFI_ALL define to regs-pmu.h.
- Merged exynos5420_set_core_flag function into powerdown_conf.
- Removed XXTI_DURATION3 register setting.
- Updated the commit message and ordered the clock registers in clock patch.
- Removed the code for SYS_DISP1_BLK_CFG handling.
- Modified SoC checks to A9 specific checks in PM code.
- Updated some comments in the code and added macros for register offsets.
- Fixed code which was changing pad retention code for older SoCs.

changes since v2:
- Addressed comments from Tomasz figa
- rebased on Pankaj's V3 patchset https://lkml.org/lkml/2014/5/2/612
- dropped patch "ARM: dts: Add node for GPIO keys on SMDK5420", will be sent separately.

changes since v1:
- Addressed comments from Tomasz figa.
- restructured/consolidated as per Tomasz figa's PM consolidations for exynos

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

Abhilash Kesavan (1):
ARM: exynos5: Add PMU support for 5420

Vikas Sajjan (1):
ARM: exynos5: Add Suspend-to-RAM support for 5420

Quoted 5 lines

 arch/arm/mach-exynos/pmu.c      |  287 +++++++++++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/regs-pmu.h |  227 +++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/suspend.c  |  151 +++++++++++++++++++-
 3 files changed, 663 insertions(+), 2 deletions(-)

Can you please pick this series.

Quoted 3 lines · Vikas Sajjan

Hi Kukjin,

Hi,

Quoted 77 lines · Vikas Sajjan

Quoted 74 lines

Rebased on
1] Kukjin Kim's tree, for-next branch
https://git.kernel.org/cgit/linux/kernel/git/kgene/linux-samsung.git/log/?h=for-next
2] Pankaj Dubey's v9 PMU patchset
http://www.spinics.net/lists/arm-kernel/msg367939.html

changes since v8:
- addressed abhilash's comments to remove the restoring of the CPU0 low power state register, since it will taken care in mcpm s2r patch from abhilash.

changes since v7:
- rebased on pankaj's latest patchset.

changes since v6:
- rebased on 3.17.rc1.

changes since v5:
- Refactored pm.c to use DT based lookup as suggested by Tomasz Figa.

changes since v4:
- Adressed comments from Tomasz figa and rebased on Pankaj Dubey's v5 PMU patchset

changes since v3:
Addressed the following comments from Pankaj Dubey, Bartlomiej Zolnierkiewicz, Tomasz Figa and Alim Akhtar:
- Moved EXYNOS5420_USE_STANDBY_WFI_ALL define to regs-pmu.h.
- Merged exynos5420_set_core_flag function into powerdown_conf.
- Removed XXTI_DURATION3 register setting.
- Updated the commit message and ordered the clock registers in clock patch.
- Removed the code for SYS_DISP1_BLK_CFG handling.
- Modified SoC checks to A9 specific checks in PM code.
- Updated some comments in the code and added macros for register offsets.
- Fixed code which was changing pad retention code for older SoCs.

changes since v2:
- Addressed comments from Tomasz figa
- rebased on Pankaj's V3 patchset https://lkml.org/lkml/2014/5/2/612
- dropped patch "ARM: dts: Add node for GPIO keys on SMDK5420", will be sent separately.

changes since v1:
- Addressed comments from Tomasz figa.
- restructured/consolidated as per Tomasz figa's PM consolidations for exynos

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

Abhilash Kesavan (1):
ARM: exynos5: Add PMU support for 5420

Vikas Sajjan (1):
ARM: exynos5: Add Suspend-to-RAM support for 5420

 arch/arm/mach-exynos/pmu.c      |  287 +++++++++++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/regs-pmu.h |  227 +++++++++++++++++++++++++++++++
 arch/arm/mach-exynos/suspend.c  |  151 +++++++++++++++++++-
 3 files changed, 663 insertions(+), 2 deletions(-)

Quoted 2 lines

Can you please pick this series.

Yes, sure I will queue this series after -rc1 release.

Thanks,
Kukjin

Quoted 4 lines

Quoted 3 lines

Can you please pick this series.

Yes, sure I will queue this series after -rc1 release.

Thank you.

Quoted 4 lines

Thanks,
Kukjin

Hello Vikas,

Quoted 1 line · Vikas Sajjan

Quoted 1 line

Quoted 1 line

Quoted 4 lines

Quoted 3 lines

Quoted 2 lines

Can you please pick this series.

Yes, sure I will queue this series after -rc1 release.

Not related to your series but did you figure out why the aclk200_disp1 and aclk300_disp1 clocks are left unused?

I'm carrying your patch to flag those clocks as CLK_IGNORE_UNUSED to prevent being disabled by the CCF. And while that allows me to test, I'm afraid that it may not be the proper solution since it could just be hiding the real cause of the issue.

Of course this should not block your series from been merged but I just wondered if you found the cause since it would be nice to have s2r properly working on exynos5420.

Quoted 6 lines

Thank you.

Quoted 4 lines

Thanks,
Kukjin

best regards,
Javier

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Quoted 4 lines

Not related to your series but did you figure out why the aclk200_disp1 and aclk300_disp1 clocks are left unused?

As mentioned earlier, I recently noticed that these clocks are preventing system from suspending.
These clocks are used by the in-flight DRM display driver and once it is merged the display driver will be handling the clock during boot-up and suspend/resume.

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

Vikas Sajjan <[email protected]> writes:

[...]

Quoted 7 lines

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Quoted 9 lines

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

I went tried to this on exynos5800-peach-pi and found first that exynos_defonfig is missing the MAX77802 kconfig options for the RTC source clock:

CONFIG_REGULATOR_MAX77802=y
CONFIG_COMMON_CLK_MAX77802=y

With those, rtc0 then comes up, but isn't waking from suspend. However, writing something to rtc0/wakealarm does result in /proc/interrupts having an interrupt for the RTC, it's just not waking the system.

Anyone else tried this on 5800/peach-pi?

Hi Kevin,

Quoted 25 lines · Kevin Hilman

Vikas Sajjan <[email protected]> writes:

[...]

Quoted 7 lines

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Quoted 9 lines

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

I went tried to this on exynos5800-peach-pi and found first that exynos_defonfig is missing the MAX77802 kconfig options for the RTC source clock:

Quoted 2 lines

CONFIG_REGULATOR_MAX77802=y
CONFIG_COMMON_CLK_MAX77802=y

I am using exynos_defconfig (no changes) with the internal SoC RTC which is enabled by default. I did not enable the MAX77802 RTC.

Quoted 6 lines

With those, rtc0 then comes up, but isn't waking from suspend. However, writing something to rtc0/wakealarm does result in /proc/interrupts having an interrupt for the RTC, it's just not waking the system.

Anyone else tried this on 5800/peach-pi?

I have tested this on a Peach-Pi and the system is resuming fine. The patches applied on kgene's for-next branch along with the aclk200_disp1 fix are:
http://lkml.org/lkml/2014/9/30/156
https://lkml.org/lkml/2014/10/6/89
http://www.spinics.net/lists/arm-kernel/msg368207.html
http://www.spinics.net/lists/linux-samsung-soc/msg37647.html

Hello Abhilash,

Quoted 28 lines · Abhilash Kesavan

Hi Kevin,

Quoted 25 lines · Kevin Hilman

Vikas Sajjan <[email protected]> writes:

[...]

Quoted 7 lines

Tested on Kukjin Kim's tree, for-next branch +
1] http://www.spinics.net/lists/linux-samsung-soc/msg33750.html
2] https://www.mail-archive.com/linux-samsung-soc at vger.kernel.org/msg37260.html 3] with the CLK_IGNORE_UNUSED flag for aclk200_disp1 CLK

on Exynos5420 based chromebook (peach-pit board)

Quoted 9 lines

Below procedures were followed to test S2R:
Procedure A:
1. make multi_v7_defconfig
2 enable MCPM for 5420
3. enable S3C RTC
5. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state
Procedure B:
1. make exynos_defconfig
4. echo +20 > /sys/class/rtc/rtc0/wakealarm && echo mem > /sys/power/state

I went tried to this on exynos5800-peach-pi and found first that exynos_defonfig is missing the MAX77802 kconfig options for the RTC source clock:

Quoted 5 lines

Quoted 2 lines

CONFIG_REGULATOR_MAX77802=y
CONFIG_COMMON_CLK_MAX77802=y

I am using exynos_defconfig (no changes) with the internal SoC RTC which is enabled by default. I did not enable the MAX77802 RTC.

That is because you probably are using kukjin's for-next branch as a base but I had the same issue when testing s2r on linux-next.

The problem is that linux-next contains commit 546b117fdf17 ("rtc:
s3c: add support for RTC of Exynos3250 SoC") that added a required "rtc_src" DT property for the Samsung's S3C RTC that is used to specify the source clock XrtcXTI.

On Peach boards, the S3C source clock is the max77802 32kHZ AP clock so I fixed that with commit 8bd43b6 ("ARM: dts: Add rtc_src clk for s3c-rtc on exynos Peach boards") but didn't send a patch to enable the max77802 subdevices on exynos_defconfig.

I'll post a patch to enable those config options today and maybe they can be considered as -rc material.

Quoted 7 lines

Quoted 6 lines

With those, rtc0 then comes up, but isn't waking from suspend. However, writing something to rtc0/wakealarm does result in /proc/interrupts having an interrupt for the RTC, it's just not waking the system.

Anyone else tried this on 5800/peach-pi?

Quoted 8 lines

I have tested this on a Peach-Pi and the system is resuming fine. The patches applied on kgene's for-next branch along with the aclk200_disp1 fix are:
http://lkml.org/lkml/2014/9/30/156
https://lkml.org/lkml/2014/10/6/89
http://www.spinics.net/lists/arm-kernel/msg368207.html
http://www.spinics.net/lists/linux-samsung-soc/msg37647.html

Kevin, I didn't test it on 5800 peach pi but I did test it on 5420 peach pit and as Abhilash said, I also needed Vikas patch [0] to flag the aclk200_disp1 and aclk300_disp1 clocks as CLK_IGNORE_UNUSED to prevent the ccf to disable them. With that patch Suspend-to-Ram worked for me and resuming using both the internal S3C RTC and the max77802 RTC.

Also I tested the pm_tests and all of them passed on my peach pit.

On peach pi you shouldn't need the clock patch AFAIU if you enable the needed kconfig options to have display working since those clocks are used by the display controller so they won't be disabled due unused.

Quoted 8 lines

On Peach boards, the S3C source clock is the max77802 32kHZ AP clock so I fixed that with commit 8bd43b6 ("ARM: dts: Add rtc_src clk for s3c-rtc on exynos Peach boards") but didn't send a patch to enable the max77802 subdevices on exynos_defconfig.

I'll post a patch to enable those config options today and maybe they can be considered as -rc material.

Great, thanks.

Trying $SUBJECT series on top of kukjin's for-next is working fine, but I'm still not able to make it work on linux-next, with or without the CLK_IGNORE_UNUSED patch.

Kevin Hilman <[email protected]> writes:

[...]

Quoted 3 lines

Trying $SUBJECT series on top of kukjin's for-next is working fine, but I'm still not able to make it work on linux-next, with or without the CLK_IGNORE_UNUSED patch.

Nevermind, turns out I didn't apply the CLK_IGNORE_UNUSED PATCH. If I actually apply it, I have it working on top of linux-next. However, it seems to have some issues with the nonboot CPUs resuming:

[   11.319833] Enabling non-boot CPUs ...
[   12.319134] CPU1: failed to come online
[   12.319235] Error taking CPU1 up: -5
[   13.319137] CPU2: failed to come online
[   13.319209] Error taking CPU2 up: -5
[   14.319133] CPU3: failed to come online
[   14.319224] Error taking CPU3 up: -5

Are you guys seeing all the non-boot CPUs coming up? They fail to come online, but the /sys/devices/.../cpu?/online reports that they are online, which means the next attempt to hotplug them will probably BUG().

Also, If I turn off the switcher (enabled by default exynos_defconfig), resume still has problems bringing all the nonboot CPUs online, and then has a bunch of I2C errors[1].

And one more..., if I try enabling CPUidle[1] with exynos_defconfig, S2R causes a reboot (seems to be on resume.)

Quoted 28 lines · Kevin Hilman

Kevin Hilman <[email protected]> writes:

[...]

Quoted 3 lines

Trying $SUBJECT series on top of kukjin's for-next is working fine, but I'm still not able to make it work on linux-next, with or without the CLK_IGNORE_UNUSED patch.

Nevermind, turns out I didn't apply the CLK_IGNORE_UNUSED PATCH. If I actually apply it, I have it working on top of linux-next. However, it seems to have some issues with the nonboot CPUs resuming:

[   11.319833] Enabling non-boot CPUs ...
[   12.319134] CPU1: failed to come online
[   12.319235] Error taking CPU1 up: -5
[   13.319137] CPU2: failed to come online
[   13.319209] Error taking CPU2 up: -5
[   14.319133] CPU3: failed to come online
[   14.319224] Error taking CPU3 up: -5

Are you guys seeing all the non-boot CPUs coming up? They fail to come online, but the /sys/devices/.../cpu?/online reports that they are online, which means the next attempt to hotplug them will probably BUG().

Also, If I turn off the switcher (enabled by default exynos_defconfig), resume still has problems bringing all the nonboot CPUs online, and then has a bunch of I2C errors[1].

Quoted 2 lines

And one more..., if I try enabling CPUidle[1] with exynos_defconfig, S2R causes a reboot (seems to be on resume.)

I tested S2R on my Peach-Pi (Exynos5800) on linux-next-20141009. My tree has:

a57db95 TEMP: Use CLK_IGNORE_UNUSED for aclk200_disp1 18158ac ARM: EXYNOS: Use MCPM call-backs to support S2R on Exynos5420

d86cb2d ARM: exynos5: Add Suspend-to-RAM support for 5420 93f1b7c ARM: exynos5: Add PMU support for 5420
8dcc2db ARM: EXYNOS: Move PMU specific definitions from common.h 493a16a ARM: EXYNOS: Add platform driver support for Exynos PMU 64e73fb mfd: syscon: Decouple syscon interface from platform devices af7b574 Add linux-next specific files for 20141009 4658bc0 Merge branch 'akpm/master'
9784a19 mm: add strictlimit knob

I have tested the following scenarios using exynos_defconfig with CONFIG_COMMON_CLK_MAX77802 enabled:

1) CONFIG_BIG_LITTLE enabled (default in exynos_defconfig)
2) CONFIG_BIG_LITTLE disabled
3) CONFIG_ARM_BIG_LITTLE_CPUIDLE enabled

In all 3 cases the secondary cores as well as the system resume fine. Can you please confirm if your kernel has a similar configuration ?

Quoted 29 lines · Abhilash Kesavan

Quoted 28 lines · Kevin Hilman

Kevin Hilman <[email protected]> writes:

[...]

Quoted 3 lines

Trying $SUBJECT series on top of kukjin's for-next is working fine, but I'm still not able to make it work on linux-next, with or without the CLK_IGNORE_UNUSED patch.

Nevermind, turns out I didn't apply the CLK_IGNORE_UNUSED PATCH. If I actually apply it, I have it working on top of linux-next. However, it seems to have some issues with the nonboot CPUs resuming:

[   11.319833] Enabling non-boot CPUs ...
[   12.319134] CPU1: failed to come online
[   12.319235] Error taking CPU1 up: -5
[   13.319137] CPU2: failed to come online
[   13.319209] Error taking CPU2 up: -5
[   14.319133] CPU3: failed to come online
[   14.319224] Error taking CPU3 up: -5

Are you guys seeing all the non-boot CPUs coming up? They fail to come online, but the /sys/devices/.../cpu?/online reports that they are online, which means the next attempt to hotplug them will probably BUG().

Also, If I turn off the switcher (enabled by default exynos_defconfig), resume still has problems bringing all the nonboot CPUs online, and then has a bunch of I2C errors[1].

Quoted 7 lines

Quoted 2 lines

And one more..., if I try enabling CPUidle[1] with exynos_defconfig, S2R causes a reboot (seems to be on resume.)

I tested S2R on my Peach-Pi (Exynos5800) on linux-next-20141009. My tree has:

a57db95 TEMP: Use CLK_IGNORE_UNUSED for aclk200_disp1 18158ac ARM: EXYNOS: Use MCPM call-backs to support S2R on Exynos5420

I was missing this one.

Quoted 12 lines

d86cb2d ARM: exynos5: Add Suspend-to-RAM support for 5420 93f1b7c ARM: exynos5: Add PMU support for 5420
8dcc2db ARM: EXYNOS: Move PMU specific definitions from common.h 493a16a ARM: EXYNOS: Add platform driver support for Exynos PMU 64e73fb mfd: syscon: Decouple syscon interface from platform devices af7b574 Add linux-next specific files for 20141009 4658bc0 Merge branch 'akpm/master'
9784a19 mm: add strictlimit knob

I have tested the following scenarios using exynos_defconfig with CONFIG_COMMON_CLK_MAX77802 enabled:

Quoted 6 lines

1) CONFIG_BIG_LITTLE enabled (default in exynos_defconfig)
2) CONFIG_BIG_LITTLE disabled
3) CONFIG_ARM_BIG_LITTLE_CPUIDLE enabled

In all 3 cases the secondary cores as well as the system resume fine. Can you please confirm if your kernel has a similar configuration ?

Adding in the missing dependency above, it's now working for me in all 3 scenarios. Looking back at the cover letter, I see it's clearly listed as a dependency. Sorry, my fault.

Feel free to add

Tested-by Kevin Hilman <[email protected]>

to $SUBJECT series.

Kevin

Kevin Hilman <[email protected]> writes:

[...]

Quoted 3 lines

Trying $SUBJECT series on top of kukjin's for-next is working fine, but I'm still not able to make it work on linux-next, with or without the CLK_IGNORE_UNUSED patch.

Nevermind, turns out I didn't apply the CLK_IGNORE_UNUSED PATCH. If I actually apply it, I have it working on top of linux-next. However, it seems to have some issues with the nonboot CPUs resuming:

[   11.319833] Enabling non-boot CPUs ...
[   12.319134] CPU1: failed to come online
[   12.319235] Error taking CPU1 up: -5
[   13.319137] CPU2: failed to come online
[   13.319209] Error taking CPU2 up: -5
[   14.319133] CPU3: failed to come online
[   14.319224] Error taking CPU3 up: -5

Are you guys seeing all the non-boot CPUs coming up? They fail to come online, but the /sys/devices/.../cpu?/online reports that they are online, which means the next attempt to hotplug them will probably BUG().

Also, If I turn off the switcher (enabled by default exynos_defconfig), resume still has problems bringing all the nonboot CPUs online, and then has a bunch of I2C errors[1].

And one more..., if I try enabling CPUidle[1] with exynos_defconfig, S2R causes a reboot (seems to be on resume.)

Kevin

[1]

[   47.644273] cros-ec-spi spi2.0: EC failed to respond in time
[   47.648535] cros-ec-spi spi2.0: spi transfer failed: -110
[   47.653866] tps65090-charger tps65090-charger: tps65090_charger_isr(): Error in reading reg 0xa

[2]
CONFIG_CPU_IDLE=y
CONFIG_ARM_BIG_LITTLE_CPUIDLE=y
CONFIG_ARM_EXYNOS_CPUIDLE=n