We currently only notify user-space about line config changes that are
made from user-space. Any kernel config changes are not signalled.
Let's improve the situation by emitting the events closer to the source.
To that end let's call the relevant notifier chain from the functions
setting direction, gpiod_set_config(), gpiod_set_consumer_name() and
gpiod_toggle_active_low(). This covers all the options that we can
inform the user-space about. We ignore events which don't have
corresponding flags exported to user-space on purpose - otherwise the
user would see a config-changed event but the associated line-info would
remain unchanged.
gpiod_direction_output/input() can be called from any context.
Fortunately, we now emit line state events using an atomic notifier
chain, so it's no longer an issue.
Let's also add non-notifying wrappers around the direction setters in
order to not emit superfluous reconfigure events when requesting the
lines as the initial config should be part of the request notification.
Use gpio_do_set_config() instead of gpiod_set_debounce() for configuring
debouncing via hardware from the character device code to avoid multiple
reconfigure events.
Reviewed-by: Kent Gibson <warthog618@gmail.com>
Link: https://lore.kernel.org/r/20241018-gpio-notify-in-kernel-events-v5-8-c79135e58a1c@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
In order to allow line state notifications to be emitted from atomic
context (for instance: from gpiod_direction_input/output()), we must
stop calling any sleeping functions in lineinfo_changed_notify(). To
that end let's use the new workqueue.
Let's atomically allocate small structures containing the required data
and fill it with information immediately upon being notified about the
change except for the pinctrl state which will be retrieved later from
process context. We can pretty reliably do this as pin functions are
typically set once per boot.
Let's make sure to bump the reference count of GPIO device and the GPIO
character device file descriptor to keep both alive until the event was
queued.
Reviewed-by: Kent Gibson <warthog618@gmail.com>
Link: https://lore.kernel.org/r/20241018-gpio-notify-in-kernel-events-v5-6-c79135e58a1c@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
We don't need to guard the GPIO chip until its first dereference in
gpio_do_set_config().
First: change the prototype of gpio_do_set_config() to take the GPIO
line descriptor as argument, then move the gpio_chip protection into it
and drop it in two places where it's done too early.
This has the added benefit of making gpio_go_set_config() safe to use
from outside of this compilation unit without taking the gdev SRCU read
lock and will come in handy when we'll want to make it available to the
character device code.
Reviewed-by: Kent Gibson <warthog618@gmail.com>
Link: https://lore.kernel.org/r/20241018-gpio-notify-in-kernel-events-v5-2-c79135e58a1c@linaro.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
After commit 0edb555a65 ("platform: Make platform_driver::remove()
return void") .remove() is (again) the right callback to implement for
platform drivers.
Convert all platform drivers below drivers/gpio/ to use .remove(), with
the eventual goal to drop struct platform_driver::remove_new(). As
.remove() and .remove_new() have the same prototypes, conversion is done
by just changing the structure member name in the driver initializer.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Link: https://lore.kernel.org/r/20241016152435.38149-2-u.kleine-koenig@baylibre.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Use devres to drop all goto labels from probe() and remove the driver
remove() callback. While at it: drop the unnecessary dev_info() message
as not only should the driver be quiet when successful, the message is
also wrong: the device was probed at this point, the driver had been
registered earlier.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20241014092227.78886-1-brgl@bgdev.pl
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
FTDI FT2232H is a USB to GPIO chip. Sealevel produces some devices
with this chip. FT2232H presents itself as a composite device with two
interfaces (each is an "MPSSE"). Each MPSSE has two banks (high and low)
of 8 GPIO each. I believe some MPSSE's have only one bank, but I don't
know how to identify them (I don't have any for testing) and as a result
are unsupported for the time being.
Additionally, this driver provides software polling-based interrupts for
edge detection. For the Sealevel device I have to test with, this works
well because there is hardware debouncing. From talking to Sealevel's
people, this is their preferred way to do edge detection.
Signed-off-by: Mary Strodl <mstrodl@csh.rit.edu>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20241009131131.1618329-1-mstrodl@csh.rit.edu
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
In the 7th generation of the SoC from Aspeed, the control logic of the
GPIO controller has been updated to support per-pin control. Each pin now
has its own 32-bit register, allowing for individual control of the pin's
value, direction, interrupt type, and other settings. The permission for
coprocessor access is supported by the hardware but hasn't been
implemented in the current patch.
Signed-off-by: Billy Tsai <billy_tsai@aspeedtech.com>
Reviewed-by: Andrew Jeffery <andrew@codeconstruct.com.au>
Link: https://lore.kernel.org/r/20241008081450.1490955-8-billy_tsai@aspeedtech.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Use module_platform_driver() to replace module_platform_driver_probe().
The former utilizes platform_driver_register(), which allows the driver to
defer probing when it doesn't acquire the necessary resources due to probe
order. In contrast, the latter uses __platform_driver_probe(), which
includes the comment "Note that this is incompatible with deferred
probing." Since our GPIO driver requires access to the clock resource, the
former is more suitable.
Reviewed-by: Andrew Jeffery <andrew@codeconstruct.com.au>
Signed-off-by: Billy Tsai <billy_tsai@aspeedtech.com>
Link: https://lore.kernel.org/r/20241008081450.1490955-4-billy_tsai@aspeedtech.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Performing a dummy read ensures that the register write operation is fully
completed, mitigating any potential bus delays that could otherwise impact
the frequency of bitbang usage. E.g., if the JTAG application uses GPIO to
control the JTAG pins (TCK, TMS, TDI, TDO, and TRST), and the application
sets the TCK clock to 1 MHz, the GPIO's high/low transitions will rely on
a delay function to ensure the clock frequency does not exceed 1 MHz.
However, this can lead to rapid toggling of the GPIO because the write
operation is POSTed and does not wait for a bus acknowledgment.
Fixes: 361b79119a ("gpio: Add Aspeed driver")
Reviewed-by: Andrew Jeffery <andrew@codeconstruct.com.au>
Signed-off-by: Billy Tsai <billy_tsai@aspeedtech.com>
Link: https://lore.kernel.org/r/20241008081450.1490955-2-billy_tsai@aspeedtech.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
In `gpiod_get_label()`, it is possible that `srcu_dereference_check()` may
return a NULL pointer, leading to a scenario where `label->str` is accessed
without verifying if `label` itself is NULL.
This patch adds a proper NULL check for `label` before accessing
`label->str`. The check for `label->str != NULL` is removed because
`label->str` can never be NULL if `label` is not NULL.
This fixes the issue where the label name was being printed as `(efault)`
when dumping the sysfs GPIO file when `label == NULL`.
Fixes: 5a646e03e9 ("gpiolib: Return label, if set, for IRQ only line")
Fixes: a86d276930 ("gpiolib: fix the speed of descriptor label setting with SRCU")
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Link: https://lore.kernel.org/r/20241003131351.472015-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
For drivers or board files that set gpio_chip->names, the links to the
GPIO attribute group created on sysfs export will be named after the
line's name set in that array. For lines that are named using device
properties, the names pointer of the gpio_chip struct is never assigned
so they are exported as if they're not named.
The ABI documentation does not mention the former behavior and given
that the majority of modern systems use device-tree, ACPI or other way
of passing GPIO names using device properties - bypassing gc->names -
it's better to make the behavior consistent by always exporting lines as
"gpioXYZ".
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20240930083029.17694-1-brgl@bgdev.pl
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
