Change the DeviceEvent InternalEvent to use doubles for its valuators,
instead of data and data_frac.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Change RawDeviceEvent to use doubles for valuators internally, rather
than data(_raw) and data(_raw)_frac.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Instead of taking pointers to x and y values to modify in
positionSprite, just modify the mask (as well as dev->last) in place.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Use doubles internally in both of these functions, eliminating most of
the remaining int co-ordinate usage.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
If the device doesn't have any valuators, or if it has less than two of
them, don't bother calling positionSprite. Users with one-dimensional
pointing devices may be upset.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Instead of passing fractional pointers around everywhere, just pass
doubles instead. Much easier.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Both these functions modify the mask and
pDev->last.{valuators,remainder} in-place now, so there's no need to
pass in pointers to local x and y values.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Instead of passing a set of int* to the acceleration code, pass it a
mask instead, which avoids an unfortunate loss of precision.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Simon Thum <simon.thum@gmx.de>
This widens almost all of the float-using code in ptrveloc.[ch] to
doubles, other than values coming from properties which are specified to
be floats by the property API.
Bumps input API to v14 as this changes the AccelScheme signature, as
used by xf86-input-synaptics.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
RawDeviceEvents have space for fractional valuator members, so might as
well start using them.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Change transformAbsolute to use doubles internally.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Change all these three to use doubles internally, though the outputs of
moveAbsolute and moveRelative are still truncated to int.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Shuffle some code around in moveRelative to make the conversion to
double easier later.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Shuffle some code around to make moving to double easier later.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Change moveAbsolute to be more symmetric with moveRelative by storing a
clipped axis value back in the mask, rather than just in
dev->last.valuators.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Add API for valuator_mask that accepts and returns doubles, rather than
ints. No double API is provided for set_range at the moment.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Switch the ValuatorMask struct to using doubles instead of ints for the
actual values. Preserve the old int API, and (attempt to) round towards
zero for values we return.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Add a flags member which will be copied wholesale into the resultant
xXIDeviceEvent.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
1) The error attribute appeared in gcc-4.3
2) The return type of __builtin_constant_p is int
3) Sun Studio 12.0 and later builtin support for __builtin_constant_p
Found by Tinderbox.
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
The xorg.conf manual uses the following convention in most of its
sections:
bold = text to be copied literally to the config file,
italic = a symbolic name to be substituted by a true value.
Some configuration keywords seem to have been changed into generic
options. Prepending Option to the manual entry swapped the
bold-italic logic. This patch restores the convention in the monitor
section and consists of
-.BI "Option " "\*qPreferredMode\*q " \*qstring\*q
+.BI "Option \*qPreferredMode\*q \*q" name \*q
modifications.
Plus a few minor changes (Modes → Mode) and a typo fix.
Signed-off-by: Servaas Vandenberghe
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Recent changes to the server change the default absolute input device
behaviour on zaphods to span the whole desktop too. Since these setups
usually use an xorg.conf, allow the transformation matrix to be specified in
the config as well.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Simon Thum <simon.thum@gmx.de>
Some extension libraries may set this bit before converting the event to
wire protocol and as such range checking the event will cause an invalid
BadValue error to result. As the documentation suggests the the bit
should be "forced on", remove it before doing range checks and continue
to force it on in the server.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The corresponding DeviceIntPtr wasn't being gotten properly,
resulting in BadDevice from dixLookupDevice().
Signed-off-by: Carlos Garnacho <carlosg@gnome.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
In all cases, the pointer was simply type-cast anyway. Let's get some
compile-time type safety going, how about that.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Squashed in:
xfree86: Move definition of xf86OptionPtr into separate header file
The pile of spaghettis that is the xfree86 include dependencies make it
rather hard to have a single typedef somewhere that's not interfering with
everything else or drags in a whole bunch of other includes.
Move the xf86OptionRec and GenericListRec declarations into a separate
header.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
As glamor's dix module already has a glamor.c, rename this
glamor.c to other name to avoid file name conflict which
may confuse gdb.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
The previous implementation is to override the CreatePixmap
firstly and assume the first call to CreatePixmap must be
screen pixmap. This is not clean. Now Refine it to normal
way. Let the Xephyr to set texture 0 to screen pixmap
during creating screen resources.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We should pass back the glXGetProcAddress to the underlying
gl dispatch initialization function to get those gl function
pointers.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
In Xephyr, screen pixmap is different from other pixmap.
It use fbo zero and texture zero. We initialize it during
creating screen resources.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
As now we use pixmap to handle cursor, and pixmaps are related to
current screen. If we close the screen, then we have to destroy all
the related cursors.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
The previous implementation call preInit again to reopen
the device if the device is closed by eglTerminate. Now
we don't need it anymore, as glamor dix doesn't call
eglTerminate any more when close screen.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
We already have a glamor egl module at glamor directory which
could initialize and create egl context and also can attach a
gbm buffer to a specified pixmap or to the screen pixmap. This
commit remove all the duplicated code here and use the glamor_dix
module directly.
As we don't want to handle egl stuffs in this layer, we have to
change the way to handle hardware cursor. Previous method is to
call egl functions to create a egl image and then bind the image
to a texutre then call gl functions to load texture image to the
cursor bo. Now we can bind the cursor bo to a cursor pixmap and
then use putimage to load new image to the cursor bo. Then we can
completely avoid to call egl or gl functions directly in this
ddx driver.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Latest mesa EGL implementation move to use gbm to manage/allocate buffers.
To keep backward compatibility, we still try to use eglGetDRMDisplayMESA
firstly, and if failed, then turn to use eglGetDisplay(gbm).
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Now we can allocate frame buffer during the screen initialization.
And set mode major when enter VT. This is the correct logic. The
previous implementation defer the set mode major to create screen
resource which will cause the switching between VTs broken.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This function is used to support dri2. In the underlying
driver, it will create a buffer object for a given pixmap.
And then call this api to create a egl image from that
buffer object, and then bind that image to a texture, and
then bind that texture to the pixmap.
Normally, this pixmap's content is shared between a dri2
client and the x server.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Create a new structure glamor_gl_dispatch to hold all the
gl function's pointer and initialize them at run time ,
rather than use them directly. To do this is to avoid
symbol conflicts.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
This commit applying the latest uxa's glyphs cache mechanism
and give up the old hash based cache algorithm. And the cache
picture now is much larger than the previous one also.
This new algorithm can avoid the hash insert/remove and also
the expensive sha1 checking. It could obtain about 10%
performance gain when rendering glyphs.
Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com>
Daniel Stone
Jeremy Huddleston
Keith Packard
vdb@picaros.org
Peter Hutterer
Sam Spilsbury
Carlos Garnacho
Zhigang Gong