Now pcf857x_irq() IRQ's dispatcher will try to run nested
IRQ handlers for each GPIO pin which state has changed.
Such IRQs are, actually, spurious and nested IRQ handlers
have to be called only for IRQs wich were enabled by users.
This is not critical issue - just  /proc/interrupts
will display counters for unused IRQS:
399:          4          0   pcf857x   0 Edge
428:          1          0   pcf857x  13 Edge
430:          1          0   pcf857x  15 Edge

Hence, fix it by adding irq_enabled field in struct pcf857x to track
enabled GPIO IRQs and corresponding callbacks in pcf857x_irq_chip.

Similar functionality was presented in pcf857x driver, commit
21fd3cd1 ('gpio: pcf857x: call the gpio user handler iff...')

and then it was removed by commit
 a39294bd ('gpio: pcf857x: Switch to use gpiolib irqchip...')

Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Fixes: a39294bd ('gpio: pcf857x: Switch to use gpiolib irqchip helpers')
Signed-off-by: NGrygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

If a driver requests a GPIO described in its _CRS but the GPIO host
controller (gpiochip) driver providing the GPIO has not been loaded yet
acpi_get_gpiod() returns -ENODEV which causes the calling driver to fail.

If the gpiochip driver is loaded afterwards the driver requesting the GPIO
will not notice this.

Better approach is to return -EPROBE_DEFER in such case. Then when the
gpiochip driver appears the driver requesting the GPIO will be probed
again. This also aligns ACPI GPIO lookup code closer to DT as it does
pretty much the same when no gpiochip driver was found.
Reported-by: NTobias Diedrich <tobiasdiedrich@gmail.com>
Signed-off-by: NMika Westerberg <mika.westerberg@linux.intel.com>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: NTobias Diedrich <ranma+kernel@tdiedrich.de>
Reviewed-by: Amos Kong <kongjianjun@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

acpi_gpiochip_request(free)_interrupts can be used for modules,
so export them. This also fixs a compile error when xgene-sb
configured as kernel module.

Fixes: 733cf014 "gpio: xgene: add ACPI support for APM X-Gene GPIO standby driver"
Reviewed-by: NMark Brown <broonie@kernel.org>
Signed-off-by: NHanjun Guo <hanjun.guo@linaro.org>

When requesting own descriptors through hogs, it is useful to
get some details about what's going on if we encounter problems.
Acked-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

These error messages are helpful to see that we fail to get
hogs. Promote them to real errors so they appear in the boot
crawl.
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Added support to Zynq Ultrascale+ MPSoC on the existing zynq
gpio driver.
Signed-off-by: NAnurag Kumar Vulisha <anuragku@xilinx.com>
Acked-by: NMichal Simek <michal.simek@xilinx.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add a GPIO driver for the General I/O block on Axis ETRAX FS SoCs.
Signed-off-by: NRabin Vincent <rabin@rab.in>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

This adds support for the GPIO IP "UPG GIO" used on
Broadcom STB SoCs (BCM7XXX and some others). Uses
basic_mmio_gpio to instantiate a gpio_chip for each bank.
The driver assumes that it handles the base set of GPIOs
on the system and that it can start its numbering sequence
from 0, so any GPIO expanders used with it must dynamically
assign GPIO numbers after this driver has finished
registering its GPIOs.

Does not implement the interrupt-controller portion yet,
will be done in a future commit.

v2:
- change include to use <linux/gpio/driver.h> instead of
  <linux/gpio.h>
- get rid of unnecessary imask member in struct bank
- rename GPIO_PER_BANK -> MAX_GPIO_PER_BANK
- always have 32 GPIOs per bank and add 'width' member in
  struct bank to hold actual number of GPIOs in use
- mark of_match table as const
List-usage-fixed-by: NBrian Norris <computersforpeace@gmail.com>
Signed-off-by: NGregory Fong <gregory.0xf0@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add ACPI support for APM X-Gene GPIO standby driver.
Signed-off-by: NY Vo <yvo@apm.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Current code uses devm_request_irq() in .probe, so drop the unneeded
free_irq() call in .remove.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NChristian Ruppert <christian.ruppert@alitech.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The CrystalCove GPIO irqchip doesn't have irq_set_wake callback defined
so we should set IRQCHIP_SKIP_SET_WAKE for it or it would cause an irq
desc's wake_depth unbalanced warning during system resume phase from the
gpio_keys driver, which is the driver for the power button of the ASUS
T100 laptop.
Signed-off-by: NAaron Lu <aaron.lu@intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

This removes some redundant code but does not have any functional impact.
Signed-off-by: NMartin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

If an interrupt controller doesn't support wake-up configuration,
irq_set_irq_wake() returns an error code.  Then any subsequent call
trying to deconfigure wake-up will cause an imbalance, and a warning
will be printed:

    WARNING: CPU: 1 PID: 1341 at kernel/irq/manage.c:540 irq_set_irq_wake+0x
    Unbalanced IRQ 26 wake disable

To fix this, refrain from any further parent interrupt controller
(de)configuration if irq_set_irq_wake() failed.
Signed-off-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

0x3 only masks two bits, but three bits have to be allowed. This fixes
GPHY0 LED2 (which is the highest bit of phy2) on my board.
Signed-off-by: NMartin Blumenstingl <martin.blumenstingl@googlemail.com>
Acked-by: NJohn Crispin <blogic@openwrt.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Interrupts were missed if an 8-bit integer overflow occurred. This was
observed when bank0,pin7 and bank1,pin7 changed simultaniously.

As the 8-bit totals were only checked against zero, replace them with
booleans. Name the booleans so that their purpose is clear.
Signed-off-by: NJoshua Scott <joshua.scott@alliedtelesis.co.nz>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The omap_gpio_irq_startup() can be called at time when:
- corresponding GPIO has been requested already and in this case
it has to be configured as input already. If not - return with -EINVAL
and do not try to re-configure it as it could be unsafe.
- corresponding GPIO is free: reconfigure GPIO as input.

In addition, call omap_enable_gpio_module directly as all needed
checks are already present inside it.
Signed-off-by: NGrygorii Strashko <grygorii.strashko@linaro.org>
Tested-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The GPIO Chip and GPIO IRQ Chip functionality are essentially orthogonal,
so GPIO Chip implementation shouldn't touch GPIO IRQ specific registers
and vise versa.

Hence, rework omap_gpio_request:
- don't reset GPIO IRQ triggering type to IRQ_TYPE_NONE, because
  GPIO irqchip should be responsible for that;
- call directly omap_enable_gpio_module as all needed checks are already
  present inside it.
Signed-off-by: NGrygorii Strashko <grygorii.strashko@linaro.org>
Tested-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The GPIO Chip and GPIO IRQ Chip functionality are essentially orthogonal,
so GPIO IRQ Chip implementation shouldn't touch GPIO specific
registers and vise versa.

Hence, rework omap_gpio_irq_shutdown and try to touch only irqs specific
registers:
- don't configure GPIO as input (it, actually, should be already configured
  as input).
- don't clear debounce configuration if GPIO is still used as GPIO.
  We need to take in to account here commit c9c55d92
  ("gpio/omap: fix off-mode bug: clear debounce settings on free/reset").

Also remove omap_reset_gpio() function as it is not used any more.
Signed-off-by: NGrygorii Strashko <grygorii.strashko@linaro.org>
Tested-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The GPIO bank will be kept powered in case if input parameters
are invalid or error occurred in omap_gpio_irq_type.

Hence, fix it by ensuring that GPIO bank will be unpowered
in case of errors and add additional check of value returned
from omap_set_gpio_triggering().
Signed-off-by: NGrygorii Strashko <grygorii.strashko@linaro.org>
Tested-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

This patch fixes following issue:
- GPIOn is used as IRQ by some dev, for example PCF8575.INT ->  gpio6.11
- PCFx driver knows nothing about type of IRQ line (GPIO or not)
  so it doesn't request gpio and just do request_irq()
- If gpio6.11 will be exported through the sysfs and then un-xeported
then IRQs from PCFx will not be received any more, because
IRQ configuration for gpio6.11 will be cleaned up unconditionally
in omap_gpio_free.

Fix this by removing all GPIO IRQ specific code from omap_gpio_free()
and also do GPIO clean up (change direction to 'in' and disable debounce)
only if corresponding GPIO is not used as IRQ too.
GPIO IRQ will be properly cleaned up by GPIO irqchip code.
Signed-off-by: NGrygorii Strashko <grygorii.strashko@linaro.org>
Tested-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

If an interrupt controller doesn't support wake-up configuration,
irq_set_irq_wake() returns an error code.  Then any subsequent call
trying to deconfigure wake-up will cause an imbalance, and a warning
will be printed:

    WARNING: CPU: 1 PID: 1341 at kernel/irq/manage.c:540 irq_set_irq_wake+0x9c/0xf8()
    Unbalanced IRQ 26 wake disable

To fix this, refrain from any further parent interrupt controller
(de)configuration if irq_set_irq_wake() failed.

Alternative fixes would be:
  - calling "gic_set_irqchip_flags(IRQCHIP_SKIP_SET_WAKE)" from the
    platform code,
  - setting "gic_chip.flags = IRQCHIP_SKIP_SET_WAKE" in the GIC driver
    code,
but these were withheld as the GIC hardware doesn't really support
wake-up interrupts.

Fixes: ab82fa7d ("gpio: rcar: Prevent module clock disable when wake-up is enabled")
Signed-off-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

There have been concerns that the function names gpiod_set_array() and
gpiod_get_array() might be confusing to users. One might expect
gpiod_get_array() to return array values, while it is actually the array
counterpart of gpiod_get(). To be consistent with the single descriptor API
we could rename gpiod_set_array() to gpiod_set_array_value(). This makes
some function names a bit lengthy: gpiod_set_raw_array_value_cansleep().
Signed-off-by: NRojhalat Ibrahim <imr@rtschenk.de>
Acked-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Fixed several spelling errors in gpio-lynxpoint, gpio-pca953x,
gpio-tegra, gpio-zynq, gpiolib-of, gpiolib.
Signed-off-by: NColin Cronin <colinpatrickcronin@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Clean up chained handler and handler data if they were set by
gpiochip_set_chained_irqchip().
Signed-off-by: NDmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The dln2 driver was initialising a gpiolib private field, which is now
gone.
Reported-by: Nkbuild test robot <fengguang.wu@intel.com>
Signed-off-by: NJohan Hovold <johan@kernel.org>
Acked-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Move irq trigger flags, which as sysfs-interface specific, to the class
device data.

This avoids accessing the gpio-descriptor flags field using non-atomic
operations without any locking, and allows for a more clear separation
of the sysfs interface from gpiolib core.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Remove FLAG_SYSFS_DIR, which is sysfs-interface specific, and store it
in the class-device data instead.

Note that the flag is only used during export.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Rename active-low helper using common prefix.

Also remove unnecessary manipulation of value argument.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Make sure to deregister the class device (and release the irq) while
holding the sysfs lock in gpio_unexport to prevent racing with
gpio_export.

Note that this requires the recently introduced per-gpio locking to
avoid a deadlock with the kernfs active protection when waiting for the
attribute operations to drain during deregistration.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add a per-gpio mutex to serialise attribute operations rather than use
one global mutex for all gpios and chips.

Having a single global lock for all gpios in a system adds unnecessary
latency to the sysfs interface, and especially when having gpio
controllers connected over slow buses.

Now that the global gpio-sysfs interrupt table is gone and with per-gpio
data in place, we can easily switch to using a more fine-grained locking
scheme.

Keep the global mutex to serialise the global (class) operations of gpio
export and unexport and chip removal.

Also document the locking assumptions made.

Note that this is also needed to fix a race between gpiod_export and
gpiod_unexport.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Drop unnecessary locking from gpiod_export_link. If the class device has
not already been unregistered, class_find_device returns the ref-counted
class device so there's no need for locking.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Remove goto from success path.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add separate helper functions for irq request and free.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Only call irq helper if actually reconfiguring interrupt state.

This is a preparatory step in introducing separate gpio-irq request and
free functions.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Store the value sysfs entry in the gpiod data rather than in a global
table accessed through an index stored in the overloaded gpio-descriptor
flag field.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Remove redundant gpio-descriptor parameters from sysfs_set_active_low and
gpio_setup_irq.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add gpiod class-device data.

This is a first step in getting rid of the insane gpio-descriptor flag
overloading, backward irq-interface implementation, and course grained
sysfs-interface locking (a single static mutex for every operation on
all exported gpios in a system).
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The attribute operations will never be called for an unregistered device
so remove redundant checks for FLAG_EXPORT.

Note that kernfs will also guarantee that any active sysfs operation has
finished before the attribute is removed during deregistration.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Make sure to release any irq only after the class device has been
deregistered.

This avoids a race between gpiod_unexport and edge_store, where an irq
could be allocated just before the gpio class device is deregistered
without relying on FLAG_EXPORT and the global sysfs lock.

Note that there is no need to hold the sysfs lock when releasing the irq
after the class device is gone as kernfs will prevent further attribute
operations.
Signed-off-by: NJohan Hovold <johan@kernel.org>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>