Implement CPU clock control for Exynos850 SoC. It follows the same
procedure which is already implemented for other SoCs in clk-cpu.c:
1. Set the correct rate for the alternate parent (if needed) before
switching to use it as the CPU clock
2. Switch to the alternate parent, so the CPU continues to get clocked
while the PLL is being re-configured
3. Adjust the dividers for the CPU related buses (ACLK, ATCLK, etc)
4. Re-configure the PLL for the new CPU clock rate. It's done
automatically, as the CPU clock rate change propagates to the PLL
clock, because the CPU clock has CLK_SET_RATE_PARENT flag set in
exynos_register_cpu_clock()
5. Once the PLL is locked, set it back as the CPU clock source
6. Set alternate parent clock rate back to max speed
As in already existing clk-cpu.c code, the divider and mux clocks are
configured in a low-level fashion (using direct register access instead
of CCF API), to avoid affecting how DIV and MUX clock flags are declared
in the actual clock driver (clk-exynos850.c).
No functional change. This patch adds support for Exynos850 CPU clock,
but doesn't enable it per se.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-13-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Make it possible to use wait_until_mux_stable() for MUX registers where
the mask is different from MUX_MASK (e.g. in upcoming CPU clock
implementation for Exynos850).
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-12-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Abstract CPU clock registers by keeping their offsets in a dedicated
chip specific structure to accommodate for oncoming Exynos850 support,
which has different offsets for cluster 0 and cluster 1. This rework
also makes it possible to use exynos_set_safe_div() for all chips, so
exynos5433_set_safe_div() is removed here to reduce the code
duplication. The ".regs" field has to be (void *) as different Exynos
chips can have very different register layout, so this way it's possible
for ".regs" to point to different structures, each representing its own
chip's layout.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-11-semen.protsenko@linaro.org
[krzysztof: drop redundant const for regs in exynos_cpuclk_chip]
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Keep chip specific data in the data structure, don't mix it with code.
It makes it easier to add more chip specific data further. Having all
chip specific data in the table eliminates possible code bloat when
adding more rate handlers for new chips, and also makes it possible to
keep some other chip related data in that array.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-10-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Use a dedicated enum field to explicitly specify which register layout
should be used for the CPU clock, instead of passing it as a bit flag.
This way it would be possible to keep the chip-specific data in some
array, where each chip structure could be accessed by its corresponding
layout index. It prepares clk-cpu.c for adding new chips support, which
might have different data for different CPU clusters.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-9-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
The documentation for struct exynos_cpuclk says .ctrl_base field should
contain the controller base address. There are two different problems
with that:
1. All Exynos clock drivers are actually passing CPU_SRC register offset
via CPU_CLK() macro, which in turn gets assigned to mentioned
.ctrl_base field. Because CPU_SRC register usually already has 0x200
offset from controller's base, all other register offsets in
clk-cpu.c (like DIVs and MUXes) are specified as offsets from CPU_SRC
offset, and not from controller's base. That makes things confusing
and inconsistent with register offsets provided in Exynos clock
drivers, also breaking the contract for .ctrl_base field as described
in struct exynos_cpuclk doc.
2. Furthermore, some Exynos chips have an additional offset for the
start of CPU clock registers block (inside of the CMU). There might
be different reasons for that, e.g.:
- The CMU contains clocks for two different CPUs (like in Exynos5420)
- The CMU contains also non-CPU clocks as well (like in Exynos4)
- The CPU CMU exists as a dedicated hardware block in the SoC layout,
but is modelled as a part of bigger CMU in the driver (like in case
of Exynos3250)
That means the .ctrl_base field is actually not a controller's base,
but instead it's a start address of the CPU clock registers inside of
the CMU.
Rework all register offsets in clk-cpu.c to be actual offsets from the
CPU clock register block start, and fix offsets provided to CPU_CLK()
macro in all Exynos clock drivers. Also clarify the .ctrl_base field
documentation and rename it to just .base, because it doesn't really
contain the CMU base.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-8-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
clk-cpu.c is going to get messy as new chips support is added.
Restructure the code by pulling related functions and definitions
together, grouping those by their relation to a particular chip or other
categories, to simplify the code navigation.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-7-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reduce the code duplication by making all chips use a single version of
exynos_cpuclk_notifier_cb() function. That will prevent the code bloat
when adding new chips support too.
Also don't pass base address to pre/post rate change functions, as it
can be easily derived from already passed cpuclk param.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-6-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Pass CPU clock data structure to exynos_register_cpu_clock() instead of
passing its fields separately there. That simplifies the signature of
exynos_register_cpu_clock() and makes it easier to add more fields to
struct samsung_cpu_clock later. This style follows the example of
samsung_clk_register_pll().
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-5-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reduce the scope of struct exynos_cpuclk, as it's only used in clk-cpu.c
internally. All drivers using clk-pll.h already include clk.h as well,
so this change doesn't break anything.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-4-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
clk-cpu.c has numerous style issues reported by checkpatch and easily
identified otherwise. Give it some love and fix those warnings where it
makes sense. Also make stabilization time a named constant to get rid of
the magic number in clk-cpu.c.
No functional change.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Link: https://lore.kernel.org/r/20240224202053.25313-3-semen.protsenko@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Convert the remaining exynos clock drivers to use
samsung_clk_register_cpu() or if possible use samsung_cmu_register_one().
With this we can now make exynos_register_cpu_clock() a static function
so that future CPU clock registration changes will use the samsung common
clock driver.
The main benefit of this change is that it standardizes the CPU clock
registration for the samsung clock drivers.
Link: https://lore.kernel.org/r/20211015190515.3760577-1-willmcvicker@google.com
Signed-off-by: Will McVicker <willmcvicker@google.com>
Tested-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
[snawrocki@kernel.org: Fixed build break in clk-exynos4.c, clk-exynos5250.c]
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Adds 'struct samsung_cpu_clock' and corresponding CPU clock registration
function to the samsung common clk driver. This allows samsung clock
drivers to register their CPU clocks with the samsung_cmu_register_one()
API.
Currently the exynos5433 apollo and atlas clks have their own custom
init functions to handle registering their CPU clocks. With this patch
we can drop their custom CLK_OF_DECLARE functions and directly call
samsung_cmu_register_one().
Signed-off-by: Will McVicker <willmcvicker@google.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20211014195347.3635601-2-willmcvicker@google.com
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
For the CPU clock registration two parent clocks are required, these
are now being passed as struct clk_hw pointers, rather than by the
global scope names. That allows us to avoid __clk_lookup() calls
and simplifies a bit the CPU clock registration function.
While at it drop unneeded extern keyword in the function declaration.
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Link: https://lore.kernel.org/r/20200826171529.23618-3-s.nawrocki@samsung.com
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Based on 2 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation #
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 4122 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Enrico Weigelt <info@metux.net>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190604081206.933168790@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
clk_notifier_register() documentation states, that the provided notifier
callbacks associated with the notifier must not re-enter into the clk
framework by calling any top-level clk APIs. Fix this by replacing
clk_get_rate() calls with clk_hw_get_rate(), which is safe in this
context.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <snawrocki@kernel.org>
Omit an extra message for a memory allocation failure
in exynos_register_cpu_clock() function.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Clock providers should use the new struct clk_hw based API, so convert
Samsung clock providers and their helper functions to the new approach.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Add new variable to avoid using clk pointer for different purposes across
the exynos_register_cpu_clock() function. This will help in future rewrite
for the new clk_hw API.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Exynos5433 uses different register layout for CPU clock registers
than earlier SoCs so add new code for handling this layout. Also
add new CLK_CPU_HAS_E5433_REGS_LAYOUT flag to request using it.
There should be no functional change resulting from this patch.
Cc: Kukjin Kim <kgene@kernel.org>
CC: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
CLK_CPU_HAS_DIV1 and CLK_CPU_NEEDS_DEBUG_ALT_DIV masks were
incorrectly used as a bit numbers. Fix it.
Tested on Exynos4210 based Origen board and on Exynos5250 based
Arndale board.
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Michael Turquette <mturquette@baylibre.com>
Cc: Thomas Abraham <thomas.ab@samsung.com>
Fixes: ddeac8d96 ("clk: samsung: add infrastructure to register cpu clocks")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@dowhile0.org>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
We're removing struct clk from the clk provider API, so switch
this code to using the clk_hw based provider APIs.
Cc: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Clock provider drivers generally shouldn't include clk.h because
it's the consumer API. Only include clk.h in files that are
using it. The clkdev.h header isn't always used either, so remove
it and add in slab.h where files were relying on it to include
slab for them.
Cc: Chanwoo Choi <cw00.choi@samsung.com>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The CPU clock provider supplies the clock to the CPU clock domain. The
composition and organization of the CPU clock provider could vary among
Exynos SoCs. A CPU clock provider can be composed of clock mux, dividers
and gates. This patch defines a new clock type for CPU clock provider and
adds infrastructure to register the CPU clock providers for Samsung
platforms.
Changes by Bartlomiej:
- fixed issue with setting lower dividers before the parent clock speed
was lowered (the issue resulted in lockup on Exynos4210 SoC based
Origen board when "ondemand" cpufreq governor was stress tested)
- fixed missing spin_unlock on error in exynos_cpuclk_post_rate_change()
problem by moving cfg_data search outside of the spin locked area
- removed leftover kfree() in exynos_register_cpu_clock() that could
result in dereferencing the NULL pointer on error
- moved spin_lock earlier in exynos_cpuclk_pre_rate_change() to cover
reading of E4210_SRC_CPU and E4210_DIV_CPU1 registers
- added missing "last chance" checks to wait_until_divider_stable() and
wait_until_mux_stable() (needed in case that IRQ handling took long
time to proceed and resulted in function printing incorrect error
message about timeout)
- moved E4210_CPU_DIV[0,1]() macros just before their only users,
this resulted in moving them from patch #2 to patch #3/6 ("clk:
samsung: exynos4: add cpu clock configuration data and instantiate
cpu clock")
- removed E5250_CPU_DIV[0,1](), E5420_EGL_DIV0() and E5420_KFC_DIV()
macros for now
- added my Copyrights to drivers/clk/samsung/clk-cpu.c
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Mike Turquette <mturquette@linaro.org>
Cc: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Thomas Abraham <thomas.ab@samsung.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
