The current XI2 mask handling is handy for copying (fixed size arrays) but a
pain to deal with otherwise. Add a struct for XI2 masks and the required
accessors.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Once grabs start having nested memory locations, we can't just use the
GrabRec on the stack anymore, we need to alloc/copy/free the grabs.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Not really needed at this point, but will be once touch support is added.
Since grabs are now expected to be allocated/freed with AllocGrab and
FreeGrab, CopyGrab must increase the refcount and duplicate the modifier
masks. Until the callers are switched to use FreeGrab, this introduces
memleaks.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Not needed since the GrabRec is a self-contained struct but will be needed
for the xi2 input mask rework.
FreeGrab already exists, make it available to other callers.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
No effective functionality change, just cleanup to make this code slightly
more sane.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
The record extension needs the major and minor opcodes in the reply
hook, but the request buffer may have been freed by the time the hook
is invoked. Saving the request major and minor codes as the request is
executed avoids fetching from the defunct request buffer.
This patch also eliminates the public MinorOpcodeOfRequest function,
inlining it into Dispatch. Usages of that function have been replaced
with direct access to the new ClientRec field.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
And let it return a boolean value, that's all the callers need anyway.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Wherever it's obvious which device we need (keyboard or pointer), use
GetMaster() instead of GetPairedDevice(). It is more reliable in actually
getting the device type we want.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Commit 587c3a2d19 fixed DoGetImage to
check windows against their backing drawables, rather than against the
screen dimensions, to prevent reading outside the bounds of redirected
windows' backing pixmaps (see bug #22804). Unfortunately, while
making that change I also removed the check that the rectangle is
contained within the bounds of the source window, which is a violation
of the specification:
If the drawable is a window, the window must be viewable, and it
must be the case that, if there were no inferiors or overlapping
windows, the specified rectangle of the window would be fully
visible on the screen *and wholly contained within the outside
edges of the window* (or a Match error results). Note that the
borders of the window can be included and read with this request.
(emphasis mine)
Reinstate the window dimension check, to return BadMatch if the
GetImage request falls outside the bounds of the window.
Fixes X Test Suite test XGetImage-15:
400|0 15 1 11:05:41|IC Start
200|0 15 11:05:41|TP Start
520|0 15 00005146 1 1|VSW5TESTSUITE PURPOSE 15
520|0 15 00005146 1 2|Assertion XGetImage-15.(A)
520|0 15 00005146 1 3|When the drawable is a window and the window is viewable
520|0 15 00005146 1 4|and it is not the case that given there were no inferiors or
520|0 15 00005146 1 5|overlapping windows the specified rectangle of the window
520|0 15 00005146 1 6|would be fully visible on the screen and wholly contained
520|0 15 00005146 1 7|within the outside edges of the window, then a BadMatch
520|0 15 00005146 1 8|error occurs.
520|0 15 00005146 1 9|METH: Create window which is not fully visible on the screen.
520|0 15 00005146 1 10|METH: Call XMapWindow to make sure the window is viewable.
520|0 15 00005146 1 11|METH: Call XGetImage with rectangle extending beyond edge of screen.
520|0 15 00005146 1 12|METH: Verify XGetImage return value is null.
520|0 15 00005146 1 13|METH: Verify that BadMatch error occurred.
520|0 15 00005146 1 14|METH: Create window which is fully visible on the screen.
520|0 15 00005146 1 15|METH: Call XMapWindow to make sure the window is viewable.
520|0 15 00005146 1 16|METH: Call XGetImage with rectangle extending beyond edge of window.
520|0 15 00005146 1 17|METH: Verify XGetImage return value is null.
520|0 15 00005146 1 18|METH: Verify that BadMatch error occurred.
520|0 15 00005146 1 19|REPORT: Got Success, Expecting BadMatch
520|0 15 00005146 1 20|REPORT: Null image not returned.
220|0 15 1 11:05:41|FAIL
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
When closing down all devices, we manually unset master for all attached
devices, but the device's sprite info still points to the master's sprite
info. This leaves us a window where the master is freed already but the
device isn't yet. A signal during that window causes dereference of the
already freed spriteInfo in mieqEnqueue's EnqueueScreen macro.
Simply block signals when removing all devices. It's not like we're really
worrying about high-responsive input at this stage.
https://bugzilla.redhat.com/show_bug.cgi?id=737031
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
If a floating device changes, the master is NULL but we must still create a
DCE for it.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
No effective functional changes, prep work for future patches.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Instead of device and master (and just using master), drop the master
argument and let the callers pass in the device the event is to be sent for.
No effective functional changes.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
3304bbff9b added smooth scrolling support for
pointer events and for XIQueryDevice but didn't add the matching parts to
XIDeviceChangedEvents.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
ptrveloc.c: In function 'QueryTrackers':
ptrveloc.c:598:34: warning: variable 'used_offset' set but not used
[-Wunused-but-set-variable]
used_offset is used, but only in the debugging code. Move the #if statement
to ignore that warning.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Use the same struct for both InputOption and XF86OptionRec so we don't need
to convert to and fro the two in the config backends.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
The protocol requires that the emulated event is marked as such. So if a
driver with smooth scrolling axis sends legacy button events, the motion
event must be marked as emulated.
Pass the real type to emulate_scroll_button_events and create the events
accordingly. For real button press or relase events, only that event must be
generated since a release event will follow or a press event has already
occured, respectively. (This fixes a bug where we'd get two release events
for each legacy button event)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Using this call simplifies callers that don't know if the mask bit is set.
Before:
if (valuator_mask_isset(mask, valnum))
value = valuator_mask_get_double(mask, valnum));
else
value = someothervalue;
Now:
if (!valuator_mask_fetch_double(mask, valnum, &value))
value = someothervalue;
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Co-authored-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
miPointerSetPosition traditionally took coordinates on a per-screen basis,
triggering a screen switch when these went out-of-bounds. For absolute
devices, this prevented screen crossing in the negative x/y direction.
This patch changes the event generation patch to handle screen coordinates
in a desktop range (i.e. all screens together). Screen switches are
triggered when these coordinates are not on the current screen.
This unifies the pointer behaviour of single ScreenRec multihead and
multiple ScreenRecs multihead in that the cursor by default moves about the
whole screen rather than be confined to one single screen. The
transformation matrix may then be used to actually confine the cursor to the
screen again.
Note: fill_pointer_events has to deal with several different coordinate
systems. Make sure you read the comment before trying to understand the code.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Allow rescaling to non-zero based axis ranges as default (for when screen
offsets are non-zero). Currently unused.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
For Zaphod mode screen crossing handling we need to know the size of all
screens together (i.e. the whole desktop size). Store that in the screenInfo to
have it readily available in events.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Don't update the MD where it's not expected, positionSprite should really
just do that - position the sprite.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Don't switch between doubles and ints in the caller, instead take doubles in
miPointerSetPosition and do the conversion there. For full feature we should
change everything down from here for doubles too.
Functional change: previously we'd restore the remainder regardless of
screen switching/confinement (despite what the comment said). Now,
screen changing or cursor constraints will cause the remainder be clipped
off. This should happen for cursor constraints but arguably not for screen
crossing.
This also corrects a currently wrong comment about miPointerSetPosition's
input coordinates.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
miPointerSetPosition may switch screens. Always return the screen the sprite
is on instead of relying on callers to call miPointerGetScreen().
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
We can just get this in the function, no effective functional changes.
Also return the screen to the caller. Though we don't use it yet, we will in
a follow-up patch.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
No functional changes.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
GPE unconditionally dereferences pDev->valuator if a mask is present. This
shouldn't really happen but if it does, don't crash, just ignore the events
with an error.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Let's be honest about what it does.
moveRelative accumulates delta _and_ clips in some cases, so that one can
keep it's name.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
This bug led to inverted scrolling axes with drivers that support smooth
scrolling axes but send legacy button events.
Signed-off-by: Max Schwarz <Max@x-quadraht.de>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
When a screen switch is triggered by PointerKeys, the device for
NewCurrentScreen is the keyboard. Submitting pointer events for this
keyboard (without valuators) has no effect as GPE ignores the event.
Force the dequeuing through the XTest device attached to this device.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
We don't actually handle the mask correctly. They're clipped and dropped
into the event but that's about it. I don't think we did since 1.4, let's
warn the user if this happens.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Switching screens relies on rootx/y to be set to the correct value. Note:
though we technically take a mask for GetKeyboardEvents we don't actually
handle it properly to move the pointer as required (and generate motion
events if needed).
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
