[PATCH v4 00/16] ARM: shmobile: Add CPG/MSTP Clock Domain

From: Geert Uytterhoeven
Date: Tue Aug 04 2015 - 08:33:59 EST


Hi all,

This patch series adds Clock Domain support to the Clock Pulse Generator
(CPG) Module Stop (MSTP) Clocks driver using the generic PM Domain, to
be used on shmobile SoCs without device power domains (R-Car Gen1 and
Gen2, RZ). This allows to power-manage the module clocks of SoC devices
that are part of the CPG/MSTP Clock Domain using Runtime PM, or for
system suspend/resume, similar to SoCs with device power domains
(SH-Mobile and R-Mobile).

SoC devices that are part of the CPG/MSTP Clock Domain and can be
power-managed through an MSTP clock are tagged in DT with a proper
"power-domains" property, also serving as a visual clue. This applies to
most on-SoC devices, which have a one-to-one mapping from SoC device to
DT device node. Notable exceptions are "display" and "sound" device
nodes, which represent multiple SoC devices, each having their own MSTP
clocks. Hence drivers for such devices still have to manage their
(multiple module) clocks themselves.

The (MSTP) clock to use for power-management is found by scanning for
clocks that are compatible with "renesas,cpg-mstp-clocks".
In V1, the "first" clock tied to each device (con_id NULL) was used,
being a bit ad-hoc. It was suggested to use the "fck" clock instead,
but this may conflict with DT bindings for devices we don't control
(e.g. GIC-400 plans to mandate "clk" for the clk-name of its single
clock). Looking for real MSTP clocks avoids this problem.

Logically, the CPG/MSTP Clock Domain operates on the SoC CPG/MSTP block.
As there's no single device node in DT representing this block (there
are separate device nodes for the CPG and for the individual MSTP
clocks), I bound the logic to the CPG device node. Perhaps this is
something we should change for future SoCs?

Finally, the legacy default PM domain hack in drivers/sh/pm_runtime.c
is no longer needed when running an ARM multi-platform kernel on an
shmobile SoC with genpd support. Please note that this hack is still
needed for legacy (SH/ARM) platforms, and for the CONFIG_PM=n case.
Perhaps we should unconditionally enable PM when building shmobile
multi-platform kernels?

Compared to the legacy default PM domain hack, the CPG/MSTP Clock Domain
has several advantages:
- It only affects on-SoC devices, not all platform devices,
- It only affects the on-SoC devices we want, as specified in DT,
- Allmost all module clocks of all on-SoC devices (barring devices
needed for wake-up) are now gated during s2ram, saving more power.

By adding a small quirk to the CPG/MSTP Clock Domain code, its functions
to attach/detach devices to a PM Domain can be reused by the pm-rmobile
driver, reducing code duplication.

Here's a list of all devices in the CPG/MSTP Clock Domain on r8a7791:

root@koelsch:~# cat /sys/kernel/debug/pm_genpd/pm_genpd_summary
domain status slaves
/device runtime status
----------------------------------------------------------------------
cpg_clocks on
/devices/platform/e61c0000.interrupt-controller active
/devices/platform/e60b0000.i2c suspended
/devices/platform/ffca0000.timer suspended
/devices/platform/e6590100.usb-phy unsupported
/devices/platform/e6050000.gpio active
/devices/platform/e6051000.gpio active
/devices/platform/e6052000.gpio active
/devices/platform/e6053000.gpio active
/devices/platform/e6054000.gpio active
/devices/platform/e6055000.gpio active
/devices/platform/e6055400.gpio active
/devices/platform/e6055800.gpio active
/devices/platform/ee090000.pci active
/devices/platform/ee0d0000.pci active
/devices/platform/fe000000.pcie unsupported
/devices/platform/e6700000.dma-controller active
/devices/platform/e6720000.dma-controller active
/devices/platform/ec700000.dma-controller active
/devices/platform/ec720000.dma-controller suspended
/devices/platform/e65a0000.dma-controller suspended
/devices/platform/e65b0000.dma-controller suspended
/devices/platform/e6e60000.serial active
/devices/platform/e6e68000.serial active
/devices/platform/ee300000.sata unsupported
/devices/platform/e6b10000.spi suspended
/devices/platform/e6e20000.spi suspended
/devices/platform/ee700000.ethernet active
/devices/platform/e6530000.i2c suspended
/devices/platform/e6ef1000.video suspended
/devices/platform/e61f0000.thermal active
/devices/platform/ee100000.sd active
/devices/platform/ee140000.sd active
/devices/platform/ee160000.sd active
root@koelsch:~#

Patch overview:
- Patch 1 adds the core CPG/MSTP Clock Domain code to the CPG MSTP
driver,
- Patches 2-5 adds CPG/MSTP Clock Domain driver support for all
Renesas SoCs that have MSTP clocks, but no device power domains, and
updates the DT binding documentation accordingly,
- Patches 6-12 add CPG/MSTP Clock Domains to the dtsi files,
- Patches 13 and 14 disable the legacy default PM domain hack on all
ARM multi-platform builds with genpd support, now it's no longer
needed nor wanted,
- Patch 15 adds a quirk for r8a73a4 and sh73a0, where the Bus State
Controller is not power-managed by an MSTP clock, but by a plain
CPG clock,
- Patch 16 makes the R-Mobile PM Domain driver use the CPG/MSTP Clock
Domain attach/detach helpers, as they're more future-proof, and to
reduce code duplication.

Changes compared to v3 ("[PATCH v3 00/16] ARM: shmobile: Add CPG/MSTP
Clock Domain", https://lkml.org/lkml/2015/7/1/408):
- clk-mstp.c needs to #include <linux/clk.h> due to recent
<linux/clk*.h> shuffling,
- Add "power-domains" properties to recently introduced device nodes
(jpu, mmcif0),
- Consistently add "power-domains" properties after clock and dma
properties,
- Add Reviewed-by,

Changes compared to v2 ("[PATCH v2 00/14] ARM: shmobile: Add CPG Clock
Domains", https://lkml.org/lkml/2015/5/28/590):
- Add Acked-by,
- Use "CPG/MSTP Clock Domain" instead of "CPG Clock Domain",
- Drop bogus addition of #includes to clk-rcar-gen2.c,
- Call pm_clk_destroy() from cpg_mstp_detach_dev() only if
cpg_mstp_attach_dev() actually added a clock,
- Add "power-domains" property to recently introduced Ethernet AVB
device node,
- Add CPG/MSTP Clock Domain to recently introduced r8a7793 SoC,
- The legacy default PM Domain runtime check must stay for the ARM
multiplatform CONFIG_PM=n case, to prevent the code from running on
non-shmobile machines,
- Provide backwards-compatibility with old DTs (mainly for R-Car
Gen2),
- Extract EMMA Mobile EV2 removal into a separate patch,
- Replace explicit platform checks in the legacy default PM Domain
code by a check for the presence of MSTP clocks,
- Drop references to legacy r8a7740/sh73a0, which are gone.

Changes compared to v1 ("[PATCH/RFC 0/5] ARM: shmobile: rcar-gen2: Add
CPG Clock Domain",
https://www.marc.info/?l=linux-pm&m=142670805530085&w=3):
- Add Acked-by and Reviewed-by.
- Move core CPG Clock Domain code from the R-Car Gen2 driver to the
CPG MSTP Clocks driver, as it's generic, and can be used on other
Renesas SoCs that have a CPG/MSTP block,
- Scan for an MSTP clock instead of using the first clock tied to the
device (con_id NULL),
- Add support for R-Car Gen1 and RZ, in addition to R-Car Gen2,
allowing to drop the legacy default PM domain hack completely in
multi-platform builds,
- Reuse the CPG Clock Domain attach/detach helpers for pm-rmobile.
More detailed change logs are available in the individual patches.

Dependencies:
- This series is against renesas-devel-20150803-v4.2-rc5 (not
renesas-drivers-2015-08-04-v4.2-rc5, as that tree has more changes
to the .dtsi files),
- As usual when involving clocks and/or PM Domains, there are stringent
dependencies between the (subsets of) patches:
- Patches 2-5 depend on patch 1,
- Patches 6-12 depend on patches 2-5,
- Patch 14 depends on patches 6-13,
- Patch 15 depends on patch 1 only,
- Patch 16 depends on patch 15.

All of this was tested on:
- r8a73a4/ape6evm,
- r8a7740/armadillo,
- r8a7791/koelsch (with and without CONFIG_PM),
- sh73a0/kzm9g.

Testing on other shmobile platforms (esp. R-Car Gen1 and RZ) would be
appreciated.

I think this series goes best in through Simon's shmobile tree.

Thanks for applying!

Geert Uytterhoeven (16):
clk: shmobile: Add CPG/MSTP Clock Domain support
clk: shmobile: r8a7778: Add CPG/MSTP Clock Domain support
clk: shmobile: r8a7779: Add CPG/MSTP Clock Domain support
clk: shmobile: rcar-gen2: Add CPG/MSTP Clock Domain support
clk: shmobile: rz: Add CPG/MSTP Clock Domain support
ARM: shmobile: r7s72100 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7778 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7779 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7790 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7791 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7793 dtsi: Add CPG/MSTP Clock Domain
ARM: shmobile: r8a7794 dtsi: Add CPG/MSTP Clock Domain
drivers: sh: Disable legacy default PM Domain on emev2
drivers: sh: Disable PM runtime for multi-platform ARM with genpd
clk: shmobile: mstp: Consider "zb_clk" suitable for power management
ARM: shmobile: R-Mobile: Use CPG/MSTP Clock Domain attach/detach
helpers

.../bindings/clock/renesas,r8a7778-cpg-clocks.txt | 29 ++++++-
.../bindings/clock/renesas,r8a7779-cpg-clocks.txt | 30 ++++++-
.../clock/renesas,rcar-gen2-cpg-clocks.txt | 26 ++++++-
.../bindings/clock/renesas,rz-cpg-clocks.txt | 29 ++++++-
arch/arm/boot/dts/r7s72100.dtsi | 19 +++++
arch/arm/boot/dts/r8a7778.dtsi | 22 ++++++
arch/arm/boot/dts/r8a7779.dtsi | 23 ++++++
arch/arm/boot/dts/r8a7790.dtsi | 86 +++++++++++++++++---
arch/arm/boot/dts/r8a7791.dtsi | 88 +++++++++++++++++++--
arch/arm/boot/dts/r8a7793.dtsi | 7 ++
arch/arm/boot/dts/r8a7794.dtsi | 29 +++++++
arch/arm/mach-shmobile/Kconfig | 2 +
arch/arm/mach-shmobile/pm-rmobile.c | 35 +--------
drivers/clk/shmobile/clk-mstp.c | 91 ++++++++++++++++++++++
drivers/clk/shmobile/clk-r8a7778.c | 2 +
drivers/clk/shmobile/clk-r8a7779.c | 2 +
drivers/clk/shmobile/clk-rcar-gen2.c | 2 +
drivers/clk/shmobile/clk-rz.c | 3 +
drivers/sh/pm_runtime.c | 19 ++---
include/linux/clk/shmobile.h | 12 +++
20 files changed, 482 insertions(+), 74 deletions(-)

--
1.9.1

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/