linux-toradex.git/drivers/input/touchscreen, branch v2.6.17.3

[PATCH] ads7846 conversion accuracy

2006-05-26T18:55:46+00:00

This improves accuracy of the touchscreen and hwmon sensor readings,
addressing an issue noted by Imre Deak: there's an extra bit written before
the sample (12 bits) gets written out.

It also catches up to various comments, and makes the /proc/interrupts
entry sensible again.

Signed-off-by: David Brownell 
Cc: Imre Deak 
Cc: Jean Delvare 
Cc: Dmitry Torokhov 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

Merge master.kernel.org:/home/rmk/linux-2.6-arm

2006-05-04T21:52:43+00:00

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] 3490/1: i.MX: move uart resources to board files
  [ARM] 3488/1: make icedcc_putc do the right thing
  [ARM] 3487/1: IXP4xx: Support non-PCI systems
  [ARM] 3486/1: Mark memory as clobbered by the ARM _syscallX() macros

[ARM] 3486/1: Mark memory as clobbered by the ARM _syscallX() macros

2006-04-30T14:34:29+00:00

Patch from Markus Gutschke

In order to prevent gcc from making incorrect optimizations, all asm()
statements that define system calls should report memory as
clobbered. Recent versions of the headers for i386 have been changed
accordingly, but the ARM headers are still defective.

This patch fixes the bug tracked at
http://bugzilla.kernel.org/show_bug.cgi?id=6205

Signed-off-by: Markus Gutschke 
Signed-off-by: Russell King

Input: ads7846 - improve filtering for thumb press accuracy

2006-04-26T04:13:18+00:00

Providing more accurate coordinates for thumb press requires additional
steps in the filtering logic:

- Ignore samples found invalid by the debouncing logic, or the ones that
  have out of bound pressure value.
- Add a parameter to repeat debouncing, so that more then two consecutive
  good readings are required for a valid sample.

Signed-off-by: Imre Deak 
Acked-by: Juha Yrjola 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - report 0 pressure value along with pen up event

2006-04-26T04:12:14+00:00

X touchscreen drivers that don't interpret the designated pen up message
assume a pen up event from a pressure value 0. For these we generate a
pressure 0 message along with the pen up message.

Signed-off-by: Imre Deak 
Acked-by: Juha Yrjola 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - handle IRQs that were latched during disabled IRQs

2006-04-12T03:44:05+00:00

The pen down IRQ will toggle during each X,Y,Z measurement cycle.
Even though the IRQ is disabled it will be latched and delivered
when after enable_irq. Thus in the IRQ handler we must avoid
starting a new measurement cycle when such an "unwanted" IRQ happens.
Add a get_pendown_state platform function, which will probably
determine this by reading the current GPIO level of the pen IRQ pin.

Move the IRQ reenabling from the SPI RX function to the timer. After
the last power down message the pen IRQ pin is still active for a
while and get_pendown_state would report incorrectly a pen down state.

When suspending we should check the ts->pending flag instead of
ts->pendown, since the timer can be pending regardless of ts->pendown.
Also if ts->pending is set we can be sure that the timer is running,
so no need to rearm it. Similarly if ts->pending is not set we can
be sure that the IRQ is enabled (and the timer is not).

Signed-off-by: Imre Deak 
Signed-off-by: David Brownell 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - miscellaneous fixes

2006-04-12T03:43:55+00:00

- Add disable attribute to support device locking mode where
  unintentional touch event shouldn't wake up the system;
- Update comments;
- Add missing spin_lock_init;
- Do device resume with the lock held;
- Do cleanup calls / free memory in the reverse order of initialization.

Signed-off-by: Imre Deak 
Signed-off-by: David Brownell 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - use msleep() instead of udelay() in suspend

2006-04-12T03:42:25+00:00

Sometimes a polling loop had a hard time changing state without
pre-emption enabled.  Use msleep instead, it's better anyway.

Signed-off-by: Juha Yrjola 
Signed-off-by: David Brownell 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - debouncing and rudimentary sample filtering

2006-04-12T03:42:03+00:00

Some touchscreens seem to oscillate heavily for a while after touching
the screen.  Implement support for sampling the screen until we get two
consecutive values that are close enough.

Signed-off-by: Imre Deak 
Signed-off-by: Juha Yrjola 
Signed-off-by: David Brownell 
Signed-off-by: Dmitry Torokhov

Input: ads7846 - power down ADC a bit later

2006-04-12T03:41:49+00:00

Submit a seperate request for powering down the ADC in ads7846,
doing it after the last read request.  Otherwise some of the read
values are incorrect.

Signed-off-by: Imre Deak 
Signed-off-by: Juha Yrjola 
Signed-off-by: David Brownell 
Signed-off-by: Dmitry Torokhov