92e0a324b5
require inputproto 1.9.99.7
2009-03-20 15:17:54 +10:00
906fe95479
Xi: don't allow OpenDevice on any MD.
...
This restores the original behaviour of XI 1. Any slave device may be opened,
but none of the MDs.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:54 +10:00
c7eb27f392
Xi: always only list VCP/VCK and the SD's in XListInputDevices.
...
If a client wants other MDs, it should use XI2's XQueryDevice() instead.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:54 +10:00
801f14d4b7
Xi: don't send attachment info down with ListInputDevices.
2009-03-20 15:17:54 +10:00
66dd8879b3
Xi: purge ExtendedGrabDevice request handling.
2009-03-20 15:17:54 +10:00
43e48900fc
dix: Force deviced ids of 2 and higher for actual devices.
...
0 is now reserved for the "AllDevices" virtual device.
1 is now reserved for the "AllMasterDevices" virtual device.
This also means that wherever we passed in (mskidx = 0), we now need to pass
in the deviceid.
2009-03-20 15:17:54 +10:00
70896f6127
dix: abstract event filters through GetEventFilters.
...
Don't let everyone acces the filters[] array directly. This is necessary once
we start dealing with GenericEvents, where the filters are a bit more
complicated.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:54 +10:00
8a8763f150
dix: restructure DeliverEventsToWindow, a little bit anyway.
...
Desperate attempt of making the code easy to read and understand. A failed
desparate attempt, possibly.
2009-03-20 15:17:54 +10:00
891cfc8020
dix: clean up DeliverDeviceEvents.
...
Move mask checking into external function and streamline the code.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:54 +10:00
6b6b660c36
Xext: remove ev_fill from GEExtensions.
2009-03-20 15:17:54 +10:00
445daa62e7
Xext: purge XGE event masks.
...
The masks were originally designed to generically handle event masks for
extensions. Since all that is in-server anyway, it's much better writing
custom event masks for those extensions that need it and not providing a
unified mechanism.
XI2 needs more than the current implementation, which is already too complex
for most other extensions. good riddance.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:53 +10:00
6fe4c28bf4
Add agressive event type checking.
...
Best to FatalError if a wrong event comes in. At least that forces me to fix
it really quickly.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 15:17:53 +10:00
836864b657
xkb: don't overrun the map index when accessing symbols.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
497a125339
dix: remove duplicate PickKeyboard() command.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
7d4df0ee44
dix: fix device sync state when calling SyncBoth during AllowEvents.
...
This did access the wrong device's sync state, potentially freezing or not
thawing the actual device that was supposed to be thawed.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
f1c7b95d83
dix: do percentage check before device check in ProcBell
...
This is just for correctness. The server should return BadValue for anything
not in [-100, 100].
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
d60391d8ca
dix: fix uncredible fail in PostSyntheticMotion.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
1a71862d33
dix/xfree86: simplified velocity approximation algorithm
...
Replace multi-stage filtering with simple linear velocity,
tracked several instances backwards. A heuristic ensures
only approximately linear motion is considered, so velocity
remains valid in any case. Numerical stability is much
better, and nothing changes to people who didn't tune the
advanced features of the previous algorithm.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
5ae129baef
Xi: check for existence of the button class before accessing it
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
e8094d8f3f
dix: ProcSendEvent shouldn't use inputInfo.keyboard directly.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-20 14:48:57 +10:00
12aeddf5ad
exa: Accept scratch pixmaps with offscreen memory as such.
2009-03-19 20:40:03 +01:00
a635bb5357
exa: round of fb_pitch to the next byte
2009-03-18 19:36:18 +01:00
ec32878359
exa: check if the pixmap is pinned for unsupported AUX indices.
2009-03-18 19:36:18 +01:00
3992dd38ca
selinux: Add support for avc_acquire_netlink_fd()
...
Requires libselinux 2.0.79 or newer. Without this, libselinux will
check for policy updates on the netlink socket on basically every policy
lookup. Statistically speaking, they never happen, and the check
translates to at least one more syscall on basically every operation.
Instead, take control of the fd from the library, and check it in
WakeupHandler if it polls readable.
2009-03-16 13:24:48 -04:00
be6dc9023b
xfree86: remove a superfluous assignment.
...
If dev is NULL, we don't have a VCP and that means we have other issues
anyway.
2009-03-16 13:29:07 +10:00
bfab422dcd
xfixes: useless (void) typecast removal
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-16 13:29:07 +10:00
4eeaee1e5a
xkb: xkbi has a pointer to the device - use this instead of inputInfo.pointer.
2009-03-16 13:29:07 +10:00
3f801ba62a
xkb: the VCP can post device events, don't stop xkb filtering on it.
2009-03-16 13:29:07 +10:00
23686e5680
xkb: Fix a mis-use of inputInfo.keyboard.
2009-03-16 13:29:07 +10:00
c97c6c3de3
xkb: fix a couple of device checks when looping through all devices.
...
Generally, we want to apply stuff to the device and to all attached slave
devices.
2009-03-16 13:29:07 +10:00
d21a546e96
Xi: VCP and VCK may register for ext. events, so delete them if necessary.
2009-03-16 13:29:07 +10:00
6ec49098f7
render: replace a wrong inputInfo.pointer with pDev
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-16 13:29:07 +10:00
0b2b330f6f
mi: remove a bunch of useless inputInfo.pointer assignments.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-16 13:29:07 +10:00
31301861c1
dix: remove wrong InputInfo.pointer usage - should be pDev instead.
2009-03-16 13:29:07 +10:00
639658b2ea
dix: use GetPairedDevice in ProcQueryPointer rather than inputInfo.keyboard.
2009-03-16 13:29:06 +10:00
f01ee198ff
dix: don't use inputInfo.keyboard to get the focus window in ActivateKbdGrab
...
I'm not sure if that's the right solution, but the other one is wronger.
2009-03-16 13:29:06 +10:00
de415743f4
dix: fix XACE checks in ProcWarpPointer
2009-03-16 13:29:06 +10:00
18430616cf
dix: use PickPointer() and PickKeyboard instead of inputInfo.pointer/keyboard.
...
Affected calls:
ProcChangeKeyboardControl
ProcBell
ProcUngrabKey
ProcUngrabButton
2009-03-16 13:29:06 +10:00
e31727158c
dix: fix ProcChangePointerControl's wrong inputInfo.pointer usage.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-16 13:29:06 +10:00
fc5e8dc4a2
Unexport xf86SetPriority
2009-03-13 14:00:37 -04:00
89d7b88f32
exa: allow exaModifyPixmapHeader to set sys_ptr for EXA_HANDLES_PIXMAPS
...
- exaModifyPixmapHeader would save sys_ptr if needed, but it would be NULL'ed afterwards.
- This is needed to support pixmaps that are not offscreen.
2009-03-13 00:44:48 +01:00
544cd9e7b5
Honour Option "DPMS" "off" on xorg.conf
...
Signed-off-by: James Cloos <cloos@jhcloos.com>
2009-03-12 14:50:59 -04:00
7abb72c5c7
xfree86: remove obsolete code
...
remove a few lines which redo part of the pointer acceleration
init. Properties is the way to go for them.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
97cb3cc28e
dix: Merge DevicePresence notify events generation into a single function.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
7f1ba804a1
dix: s/numEvents/num_events/ in GetKeyboardValuatorEvents
...
GPE and GProxE use the same.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
067f26f7ba
Xi: set the modifier + group state during POE.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
da724e8f1d
Xi: Set the button state on the event during POE.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
b368dd1720
dix: remove a few pointless (void) casts of return values.
...
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-03-12 11:03:29 +10:00
cadf65a6e1
randr: Nuke broken set_origin shortcut
...
Shortcut is impossible to implement this way, because we don't know for sure
whether the crtc of an output has changed or not.
2009-03-10 15:49:49 +01:00
6544490700
Correct access mode in call to dixLookupWindow() within RRSelectInput.
...
Reported by Alan Coopersmith.
2009-03-09 17:29:24 -04:00
Peter Hutterer
Simon Thum
Maarten Maathuis
Adam Jackson
Ander Conselvan de Oliveira
Matthias Hopf
Eamon Walsh