XI1 doesn't cater for mixed mode devices, so bail out on the first valuator
that has a different mode.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
We only skip relative events for proximity, not absolute ones. Now with
mixed mode, just unset those axes that are relative.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
The XI2 protocol supports per-axis modes, but the server so far does
not. This change adds support in the server.
A complication is the fact that XI1 does not support per-axis modes.
The solution provided here is to set a per-device mode that defines the
mode of at least the first two valuators (X and Y). Note that initializing
the first two axes to a different mode than the device mode will fail.
For XI1 events, any axes following the first two that have the same mode
will be sent to clients, up to the first axis that has a different mode.
Thus, if a device has relative, then absolute, then relative mode axes,
only the first block of relative axes will be sent over XI1.
Since the XI2 protocol supports per-axis modes, all axes are sent to the
client.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
The valuators are stored inside the mask, use it from there. are stored
inside the mask, use it from there. are stored inside the mask, use it from
there. are stored inside the mask, use it from there.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
This commit introduces an abstraction API for handling masked valuators. The
intent is that drivers just allocate a mask, set the data and pass the mask
to the server. The actual storage type of the mask is hidden from the
drivers.
The new calls for drivers are:
valuator_mask_new() /* to allocate a valuator mask */
valuator_mask_zero() /* to reset a mask to zero */
valuator_mask_set() /* to set a valuator value */
The new interface to the server is
xf86PostMotionEventM()
xf86PostButtonEventM()
xf86PostKeyboardEventM()
xf86PostProximityEventM()
all taking a mask instead of the valuator array.
The ValuatorMask is currently defined for MAX_VALUATORS fixed size due to
memory allocation restrictions in SIGIO handlers.
For easier review, a lot of the code still uses separate valuator arrays.
This will be fixed in a later patch.
This patch was initially written by Chase Douglas.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Without this patch, any negative valuator value is wrong when returned
from XQueryDeviceState(). This is a regression from at least xserver
1.4.
Valuator data is set in dix/getevents.c:set_valuators() by copying
signed int values into an unsigned int field
DeviceEvent.valuators.data.
That data is converted into a double with an implicit cast by
assignment to axisVal[i] in Xi/exevents.c:UpdateDeviceState().
That double is converted back to a signed int in
queryst.c:ProcXQueryDeviceState(). If the original value in
set_valuators() is negative, the double value will be > 2^31 and the
conversion back to a signed int is undefined. (Although I
consistently see the value -2^31.)
Fix this by changing the definition of DeviceEvent.valuators.data from
uint32_t to int32_t.
Signed-off-by: Joe Shaw <joeshaw@litl.com>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
All these now generate InternalEvents, point this out. Remove XKB/XI
references, that's just confusing. This comment referred to the old-style
event generation code from server 1.4 to including 1.6 but is now just
confusing to newcomers.
Remove comment about SwitchCoreKeyboard() for the same reason.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
GPE and friends modify the valuators array passed in. Which means any driver
using e.g. xf86PostButtonEventP(..., valuators) twice to emulate a button
click will provide garbage data on the second run.
This is currently affecting the wacom driver, xf86PostButtonEventP() with
valuators is required to have input events with device-specific axis values.
Passing the same valuators in twice, once with press, once with release,
will see the valuators modified in the first call and garbage submitted in
the next one.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
commit 1432785839
xkb: release XTEST pointer buttons on physical releases. (#28808)
revealed a bug with the XTEST/PointerKeys interaction.
Events resulting from PointerKeys are injected into the event processing
stream, not appended to the event queue. The events generated for the fake
button press include a DeviceChangedEvent (DCE), a raw button event and the
button event itself. The DCE causes the master to switch classes to the
attached XTEST pointer device.
Once the fake button is processed, normal event processing continues with
events in the EQ. The master still contains the XTEST classes, causing some
events to be dropped if e.g. the number of valuators of the event in the
queue exceeds the XTEST device's number of valuators.
Example: the EQ contains the following events, processed one-by-one, left to
right.
[DCE (dev)][Btn down][Btn up][Motion][Motion][...]
^ XkbFakeDeviceButton injects [DCE (XTEST)][Btn up]
Thus the event sequence processed looks like this:
[DCE (dev)][Btn down][Btn up][DCE (XTEST)][Btn up][Motion][Motion][...]
The first DCE causes the master to switch to the device. The button up event
injects a DCE to the XTEST device, causing the following Motion events to be
processed with the master still being on XTEST classes.
This patch post-fixes the injected event sequence with a DCE to restore the
classes of the original slave device, resulting in an event sequence like
this:
[DCE (dev)][Btn down][Btn up][DCE (XTEST)][Btn up][DCE (dev)][Motion][Motion]
Note that this is a simplified description. The event sequence injected by
the PointerKeys code is injected for the master device only and the matching
slave device that caused the injection has already finished processing on
the slave. Furthermore, the injection happens as part of the the XKB layer,
before the unwrapping of the processInputProc takes us into the DIX where
the DCE is actually handled.
Bug reproducible with a device that reports more than 2 valuators. Simply
cause button releases on the device and wait for a "too many valuators"
warning message.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Keith Packard <keithp@keithp.com>
Same as the matching key functions. Buttons, like keys, can have two states
for down/up - one posted, one processed. Posted is set during event
generation (usually in the signal handler). Processed is set during event
processing when the event queue is emptied and events are being delivered to
the client.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
This patch has been generated by the following Coccinelle semantic patch:
@@
expression E;
@@
-if(E) { free(E); }
+free(E);
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Fernando Carrijo <fcarrijo@yahoo.com.br>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Many references to the dixScreenOrigins array already had the
corresponding screen pointer handy, which meant they usually looked like
"dixScreenOrigins[pScreen->myNum]". Adding a field to ScreenRec instead
of keeping this information in a parallel array simplifies those
expressions, and eliminates a MAXSCREENS-sized array.
Since dix declared the dixScreenOrigins array, I figure allocating a
screen private for these values is overkill.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com> (i686 GNU/Linux)
For absolute input devices (E.G. touchscreens) in multi-head setups,
we need a way to bind the device to an randr output. This adds the
infrastructure to the server to allow us to do so.
positionSprite() scales input coordinates to the dimensions of the shared
(total) screen frame buffer, so to restrict motion to an output we need to
scale/rotate/translate device coordinates to a subset of the frame buffer
before passing them on to positionSprite.
This is done here using a 3x3 transformation matrix, which is applied to
the device coordinates using homogeneous coordinates, E.G.:
[ c0 c1 c2 ] [ x ]
[ c3 c4 c5 ] * [ y ]
[ c6 c7 c8 ] [ 1 ]
Notice: As input devices have varying input ranges, the coordinates are
first scaled to the [0..1] range for generality, and afterwards scaled
back up.
E.G. for a dual head setup (using same resolution) next to each other, you
would want to scale the X coordinates of the touchscreen connected to the
both heads by 50%, and translate (offset) the coordinates of the rightmost
head by 50%, or in matrix form:
left: right:
[ 0.5 0 0 ] [ 0.5 0 0.5 ]
[ 0 1 0 ] [ 0 1 0 ]
[ 0 0 1 ] [ 0 0 0 ]
Which can be done using xinput:
xinput set-prop <left> --type=float "Coordinate Transformation Matrix" \
0.5 0 0 0 1 0 0 0 1
xinput set-prop <right> --type=float "Coordinate Transformation Matrix" \
0.5 0 0.5 0 1 0 0 0 1
Likewise more complication setups involving more heads, rotation or
different resolution can be handled.
Signed-off-by: Peter Korsgaard <peter.korsgaard@barco.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The only remaining X-functions used in server are XNF*, the rest is converted to
plain alloc/calloc/realloc/free/strdup.
X* functions are still exported from server and x* macros are still defined in
header file, so both ABI and API are not affected by this change.
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
An absolute device in relative mode may provide valuators outside of the
axis range. Clipping back into the range prevents screen crossings in a
multi-screen (Xinerama) setup as the required screen edge for crossing is
never met: miPointerSetPosition crosses the screen conditional to the X
coordinate being equal to the screen width or _less than_ 0. While the
former can be met when clipping into the coordinate range and scaling, the
latter cannot, resulting in a mouse pointer that gets stuck on the rightmost
screen.
This patch only applies axis clipping for valuators in mode Absolute. If
relative, we allow the values to get above/below the axis ranges. Doesn't
matter, miPointerSetPosition will reset the values to the allowed range even
if no screen was crossed.
This leads to interesting values provided to clients, the valuator range of
the device resets once a screen is crossed and essentially reflects
the position of the cursor on the screen - scaled into the valuator range.
The values themselves are valid given the range though.
In theory, the XI1 specs require that a relative device has a min/max range
of 0/0. This doesn't really go well with devices that actually can switch
mode between relative and absolute since they would have to reset their axis
range when switching. If multiple XI clients are in use, we have no method
of notifying them about the changes, so other clients may continue to use
the wrong axis ranges (note: XI1 wasn't really designed to have multiple
clients use a device). Expecting all relative devices to have this min/max
of 0 is unrealistic at this point.
So pick what is possibly the lesser of all evils, pass the beer and despair.
X.Org Bug 26543 <http://bugs.freedesktop.org/show_bug.cgi?id=26543>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
The index [0] for the second valuator looks bogus; fix it.
Signed-off-by: Oldřich Jedlička <oldium.pro@seznam.cz>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
A keyboard event from a device with both valuators and keys will be posted
through the VCK. In this case, do not update the slave device coordinates
from the VCK - they're always 0/0. Leave them as-is, for the next pointer
event will continue where it left.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Currently the root coordinates may fall into ]-1..0] if the subpixel
remainder is less than 0. Screen coordinates mustn't go below 0, so use
miPointerSetPosition to cap off the remainder if the coordinates are below
0.
This is cheating a bit, a more comprehensive solution to deal with subpixels
correctly when crossing screens is needed. For now, this'll do.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Simon Thum <simon.thum@gmx.de>
Now that all event queues hold internal events only, they never need
to be resized. Resizing them led to memory corruption as they would
get sized for an appropriate xEvent, not an internal event.
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
DCEs are now processed when sent throught the master device, not when sent
through the slave device. This includes a removal of some un-used (or partly
used) fields in the DCE itself to something more self-explanatory.
TODO: if a device has events queued and its attachment is changed, the DCE
is silently dropped now. Instead, it should be generated as soon as the
first event after the attachment is sent.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
GPE and friends now use internal events so they may generate up to 3 events.
One (optional) DeviceChanged event and one raw event plus a device event.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
InternalEvents shouldn't be used anywhere outside the X server itself. Split
up into events.h for opaque typedefs for the events needed by various
headers and eventstr.h for the actual struct definitions.
eventstr.h must only be included by code that requires internal events and
is not part of the SDK.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
internal events keep valuator data at the index for the valuator, not like
the wire events that start with first_valuator.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
it's unclear whether there actually is a problem, but in a very similar
case there is (bug#21456). Also, integer addition is generally faster.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
With internal events, we only have one event for all the data, no need to
calculate for extra events.
Reported-by: Thomas Jaeger
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Copying all classes into the master device has drawbacks for hybrid devices
(devices that are both mice and keyboards). If such a device posts an event,
it's key classes are moved into the VCP. The key event itself is unaffected
by keyboard grabs and the like.
Partial class copying copies depending on the event and copies the classes
into the right master device (i.e. the VCK for key events, the VCP for
pointer events).
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
For hybrid devices (keys + buttons/axes) the attached master device is
generally the wrong one. One shouldn't post a button event through a
keyboard and vice versa.
GetMaster(dev) returns the right master device for the given type needed.
This may be the MD paired with this device's MD.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
ChangeDeviceId would actually overwrite the flags field if deviceid wasn't
present. Aside from the event of course not telling which device generated
it in the first place.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
isMaster is not enough as long as we differ between master pointers and
keyboard. With flexible device classes, the usual checks for whether a
master device is a pointer (currently check for ->button, ->valuators or
->key) do not work as an SD may post an event through a master and mess this
check up.
Example, a device with valuators but no buttons would remove the button
class from the VCP and thus result in the
IsPointerDevice(inputInfo.pointer) == FALSE.
This will become worse in the future when new device classes are introduced
that aren't provided in the current system (e.g. a switch class).
This patch replaces isMaster with "type", one of SLAVE, MASTER_POINTER and
MASTER_KEYBOARD. All checks for dev->isMaster are replaced with an
IsMaster(dev).
dev->u.lastSlave was not signal safe since it was accessed by the DIX and
during signal handling.
Replaced with:
'dev->last.slave' for the signal handler's lastSlave (used to generate
DeviceChangedEvents), .
'dev->u.lastSlave' for the DIX lastSlave (currently only used in
change_modmap)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If the device is disabled ("off"), it must not send events to a client.
The driver shouldn't send events in that case anyway, but just to make sure
we simply drop events coming while the device is disabled.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>