The TCA8418 might be used using different interrupt triggers on various
boards. This is not working so far because the current code forces a
falling edge trigger.

The device tree already provides a trigger type, so let's use whatever it
sets up, and since we can be loaded without DT, keep the old behaviour for
the non-DT case.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

ulseep_range() uses hrtimers and provides no advantage over msleep()
for larger delays. Fix up the 50ms delays here to use msleep() and
reduce the load on the hrtimer subsystem.
Signed-off-by: NNicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Using counter based retry loops for peripherals results in the delay
being significantly overrun during high-load situations where delay
functions tend to be vary imprecise and overrun there timeouts. So
condition the termination on the actual condition of 2s for the
re-calibration to have been successful.

As this is a very long delay there is no advantage in using
high-resolution timers thus switching this to msleep().
Signed-off-by: NNicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

This driver currently only supports legacy platform data probe.  This
change adds device tree support and gets rid of platform data probe code
since no one is actually using mpr121 platform data in the mainline.

The device tree property parsing code is based on the work of
atmel_captouch driver.
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

This driver reports input events on their interrupts which are triggered
by the sensor's status register changes.  But only single bit change is
reported in the interrupt handler.  So if there are multiple bits are
changed at almost the same time, other press or release events are ignored.

This fixes it by detecting all changed bits in the status register.
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

This driver reports misc scan input events on the sensor's status
register changes.  But the event capability for them was not set in the
device initialization, so these events were ignored.

This change adds the missing event capability.
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Remove unused key_val field in struct mpr121_touchkey.
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Instead of using #ifdef, let's mark suspend and resume methods as
__maybe_unused to provide better compile coverage.
Suggested-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

The Moorestown support was removed by commit 1a8359e4 ("x86/mid: Remove
Intel Moorestown").

Remove this leftover.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

msleep(1~20) may not do what the caller intends, and will often sleep
longer (~20 ms actual sleep for any value given in the 1~20ms range).  This
is not the desired behaviour for many cases like device resume time, device
suspend time, device enable time, retry logic, etc.  Thus, change msleep to
usleep_range for precise wakeups.
Signed-off-by: NAniroop Mathur <a.mathur@samsung.com>
Acked-by: NSimon Budig <simon.budig@kernelconcepts.de>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

The order of endpoints is well defined on official Xbox pads, but
we have found at least one 3rd-party pad that doesn't follow the
standard ("Titanfall 2 Xbox One controller" 0e6f:0165).

Fortunately, we get lucky with this specific pad because it uses
endpoint addresses that differ only by direction. We know that
there are other pads out where this is not true, so let's go
ahead and fix this.
Signed-off-by: NCameron Gutman <aicommander@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Replace first goto with simple returns as we really are just returning
one error code.
Signed-off-by: NPavel Rojtberg <rojtberg@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

xbox one was the only device that has a *_process_buttons routine.
Signed-off-by: NPavel Rojtberg <rojtberg@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Signed-off-by: NDaniel Tobias <dan.g.tob@gmail.com>
Signed-off-by: NPavel Rojtberg <rojtberg@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

msleep (1~20) may not do what the caller intends, and will often sleep
longer.  (~20 ms actual sleep for any value given in the 1~20ms range) This
is not the desired behaviour for many cases like device resume time, device
suspend time, device enable time, etc.  Thus, change msleep to usleep_range
for precise wakeups.
Signed-off-by: NAniroop Mathur <a.mathur@samsung.com>
Acked by: Albert Zhang <xu.zhang@bosch-sensortec.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

We rely on the result of scanning PNP bus for keyboard and mouse devices to
decide whether we should continue initialization of i8042 driver or not, so
the probes can not be asynchronous.

Also, it is not a good idea to unbind either of the PNP devices while i8042
is running, so let's disable bind/unbind for the PNP drivers through sysfs
and rely on i8042 cleaning after itself properly. User can still unbind and
rebind entire i8042 and that will take care of PNP drivers as well.
Reviewed-by: NLuis R. Rodriguez <mgrof@kernel.org>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Merge updated Synaptics RMI4 support, including support for SMBus
controllers and flashing firmware.

Make the avarage-samples property a general touchscreen property
rather than imx6ul device specific.
Signed-off-by: NGuy Shapiro <guy.shapiro@mobi-wize.com>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Update driver drv260x to use generic device properties so that it can be
used on non-DT systems. We also remove platform data as generic device
properties work on static board code as well.
Signed-off-by: NJingkui Wang <jkwang@google.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Let's introduce a temporary for "client->dev" is probe() as we use
it quite a few times and "dev" is shorter.
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

We were assigning I2C bus controller instead of client as parent device.
Besides being logically wrong, it messed up with devm handling of input
device. As a result we were leaving input device and event node behind
after rmmod-ing the driver, which lead to a kernel oops if one were to
access the event node later.

Let's remove the assignment and rely on devm_input_allocate_device() to
set it up properly for us.
Signed-off-by: NJingkui Wang <jkwang@google.com>
Fixes: 7132fe4f ("Input: drv260x - add TI drv260x haptics driver")
Cc: stable@vger.kernel.org
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Port firmware update code from Samsung Galaxy S7 driver into
mainline framework.

This patch has been tested on Synaptics S7813.
Signed-off-by: NNick Dyer <nick@shmanahar.org>
Tested-by: NChris Healy <cphealy@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

We were accidentally initializing haptics->rated_voltage twice, and did not
initialize overdrive voltage.
Acked-by: NDan Murphy <dmurphy@ti.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Signed-off-by: NMarcos Paulo de Souza <marcos.souza.org@gmail.com>
Reviewed-by: NPali Rohár <pali.rohar@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

As this define check if huge, this makes easier to read the code.
Signed-off-by: NMarcos Paulo de Souza <marcos.souza.org@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Semantics of NR_IRQS is different on machines with SPARSE_IRQ option
disabled or enabled, in the latter case IRQs are allocated starting
at least from the value specified by NR_IRQS and going upwards, so
the check of (irq >= NR_IRQ) to decide about an error code returned by
platform_get_irq() is completely invalid, don't attempt to overrule
irq subsystem in the driver.

The change fixes lpc32xx_keys driver initialization on boot:

  lpc32xx_keys 40050000.key: failed to get platform irq
  lpc32xx_keys: probe of 40050000.key failed with error -22
Signed-off-by: NVladimir Zapolskiy <vz@mleia.com>
Acked-by: NSylvain Lemieux <slemieux.tyco@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

The debug would only ever output zero for the clip information.
Signed-off-by: NNick Dyer <nick@shmanahar.org>
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Tested-by: NChris Healy <cphealy@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Now that we have a proper API to set the attention data, there is
no point in keeping it in the transport driver.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

The HID implementation of RMI4 provides the data during
the interrupt (in the input report). We need to provide
a way for this transport driver to provide the attention
data while calling an IRQ.

We use a fifo in rmi_core to not lose any incoming event.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

First check if there are data available passed by the transport device.
If data available use these data. If there are no data available
try to read the rmi block if dsata are passed this way.

This is the way the other rmi function handlers will do this.

This patch is needed on HID devices because the firmware reads F03 data
registers and adds them to the HID attention report. Reading those
registers from the driver after the firmware read them will result in
invalid data.
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Signed-off-by: NDennis Wassenberg <dennis.wassenberg@secunet.com>
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

This adds basic functionality for PS/2 passthrough on Synaptics
Touchpads using RMI4 through smbus.
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Signed-off-by: NLyude Paul <thatslyude@gmail.com>
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

The code uses of_property_read_u32 and expects positive values.  However,
the values are stored in signed int variables.  Additionally, the registers
values are also stored in signed variables without a good reason
(readl/writel expect u32).

The only time this caused a real bug was in the new average-samples
property, in which the property is numerically compared and implicitly
expected to be positive.

I believe it's better to change all the properties and registers to u32,
for consistency and warnings reduction.
Signed-off-by: NGuy Shapiro <guy.shapiro@mobi-wize.com>
Reported-by: NBjørn Forsman <bjorn.forsman@gmail.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Add mask of each function bits of REG_ADC_CFG, and clear these
function bits first, otherwise use '|=' operation may get the
wrong setting which depends on the original value of REG_ADC_CFG.
Signed-off-by: NHaibo Chen <haibo.chen@nxp.com>
Reviewed-by: NGuy Shapiro <guy.shapiro@mobi-wize.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

If struct rmi_device_platform_data contains pointers to other struct,
it gets difficult to allocate a fixed size struct and copy it over between
drivers.

Change the pointers into a struct and change the code in rmi4 accordingly.
Reviewed-by: NAndrew Duggan <aduggan@synaptics.com>
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

those functions should not be used outside of rmi_core.ko.
There is no point in exporting them to the world.

It looks like rmi_read_pdt_entry() should be static too.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

This partially reverts commit 29fd0ec2 ("Input: synaptics-rmi4 -
add support for F34 device reflash")

irq_mutex should be used only to protect data->current_irq_mask, not
preventing incoming input to be processed while the upgrade of the
firmware is happening. We can simply disable the irqs when we don't
want them to interfere with the upgrade process.

Tested on S7300 and S7800 (with F34 v7 patch added)
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NNick Dyer <nick@shmanahar.org>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Using regmap_update_bits(..., mask, 1) with 'mask' following (1 << k)
and k greater than 0 is wrong. Indeed, _regmap_update_bits will perform
(mask & 1), which results in 0 if LSB of mask is 0. Thus the call
regmap_update_bits(..., mask, 1) is in reality equivalent to
regmap_update_bits(..., mask, 0).

In such a case, the correct use is regmap_update_bits(..., mask, mask).

This driver is performing such a mistake with the DRV2667_STANDBY mask,
which equals (1 << 6). Fix the driver to make it consistent with the
API, and fix the alignment problem at the same time. Please note that
this change is untested, as I do not have this piece of hardware.
Testers are welcome!
Signed-off-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Using regmap_update_bits(..., mask, 1) with 'mask' following (1 << k)
and k greater than 0 is wrong. Indeed, _regmap_update_bits will perform
(mask & 1), which results in 0 if LSB of mask is 0. Thus the call
regmap_update_bits(..., mask, 1) is in reality equivalent to
regmap_update_bits(..., mask, 0).

In such a case, the correct use is regmap_update_bits(..., mask, mask).

This driver is performing such a mistake with the DRV2665_STANDBY mask,
which equals BIT(6). Fix the driver to make it consistent with the API,
and fix the alignment problem at the same time. Please note that this
change is untested, as I do not have this piece of hardware. Testers
are welcome!
Signed-off-by: NFlorian Vaussard <florian.vaussard@heig-vd.ch>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

Set the .enabled boolean and trigger an event processing when enabling
for edge-triggered systems.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>

These fields are not used anywhere, there is no point in carrying them.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NDmitry Torokhov <dmitry.torokhov@gmail.com>