pwm: sifive: Make use of devm_pwmchip_alloc() function

This prepares the pwm-sifive driver to further changes of the pwm core
outlined in the commit introducing devm_pwmchip_alloc(). There is no
intended semantical change and the driver should behave as before.

Link: https://lore.kernel.org/r/30a4cacafe2c323f2531dd1c1126f0bf0fe5e03c.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

pwm: sifive: Make use of pwmchip_parent() accessor

struct pwm_chip::dev is about to change. To not have to touch this
driver in the same commit as struct pwm_chip::dev, use the accessor
function provided for exactly this purpose.

Link: https://lore.kernel.org/r/170c3c61707992cbc57dce31f70c168fcd3dbe51.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

pwm: sifive: Prepare removing pwm_chip from driver data

This prepares the driver for further changes that will drop struct
pwm_chip chip from struct pwm_sifive_ddata. Use the pwm_chip as driver
data instead of the pwm_sifive_ddata to get access to the pwm_chip in
pwm_sifive_remove() without using ddata->chip. In the clock rate
notifier it's not possible to get the pwm_chip without adding a pointer
to this to struct pwm_sifive_ddata. Instead of that add a parent device
pointer which is all that is needed there.

Link: https://lore.kernel.org/r/b7b7985f4dc746f6a36c5048d428c4ed0a2d42dc.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

pwm: sifive: Simplify code to determine the pwmchip's parent device

There is already a pointer to the pwmchip, make use of it directly
instead of using the struct pwm_sifive_ddata *ddata just obtained from
it. This also has the advantage of not using struct
pwm_sifive_ddata::chip any more which will be dropped soon.

Link: https://lore.kernel.org/r/78da5070d3ff5767da6f1c053d069c1e25229375.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

pwm: Manage owner assignment implicitly for drivers

Instead of requiring each driver to care for assigning the owner member
of struct pwm_ops, handle that implicitly using a macro. Note that the
owner member has to be moved to struct pwm_chip, as the ops structure
usually lives in read-only memory and so cannot be modified.

The upside is that new low level drivers cannot forget the assignment and
save one line each. The pwm-crc driver didn't assign .owner, that's not
a problem in practice though as the driver cannot be compiled as a
module.

Acked-by: Andy Shevchenko <andy.shevchenko@gmail.com> # Intel LPSS
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> # pwm-{bcm,brcm}*.c
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com> # sun4i
Acked-by: Andi Shyti <andi.shyti@kernel.org>
Acked-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp> # pwm-visconti
Acked-by: Heiko Stuebner <heiko@sntech.de> # pwm-rockchip
Acked-by: Michael Walle <michael@walle.cc> # pwm-sl28cpld
Acked-by: Neil Armstrong <neil.armstrong@linaro.org> # pwm-meson
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230804142707.412137-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: Explicitly include correct DT includes

The DT of_device.h and of_platform.h date back to the separate
of_platform_bus_type before it as merged into the regular platform bus.
As part of that merge prepping Arm DT support 13 years ago, they
"temporarily" include each other. They also include platform_device.h
and of.h. As a result, there's a pretty much random mix of those include
files used throughout the tree. In order to detangle these headers and
replace the implicit includes with struct declarations, users need to
explicitly include the correct includes.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Consistently name pwm_chip variables "chip"

Most variables holding a pointer to a pwm_chip are called "chip" which
is also the usual name in most other PWM drivers. Rename the single
variable that have a different name to be called "chip", too, for
consistency.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Simplify using devm_clk_get_prepared()

Instead of preparing the clk after it was requested and unpreparing in
.probe()'s error path and .remove(), use devm_clk_get_prepared() which
copes for unpreparing automatically.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Always let the first pwm_apply_state succeed

Commit 2cfe9bbec56ea579135cdd92409fff371841904f added support for the
RGB and green PWM controlled LEDs on the HiFive Unmatched board
managed by the leds-pwm-multicolor and leds-pwm drivers respectively.
All three colours of the RGB LED and the green LED run from different
lines of the same PWM, but with the same period so this works fine when
the LED drivers are loaded one after the other.

Unfortunately it does expose a race in the PWM driver when both LED
drivers are loaded at roughly the same time. Here is an example:

| Thread A | Thread B |
| led_pwm_mc_probe | led_pwm_probe |
| devm_fwnode_pwm_get | |
| pwm_sifive_request | |
| ddata->user_count++ | |
| | devm_fwnode_pwm_get |
| | pwm_sifive_request |
| | ddata->user_count++ |
| ... | ... |
| pwm_state_apply | pwm_state_apply |
| pwm_sifive_apply | pwm_sifive_apply |

Now both calls to pwm_sifive_apply will see that ddata->approx_period,
initially 0, is different from the requested period and the clock needs
to be updated. But since ddata->user_count >= 2 both calls will fail
with -EBUSY, which will then cause both LED drivers to fail to probe.

Fix it by letting the first call to pwm_sifive_apply update the clock
even when ddata->user_count != 1.

Fixes: 9e37a53eb051 ("pwm: sifive: Add a driver for SiFive SoC PWM")
Signed-off-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: Make .get_state() callback return an error code

.get_state() might fail in some cases. To make it possible that a driver
signals such a failure change the prototype of .get_state() to return an
error code.

This patch was created using coccinelle and the following semantic patch:

@p1@
identifier getstatefunc;
identifier driver;
@@
struct pwm_ops driver = {
...,
.get_state = getstatefunc
,...
};

@p2@
identifier p1.getstatefunc;
identifier chip, pwm, state;
@@
-void
+int
getstatefunc(struct pwm_chip *chip, struct pwm_device *pwm, struct pwm_state *state)
{
...
- return;
+ return 0;
...
}

plus the actual change of the prototype in include/linux/pwm.h (plus some
manual fixing of indentions and empty lines).

So for now all drivers return success unconditionally. They are adapted
in the following patches to make the changes easier reviewable.

Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Reviewed-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Acked-by: Douglas Anderson <dianders@chromium.org>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20221130152148.2769768-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Call pwm_sifive_update_clock() while mutex is held

As was documented in commit 0f02f491b786 ("pwm: sifive: Reduce time the
controller lock is held") a caller of pwm_sifive_update_clock() must
hold the mutex. So fix pwm_sifive_clock_notifier() to grab the lock.

While this necessity was only documented later, the race exists since
the driver was introduced.

Fixes: 9e37a53eb051 ("pwm: sifive: Add a driver for SiFive SoC PWM")
Reported-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Reviewed-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Link: https://lore.kernel.org/r/20221018061656.1428111-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Shut down hardware only after pwmchip_remove() completed

The PWMs are expected to be functional until pwmchip_remove() is called.
So disable the clks only afterwards.

Fixes: 9e37a53eb051 ("pwm: sifive: Add a driver for SiFive SoC PWM")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Ensure the clk is enabled exactly once per running PWM

.apply() assumes the clk to be for a given PWM iff the PWM is enabled.
So make sure this is the case when .probe() completes. And in .remove()
disable the according number of times.

This fixes a clk enable/disable imbalance, if some PWMs are already running
at probe time.

Fixes: 9e37a53eb051 (pwm: sifive: Add a driver for SiFive SoC PWM)
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Simplify clk handling

The clk is necessary for both register access and (enabled) operation of
the PWM. Instead of

clk_enable()
update_hw()
if pwm_got_enabled():
clk_enable()
elif pwm_got_disabled():
clk_disable()
clk_disable()

which is some cases only calls clk_enable() to immediately afterwards
call clk_disable again, do:

if (!prev_state.enabled)
clk_enable()

# clk enabled exactly once

update_hw()

if (!next_state.enabled)
clk_disable()

which is much easier.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Enable clk only after period check in .apply()

For the period check and the initial calculations of register values there
is no hardware access needed. So delay enabling the clk a bit to simplify
the code flow a bit.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Reduce time the controller lock is held

The lock is only to serialize access and update to user_count and
approx_period between different PWMs served by the same pwm_chip.
So the lock needs only to be taken during the check if the (chip global)
period can and/or needs to be changed.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Fold pwm_sifive_enable() into its only caller

There is only a single caller of pwm_sifive_enable() which only enables
or disables the clk. Put this implementation directly into
pwm_sifive_apply() which allows further simplification in the next
change.

There is no change in behaviour.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Simplify offset calculation for PWMCMP registers

Instead of explicitly using PWM_SIFIVE_PWMCMP0 + pwm->hwpwm *
PWM_SIFIVE_SIZE_PWMCMP for each access to one of the PWMCMP registers,
introduce a macro that takes the hwpwm id as parameter.

For the register definition using a plain 4 instead of the cpp constant
PWM_SIFIVE_SIZE_PWMCMP is easier to read, so define the offset macro
without the constant. The latter can then be dropped as there are no
users left.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Emil Renner Berthing <emil.renner.berthing@canonical.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Simplify if-if to if-else

Use if and else instead of if(A) and if (!A).

Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Don't check the return code of pwmchip_remove()

pwmchip_remove() returns always 0. Don't use the value to make it
possible to eventually change the function to return void. Also the
driver core ignores the return value of pwm_sifive_remove().

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: Simplify all drivers with explicit of_pwm_n_cells = 3

With the previous commit there is no need for the lowlevel driver any
more to specify it it uses two or three cells. So simplify accordingly.

The only non-trival change affects the pwm-rockchip driver: It used to only
support three cells if the hardware supports polarity. Now the default
number depends on the device tree which has to match hardware anyhow
(and if it doesn't the error is just a bit delayed as a PWM handle with
an inverted setting is catched when pwm_apply_state() is called).

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: Always allocate PWM chip base ID dynamically

Since commit 5e5da1e9fbee ("pwm: ab8500: Explicitly allocate pwm chip
base dynamically") all drivers use dynamic ID allocation explicitly. New
drivers are supposed to do the same, so remove support for driver
specified base IDs and drop all assignments in the low-level drivers.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Convert to devm_platform_ioremap_resource()

Use devm_platform_ioremap_resource() to simplify code.

Signed-off-by: Yangtao Li <tiny.windzz@gmail.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Simplify with dev_err_probe()

Common pattern of handling deferred probe can be simplified with
dev_err_probe(). Less code and also it prints the error value.

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Use 64-bit division macro

Since the PWM framework is switching struct pwm_args.period's datatype
to u64, prepare for this transition by using DIV64_U64_ROUND_CLOSEST to
handle a 64-bit divisor.

Signed-off-by: Guru Das Srinagesh <gurus@codeaurora.org>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: Ensure pwm_apply_state() doesn't modify the state argument

It is surprising for a PWM consumer when the variable holding the
requested state is modified by pwm_apply_state(). Consider for example a
driver doing:

#define PERIOD 5000000
#define DUTY_LITTLE 10
...
struct pwm_state state = {
.period = PERIOD,
.duty_cycle = DUTY_LITTLE,
.polarity = PWM_POLARITY_NORMAL,
.enabled = true,
};

pwm_apply_state(mypwm, &state);
...
state.duty_cycle = PERIOD / 2;
pwm_apply_state(mypwm, &state);

For sure the second call to pwm_apply_state() should still have
state.period = PERIOD and not something the hardware driver chose for a
reason that doesn't necessarily apply to the second call.

So declare the state argument as a pointer to a const type and adapt all
drivers' .apply callbacks.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Remove redundant error message

devm_ioremap_resource() already outputs an error message, so remove the
extra error message on failure.

Signed-off-by: Ding Xiang <dingxiang@cmss.chinamobile.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>

pwm: sifive: Add a driver for SiFive SoC PWM

Adds a PWM driver for PWM chip present in SiFive's HiFive Unleashed SoC.

Signed-off-by: Wesley W. Terpstra <wesley@sifive.com>
[Atish: Various fixes and code cleanup]
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Signed-off-by: Yash Shah <yash.shah@sifive.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>