The ADP5585 family keypad decoder and IO expander is similar to the ADP5589,
however it features less IO pins, and lacks hardware assisted key-lock
functionality. Unfortunately the register addresses are different, as well as
the event codes and bit organization within the port related registers.

Move ADP5589 Register defines from the header file into the main source file.
Add new defines while making sure we don't break existing platform_data.
Add register address translation, and turn device specific defines into variables.
Introduce some helper functions and disable functions that doesn't
exist on the added devices.
Signed-off-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

The imx_keypad driver was indicating that it was wakeup capable in
imx_keypad_probe(), but it didn't implement suspend or resume methods.

According to the i.MX series MCU Reference Manual, the kpp (keypad
port) is a major wake up source which can detect any key press even
in low power mode and even when there is no clock.
Signed-off-by: NHui Wang <jason77.wang@gmail.com>
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

The problem here is that max_effects can wrap on 32 bits systems.
We'd allocate a smaller amount of data than sizeof(struct ff_device).
The call to kcalloc() on the next line would fail but it would write
the NULL return outside of the memory we just allocated causing data
corruption.

The call path is that uinput_setup_device() get ->ff_effects_max from
the user and sets the value in the ->private_data struct.  From there
it is:
-> uinput_ioctl_handler()
   -> uinput_create_device()
      -> input_ff_create(dev, udev->ff_effects_max);

I've also changed ff_effects_max so it's an unsigned int instead of
a signed int as a cleanup.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Abort driver initialization if X plate resistance was not specified in
platform data as it will cause pressure to be always calculated as 0,
and making userspace ignore touch coordinates.
Signed-off-by: NPhilip Rakity <prakity@marvell.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Apparently we never freed memory allocated when users open our char
devices nor removed old users from the list of connected clients.

Also unregister misc device immediately upon disconnecting the port
instead of waiting until last user drops off (refcounting in misc
device code will make sure needed pieces stay around while they
are needed) and make sure we are not holing holding serio_raw_mutex
when registering/unregistering misc device. This should fix potential
deadlock between serio_raw and misc device code uncovered by lockdep
and reported by Thomas Tuttle.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Send SIGIO/POLL_HUP and otherwise wake up waiters when corresponding serio
port is being disconnected. Also check if port is dead in serio_raw_poll
and signal POLLHUP|POLLERR.

This should speed up process of releasing dead devices by userspace
applications.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Instead of relying on setting serio_raw->serio to NULL upon disconnecting
ports mark them explicitly as "dead". Also take and carry reference to
underlying serio port to make sure it does not go away until we are done
with it.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This makes checkpatch.pl happy with the driver
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This will ensure our reporting is consistent with the rest of the system.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Make sure we hold serio lock when adding clients to client list so that
we do not race with serio_raw_release() removing clients from the same
list.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

'serio_raw_list' and 'list' names do not accurately represent their objects
and are extremely confusing when reading the code. Let's use better suited
names.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This will ensure our reporting is consistent with the rest of the system
and we do not refer to obsolete source file names.
Reviewed-by: NWanlong Gao <gaowanlong@cn.fujitsu.com>
Reviewed-by: NJJ Ding <dgdunix@gmail.com>
Reviewed-by: NDaniel Kurtz <djkurtz@chromium.org>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Medium size Bamboo P&T driver reused max X/Y form older Bamboo 1
medium size tablets and never updated to real value. Actual
active area of tablet is slightly larger in Y direction.
Signed-off-by: NChris Bagwell <chris@cnpbagwell.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Tablet reports a distance of 0 right at highest point possible
to be in proximity and distance_max when touching tablet. Inverse
the distance since user land has no way of knowing ABS_DISTANCE
is not distance from tablet.
Signed-off-by: NChris Bagwell <chris@cnpbagwell.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

These were left in during removal of touch pressure reports but not
needed now.
Signed-off-by: NChris Bagwell <chris@cnpbagwell.com>
Reviewed-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

CONFIG_PM is defined when CONFIG_PM_SLEEP or CONFIG_PM_RUNTIME is defined,
however suspend and resume methods are only valid in the context of
CONFIG_PM_SLEEP. If only CONFIG_PM_RUNTIME is defined we get the following
warning (courtesy of Geerts randconfig builds):

	lm8323.c: warning: 'lm8323_resume' defined but not used
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

CONFIG_PM is defined when CONFIG_PM_SLEEP or CONFIG_PM_RUNTIME is defined,
however suspend and resume methods are only valid in the context of
CONFIG_PM_SLEEP. If only CONFIG_PM_RUNTIME is defined we get the following
warning (courtesy of Geerts randconfig builds):

	ad7879-i2c.c: warning: 'ad7879_i2c_resume' defined but not used
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

CONFIG_PM is defined when CONFIG_PM_SLEEP or CONFIG_PM_RUNTIME is defined,
however suspend and resume methods are only valid in context of
CONFIG_PM_SLEEP. If only CONFIG_PM_RUNTIME is defined we get the following
warning (courtesy of Geerts randconfig builds):

	synaptics_i2c.c: warning: 'synaptics_i2c_resume' defined but not used
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Sometimes GCC is not smart enough to recognize that x, y and z are
always used properly initialized in mma8450_poll(). Let's rearrange
the code a bit to help GCC.
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Sysfs attribute show methods are always passed a buffer of length
PAGE_SIZE.  To keep from overwriting this buffer and causing havoc, use
snprintf() to guarantee we never write more than the buffer can hold.

In addition, at least for my touchscreen, the number and size of objects
was far too big to fit in a single 4K page.  Therefore, this patch also
trims some redundant framing text to leave more room for actual data.
Signed-off-by: NDaniel Kurtz <djkurtz@chromium.org>
Acked-by: NNick Dyer <nick.dyer@itdev.co.uk>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Reviewed-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Tested-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Signed-off-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Cintiq 21ux2 has two sets of four LEDs on right and left side of
the tablet, respectively.
Reviewed-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Tested-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Signed-off-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

The LED also indicates the status of the tablet. Don't turn it off.
Reviewed-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Tested-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Signed-off-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

The LED luminance level is normally lower when no button is pressed.
Reviewed-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Tested-by: NEduard Hasenleithner <eduard@hasenleithner.at>
Signed-off-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This reverts commit 71c86ce5.
The 21UX2 does have touchstrips, but they are in a somewhat-
hidden location.
Signed-off-by: NJason Gerecke <killertofu@gmail.com>
Acked-by: NPing Cheng <pinglinux@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This patch adds the TSC-40 serial touchscreen driver and should be
compatible with TSC-10 and TSC-25.

The driver was written by Linutronix on behalf of Bachmann electronic GmbH.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NChristian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This patch clarifies a few bits of documentation in the header file
for the adxl34x driver.
Signed-off-by: NMichael Tandy <lkml@mkt.me.uk>
Acked-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Use the of_property_read_u32() helper function to retrieve u32 values
from the device tree. Also do not pass the len parameter to
of_get_property if it isn't checked afterwards.
Signed-off-by: NTobias Klauser <tklauser@distanz.ch>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Since touchscreen driver does not handle any events to be sent to the
device we can close serio port first and then unregister the input device.
Tested-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Instead of having one large switch based on product ID use pointer to
function actually doing protocol decoding.
Tested-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Add multi touch support for PenMount 6250 touch controller.
Signed-off-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Add dual touch support for PenMount 3000 touch controller.
Signed-off-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Add support for PenMount 6000 touch controller.
Signed-off-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

The MSB and LSB of the XY axis value are switched according to the PenMount 9000 protocol. The driver name is also changed from penmountlpc, since it is not for LPC interface at all.
Signed-off-by: NJohn Sung <penmount.touch@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

This essentially reverts commit f81bc788.

With recent work on elantech driver, I believe we now have complete support
for all elantech touchpads. So remove this hack.
Signed-off-by: NJJ Ding <jj_ding@emc.com.tw>
Reviewed-by: NÉric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

V2 hardware has many variants. This patch adddresses two issues:

 - some model also has debounce packets, but with a different signature
   than v3. Now we just check debounce for all v2 hardware.

 - due to different scanning methods the hardware uses, x and y ranges have
   to be calculated differently. And for some specific versions, we can just
   see them as custom-made, so set {x, y} the same values as Windows driver
   does.
Signed-off-by: NJJ Ding <jj_ding@emc.com.tw>
Tested-by: NRichard Schütz <r.schtz@t-online.de>
Reviewed-by: NÉric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>

Bamboos with Product ID's > 0xD4 return values unrelated to pressure
in touch 1 pressure field.  They also report 2nd touch X/Y values
shifted down 1 byte (where pressure was).  This results in jumpy
1 finger touch and totally invalid 2nd finger data.

For touch detection, switch to a Touch Present single bit that
all versions of Bamboo support.

For touch 2 offset, calculate offset based on a bit that is set
different between the two packet layouts.

Since touch pressure reports were removed from driver, there was
no need to be reading pressure any more.
Signed-off-by: NChris Bagwell <chris@cnpbagwell.com>
Reviewed-by: NPing Cheng <pinglinux@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>