[git pull] Input updates for 2.6.35-rc0

[git pull] Input updates for 2.6.35-rc0

[Dmitry Torokhov]

Hi Linus,

Please pull from:

	git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git for-linus
or
	master.kernel.org:/pub/scm/linux/kernel/git/dtor/input.git for-linus

to receive updates for the input subsystem.

Changelog:
---------

Adam Bennett (1):
      Input: add driver for hampshire serial touchscreens

Axel Lin (2):
      Input: acecad - fix incorrect size parameter in usb_buffer_free
      Input: kbtab - fix incorrect size parameter in usb_buffer_free

Barry Song (1):
      Input: ad714x - add support for the AD7143/8/7A parts

Bryan Wu (2):
      Input: add PCF8574 I2C keypad input device driver
      Input: add Analog Devices AD714x captouch input driver

Dmitry Torokhov (17):
      Input: wacom - do not allocate wacom_wac separately
      Input: wacom - fix some formatting issues
      Input: wacom - get rid of input event wrappers
      Input: wacom - get rid of wacom_combo structure
      Input: wacom - use get_unaligned to access unaligned data
      Input: implement SysRq as a separate input handler
      Input: keyboard - switch to using pr_err() and friends
      Input: keyboard - fix formatting issues
      Input: acecad - simplify usb_acecad_disconnect()
      Input: acecad - fix some formatting issues
      Input: acecad - don't advertise mouse events
      Input: kbtab - simplify kbtab_disconnect()
      Input: kbtab - do not advertise unsupported events
      Input: elantech - enforce common prefix on messages
      Input: elantech - relax signature checks
      Input: synaptics - set dimensions as reported by firmware
      Input: psmouse - small formatting changes to better follow coding style

Eric Miao (2):
      Input: remove now deprecated corgi_ts.c touchscreen driver
      Input: remove obsolete {corgi,spitz,tosa}kbd.c

Haojian Zhuang (1):
      Input: 88pm860x_onkey - remove invalid irq number assignment

Jacob Pan (1):
      Input: do not force selecting i8042 on Moorestown

Julia Lawall (2):
      Input: usbtouchscreen - switch to using kmemdup()
      Input: wistron_btns - switch to using kmemdup()

Manuel Lauss (1):
      Input: wm9712 - fix wm97xx_set_gpio() logic

Matthew Garrett (2):
      Input: i8042 - add some extra PNP keyboard types
      Input: i8042 - add a PNP entry to the aux device list

Paul Fox (1):
      Input: Documentation/sysrq.txt - update KEY_SYSRQ info

Ping Cheng (4):
      Input: wacom - switch mode upon system resume
      Input: wacom - merge out and in prox events
      Input: wacom - share pen info with touch of the same ID
      Input: wacom - streamline 2-finger touch support

Sriramakrishnan Govindarajan (1):
      Input: add keypad driver for keys interfaced to TCA6416

Wolfram Sang (2):
      Input: tsc2007 - do not leave dangling clientdata pointer
      Input: lm8323 - do not leave dangling client data pointer


Diffstat:
--------

 Documentation/sysrq.txt                    |   14 +-
 drivers/char/keyboard.c                    |  325 ++++----
 drivers/char/sysrq.c                       |  243 +++++-
 drivers/input/keyboard/Kconfig             |   18 +-
 drivers/input/keyboard/Makefile            |    1 +
 drivers/input/keyboard/lm8323.c            |    6 +-
 drivers/input/keyboard/tca6416-keypad.c    |  349 +++++++
 drivers/input/misc/88pm860x_onkey.c        |    1 -
 drivers/input/misc/Kconfig                 |   40 +
 drivers/input/misc/Makefile                |    4 +
 drivers/input/misc/ad714x-i2c.c            |  140 +++
 drivers/input/misc/ad714x-spi.c            |  103 +++
 drivers/input/misc/ad714x.c                | 1347 ++++++++++++++++++++++++++++
 drivers/input/misc/ad714x.h                |   26 +
 drivers/input/misc/pcf8574_keypad.c        |  227 +++++
 drivers/input/misc/wistron_btns.c          |    4 +-
 drivers/input/mouse/Kconfig                |    2 +-
 drivers/input/mouse/elantech.c             |   87 ++-
 drivers/input/mouse/hgpk.c                 |    4 +-
 drivers/input/mouse/logips2pp.c            |   98 ++-
 drivers/input/mouse/psmouse-base.c         |   72 +-
 drivers/input/mouse/synaptics.c            |   53 +-
 drivers/input/mouse/synaptics.h            |    6 +-
 drivers/input/serio/i8042-x86ia64io.h      |   14 +
 drivers/input/tablet/acecad.c              |   86 +-
 drivers/input/tablet/kbtab.c               |   53 +-
 drivers/input/tablet/wacom.h               |   36 +-
 drivers/input/tablet/wacom_sys.c           |  316 +++-----
 drivers/input/tablet/wacom_wac.c           | 1168 +++++++++++++-----------
 drivers/input/tablet/wacom_wac.h           |   13 +-
 drivers/input/touchscreen/Kconfig          |   12 +
 drivers/input/touchscreen/Makefile         |    1 +
 drivers/input/touchscreen/hampshire.c      |  205 +++++
 drivers/input/touchscreen/tsc2007.c        |    2 +
 drivers/input/touchscreen/usbtouchscreen.c |    5 +-
 drivers/input/touchscreen/wm97xx-core.c    |    2 +-
 include/linux/input.h                      |    1 +
 include/linux/input/ad714x.h               |   63 ++
 include/linux/serio.h                      |    1 +
 include/linux/sysrq.h                      |   23 +-
 include/linux/tca6416_keypad.h             |   34 +
 kernel/sysctl.c                            |   23 +-
 42 files changed, 3995 insertions(+), 1233 deletions(-)
 create mode 100644 drivers/input/keyboard/tca6416-keypad.c
 create mode 100644 drivers/input/misc/ad714x-i2c.c
 create mode 100644 drivers/input/misc/ad714x-spi.c
 create mode 100644 drivers/input/misc/ad714x.c
 create mode 100644 drivers/input/misc/ad714x.h
 create mode 100644 drivers/input/misc/pcf8574_keypad.c
 create mode 100644 drivers/input/touchscreen/hampshire.c
 create mode 100644 include/linux/input/ad714x.h
 create mode 100644 include/linux/tca6416_keypad.h

-- 
Dmitry

[git pull] Input updates for 2.6.35-rc0

[Dmitry Torokhov]

Hi Linus,

Please pull from:

	git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git for-linus
or
	master.kernel.org:/pub/scm/linux/kernel/git/dtor/input.git for-linus

to receive some more updates for the input subsystem.

Changelog:
---------

Daniel Mack (1):
      Input: use ABS_CNT rather than (ABS_MAX + 1)

Haojian Zhuang (1):
      Input: enable onkey driver of max8925

Kevin Hilman (1):
      Input: ads7846 - return error on regulator_get() failure

Ondrej Zary (1):
      Input: usbtouchscreen - support bigger iNexio touchscreens

Peter Ujfalusi (1):
      Input: twl4030-vibra - correct the power down sequence


Diffstat:
--------

 drivers/hid/hid-debug.c                    |    2 +-
 drivers/input/joydev.c                     |   10 +-
 drivers/input/misc/Kconfig                 |   10 ++
 drivers/input/misc/Makefile                |    1 +
 drivers/input/misc/max8925_onkey.c         |  148 ++++++++++++++++++++++++++++
 drivers/input/misc/twl4030-vibra.c         |    2 +-
 drivers/input/misc/uinput.c                |    4 +-
 drivers/input/touchscreen/ads7846.c        |    4 +-
 drivers/input/touchscreen/usbtouchscreen.c |    7 +-
 include/linux/input.h                      |   12 +-
 include/linux/joystick.h                   |    4 +-
 include/linux/uinput.h                     |   10 +-
 12 files changed, 189 insertions(+), 25 deletions(-)
 create mode 100644 drivers/input/misc/max8925_onkey.c

-- 
Dmitry

Re: [git pull] Input updates for 2.6.35-rc0

[Linus Torvalds]

On Thu, 20 May 2010, Dmitry Torokhov wrote:
> 
> Dmitry Torokhov (17):
>       Input: elantech - enforce common prefix on messages
>       Input: elantech - relax signature checks

These two commits (I _think_ - I didn't actually bisect it) seem to result 
in a Dell Inspiron bootup message changing from:

	input: ImPS/2 Logitech Wheel Mouse as /devices/platform/i8042/serio1/input/input5

to

	elantech: assuming hardware version 2, firmware version 4.3.1
	elantech: Synaptics capabilities query result 0x08, 0x18, 0x0c.
	input: ETPS/2 Elantech Touchpad as /devices/platform/i8042/serio1/input/input5

and the end result is a total disaster. In the words of my daughter, whose 
laptop this is:

	Subject: Computer=awful

	I am not able to click by tapping, I actually have to press down.. It's
	really frustrating and I want it fixed.

and having tested it, I have to agree. It's beyond awful. It's absolutely 
disgusting, and the touchpad went from "quite usable" (with tapping, 
double-tapping, scroll-wheel on the right side etc) to "totally unusable". 
The buttons you have to press (since tapping doesn't work) are on the 
touch-pad itself, and it's basically a total pain.

The scroll-wheel emulation still seems to work, but tapping has gone away. 
I'm reading the docs, and they say that you can do magic things with the 
raw registers, but if I'm looking at the register contents they seem to 
say that tapping _is_ enabled. It just doesn't work.

So it would seem that the "non-native" PS/2 emulation of that thing is way 
better than the native driver, and we should _not_ try to treat it as a 
touchpad. We're better off leaving it in ImPS/2 mode.

		Linus

Re: [git pull] Input updates for 2.6.35-rc0

[Dmitry Torokhov]

On Thu, May 27, 2010 at 06:53:20PM -0700, Linus Torvalds wrote:
> 
> 
> On Thu, 20 May 2010, Dmitry Torokhov wrote:
> > 
> > Dmitry Torokhov (17):
> >       Input: elantech - enforce common prefix on messages
> >       Input: elantech - relax signature checks
> 
> These two commits (I _think_ - I didn't actually bisect it) seem to result 
> in a Dell Inspiron bootup message changing from:
> 
> 	input: ImPS/2 Logitech Wheel Mouse as /devices/platform/i8042/serio1/input/input5
> 
> to
> 
> 	elantech: assuming hardware version 2, firmware version 4.3.1
> 	elantech: Synaptics capabilities query result 0x08, 0x18, 0x0c.
> 	input: ETPS/2 Elantech Touchpad as /devices/platform/i8042/serio1/input/input5
> 
> and the end result is a total disaster. In the words of my daughter, whose 
> laptop this is:
> 
> 	Subject: Computer=awful
> 
> 	I am not able to click by tapping, I actually have to press down.. It's
> 	really frustrating and I want it fixed.
> 
> and having tested it, I have to agree. It's beyond awful. It's absolutely 
> disgusting, and the touchpad went from "quite usable" (with tapping, 
> double-tapping, scroll-wheel on the right side etc) to "totally unusable". 
> The buttons you have to press (since tapping doesn't work) are on the 
> touch-pad itself, and it's basically a total pain.
> 
> The scroll-wheel emulation still seems to work, but tapping has gone away. 
> I'm reading the docs, and they say that you can do magic things with the 
> raw registers, but if I'm looking at the register contents they seem to 
> say that tapping _is_ enabled. It just doesn't work.
> 
> So it would seem that the "non-native" PS/2 emulation of that thing is way 
> better than the native driver, and we should _not_ try to treat it as a 
> touchpad. We're better off leaving it in ImPS/2 mode.
>

*sigh*

I do hope we will sort it out because there is a demand to make it use
Synaptics driver so it is configurable. Any chance I could persuade you
to stick "options psmouse proto=imps" in your daughter's laptop for the
time being? If we can't make it work reasonably in the next 2-3 rcs
we'll revert autodetection to what it was before.

Is it Mini 10 BTW?

Thanks.

-- 
Dmitry

Re: [git pull] Input updates for 2.6.35-rc0

[Dave Airlie]

On Fri, May 28, 2010 at 11:53 AM, Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
>
> On Thu, 20 May 2010, Dmitry Torokhov wrote:
>>
>> Dmitry Torokhov (17):
>>       Input: elantech - enforce common prefix on messages
>>       Input: elantech - relax signature checks
>
> These two commits (I _think_ - I didn't actually bisect it) seem to result
> in a Dell Inspiron bootup message changing from:
>
>        input: ImPS/2 Logitech Wheel Mouse as /devices/platform/i8042/serio1/input/input5
>
> to
>
>        elantech: assuming hardware version 2, firmware version 4.3.1
>        elantech: Synaptics capabilities query result 0x08, 0x18, 0x0c.
>        input: ETPS/2 Elantech Touchpad as /devices/platform/i8042/serio1/input/input5
>
> and the end result is a total disaster. In the words of my daughter, whose
> laptop this is:
>
>        Subject: Computer=awful
>
>        I am not able to click by tapping, I actually have to press down.. It's
>        really frustrating and I want it fixed.
>
> and having tested it, I have to agree. It's beyond awful. It's absolutely
> disgusting, and the touchpad went from "quite usable" (with tapping,
> double-tapping, scroll-wheel on the right side etc) to "totally unusable".
> The buttons you have to press (since tapping doesn't work) are on the
> touch-pad itself, and it's basically a total pain.
>
> The scroll-wheel emulation still seems to work, but tapping has gone away.
> I'm reading the docs, and they say that you can do magic things with the
> raw registers, but if I'm looking at the register contents they seem to
> say that tapping _is_ enabled. It just doesn't work.
>
> So it would seem that the "non-native" PS/2 emulation of that thing is way
> better than the native driver, and we should _not_ try to treat it as a
> touchpad. We're better off leaving it in ImPS/2 mode.
>

Was this in GNOME?

Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?

Dave.
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [git pull] Input updates for 2.6.35-rc0

[Linus Torvalds]

On Thu, 27 May 2010, Dmitry Torokhov wrote:
> 
> I do hope we will sort it out because there is a demand to make it use
> Synaptics driver so it is configurable. Any chance I could persuade you
> to stick "options psmouse proto=imps" in your daughter's laptop for the
> time being?

You mean "psmouse.proto=imps" on the command line, I presume.

> Is it Mini 10 BTW?

No. It's a Inspiron 11z. Much better CPU, better screen, yet still pretty 
small.

		Linus

Re: [git pull] Input updates for 2.6.35-rc0

[Linus Torvalds]

On Fri, 28 May 2010, Dave Airlie wrote:
> 
> Was this in GNOME?
> 
> Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?

Bingo! Yup, that turns a totally useless piece-of-crap touchpad into a 
perfectly cromulent one!

Too bad that the default behavior made it act so horribly bad. 

		Linus

Re: [git pull] Input updates for 2.6.35-rc0

[Dmitry Torokhov]

On Thu, May 27, 2010 at 09:12:27PM -0700, Linus Torvalds wrote:
> 
> 
> On Fri, 28 May 2010, Dave Airlie wrote:
> > 
> > Was this in GNOME?
> > 
> > Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?
> 
> Bingo! Yup, that turns a totally useless piece-of-crap touchpad into a 
> perfectly cromulent one!
> 
> Too bad that the default behavior made it act so horribly bad. 
> 

So I take it we can leave the touchpad in Enaltech mode for now?

-- 
Dmitry

Re: [git pull] Input updates for 2.6.35-rc0

[Dave Airlie]

On Fri, May 28, 2010 at 2:12 PM, Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
>
> On Fri, 28 May 2010, Dave Airlie wrote:
>>
>> Was this in GNOME?
>>
>> Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?
>
> Bingo! Yup, that turns a totally useless piece-of-crap touchpad into a
> perfectly cromulent one!
>
> Too bad that the default behavior made it act so horribly bad.
>

I've cc'ed Peter, but I think thats the GNOME default behaviour, which
might make sense for
older style touchpads, but maybe not for these newer the touchpad is a
button type.

Dave.

Re: [git pull] Input updates for 2.6.35-rc0

[Peter Hutterer]

On Fri, May 28, 2010 at 03:44:18PM +1000, Dave Airlie wrote:
> On Fri, May 28, 2010 at 2:12 PM, Linus Torvalds
> <torvalds@linux-foundation.org> wrote:
> >
> >
> > On Fri, 28 May 2010, Dave Airlie wrote:
> >>
> >> Was this in GNOME?
> >>
> >> Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?
> >
> > Bingo! Yup, that turns a totally useless piece-of-crap touchpad into a
> > perfectly cromulent one!
> >
> > Too bad that the default behavior made it act so horribly bad.
> >
> 
> I've cc'ed Peter, but I think thats the GNOME default behaviour, which
> might make sense for older style touchpads, but maybe not for these newer
> the touchpad is a button type.

Default behaviour is tapping on for no-button touchpads (like the ones on
the apples) and tapping off for everything else. <insert flamewar here>
If you're suggesting to change the defaults, I think you're best off just
filing a GNOME bug - the number of multi-finger tap-capable touchpads has
increased enough in the last years that this default is worth rethinking.

For the rest - please leave the touchpad in touchpad mode, because while the
emulation may look nice if you like tapping, it removes a lot of the other
configuration features that the driver can otherwise provide (edge
scrolling, multifinger scrolling, clickfinger, to name just a few).

Cheers,
  Peter

Re: [git pull] Input updates for 2.6.35-rc0

[Éric Piel]

Op 28-05-10 06:12, Linus Torvalds schreef:
> 
> 
> On Fri, 28 May 2010, Dave Airlie wrote:
>>
>> Was this in GNOME?
>>
>> Does System->Preferences->Mouse have a touchpad tab? do it have tapping off?
> 
> Bingo! Yup, that turns a totally useless piece-of-crap touchpad into a 
> perfectly cromulent one!
> 
> Too bad that the default behavior made it act so horribly bad. 
Hi,
Good to hear it was just a userspace problem. Actually on my Mini 10, to
get it with all the nice feature I have this in a script:

for DEVICE in "SynPS/2 Synaptics TouchPad" "ETPS/2 Elantech Touchpad"; do
if xinput --list "$DEVICE" > /dev/null 2>/dev/null ; then
xinput --set-prop "$DEVICE" "Synaptics Two-Finger Scrolling" 1 1
xinput --set-prop "$DEVICE" "Synaptics Tap Action" 0 0 0 0 1 2 3
xinput --set-prop "$DEVICE" "Synaptics Tap Move" 1000
xinput --set-prop "$DEVICE" "Synaptics Edge Scrolling" 1 1 0
xinput --set-prop "$DEVICE" "Synaptics Scrolling Distance" 100 100
xinput --set-prop "$DEVICE" "Synaptics Circular Scrolling" 1
xinput --set-prop "$DEVICE" "Synaptics Circular Scrolling Trigger" 3
fi
done


Currently, using the native mode, compared to the emulated imps/2, just
gives you the circular scrolling and the horizontal two-finger scrolling
(which is for me already enough to make it worth). However, once we
fully understand the protocol, and more precisely, on which device to
use one of the three sub-protocols, it should be able to also provide:
 * gesture with three fingers (soon)
 * palm detection thanks to touch width measurement (soon)
 * pitch-to-zoom (needs userspace update, so not so soon)

Cheers,
Eric

Re: [git pull] Input updates for 2.6.35-rc0

[Linus Torvalds]

On Thu, 27 May 2010, Dmitry Torokhov wrote:
> 
> So I take it we can leave the touchpad in Enaltech mode for now?

Oh yes, no problem.  Now it's more of a gnome issue than anything else, 
but I don't care deeply. It was the change of behavior that threw me and 
made it look like a kernel regression. If it had always worked that way, 
I'd probably even have noticed the touchpad configuration when installing 
the machine.

		Linus

Re: [git pull] Input updates for 2.6.35-rc0

[Linus Torvalds]

On Fri, 28 May 2010, Peter Hutterer wrote:
> 
> Default behaviour is tapping on for no-button touchpads (like the ones on
> the apples) and tapping off for everything else.

It may well be a good default when the buttons are physically separate 
(like on a lot of touchpads). But when the bottom of the touchpad itself 
is the button (ie the touchpad has a rocker, and you have to press down on 
the touchpad itself), the buttons technically _exist_, but they are 
basically useless.

That Dell machine got a lot of bad reviews for its horribe touchpad, and I 
used to not understand why people disliked the touchpad so much - it 
didn't feel horrible to me. But now I wonder if Windows has the same 
logic, and reviewers - like me - hadn't realized to enable the tap 
feature.

			Linus

Re: [git pull] Input updates for 2.6.35-rc0

[Bastien Nocera]

On Fri, 2010-05-28 at 07:28 -0700, Linus Torvalds wrote:
> 
> On Fri, 28 May 2010, Peter Hutterer wrote:
> > 
> > Default behaviour is tapping on for no-button touchpads (like the ones on
> > the apples) and tapping off for everything else.
> 
> It may well be a good default when the buttons are physically separate 
> (like on a lot of touchpads). But when the bottom of the touchpad itself 
> is the button (ie the touchpad has a rocker, and you have to press down on 
> the touchpad itself), the buttons technically _exist_, but they are 
> basically useless.

I had one of those machines, and yes, the touchpad is horrible :)

> That Dell machine got a lot of bad reviews for its horribe touchpad, and I 
> used to not understand why people disliked the touchpad so much - it 
> didn't feel horrible to me. But now I wonder if Windows has the same 
> logic, and reviewers - like me - hadn't realized to enable the tap 
> feature.

On windows, it's basically as horrible as you suggested if you don't
install the elantech drivers.

For the default GNOME configuration, we could certainly switch it again,
but we would break people with older laptops.

Seeing as we already know if there's physical buttons, we would just
need to know whether the physical buttons are embedded in the touchpad
like on yours. Then we could make an informed decision about the
defaults.

I'm sure that Peter would love the challenge of exporting that to the
desktops :)

Cheers

Re: [git pull] Input updates for 2.6.35-rc0

[Peter Hutterer]

On Fri, May 28, 2010 at 04:01:25PM +0100, Bastien Nocera wrote:
> > That Dell machine got a lot of bad reviews for its horribe touchpad, and I 
> > used to not understand why people disliked the touchpad so much - it 
> > didn't feel horrible to me. But now I wonder if Windows has the same 
> > logic, and reviewers - like me - hadn't realized to enable the tap 
> > feature.
> 
> On windows, it's basically as horrible as you suggested if you don't
> install the elantech drivers.
> 
> For the default GNOME configuration, we could certainly switch it again,
> but we would break people with older laptops.
> 
> Seeing as we already know if there's physical buttons, we would just
> need to know whether the physical buttons are embedded in the touchpad
> like on yours. Then we could make an informed decision about the
> defaults.
> 
> I'm sure that Peter would love the challenge of exporting that to the
> desktops :)

Yeah. exciting...
At the moment we rely on udev to tag the device based on the the dmi product
name. Unless I've missed something else, this is the _only_ thing we can go
on. The touchpad hardware itself looks sane and even the driver doesn't know
that this touchpad is crap, it simply applies an option to disable part of
the touchpad for movements because that's what our xorg.conf.d snippets say.
I'd love to have an EVIOISTHISDEVICEONCRACK ioctl, but for now we have to
work around it.

For clients, well, I just don't know yet how to export this.

Cheers,
  Peter