xorg-next:xserver-u8500.git
7 years agoRevert "debian/xorg.xonf: Enabled multitouch driver."
Rami Ylimaki [Thu, 18 Mar 2010 06:04:49 +0000 (08:04 +0200)]
Revert "debian/xorg.xonf: Enabled multitouch driver."

This reverts commit 4b216d7c4c62bf7e6f012b541f6c303c5f8ca6b6.

7 years agodebian/changelog: Updated version to 2:1.7.993-12. maemo/2_1.7.993-12
Rami Ylimäki [Thu, 25 Mar 2010 12:27:43 +0000 (14:27 +0200)]
debian/changelog: Updated version to 2:1.7.993-12.

Signed-off-by: Rami Ylimäki <ext-rami.ylimaki@nokia.com>
7 years agoos: Prevent backtrace from being stopped in noreturn functions.
Rami Ylimaki [Wed, 17 Mar 2010 10:16:57 +0000 (10:16 +0000)]
os: Prevent backtrace from being stopped in noreturn functions.

There are two noreturn functions in the X server: FatalError and
AbortServer. Having any of those two functions in the middle of a call
stack will prevent unwinding the program properly and stops the
backtrace at those functions in gdb.

The file containing FatalError and AbortServer, os/log.c, has to be
compiled with the -mapcs-frame option on ARM to get proper
backtraces. Automake imposes its own restrictions on compiling
individual source files with different options. The recommended way to
do this is to put os/log.c into a convenience library and add this
library inside os/libos.la. See the documentation of GNU Automake
manual, version 1.11.1, section 27.8 Per-Object Flags Emulation, for
details.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agodebian/changelog: Updated version to 2:1.7.993-11. maemo/2_1.7.993-11
Rami Ylimaki [Fri, 12 Mar 2010 14:48:27 +0000 (16:48 +0200)]
debian/changelog: Updated version to 2:1.7.993-11.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
7 years agodebian/xorg.xonf: Enabled multitouch driver.
Rami Ylimaki [Fri, 12 Mar 2010 14:45:40 +0000 (16:45 +0200)]
debian/xorg.xonf: Enabled multitouch driver.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
7 years agodebian/control: added dependency on xserver-xorg-input-mtev maemo/2_1.7.993-10
Oliver McFadden [Fri, 5 Mar 2010 14:13:30 +0000 (16:13 +0200)]
debian/control: added dependency on xserver-xorg-input-mtev

7 years agodebian/changelog: debian/control: corrected a missing comma causing build failures. maemo/2_1.7.993-9
Oliver McFadden [Thu, 25 Feb 2010 09:31:09 +0000 (11:31 +0200)]
debian/changelog: debian/control: corrected a missing comma causing build failures.

7 years agodebian/changelog: mi: extended the -nocursor option to disable drawing of the cursor... maemo/2_1.7.993-8
Oliver McFadden [Wed, 24 Feb 2010 08:08:37 +0000 (10:08 +0200)]
debian/changelog: mi: extended the -nocursor option to disable drawing of the cursor sprite.

7 years agomi: extended the -nocursor option to disable drawing of the cursor sprite.
Mcfadden Oliver (Nokia-D/Helsinki) [Wed, 17 Feb 2010 10:38:51 +0000 (11:38 +0100)]
mi: extended the -nocursor option to disable drawing of the cursor sprite.

Signed-off-by: Oliver McFadden <oliver.mcfadden@nokia.com>
7 years agodebian/changelog: Rebased to Freedesktop.org commit db4f676f25c6d8e58263d5151942be730... maemo/2_1.7.993-7
Oliver McFadden [Tue, 23 Feb 2010 18:41:21 +0000 (20:41 +0200)]
debian/changelog: Rebased to Freedesktop.org commit db4f676f25c6d8e58263d5151942be730592d444

7 years agodebian/changelog: debian/rules: always install the Xorg configuration file
Oliver McFadden [Tue, 16 Feb 2010 08:54:10 +0000 (10:54 +0200)]
debian/changelog: debian/rules: always install the Xorg configuration file

7 years agodebian/rules: always install the Xorg configuration file
Oliver McFadden [Mon, 15 Feb 2010 16:23:34 +0000 (18:23 +0200)]
debian/rules: always install the Xorg configuration file

7 years agodebian/changelog: debian/xorg.conf: updated the input devices to use symbolic names.
Oliver McFadden [Mon, 15 Feb 2010 16:01:10 +0000 (18:01 +0200)]
debian/changelog: debian/xorg.conf: updated the input devices to use symbolic names.

7 years agodebian/xorg.conf: updated the input devices to use symbolic names.
Oliver McFadden [Mon, 15 Feb 2010 16:00:42 +0000 (18:00 +0200)]
debian/xorg.conf: updated the input devices to use symbolic names.

7 years agodebian/changelog: debian/xorg.conf: Fix input device configuration.
Oliver McFadden [Thu, 11 Feb 2010 14:32:53 +0000 (16:32 +0200)]
debian/changelog: debian/xorg.conf: Fix input device configuration.

7 years agodebian/xorg.conf: Fix input device configuration.
Ylimaki Rami (EXT-Vincit/Tampere) [Wed, 10 Feb 2010 17:17:58 +0000 (18:17 +0100)]
debian/xorg.conf: Fix input device configuration.

Proper input device configuration is needed for a Dali demo that is
going to be held soon. A manual configuration is provided temporarily
and a proper UDEV based configuration later.

This temporary manual configuration enables the following devices:

touchscreen: Obviously the device is unusable without this. Some minor
fixes to the configuration have been made.

keypad: There is no need to delay internal keyboard support.

camera/switches: X won't be the official source for camera keys and
GPIO switches. Those are provided by libqmsystem for
applications. However, there isn't harm in exposing this device also
from X.

BT device: Until there is no more shortage of prototypes with internal
keyboards, testing HW keyboard functionality is easier with a BT
keyboard. This device doesn't have to be configured, but having it
there makes my life easier until a proper UDEV based automatic
configuration has been integrated.

Reviewed-by: Oliver McFadden <oliver.mcfadden@nokia.com>
Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
7 years agodebian/changelog: Fixes: NB#150409 - smooth X startup without waitx
Oliver McFadden [Thu, 11 Feb 2010 11:27:27 +0000 (13:27 +0200)]
debian/changelog: Fixes: NB#150409 - smooth X startup without waitx

7 years agoOS: Added the -upstart option for compatibility with Upstart's signaling method.
Oliver McFadden [Mon, 8 Feb 2010 15:30:17 +0000 (17:30 +0200)]
OS: Added the -upstart option for compatibility with Upstart's signaling method.

This is very similar to the RunFromSmartParent (implicit) option, except
we do not send the signal to our parent process, but our own process
instead, and that signal is SIGSTOP, not SIGUSR1.

Upstart will detect this, realize that we are ready to accept clients
now, send us SIGCONT and move our job status from SPAWNED to RUNNING.

Signed-off-by: Oliver McFadden <oliver.mcfadden@nokia.com>
7 years agodebian/changelog: Fixes: NB#156263 - X server config parser should not be case sensitive
Oliver McFadden [Wed, 10 Feb 2010 00:26:45 +0000 (02:26 +0200)]
debian/changelog: Fixes: NB#156263 - X server config parser should not be case sensitive

7 years agodebian/changelog: upstream rebase
Tiago Vignatti [Tue, 26 Jan 2010 13:47:22 +0000 (15:47 +0200)]
debian/changelog: upstream rebase

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian/control: xserver now depends xfonts-utils and dri2proto 2.2
Tiago Vignatti [Tue, 26 Jan 2010 14:48:30 +0000 (16:48 +0200)]
debian/control: xserver now depends xfonts-utils and dri2proto 2.2

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian/rules: install xorg.conf only on kernels >= 2.6.32
Oliver Mcfadden [Fri, 22 Jan 2010 12:39:56 +0000 (14:39 +0200)]
debian/rules: install xorg.conf only on kernels >= 2.6.32

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian/changelog: fix touchscreen fd configuration
Tiago Vignatti [Wed, 20 Jan 2010 14:50:48 +0000 (16:50 +0200)]
debian/changelog: fix touchscreen fd configuration

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian: reintroduce manual configuration of input devices
Ylimaki Rami (EXT-Vincit/Tampere) [Wed, 20 Jan 2010 14:40:53 +0000 (15:40 +0100)]
debian: reintroduce manual configuration of input devices

This commit will be re-worked as soon as proper HAL/UDEV configuration
is written. Manual configuration is done temporarily so that a new
kernel can be integrated.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian/changelog: libdrm removal
Tiago Vignatti [Tue, 19 Jan 2010 15:26:57 +0000 (17:26 +0200)]
debian/changelog: libdrm removal

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agodebian: disable libdrm support
Tiago Vignatti [Tue, 26 Jan 2010 13:43:41 +0000 (15:43 +0200)]
debian: disable libdrm support

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agoxfree: dri2: wrap drm bits with macros and change drm_magic type
Tiago Vignatti [Fri, 22 Jan 2010 16:34:36 +0000 (18:34 +0200)]
xfree: dri2: wrap drm bits with macros and change drm_magic type

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agoxfree: dri2: add AuthMagic hook for driver side support
Tiago Vignatti [Fri, 22 Jan 2010 15:58:17 +0000 (17:58 +0200)]
xfree: dri2: add AuthMagic hook for driver side support

With this new hook drmAuthMagic becomes useless and should be deprecated.
You might want to implement AuthMagic on driver side instead.

Attention: ABI being break.

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agoconfigure: introduce --{enable,disable}-libdrm
Tiago Vignatti [Fri, 22 Jan 2010 15:30:48 +0000 (17:30 +0200)]
configure: introduce --{enable,disable}-libdrm

Some drivers use DRI protocol but implement their own kernel rendering
manager. For these drivers, libdrm becomes useless.

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Gaetan Nadon <memsize@videotron.ca>
7 years agodebian: upstart removal
Tiago Vignatti [Tue, 19 Jan 2010 17:46:44 +0000 (19:46 +0200)]
debian: upstart removal

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agoNOKIA: debianization
Tiago Vignatti [Thu, 10 Sep 2009 17:21:40 +0000 (20:21 +0300)]
NOKIA: debianization

Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
7 years agoxselinux: Bump extension minor version.
Eamon Walsh [Wed, 3 Feb 2010 20:40:47 +0000 (15:40 -0500)]
xselinux: Bump extension minor version.

Changes introduced in this version:
- 3 window-related requests now handle pixmaps also.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoxselinux: Rename window-related requests that now support pixmaps.
Eamon Walsh [Wed, 3 Feb 2010 20:38:57 +0000 (15:38 -0500)]
xselinux: Rename window-related requests that now support pixmaps.

Renamed requests:
SetWindowCreateContext -> SetDrawableCreateContext
GetWindowCreateContext -> GetDrawableCreateContext
GetWindowContext -> GetDrawableContext

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoxselinux: Allow GetWindowContext to be used for pixmaps as well.
Eamon Walsh [Wed, 3 Feb 2010 20:08:40 +0000 (15:08 -0500)]
xselinux: Allow GetWindowContext to be used for pixmaps as well.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoxselinux: Remove reference counting calls for SID objects.
Eamon Walsh [Wed, 6 Jan 2010 17:52:51 +0000 (12:52 -0500)]
xselinux: Remove reference counting calls for SID objects.

Starting with libselinux 2.0.86, SID objects are no longer
reference counted and the sidput() and sidget() calls are no-ops.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoDon't print a failure message when XACE denies an input event delivery.
Eamon Walsh [Thu, 22 Oct 2009 01:59:01 +0000 (21:59 -0400)]
Don't print a failure message when XACE denies an input event delivery.

A denial is normal and the behavior should be to drop the event.
Having the log message creates excessive log spam.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoRevert "Remove some debug messages that trigger on XACE event delivery failure."
Eamon Walsh [Thu, 22 Oct 2009 01:46:50 +0000 (21:46 -0400)]
Revert "Remove some debug messages that trigger on XACE event delivery failure."

The log messages still need to be there for non-XACE failures.

This reverts commit 4be354c4c2da5168b302601b91bd80cfaca7e193.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agolibselinux now has a pkgconfig file. Use it.
Eamon Walsh [Wed, 21 Oct 2009 23:28:26 +0000 (19:28 -0400)]
libselinux now has a pkgconfig file.  Use it.

Also remove HAVE_NETLINK_AVC_ACQUIRE_FD tests, because we now
require a version of libselinux that has it.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoxselinux: Allow SetWindowCreateContext to be used for pixmaps as well.
Eamon Walsh [Thu, 15 Oct 2009 21:32:21 +0000 (17:32 -0400)]
xselinux: Allow SetWindowCreateContext to be used for pixmaps as well.

Signed-off-by: Eamon Walsh <ewalsh@tycho.nsa.gov>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agoAlways enable outputs that have been forced on in the configuration file
Simon Farnsworth [Mon, 15 Feb 2010 12:18:00 +0000 (14:18 +0200)]
Always enable outputs that have been forced on in the configuration file

If the user has gone to the effort of manually enabling an output in
the configuration file assume that they know what they're doing.

X.org Bug 14611 <http://bugs.freedesktop.org/show_bug.cgi?id=14611>

Signed-off-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoMerge remote branch 'whot/for-keith'
Keith Packard [Mon, 22 Feb 2010 21:26:06 +0000 (13:26 -0800)]
Merge remote branch 'whot/for-keith'

7 years agoXi: assume BadMode for non-BadMatch errors returned from SetDeviceMode.
Peter Hutterer [Thu, 18 Feb 2010 05:49:09 +0000 (15:49 +1000)]
Xi: assume BadMode for non-BadMatch errors returned from SetDeviceMode.

The XI protocol spec only allows for two errors on the SetDeviceMode
requests: BadMatch or BadMode. BadMode however is a dynamically assigned
extension error and the driver doesn't have access to the actual error
number. Hence, if a SetDeviceMode driver returns an error other than
BadMatch, assume BadMode.

The two exceptions are BadAlloc and BadImplementations, pass these on to the
client (any request is allowed to return either of those).

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoXi: assume BadMode for non-BadMatch errors returned from SetDeviceMode.
Peter Hutterer [Mon, 11 Jan 2010 06:35:36 +0000 (16:35 +1000)]
Xi: assume BadMode for non-BadMatch errors returned from SetDeviceMode.

The XI protocol spec only allows for two errors on the SetDeviceMode
requests: BadMatch or BadMode. BadMode however is a dynamically assigned
extension error and the driver doesn't have access to the actual error
number. Hence, if a SetDeviceMode driver returns an error other than
BadMatch, assume BadMode.

The two exceptions are BadAlloc and BadImplementations, pass these on to the
client (any request is allowed to return either of those).

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
7 years agoRe-enable RECORD extension.
Chris Dekter [Thu, 11 Feb 2010 06:34:09 +0000 (16:34 +1000)]
Re-enable RECORD extension.

RECORD was disabled during the switch to internal events. This patch
modifies the record callback to work with internal events instead of
xEvents. The InternalEvents are converted to core/Xi events as needed.

Since record is a loadable extension, the EventTo* calls must be externed.

Signed-off-by: Chris Dekter <cdekter@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoEXA/mixed: Clean up exaPrepareAccessReg_mixed() a little.
Michel Dänzer [Wed, 17 Feb 2010 10:40:44 +0000 (11:40 +0100)]
EXA/mixed: Clean up exaPrepareAccessReg_mixed() a little.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Acked-by: Maarten Maathuis <madman2003@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoexa/mixed: fix gnome-panel corruption
Maarten Maathuis [Wed, 17 Feb 2010 10:40:43 +0000 (11:40 +0100)]
exa/mixed: fix gnome-panel corruption

- A mapped pixmap can't be used for acceleration, any decent memory manager
will refuse this.
- Source pixmaps migrated with a bounding region are incomplete (from the
gpu point of view), so do the upload unconditionally, instead of just for
deferred destination pixmaps.
- Fixes fd.o bug #26076.

Signed-off-by: Maarten Maathuis <madman2003@gmail.com>
Acked-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoos: Prevent core dump from being truncated.
Rami Ylimaki [Thu, 28 Jan 2010 09:08:34 +0000 (11:08 +0200)]
os: Prevent core dump from being truncated.

The problem fixed by this patch can be reproduced on Linux with the
following steps.
- Access NULL pointer intentionally in ProcessOtherEvent on key press.
- Instead of saving core dump to a file, write it into a pipe.
  echo "|/usr/sbin/my-core-dumper" > /proc/sys/kernel/core_pattern
- Dump the core by pressing a key.

While the core is being dumped into the pipe, the smart schedule timer
will cause a pending SIGALRM. Linux kernel stops writing data to the
pipe when there are pending signals. This causes the core dump to be
truncated. On my system I'm expecting a 6 MB dump but the size will be
60 kB instead. The problem is solved if we block the SIGALRM caused by
expired smart schedule timer.

I haven't been able to reproduce this problem in the following cases.
- Save core dump to a file instead of a pipe.
- kill -SEGV `pidof Xorg`
- Press a key to dump core while gdb is attached to Xorg.
- Give option -dumbSched to Xorg.

Also note that the fix works only when NoTrapSignals has the default
value FALSE. The problem can still be reproduced if error signals
aren't trapped. In addition to pending SIGALRM, there is a similar
problem with pending SIGIO from the keyboard driver during core dump.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoos: Introduce OsAbort for proper core dumps.
Rami Ylimaki [Thu, 28 Jan 2010 09:08:33 +0000 (11:08 +0200)]
os: Introduce OsAbort for proper core dumps.

Signed-off-by: Rami Ylimaki <ext-rami.ylimaki@nokia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoparser: corrected xf86getBoolValue to use case insensitive compare
Oliver McFadden [Wed, 3 Feb 2010 08:05:20 +0000 (10:05 +0200)]
parser: corrected xf86getBoolValue to use case insensitive compare

commit c6e8637e29e0ca11dfb35c02da7ca6002ac8c597 introduced this
regression; it can cause existing config files to be parsed incorrectly.

Acked-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Signed-off-by: Oliver McFadden <oliver.mcfadden@nokia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoSolaris xf86OSRingBell() off-by-one error in filling iov[] array
Alan Coopersmith [Tue, 16 Feb 2010 01:42:11 +0000 (17:42 -0800)]
Solaris xf86OSRingBell() off-by-one error in filling iov[] array

When generating sound buffers for /dev/audio bells, insert waveform
for beep *or* silence, but not both, so we don't write one entry past
the end of the iov buffer when the final bit of soundwave ends up in
the final entry allocated in the iov array.

Fixes OpenSolaris bug 6894890:
 http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6894890

Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoUse C-style comments in x86emu
Matt Turner [Tue, 16 Feb 2010 01:05:24 +0000 (20:05 -0500)]
Use C-style comments in x86emu

Signed-off-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoXQuartz: Fix a possible buffer overrun in quartzAudio
Jeremy Huddleston [Tue, 16 Feb 2010 00:46:21 +0000 (16:46 -0800)]
XQuartz: Fix a possible buffer overrun in quartzAudio

Also dropped deprecated API while there

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
7 years agoXQuartz: clang static analysis fixes
Jeremy Huddleston [Thu, 11 Feb 2010 01:56:30 +0000 (17:56 -0800)]
XQuartz: clang static analysis fixes

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
7 years agoXQuartz: Fix linking (CloseInput())
Jeremy Huddleston [Mon, 15 Feb 2010 23:03:01 +0000 (15:03 -0800)]
XQuartz: Fix linking (CloseInput())

Fixes regresison from d33adcdf03c69407d151e732fa0cf9947151eb19

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
7 years agoudev: Don't filter subsystem "input"
Thomas Jaeger [Mon, 4 Jan 2010 20:00:49 +0000 (15:00 -0500)]
udev: Don't filter subsystem "input"

This allows serial wacom devices to work, whose subsystem is "tty".

Signed-off-by: Thomas Jaeger <ThJaeger@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxfree86: Reorder InputClass option priorities
Dan Nicholson [Sat, 13 Feb 2010 18:40:40 +0000 (10:40 -0800)]
xfree86: Reorder InputClass option priorities

Currently the config and InputClasses are merged together so that the
options from the config backend have the highest priority. This is bad
since it means options such as a default XKB layout set by the backend
cannot be changed by the user.

This patch changes order of precedence to be:

1. xorg.conf
2. xorg.conf.d (later files have higher priority)
3. config backend

In order to allow this ordering, the config parsing has been changed to
read the xorg.conf.d files before xorg.conf. This has the consequence
that the core device picking which looks for the first InputDevice may
not find it in xorg.conf.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoXi: reset the sli pointers after copying device classes. (#25640)
Peter Hutterer [Wed, 13 Jan 2010 05:03:45 +0000 (15:03 +1000)]
Xi: reset the sli pointers after copying device classes. (#25640)

If the indicator flags have the XkbSLI_IsDefault bit set, the indicator map
and names aren't their own bit of memory but rather point into the
device->key->xkbInfo->desc structure. XkbCopySrvLedInfo knows about this and
leaves the pointers alone.

When copying the classes from the slave to the master, these pointers are
copied and still point to the dev->key class of the slave device. If the
slave device is removed, the memory becomes invalid and a call to modify
this data (e.g. XkbSetIndicators) may cause a deadlock.

The copying of dev->key relies on dev->kbdfeed to be already set up. Hence
the pointers need to be reset once _both_ kbdfeed and key have been copied
into the master device.

X.Org Bug 25640 <http://bugs.freedesktop.org/show_bug.cgi?id=25640>
Fedora Bug 540584 <https://bugzilla.redhat.com/show_bug.cgi?id=540584>

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
7 years agodix: move config_init into the DDX.
Peter Hutterer [Fri, 15 Jan 2010 05:22:02 +0000 (15:22 +1000)]
dix: move config_init into the DDX.

The only DDX currently using hotplugging is the xfree86 one and it looks
like it'll stay that way for a bit. Move the initialization to the DDX,
since Xephyr, Xnest, and friends don't need HAL or udev notifications.

Add CloseInput (counterpart to InitInput) to be able to clean up the config
initialization from the DDX as well.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
7 years agodix: restore lastDeviceEventTime update in dixSaveScreens
Julien Cristau [Fri, 12 Feb 2010 22:34:57 +0000 (23:34 +0100)]
dix: restore lastDeviceEventTime update in dixSaveScreens

This was removed in 6b5978dcf1f7ac3ecc2f22df06f7000f360e2066 (Do not
reset lastDeviceEventTime when we do dixSaveScreens), but caused a
regression for XResetScreenSaver.  Add the lastDeviceEventTime update
back, but restrict it to that case.

X.Org bug#25855 <http://bugs.freedesktop.org/25855>

Reported-by: Lubos Lunak <l.lunak@suse.cz>
Tested-by: Lubos Lunak <l.lunak@suse.cz>
Signed-off-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoDon't double-swap the RandR PropertyNotify event
Peter Harris [Fri, 12 Feb 2010 20:36:30 +0000 (15:36 -0500)]
Don't double-swap the RandR PropertyNotify event

The event is already swapped in randr.c/SRROutputPropertyNotifyEvent, so
it should not be swapped here.

X.Org Bugzilla #26511: http://bugs.freedesktop.org/show_bug.cgi?id=26511

Tested-by: Leonardo Chiquitto <leonardo@ngdn.org>
Acked-by: Adam Jackson <ajax at redhat.com>
Reviewed-by: Julien Cristau <jcristau at debian.org>
Signed-off-by: Peter Harris <pharris@opentext.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoxfree86: Add qxl driver to the autoconfig logic
Soeren Sandmann [Fri, 12 Feb 2010 19:04:53 +0000 (20:04 +0100)]
xfree86: Add qxl driver to the autoconfig logic

The qxl driver is for the QXL virtualized graphics device.

Signed-off-by: Søren Sandmann Pedersen <ssp@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
7 years agoBump to 1.7.99.901 -- 1.8 RC1 xorg-server-1.7.99.901
Keith Packard [Fri, 12 Feb 2010 21:46:03 +0000 (13:46 -0800)]
Bump to 1.7.99.901 -- 1.8 RC1

7 years agoMerge remote branch 'jturney/master'
Keith Packard [Fri, 12 Feb 2010 21:36:32 +0000 (13:36 -0800)]
Merge remote branch 'jturney/master'

7 years agoAdd tag matching to input attributes.
Peter Hutterer [Wed, 10 Feb 2010 05:36:50 +0000 (15:36 +1000)]
Add tag matching to input attributes.

Tags may be a list of comma-separated strings that match against a MatchTag
InputClass section. If any of the tags specified for a device match against
the MatchTag of the section, this match is evaluated true and passed on to
the next match condition.

Tags are specified as "input.tags" (hal) or "ID_INPUT.tags" (udev), the
value of the tags is case-sensitive and require an exact match (not a
substring match).

i.e. "quirk" will not match "QUIRK", "need_quirk" or "quirk_needed".

Example configuration:
udev:
    ENV{ID_INPUT.tags}="foo,bar"

hal:
    <merge key="input.tags" type="string">foo,bar</merge>

xorg.conf:
    Section "InputClass"
            Identifier "foobar quirks"
            MatchTag "foo|foobar"
            Option "Foobar" "on"
    EndSection

Where the xorg.conf section matches against any device with the tag "foo"
or tag "foobar" set.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
7 years agoAdd xstrtokenize to the dix.
Peter Hutterer [Wed, 10 Feb 2010 05:36:49 +0000 (15:36 +1000)]
Add xstrtokenize to the dix.

Move tokenize out of the parser, make it a dix util function instead.
Splitting a string into multiple substrings is useful by other places, so
let's use it across the line. Future users include config/hal, config/udev
and of course the parser.

Example usage:
char **substrings = xstrtokenize(my_string, "\n");

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
7 years agoxfree86: Set fnmatch pathname flag for InputClass device matching
Dan Nicholson [Wed, 10 Feb 2010 05:36:48 +0000 (15:36 +1000)]
xfree86: Set fnmatch pathname flag for InputClass device matching

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxfree86: Allow multiple arguments to InputClass matches
Dan Nicholson [Wed, 10 Feb 2010 05:36:47 +0000 (15:36 +1000)]
xfree86: Allow multiple arguments to InputClass matches

In order to keep the number of InputClass sections manageable, allow
matches to contain multiple arguments. The arguments will be separated
by the '|' character. This allows a policy to apply to multiple types of
devices. For example:

    Section "InputClass"
        Identifier "Inverted Mice"
        MatchProduct "Crazy Mouse|Silly Mouse"
        Option "InvertX" "yes"
    EndSection

This applies to the MatchProduct, MatchVendor and MatchDevicePath
entries. Currently there is no way to escape characters, so names or
patterns cannot contain '|'.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxfree86: Use "Ignore" option in InputClass to skip devices
Dan Nicholson [Wed, 10 Feb 2010 05:36:46 +0000 (15:36 +1000)]
xfree86: Use "Ignore" option in InputClass to skip devices

Sometimes it is desirable to skip adding specific input devices to the
server. The "Ignore" option is used similarly to Monitor sections so
that matched devices will not be added. BadIDChoice is returned to the
config backend so that it will clean up all resources.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxfree86: Handle config files ending without newline
Dan Nicholson [Wed, 10 Feb 2010 05:36:45 +0000 (15:36 +1000)]
xfree86: Handle config files ending without newline

The config parser expects to find a newline at the end of each line, so
files ending without one would confuse it. A newline is inserted at the
end of the buffer in these situations. Additionally, switching to the
next config file is moved to the higher level to allow parsing of the
last line of the previous file to complete before shifting the index and
resetting the line number.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Tested-by: Stephan Raue<stephan.raue@gmx.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxfree86: Make InputClass docs and comments match reality
Dan Nicholson [Wed, 10 Feb 2010 05:36:44 +0000 (15:36 +1000)]
xfree86: Make InputClass docs and comments match reality

Drivers and options specified in InputClass sections work on a "first
match wins" strategy. Let's be consistent when documenting it.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoAdd labels for multitouch valuators
Benjamin Tissoires [Mon, 8 Feb 2010 12:50:43 +0000 (13:50 +0100)]
Add labels for multitouch valuators

Thoses definitions have been included in the kernel but the X server is not updated accordingly.
Without these definitions, the multitouch axes are not correctly labelled.

Signed-off-by: Benjamin Tissoires <tissoire@cena.fr>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoCygwin/X: Avoid a collision between DEBUG and a token name
Jon TURNEY [Tue, 2 Feb 2010 20:53:44 +0000 (20:53 +0000)]
Cygwin/X: Avoid a collision between DEBUG and a token name

Rename a token to avoid a collision between DEBUG defined via AC_DEFINE
if --enable-debug is configured, and the token for the 'debug' instruction
in the XWin preferences file

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Avoid cursor size log spam
Jon TURNEY [Sun, 22 Nov 2009 23:35:09 +0000 (23:35 +0000)]
Cygwin/X: Avoid cursor size log spam

Fedora 12 likes to use a 39x26 animated wait cursor.  Avoid spamming
the log with warnings that each frame can't be completely contained
in the 32x32 native cursor

Also reformat log message so it doesn't contain a '\n\t'. I mean,
it's not like we might want to grep the log or something...

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoRepair '-nolock'
Jon TURNEY [Sat, 28 Nov 2009 17:51:33 +0000 (17:51 +0000)]
Repair '-nolock'

commit 446fe9eecddd1337f9d5164dd7c301e1ba3dfe32 removes the AC_DEFINE for
SERVER_LOCK and conditional compilation checking it, making it always on
everywhere, except in os/utils.c where code is left under SERVER_LOCK, which
now never gets built, making the '-nolock' option non-functional...

This seems to have been broken since Xserver 1.7.0, but this option is
actually of some slight use on cygwin, as if /tmp resides on a FAT filesystem
(yes, I know...), hard links aren't supported.

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Make transient windows resizable again
Jon TURNEY [Fri, 16 Jan 2009 13:12:12 +0000 (13:12 +0000)]
Cygwin/X: Make transient windows resizable again

Reverts the change which makes parented windows non-resizeable

Because this was trying to do something which we should be doing, as an alternative we
try checking WM_NORMAL_HINTS for windows which shouldn't  be resizable

If a window has a maximum size specified, no maximize box
If a window has a fixed size (max size == min size, per EWMH 1.3 Implementation Notes), no resizing frame

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Discourage other WMs in multiwindow mode
Jon TURNEY [Tue, 10 Nov 2009 13:24:40 +0000 (13:24 +0000)]
Cygwin/X: Discourage other WMs in multiwindow mode

Tidying up of other WM detection code when ading SWT/Motif
fix in commit 71519a572fe15b85c0eb2b02636c9e871f2c858f
was rather over-agressive and now allows other WMs to think
they can start when the internal WM is running.

Revert to the behaviour of selecting ButtonPressMask events
as well on the root window, so other WMs will be dissuaded
from starting

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Update XWin man page
Jon TURNEY [Sat, 31 Oct 2009 17:19:17 +0000 (17:19 +0000)]
Cygwin/X: Update XWin man page

Restructure to group similar options and offer some commentary on those groups
Review option descriptions, clarify and bring up-to-date

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Update DDX specific -help text
Jon TURNEY [Mon, 2 Nov 2009 17:37:14 +0000 (17:37 +0000)]
Cygwin/X: Update DDX specific -help text

Alphabetize options
Review option descriptions, clarify and bring up to date
Add missing option descriptions

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: XSupportsLocale() failure is non-critical
Jon TURNEY [Thu, 29 Oct 2009 18:50:23 +0000 (18:50 +0000)]
Cygwin/X: XSupportsLocale() failure is non-critical

Treat XSupportsLocale() returning false as non-critical to internal client
theads startup, and issue a warning, not an error

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Report BUILDERSTRING with version information
Yaakov Selkowitz [Mon, 1 Feb 2010 17:13:00 +0000 (17:13 +0000)]
Cygwin/X: Report BUILDERSTRING with version information

Report BUILDERSTRING with version information

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoDefine FD_SETSIZE on Cygwin regardless of XWin DDX
Yaakov Selkowitz [Mon, 1 Feb 2010 17:12:57 +0000 (17:12 +0000)]
Define FD_SETSIZE on Cygwin regardless of XWin DDX

All DDXs segfault on Cygwin unless -DFD_SETSIZE=256 is set, so make sure
we do so whether or not we are building XWin.

FD_SETSIZE must be at least XFD_SETSIZE for uses of select() to be correct.
The Cygwin default is only 64, so it must be increased to 256

Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Copy the state of the Windows keyboard device to the Virtual Core Keyboard...
Jon TURNEY [Fri, 19 Jun 2009 20:14:47 +0000 (21:14 +0100)]
Cygwin/X: Copy the state of the Windows keyboard device to the Virtual Core Keyboard at startup.

Otherwise, this happens lazily after the first keypress, which can lead
to applications which are started from a shell window and inspect the
keyboard state before a character is typed getting the wrong idea about
the desired keymap (e.g. xemacs shows this behaviour)

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Tidy up system.Xwinrc
Jon TURNEY [Sun, 4 Oct 2009 14:55:40 +0000 (15:55 +0100)]
Cygwin/X: Tidy up system.Xwinrc

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Look for system.Xwinrc in SYSCONFDIR/X11
Yaakov Selkowitz [Mon, 1 Feb 2010 17:11:54 +0000 (17:11 +0000)]
Cygwin/X: Look for system.Xwinrc in SYSCONFDIR/X11

Look for system.Xwinrc in SYSCONFDIR/X11 (usually /etc/X11)
Rename sample system.Xwinrc file not to have a X11R6 path in it's name
Add makefile install rule for system.XWinrc

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Update resources file and About dialog
Yaakov Selkowitz [Tue, 2 Feb 2010 16:21:44 +0000 (16:21 +0000)]
Cygwin/X: Update resources file and About dialog

Use the configured vendor web address in the About dialog

Update resources file: rework About dialog, use 'MS Shell Dlg 2'
logical font for all dialogs, add ellipsis to exit option in tray
menu as it (may) trigger another dialog.

Get short vendor name from xwin-config.h, like other configuration
data presented in the About dialog box, rather than creating the
PROJECT_NAME define

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoXming: Use 8 point font for URL in About dialog
Colin Harrison [Tue, 2 Feb 2010 16:22:39 +0000 (16:22 +0000)]
Xming: Use 8 point font for URL in About dialog

Use an 8 point font for URL in About dialog, to match the rest
of the text in that dialog

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
7 years agoXming: Add a manifest file
Colin Harrison [Tue, 2 Feb 2010 16:08:45 +0000 (16:08 +0000)]
Xming: Add a manifest file

Use manifests to enable XP style controls
(only effective for XP and later and when themes are enabled).
The addition of manifests shouldn't cause compatibility problems with older Windows versions.

Manifest must have execute permissions, otherwise attempts to
execute XWin.exe in the same directory will fail...

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
7 years agoCygwin/X: Window placement refinement for multiwindow mode
Jon TURNEY [Fri, 19 Jun 2009 20:14:39 +0000 (21:14 +0100)]
Cygwin/X: Window placement refinement for multiwindow mode

Window placement refinement for multiwindow mode, ensure a window actually
ends up somewhere visible if it tries to create itself offscreen (which
can happen for e.g. if it has a stored position from a different sized
display)

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoCygwin/X: Respect the system's ownership of the clipping region
Jon TURNEY [Tue, 2 Feb 2010 15:42:12 +0000 (15:42 +0000)]
Cygwin/X: Respect the system's ownership of the clipping region

Respect the system's ownership of the clipping region used for shaped windows

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
7 years agoXming: Always update the Windows title Unicode (wide-character) in -multiwindow mode.
Colin Harrison [Mon, 1 Feb 2010 17:11:22 +0000 (17:11 +0000)]
Xming: Always update the Windows title Unicode (wide-character) in -multiwindow mode.

Apply the Windows title wide-character in -multiwindow mode. Windows should now
display correct client X Window titles for locales with wide characters.

Copyright (C) Colin Harrison 2005-2008
http://www.straightrunning.com/XmingNotes/
http://sourceforge.net/projects/xming/

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
7 years agomain: move config_init() after InitInput()
Julien Cristau [Mon, 11 Jan 2010 00:31:27 +0000 (00:31 +0000)]
main: move config_init() after InitInput()

With the udev backend, config_init() calls NewInputDeviceRequest(),
which enables devices.  They can then start sending events, even though
the event queue is only initialized later in InitInput().  Oops.

Debian bug#564256 <http://bugs.debian.org/564256>

Reported-by: Sedat Dilek <sedat.dilek@googlemail.com>
Signed-off-by: Julien Cristau <jcristau@debian.org>
Tested-by: Sedat Dilek <sedat.dilek@googlemail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoDon't use AC_CHECK_FILE for fontpath checks when cross compiling
Dan Nicholson [Thu, 14 Jan 2010 06:42:16 +0000 (16:42 +1000)]
Don't use AC_CHECK_FILE for fontpath checks when cross compiling

AC_CHECK_FILE chokes when cross compiling, so instead we set the default
to the standard FONTROOTDIR directories in that case.

Signed-off-by: Dan Nicholson<dbn.lists@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxkb: sed True -> TRUE and False -> FALSE
Peter Hutterer [Fri, 15 Jan 2010 00:07:38 +0000 (10:07 +1000)]
xkb: sed True -> TRUE and False -> FALSE

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>
7 years agoxkb: Add XKM file format description.
Peter Hutterer [Fri, 18 Dec 2009 10:07:30 +0000 (20:07 +1000)]
xkb: Add XKM file format description.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Dan Nicholson <dbn.lists@gmail.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
7 years agoAllow driver to call DeleteInputDeviceRequest during UnInit
Oldřich Jedlička [Sun, 17 Jan 2010 16:59:03 +0000 (17:59 +0100)]
Allow driver to call DeleteInputDeviceRequest during UnInit

When the input driver (like xf86-input-wacom) removes it's devices
during a call to UnInit, the CloseDownDevices() cannot handle it. The
"next" variable can become a pointer to freed memory.

The patch introduces order-independent device freeing mechanism by
remembering the already freed device ids. The devices can reorder any
time during freeing. No device will be double-freed - if the removing
failed for any reason; some implementations of DeleteInputDeviceRequest
don't free the devices already.

Signed-off-by: Oldřich Jedlička <oldium.pro@seznam.cz>
Reviewed-by: Simon Thum <simon.thum@gmx.de>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoxkb: make ctrl+alt+keypad + / ctrl+alt+keypad - work again (#25743)
Horst Wente [Wed, 30 Dec 2009 18:35:20 +0000 (19:35 +0100)]
xkb: make ctrl+alt+keypad + / ctrl+alt+keypad - work again (#25743)

Video mode switching via keypad keys did not work

Signed-off-by: Horst Wente <horst.wente@acm.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
7 years agoFix typo in updateSlaveDeviceCoords
Oldřich Jedlička [Tue, 12 Jan 2010 18:15:06 +0000 (19:15 +0100)]
Fix typo in updateSlaveDeviceCoords

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>
7 years agoXQuartz: Attatch a stub display when CoreGraphics reports no displays.
Jeremy Huddleston [Sat, 30 Jan 2010 22:49:02 +0000 (14:49 -0800)]
XQuartz: Attatch a stub display when CoreGraphics reports no displays.

This is half of the required changes to address the "stuck mouse pointer"
bug that occurs when X11 launches while the displays are asleep.  The
remainder of the fix is part of libXplugin which needs to be updated to
send XP_EVENT_DISPLAY_CHANGED when the display wakes up.  If you don't
have a recent enough libXplugin (expected in 2.5.0_beta2 or later), you
can cause this event to be sent by changing your display resolution (or
you could just start X11.app with your screens awake).

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
7 years agoXQuartz: Dead code removal
Jeremy Huddleston [Thu, 28 Jan 2010 01:12:12 +0000 (17:12 -0800)]
XQuartz: Dead code removal

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
7 years agoXQuartz: Add some .gitignore magic
Jeremy Huddleston [Sat, 30 Jan 2010 18:55:38 +0000 (10:55 -0800)]
XQuartz: Add some .gitignore magic

Signed-off-by: Jeremy Huddleston <jeremyhu@freedesktop.org>