The function pwm_apply_args() was introduced with the concept of atomic
PWM configuration and needed for drivers not using this concept yet. Now
all drivers are converted accordingly and so no callers are left which
allows to remove this function.
Signed-off-by: Uwe Kleine-König <[email protected]>
Link: https://patch.msgid.link/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
The upcoming Rust abstraction layer for the PWM subsystem uses a custom
`dev->release` handler to safely manage the lifetime of its driver
data.
To prevent leaking the memory of the `struct pwm_chip` (allocated by
`pwmchip_alloc`), this custom handler must also call the original
`pwmchip_release` function to complete the cleanup.
Make `pwmchip_release` a global, exported function so that it can be
called from the Rust FFI bridge. This involves removing the `static`
keyword, adding a prototype to the public header, and exporting the
symbol.
Reviewed-by: Elle Rhumsaa <[email protected]>
Signed-off-by: Michal Wilczynski <[email protected]>
Link: https://patch.msgid.link/20251016-rust-next-pwm-working-fan-for-sending-v16-1-a5df2405d2bd@samsung.com
Signed-off-by: Uwe Kleine-König <[email protected]>
A PWM is a more general concept than an output-only GPIO. When using
duty_length = period_length the PWM looks like an active GPIO, with
duty_length = 0 like an inactive GPIO. With the waveform abstraction
there is enough control over the configuration to ensure that PWMs that
cannot generate a constant signal at both levels error out.
The pwm-pca9685 driver already provides a gpio chip. When this driver is
converted to the waveform callbacks, the gpio part can just be dropped.
Signed-off-by: Uwe Kleine-König <[email protected]>
Reviewed-by: Bartosz Golaszewski <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
The WFHWSIZE constant defines the maximum size for the hardware-specific
waveform representation buffer. It is currently local to
drivers/pwm/core.c, which makes it inaccessible to external tools like
bindgen.
Move the constant to include/linux/pwm.h to make it part of the public
API. As part of this change, rename it to PWM_WFHWSIZE to follow
standard kernel conventions for namespacing macros in public headers.
This allows bindgen to automatically generate a corresponding constant
for the Rust PWM abstractions, ensuring the value remains synchronized
between the C core and Rust code and preventing future maintenance
issues.
Signed-off-by: Michal Wilczynski <[email protected]>
Link: https://lore.kernel.org/r/20250702-rust-next-pwm-working-fan-for-sending-v7-1-67ef39ff1d29@samsung.com
Signed-off-by: Uwe Kleine-König <[email protected]>
With this change each pwmchip defining the new-style waveform callbacks
can be accessed from userspace via a character device. Compared to the
sysfs-API this is faster and allows to pass the whole configuration in a
single ioctl allowing atomic application and thus reducing glitches.
On an STM32MP13 I see:
root@DistroKit:~ time pwmtestperf
real 0m 1.27s
user 0m 0.02s
sys 0m 1.21s
root@DistroKit:~ rm /dev/pwmchip0
root@DistroKit:~ time pwmtestperf
real 0m 3.61s
user 0m 0.27s
sys 0m 3.26s
pwmtestperf does essentially:
for i in 0 .. 50000:
pwm_set_waveform(duty_length_ns=i, period_length_ns=50000, duty_offset_ns=0)
and in the presence of /dev/pwmchip0 is uses the ioctls introduced here,
without that device it uses /sys/class/pwm/pwmchip0.
Signed-off-by: Uwe Kleine-König <[email protected]>
Link: https://lore.kernel.org/r/ad4a4e49ae3f8ea81e23cac1ac12b338c3bf5c5b.1746010245.git.u.kleine-koenig@baylibre.com
Signed-off-by: Uwe Kleine-König <[email protected]>
Preparing CONFIG_PWM becoming tristate the right magic to check for the
availability of the pwm functions is using IS_REACHABLE() and not
IS_ENABLED(). The latter gives the wrong result for built-in code with
CONFIG_PWM=m.
Signed-off-by: Uwe Kleine-König <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
Pull pwm fixes from Uwe Kleine-König:
"Two fixes.
Conor Dooley found and fixed a problem in the pwm-microchip-core
driver that existed since the driver's birth in v6.5-rc1. It's about a
corner case that only happens if two pwm devices of the same chip are
set to the same long period.
The other problem is about the new pwm API that currently is only
supported by two hardware drivers. The fix prevents a NULL pointer
exception if one of the new functions is called for a pwm device with
a driver that only provides the old callbacks"
* tag 'pwm/for-6.14-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux:
pwm: Ensure callbacks exist before calling them
pwm: microchip-core: fix incorrect comparison with max period
Export the pwm_get_state_hw() function. This is useful in cases where
we want to know what the hardware is actually doing, rather than what
what we requested it should do.
Locking had to be rearranged to ensure that the chip is still
operational before trying to access ops now that this can be called
from outside the pwm core.
Signed-off-by: David Lechner <[email protected]>
Link: https://lore.kernel.org/r/20241029-pwm-export-pwm_get_state_hw-v2-1-03ba063a3230@baylibre.com
[ukleinek: Add dummy for !CONFIG_PWM]
Signed-off-by: Uwe Kleine-König <[email protected]>
Up to now the configuration of a PWM setting is described exclusively by
a struct pwm_state which contains information about period, duty_cycle,
polarity and if the PWM is enabled. (There is another member usage_power
which doesn't completely fit into pwm_state, I ignore it here for
simplicity.)
Instead of a polarity the new abstraction has a member duty_offset_ns
that defines when the rising edge happens after the period start. This
is more general, as with a pwm_state the rising edge can only happen at
the period's start or such that the falling edge is at the end of the
period (i.e. duty_offset_ns == 0 or duty_offset_ns == period_length_ns -
duty_length_ns).
A disabled PWM is modeled by .period_length_ns = 0. In my eyes this is a
nice usage of that otherwise unusable setting, as it doesn't define
anything about the future which matches the fact that consumers should
consider the state of the output as undefined and it's just there to say
"No further requirements about the output, you can save some power.".
Further I renamed period and duty_cycle to period_length_ns and
duty_length_ns. In the past there was confusion from time to time about
duty_cycle being measured in nanoseconds because people expected a
percentage of period instead. With "length_ns" as suffix the semantic
should be more obvious to people unfamiliar with the pwm subsystem.
period is renamed to period_length_ns for consistency.
The API for consumers doesn't change yet, but lowlevel drivers can
implement callbacks that work with pwm_waveforms instead of pwm_states.
A new thing about these callbacks is that the calculation of hardware
settings needed to implement a certain waveform is separated from
actually writing these settings. The motivation for that is that this
allows a consumer to query the hardware capabilities without actually
modifying the hardware state.
The rounding rules that are expected to be implemented in the
round_waveform_tohw() are: First pick the biggest possible period not
bigger than wf->period_length_ns. For that period pick the biggest
possible duty setting not bigger than wf->duty_length_ns. Third pick the
biggest possible offset not bigger than wf->duty_offset_ns. If the
requested period is too small for the hardware, it's expected that a
setting with the minimal period and duty_length_ns = duty_offset_ns = 0
is returned and this fact is signaled by a return value of 1.
Signed-off-by: Uwe Kleine-König <[email protected]>
Tested-by: Trevor Gamblin <[email protected]>
Link: https://lore.kernel.org/r/df0faa33bf9e7c9e2e5eab8d31bbf61e861bd401.1726819463.git.u.kleine-koenig@baylibre.com
[ukleinek: Update pwm_check_rounding() to return bool instead of int.]
Signed-off-by: Uwe Kleine-König <[email protected]>
This ensures that a pwm_chip that has no corresponding driver isn't used
and that a driver doesn't go away while a callback is still running.
In the presence of device links this isn't necessary yet (so this is no
fix) but for pwm character device support this is needed.
To not serialize all pwm_apply_state() calls, this introduces a per chip
lock. An additional complication is that for atomic chips a mutex cannot
be used (as pwm_apply_atomic() must not sleep) and a spinlock cannot be
held while calling an operation for a sleeping chip. So depending on the
chip being atomic or not a spinlock or a mutex is used.
An additional change implemented here is that on driver remove the
.free() callback is called for each requested pwm_device. This is the
right time because later (e.g. when the consumer calls pwm_put()) the
free function is (maybe) not available any more.
Signed-off-by: Uwe Kleine-König <[email protected]>
Link: https://lore.kernel.org/r/026aa891c8270a11723a1ba7e4256f456f7e1e86.1726819463.git.u.kleine-koenig@baylibre.com
Signed-off-by: Uwe Kleine-König <[email protected]>
This function is not supposed to be used any more since commit
c748a6d77c ("pwm: Rename pwm_apply_state() to
pwm_apply_might_sleep()") that is included in v6.8-rc1. Two kernel
releases should be enough for everyone to adapt, so drop the old
function that was introduced as a compatibility stub for the transition.
Signed-off-by: Uwe Kleine-König <[email protected]>
Define all pwm core's symbols in the namespace "PWM". The necessary
module import statement is just added to the main header, this way every
file that knows about the public functions automatically has this
namespace available.
Thanks to Biju Das for pointing out a cut'n'paste failure in my initial
patch.
Signed-off-by: Uwe Kleine-König <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
It's required to not free the memory underlying a requested PWM
while a consumer still has a reference to it. While currently a pwm_chip
doesn't live long enough in all cases, linking the struct pwm to the
pwm_chip results in the right lifetime as soon as the pwmchip is living
long enough. This happens with the following commits.
Note this is a breaking change for all pwm drivers that don't use
pwmchip_alloc().
Reviewed-by: Gustavo A. R. Silva <[email protected]> # for struct_size() and __counted_by()
Link: https://lore.kernel.org/r/7e9e958841f049026c0023b309cc9deecf0ab61d.1710670958.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <[email protected]>
With the upcoming restructuring having all in a single file simplifies
things a bit. The relevant and somewhat visible changes are:
- Some dropped prototypes from include/linux/pwm.h that were only
necessary that core.c has a declaration of the symbols defined in
sysfs.c. The respective functions are static now.
- The pwm class now also exists if CONFIG_SYSFS isn't enabled. Having
CONFIG_SYSFS is not very relevant today, but even without it the
class and device stuff still provides lifetime tracking.
- Both files had an initcall, these are merged into a single one now.
Instead of a big #ifdef block for CONFIG_DEBUG_FS, a single
if (IS_ENABLED(CONFIG_DEBUG_FS)) is used now. This increases compile
coverage a bit and is a tad nicer on the eyes.
Link: https://lore.kernel.org/r/9e2d39a5280d7dda5bfc6682a8aef510148635b2.1710670958.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <[email protected]>
This function allocates a struct pwm_chip and driver data. Compared to
the status quo the split into pwm_chip and driver data is new, otherwise
it doesn't change anything relevant (yet).
The intention is that after all drivers are switched to use this
allocation function, its possible to add a struct device to struct
pwm_chip to properly track the latter's lifetime without touching all
drivers again. Proper lifetime tracking is a necessary precondition to
introduce character device support for PWMs (that implements atomic
setting and doesn't suffer from the sysfs overhead of the /sys/class/pwm
userspace support).
The new function pwmchip_priv() (obviously?) only works for chips
allocated with pwmchip_alloc().
Link: https://lore.kernel.org/r/9577d6053a5a52536057dc8654ff567181c2da82.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <[email protected]>
Currently a pwm_chip stores in its struct device *dev member a pointer
to the parent device. Preparing a change that embeds a full struct
device in struct pwm_chip, this accessor function should be used in all
drivers directly accessing chip->dev now. This way struct pwm_chip and
this new function can be changed without having to touch all drivers in
the same change set.
Make use of this function in the framework's core sources.
Reviewed-by: Greg Kroah-Hartman <[email protected]>
Link: https://lore.kernel.org/r/cc30090d2f9762bed9854a55612144bccc910781.1707900770.git.u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <[email protected]>
In order to make the transition to the new pwm_apply_might_sleep() a bit
smoother, add a compatibility stub. This will prevent new calls to the
old function introduced via other subsystems from breaking builds. Once
the next merge window has closed we can take another stab at removing
the stub.
Reviewed-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
These functions are unused. Also I think there is no valid use case
where these are correct to be called. So drop them.
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Some PWM devices require sleeping, for example if the pwm device is
connected over I2C. However, many PWM devices could be used from atomic
context, e.g. memory mapped PWM. This is useful for, for example, the
pwm-ir-tx driver which requires precise timing. Sleeping causes havoc
with the generated IR signal.
Since not all PWM devices can support atomic context, we also add a
pwm_might_sleep() function to check if is not supported.
Signed-off-by: Sean Young <[email protected]>
Reviewed-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
According to Documentation/dev-tools/checkpatch.rst ENOTSUPP is
not recommended and EOPNOTSUPP should be used instead.
Signed-off-by: Sean Young <[email protected]>
Acked-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
In order to introduce a pwm api which can be used from atomic context,
we will need two functions for applying pwm changes:
int pwm_apply_might_sleep(struct pwm *, struct pwm_state *);
int pwm_apply_atomic(struct pwm *, struct pwm_state *);
This commit just deals with renaming pwm_apply_state(), a following
commit will introduce the pwm_apply_atomic() function.
Acked-by: Uwe Kleine-König <[email protected]>
Acked-by: Guenter Roeck <[email protected]>
Acked-by: Mark Brown <[email protected]>
Acked-by: Dmitry Torokhov <[email protected]> # for input
Acked-by: Hans de Goede <[email protected]>
Acked-by: Jani Nikula <[email protected]>
Acked-by: Lee Jones <[email protected]>
Signed-off-by: Sean Young <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Commit c572f3b9c8b7 ("pwm: Replace PWM chip unique base by unique ID")
changed the members of struct pwm_chip, but failed to update the
documentation accordingly. Catch up and document the new member and drop
description for the two removed ones.
Reported-by: Stephen Rothwell <[email protected]>
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Traditionally each PWM device had a unique ID stored in the "pwm" member
of struct pwm_device. However this number was hardly used and dropped
in the previous commit. To identify a certain PWM you're supposed to use
the chip's ID and the hwpwm of the PWM device now.
With the PWM chip base gone PWM chips can get their IDs better and
simpler using an idr.
This is expected to change the numbering of PWM chips, but nothing
should rely on the numbering anyhow.
Other than that the side effects are:
- The PWM chip IDs are smaller and in most cases consecutive.
- The ordering in /sys/kernel/debug/pwm is ordered by ascending PWM
chip ID.
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
This member is only assigned to and never read. So drop it.
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
These enums are passed to set/test_bit(). The set/test_bit() functions
take a bit number instead of a shifted value. Passing a shifted value
is a double shift bug like doing BIT(BIT(1)). The double shift bug
doesn't cause a problem here because we are only checking 0 and 1 but
if the value was 5 or above then it can lead to a buffer overflow.
Signed-off-by: Dan Carpenter <[email protected]>
Reviewed-by: Uwe Kleine-König <[email protected]>
Reviewed-by: Sam Protsenko <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
The semantic of chip_data is a bit surprising as it's cleared when
pwm_put() is called. Also there is a big overlap with the standard driver
data.
All drivers were adapted to not make use of chip_data any more, so it can
go away.
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
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 <[email protected]> # Intel LPSS
Reviewed-by: Florian Fainelli <[email protected]> # pwm-{bcm,brcm}*.c
Acked-by: Jernej Skrabec <[email protected]> # sun4i
Acked-by: Andi Shyti <[email protected]>
Acked-by: Nobuhiro Iwamatsu <[email protected]> # pwm-visconti
Acked-by: Heiko Stuebner <[email protected]> # pwm-rockchip
Acked-by: Michael Walle <[email protected]> # pwm-sl28cpld
Acked-by: Neil Armstrong <[email protected]> # pwm-meson
Reviewed-by: Linus Walleij <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Most variables of type struct pwm_chip * are named "chip", there are
only three outliers called "pc". Change these three to "chip", too, for
consistency.
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Since commit 5a7fbe452a ("backlight: pwm_bl: Drop support for legacy PWM
probing") the last user of pwm_request() and pwm_free() is gone. So remove
these functions that were deprecated over 10 years ago in commit
8138d2ddbc ("pwm: Add table-based lookup for static mappings").
Signed-off-by: Uwe Kleine-König <[email protected]>
[[email protected]: clean up a bit after removal]
Signed-off-by: Thierry Reding <[email protected]>
Move the dummy pwm_capture(), to make the declaration order of all
dummies to match the declaration order of the real functions.
Signed-off-by: Geert Uytterhoeven <[email protected]>
Acked-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
Pull pwm updates from Thierry Reding:
"Various changes across the board, mostly improvements and cleanups"
* tag 'pwm/for-6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm: (42 commits)
pwm: pca9685: Convert to i2c's .probe_new()
pwm: sun4i: Propagate errors in .get_state() to the caller
pwm: Handle .get_state() failures
pwm: sprd: Propagate errors in .get_state() to the caller
pwm: rockchip: Propagate errors in .get_state() to the caller
pwm: mtk-disp: Propagate errors in .get_state() to the caller
pwm: imx27: Propagate errors in .get_state() to the caller
pwm: cros-ec: Propagate errors in .get_state() to the caller
pwm: crc: Propagate errors in .get_state() to the caller
leds: qcom-lpg: Propagate errors in .get_state() to the caller
drm/bridge: ti-sn65dsi86: Propagate errors in .get_state() to the caller
pwm/tracing: Also record trace events for failed API calls
pwm: Make .get_state() callback return an error code
pwm: pxa: Enable for MMP platform
pwm: pxa: Add reference manual link and limitations
pwm: pxa: Use abrupt shutdown mode
pwm: pxa: Remove clk enable/disable from pxa_pwm_config
pwm: pxa: Set duty cycle to 0 when disabling PWM
pwm: pxa: Remove pxa_pwm_enable/disable
pwm: mediatek: Add support for MT7986
...
.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 <[email protected]>
Reviewed-by: Baolin Wang <[email protected]>
Reviewed-by: Tzung-Bi Shih <[email protected]>
Reviewed-by: Neil Armstrong <[email protected]>
Reviewed-by: Nobuhiro Iwamatsu <[email protected]>
Reviewed-by: Andre Przywara <[email protected]>
Reviewed-by: Dave Stevenson <[email protected]>
Acked-by: Douglas Anderson <[email protected]>
Acked-by: Jernej Skrabec <[email protected]>
Acked-by: Pavel Machek <[email protected]>
Acked-by: Conor Dooley <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Thierry Reding <[email protected]>
The devm_pwmchip_add() can be called by a module that optionally
instantiates PWM chip. In the case of CONFIG_PWM=n, the compilation
can't be performed. Hence, add a necessary stub.
Signed-off-by: Andy Shevchenko <[email protected]>
Acked-by: Thierry Reding <[email protected]>
Reviewed-by: Mika Westerberg <[email protected]>
Acked-by: Uwe Kleine-König <[email protected]>
Uwe Kleine-König