linux-toradex.git/drivers/clk/sunxi-ng, branch v5.17-rc4 Linux kernel for Apalis and Colibri modules Merge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux 2022-01-13T01:02:27+00:00 Linus Torvalds torvalds@linux-foundation.org 2022-01-13T01:02:27+00:00 455e73a07f6e288b0061dfcf4fcf54fa9fe06458 Pull clk updates from Stephen Boyd: "We have a couple patches in the framework core this time around but they're mostly minor cleanups and some debugfs stuff. The real work that's in here is the typical pile of clk driver updates and new SoC support. Per usual (or maybe just recent trends), Qualcomm gains a handful of SoC drivers additions and has the largest diffstat. After that there are quite a few updates to the Allwinner (sunxi) drivers to support modular drivers and Renesas is heavily updated to add more support for various clks. Overall it looks pretty normal. New Drivers: - Add MDMA and BDMA clks to Ingenic JZ4760 and JZ4770 - MediaTek mt7986 SoC basic support - Clock and reset driver for Toshiba Visconti SoCs - Initial clock driver for the Exynos7885 SoC (Samsung Galaxy A8) - Allwinner D1 clks - Lan966x Generic Clock Controller driver and associated DT bindings - Qualcomm SDX65, SM8450, and MSM8976 GCC clks - Qualcomm SDX65 and SM8450 RPMh clks Updates: - Set suppress_bind_attrs to true for i.MX8ULP driver - Switch from do_div to div64_ul for throughout all i.MX drivers - Fix imx8mn_clko1_sels for i.MX8MN - Remove unused IPG_AUDIO_ROOT from i.MX8MP - Switch parent for audio_root_clk to audio ahb in i.MX8MP driver - Removal of all remaining uses of __clk_lookup() in drivers/clk/samsung - Refactoring of the CPU clocks registration to use common interface - An update of the Exynos850 driver (support for more clock domains) required by the E850-96 development board - Prep for runtime PM and generic power domains on Tegra - Support modular Allwinner clk drivers via platform bus - Lan966x clock driver extended to support clock gating - Add serial (SCI1), watchdog (WDT), timer (OSTM), SPI (RSPI), and thermal (TSU) clocks and resets on Renesas RZ/G2L - Rework SDHI clock handling in the Renesas R-Car Gen3 and RZ/G2 clock drivers, and in the Renesas SDHI driver - Make the Cortex-A55 (I) clock on Renesas RZ/G2L programmable - Document support for the new Renesas R-Car S4-8 (R8A779F0) SoC - Add support for the new Renesas R-Car S4-8 (R8A779F0) SoC - Add GPU clock and resets on Renesas RZ/G2L - Add clk-provider.h to various Qualcomm clk drivers - devm version of clk_hw_register_gate() - kerneldoc fixes in a couple drivers" * tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (131 commits) clk: visconti: Remove pointless NULL check in visconti_pll_add_lookup() clk: mediatek: add mt7986 clock support clk: mediatek: add mt7986 clock IDs dt-bindings: clock: mediatek: document clk bindings for mediatek mt7986 SoC clk: mediatek: clk-gate: Use regmap_{set/clear}_bits helpers clk: mediatek: clk-gate: Shrink by adding clockgating bit check helper clk: x86: Fix clk_gate_flags for RV_CLK_GATE clk: x86: Use dynamic con_id string during clk registration ACPI: APD: Add a fmw property clk-name drivers: acpi: acpi_apd: Remove unused device property "is-rv" x86: clk: clk-fch: Add support for newer family of AMD's SOC clk: ingenic: Add MDMA and BDMA clocks dt-bindings: clk/ingenic: Add MDMA and BDMA clocks clk: bm1880: remove kfrees on static allocations clk: Drop unused COMMON_CLK_STM32MP157_SCMI config clk: st: clkgen-mux: search reg within node or parent clk: st: clkgen-fsyn: search reg within node or parent clk: Enable/Disable runtime PM for clk_summary MAINTAINERS: Add entries for Toshiba Visconti PLL and clock controller clk: visconti: Add support common clock driver and reset driver ... 
Pull clk updates from Stephen Boyd:
 "We have a couple patches in the framework core this time around but
  they're mostly minor cleanups and some debugfs stuff. The real work
  that's in here is the typical pile of clk driver updates and new SoC
  support.

  Per usual (or maybe just recent trends), Qualcomm gains a handful of
  SoC drivers additions and has the largest diffstat. After that there
  are quite a few updates to the Allwinner (sunxi) drivers to support
  modular drivers and Renesas is heavily updated to add more support for
  various clks.

  Overall it looks pretty normal.

  New Drivers:
   - Add MDMA and BDMA clks to Ingenic JZ4760 and JZ4770
   - MediaTek mt7986 SoC basic support
   - Clock and reset driver for Toshiba Visconti SoCs
   - Initial clock driver for the Exynos7885 SoC (Samsung Galaxy A8)
   - Allwinner D1 clks
   - Lan966x Generic Clock Controller driver and associated DT bindings
   - Qualcomm SDX65, SM8450, and MSM8976 GCC clks
   - Qualcomm SDX65 and SM8450 RPMh clks

  Updates:
   - Set suppress_bind_attrs to true for i.MX8ULP driver
   - Switch from do_div to div64_ul for throughout all i.MX drivers
   - Fix imx8mn_clko1_sels for i.MX8MN
   - Remove unused IPG_AUDIO_ROOT from i.MX8MP
   - Switch parent for audio_root_clk to audio ahb in i.MX8MP driver
   - Removal of all remaining uses of __clk_lookup() in
     drivers/clk/samsung
   - Refactoring of the CPU clocks registration to use common interface
   - An update of the Exynos850 driver (support for more clock domains)
     required by the E850-96 development board
   - Prep for runtime PM and generic power domains on Tegra
   - Support modular Allwinner clk drivers via platform bus
   - Lan966x clock driver extended to support clock gating
   - Add serial (SCI1), watchdog (WDT), timer (OSTM), SPI (RSPI), and
     thermal (TSU) clocks and resets on Renesas RZ/G2L
   - Rework SDHI clock handling in the Renesas R-Car Gen3 and RZ/G2
     clock drivers, and in the Renesas SDHI driver
   - Make the Cortex-A55 (I) clock on Renesas RZ/G2L programmable
   - Document support for the new Renesas R-Car S4-8 (R8A779F0) SoC
   - Add support for the new Renesas R-Car S4-8 (R8A779F0) SoC
   - Add GPU clock and resets on Renesas RZ/G2L
   - Add clk-provider.h to various Qualcomm clk drivers
   - devm version of clk_hw_register_gate()
   - kerneldoc fixes in a couple drivers"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (131 commits)
  clk: visconti: Remove pointless NULL check in visconti_pll_add_lookup()
  clk: mediatek: add mt7986 clock support
  clk: mediatek: add mt7986 clock IDs
  dt-bindings: clock: mediatek: document clk bindings for mediatek mt7986 SoC
  clk: mediatek: clk-gate: Use regmap_{set/clear}_bits helpers
  clk: mediatek: clk-gate: Shrink by adding clockgating bit check helper
  clk: x86: Fix clk_gate_flags for RV_CLK_GATE
  clk: x86: Use dynamic con_id string during clk registration
  ACPI: APD: Add a fmw property clk-name
  drivers: acpi: acpi_apd: Remove unused device property "is-rv"
  x86: clk: clk-fch: Add support for newer family of AMD's SOC
  clk: ingenic: Add MDMA and BDMA clocks
  dt-bindings: clk/ingenic: Add MDMA and BDMA clocks
  clk: bm1880: remove kfrees on static allocations
  clk: Drop unused COMMON_CLK_STM32MP157_SCMI config
  clk: st: clkgen-mux: search reg within node or parent
  clk: st: clkgen-fsyn: search reg within node or parent
  clk: Enable/Disable runtime PM for clk_summary
  MAINTAINERS: Add entries for Toshiba Visconti PLL and clock controller
  clk: visconti: Add support common clock driver and reset driver
  ...
dt-bindings: clock: sunxi: Export CLK_DRAM for devfreq 2021-11-23T10:29:35+00:00 Samuel Holland samuel@sholland.org 2021-11-18T03:18:36+00:00 71b597ef5d46a326fb0d5cbfc1c6ff1d73cdc7f9 The MBUS node needs to reference the CLK_DRAM clock, as the MBUS hardware implements memory dynamic frequency scaling using this clock. Export this clock for SoCs which will be getting a devfreq driver. Acked-by: Rob Herring <robh@kernel.org> Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com> Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211118031841.42315-2-samuel@sholland.org 
The MBUS node needs to reference the CLK_DRAM clock, as the MBUS
hardware implements memory dynamic frequency scaling using this clock.

Export this clock for SoCs which will be getting a devfreq driver.

Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211118031841.42315-2-samuel@sholland.org
clk: sunxi-ng: Add support for the D1 SoC clocks 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T04:35:44+00:00 35b97bb941110a20fea1f3125837a98fdf8de054 The D1 SoC contains a CCU and a R_CCU (PRCM CCU). Add support for them. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119043545.4010-7-samuel@sholland.org 
The D1 SoC contains a CCU and a R_CCU (PRCM CCU). Add support for them.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119043545.4010-7-samuel@sholland.org
clk: sunxi-ng: gate: Add macros for gates with fixed dividers 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T04:35:43+00:00 b30fc68e6ce543130cfcf69fb02bafbcf961a7d4 It is possible to declare a gate with a fixed divider, by using the CCU_FEATURE_ALL_PREDIV flag. Since this is not obvious, add a macro for declaring this type of clock. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119043545.4010-6-samuel@sholland.org 
It is possible to declare a gate with a fixed divider, by using the
CCU_FEATURE_ALL_PREDIV flag. Since this is not obvious, add a macro
for declaring this type of clock.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119043545.4010-6-samuel@sholland.org
clk: sunxi-ng: mux: Add macros using clk_parent_data and clk_hw 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T04:35:42+00:00 8107c859a39159872ca45f751c4a70a89abe3321 Referencing parents with clk_hw pointers is more efficient and removes the dependency on global clock names. clk_parent_data is needed when some parent clocks are provided from another driver. Add macros for declaring muxes that take advantage of these. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119043545.4010-5-samuel@sholland.org 
Referencing parents with clk_hw pointers is more efficient and removes
the dependency on global clock names. clk_parent_data is needed when
some parent clocks are provided from another driver. Add macros for
declaring muxes that take advantage of these.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119043545.4010-5-samuel@sholland.org
clk: sunxi-ng: mp: Add macros using clk_parent_data and clk_hw 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T04:35:41+00:00 639e1acb69b5fc0db3621799278033e954e84d25 Referencing parents with clk_hw pointers is more efficient and removes the dependency on global clock names. clk_parent_data is needed when some parent clocks are provided from another driver. Add macros for declaring dividers that take advantage of these. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119043545.4010-4-samuel@sholland.org 
Referencing parents with clk_hw pointers is more efficient and removes
the dependency on global clock names. clk_parent_data is needed when
some parent clocks are provided from another driver. Add macros for
declaring dividers that take advantage of these.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119043545.4010-4-samuel@sholland.org
clk: sunxi-ng: div: Add macros using clk_parent_data and clk_hw 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T04:35:40+00:00 3317cb17d5da8416e0a0d9fe3c87dc846f3eb953 Referencing parents with clk_hw pointers is more efficient and removes the dependency on global clock names. clk_parent_data is needed when some parent clocks are provided from another driver. Add macros for declaring dividers that take advantage of these. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119043545.4010-3-samuel@sholland.org 
Referencing parents with clk_hw pointers is more efficient and removes
the dependency on global clock names. clk_parent_data is needed when
some parent clocks are provided from another driver. Add macros for
declaring dividers that take advantage of these.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119043545.4010-3-samuel@sholland.org
clk: sunxi-ng: Allow the CCU core to be built as a module 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T03:33:37+00:00 91389c390521a02ecfb91270f5b9d7fae4312ae5 Like the individual CCU drivers, it can be beneficial for memory consumption of cross-platform configurations to only load the CCU core on the relevant platform. For example, a generic arm64 kernel sees the following improvement when building the CCU core and drivers as modules: before: text data bss dec hex filename 13882360 5251670 360800 19494830 12977ae vmlinux after: text data bss dec hex filename 13734787 5086442 360800 19182029 124b1cd vmlinux So the result is a 390KB total reduction in kernel image size. The one early clock provider (sun5i) requires the core to be built in. Now that loading the MMC driver will trigger loading the CCU core, the MMC timing mode functions do not need a compile-time fallback. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119033338.25486-5-samuel@sholland.org 
Like the individual CCU drivers, it can be beneficial for memory
consumption of cross-platform configurations to only load the CCU core
on the relevant platform. For example, a generic arm64 kernel sees the
following improvement when building the CCU core and drivers as modules:

  before:
    text      data     bss     dec       hex      filename
    13882360  5251670  360800  19494830  12977ae  vmlinux

  after:
    text      data     bss     dec       hex      filename
    13734787  5086442  360800  19182029  124b1cd  vmlinux

So the result is a 390KB total reduction in kernel image size.

The one early clock provider (sun5i) requires the core to be built in.

Now that loading the MMC driver will trigger loading the CCU core, the
MMC timing mode functions do not need a compile-time fallback.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119033338.25486-5-samuel@sholland.org
clk: sunxi-ng: Convert early providers to platform drivers 2021-11-23T09:29:05+00:00 Samuel Holland samuel@sholland.org 2021-11-19T03:33:36+00:00 7ec03b588d2214d11bea91d5f9ffa444094d3cbd The PRCM CCU drivers depend on clocks provided by other CCU drivers. For example, the sun8i-r-ccu driver uses the "pll-periph" clock provided by the SoC's main CCU. However, sun8i-r-ccu is an early OF clock provider, and many of the main CCUs (e.g. sun50i-a64-ccu) use platform drivers. This means that the consumer clocks will be orphaned until the supplier driver is bound. This can be avoided by converting the remaining CCUs to use platform drivers. Then fw_devlink will ensure the drivers are bound in the optimal order. The sun5i CCU is the only one which actually needs to be an early clock provider, because it provides the clock for the system timer. That one is left alone. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119033338.25486-4-samuel@sholland.org 
The PRCM CCU drivers depend on clocks provided by other CCU drivers. For
example, the sun8i-r-ccu driver uses the "pll-periph" clock provided by
the SoC's main CCU.

However, sun8i-r-ccu is an early OF clock provider, and many of the
main CCUs (e.g. sun50i-a64-ccu) use platform drivers. This means that
the consumer clocks will be orphaned until the supplier driver is bound.
This can be avoided by converting the remaining CCUs to use platform
drivers. Then fw_devlink will ensure the drivers are bound in the
optimal order.

The sun5i CCU is the only one which actually needs to be an early clock
provider, because it provides the clock for the system timer. That one
is left alone.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119033338.25486-4-samuel@sholland.org
clk: sunxi-ng: Allow drivers to be built as modules 2021-11-22T09:02:21+00:00 Samuel Holland samuel@sholland.org 2021-11-19T03:33:35+00:00 c8c525b06f532923d21d99811a7b80bf18ffd2be While it is useful to build all of the CCU drivers at once, only 1-3 of them will be loaded at a time, or possibly none of them if the kernel is booted on a non-sunxi platform. These CCU drivers are relatively large; 32-bit drivers have 30-50k of data each, while the 64-bit ones are 50-75k due to the increased pointer overhead. About half of that data comes from relocations. Let's allow the user to build these drivers as modules so only the necessary data is loaded. As a first step, convert the CCUs that are already platform drivers. When the drivers are built as modules, normally the file name becomes the module name. However, the current file names are inconsistent with the <platform>-<peripheral> name used everywhere else: the devicetree bindings, the platform driver names, and the Kconfig symbols. Use Makfile logic to rename the modules so they follow the usual pattern. Signed-off-by: Samuel Holland <samuel@sholland.org> Signed-off-by: Maxime Ripard <maxime@cerno.tech> Link: https://lore.kernel.org/r/20211119033338.25486-3-samuel@sholland.org 
While it is useful to build all of the CCU drivers at once, only 1-3 of
them will be loaded at a time, or possibly none of them if the kernel is
booted on a non-sunxi platform. These CCU drivers are relatively large;
32-bit drivers have 30-50k of data each, while the 64-bit ones are
50-75k due to the increased pointer overhead. About half of that data
comes from relocations. Let's allow the user to build these drivers as
modules so only the necessary data is loaded.

As a first step, convert the CCUs that are already platform drivers.

When the drivers are built as modules, normally the file name becomes
the module name. However, the current file names are inconsistent with
the <platform>-<peripheral> name used everywhere else: the devicetree
bindings, the platform driver names, and the Kconfig symbols. Use
Makfile logic to rename the modules so they follow the usual pattern.

Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119033338.25486-3-samuel@sholland.org