fromDIX is neither exactly true nor particularly helpful in understanding
what this parameter triggers. Rename to set_dequeue_screen, because that's
exactly what happens.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Passive core grabs are mostly device-independent. In an MPX scenario, they
may change to reflect whichever master pair activated the grab last. For
adding new grabs to the list, ignore the device for core grabs to return
failures when trying to set the same grab combo twice on a window.
X.Org Bug 39545 <http://bugs.freedesktop.org/show_bug.cgi?id=39545>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Core grabs may change device when they're activated to reflect the master
they apply to. If the device is a keyboard, modifierDevice is erroneously
set to the Virtual Core Pointer.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
_source was being allocated manually, with all other options added to that
list through add_option. Skip the manual part, allocate the first option
_source with add_option too.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Change add_option to return the new InputOption on success, or NULL
failure. This way we can at least check for errors in callers.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
PRODUCT was taken from the parent, hence ppath.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
This reverts commit 43d9edd31e.
This commit was introduced in the 1.2 cycle when hotplugging was less than
ideal (i.e. it didn't exist). From the commit message:
Always add a mouse driver instance configured to send core events, unless
a core pointer already exists using either the mouse or void drivers. This
handles the laptop case where the config file only specifies, say,
synaptics, which causes the touchpad to work but not the pointing stick.
We don't double-instantiate the mouse driver to avoid the mouse moving twice
as fast, and we skip this logic when the user asked for a void core pointer
since that probably means they want to run with no pointer at all.
To get this case above, a user would need to disable hotplugging _and_ have a
xorg.conf that only references one device. This is possible, but not a use-case
we should worry about too much now.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Add support for multi-seat-aware input device hotplugging. This
implements the multi-seat scheme explained here:
http://www.freedesktop.org/wiki/Software/systemd/multiseat
This introduces a new X server switch "-seat" which allows configuration
of the seat to enumerate hotplugging devices on. If specified the value
of this parameter will also be exported as root window property
Xorg_Seat.
To properly support input hotplugging devices need to be tagged in udev
according to the seat they are on. Untagged devices are assumed to be on
the default seat "seat0". If no "-seat" parameter is passed only devices
on "seat0" are used. This means that the new scheme is perfectly
compatible with existing setups which have no tagged input devices.
Note that the -seat switch takes a completely generic identifier, and
that it has no effect on non-Linux systems. In fact, on other OSes a
completely different identifier scheme for seats could be used but still
be exposed with the Xorg_Seat and -seat.
I tried to follow the coding style of the surrounding code blocks if
there was any one could follow.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Slowly merging the vastly different code-paths.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
No functional changes, just readability improvements. This also gets rid of
the count variable. Count was just used for resizing the null-terminated
list. Since we're not in a time-critical path here at all we can afford to
loop the list multiple times instead of keeping an extra variable around.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
If we find the core device, move all other device pointers forward right
then and there. The break will jump out of the top loop.
They had a special on braces today, so I added some for readability (and
fixed up tab vs space indentation.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
No functional changes.
The options we assign are the ones from the Pointer/Keyboard device so we
might as well use those readable names instead of dev[count-1]->options.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Devices are core pointers/keyboards by default now anyway, but let's set the
option to some value instead of just NULL.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
It doesn't matter. All devices are core pointer devices by default now
anyway.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
LogHdrMessageVerb allows passing a parameterized header to insert in a log
message between MessageType and the formatted message body string.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
LogHdrMessageVerb allows passing a parameterized header to insert in a log
message between MessageType and the formatted message body string.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
LogVHdrMessageVerb allows a custom header to be inserted in a log message,
between the Log system's MessageType string, and a formatted variable
message body. The custom header can itself be a formatted variable string.
These functions can be used, for example, by driver abstraction layers to
format specific driver messages in a standard format, but do it in a way
that is efficient, obeys the log-layers verbosity settings, and is safe
to use in signal handlers (because they don't call malloc), even for
types besides X_NONE.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Also, optimize how the type and format strings are combined.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Guillem Jover <guillem@hadrons.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The memcpy fast path implicitly assumes that the copy walks
left-to-right. That's not something memcpy guarantees, and newer glibc
on some processors will indeed break that assumption. Since we walk a
line at a time, check the source and destination against the width of
the blit to determine whether we can be sloppy enough to allow memcpy.
(Having done this, we can remove the check for !reverse as well.)
On an Intel Core i7-2630QM with an NVIDIA GeForce GTX 460M running in
NoAccel, the broken code and various fixes for -copywinwin{10,100,500}
gives (edited to fit in 80 columns):
1: Disable the fastpath entirely
2: Replace memcpy with memmove
3: This fix
4: The code before this fix
1 2 3 4 Operation
------ --------------- --------------- --------------- ------------
258000 269000 ( 1.04) 544000 ( 2.11) 552000 ( 2.14) Copy 10x10
21300 23000 ( 1.08) 43700 ( 2.05) 47100 ( 2.21) Copy 100x100
960 962 ( 1.00) 1990 ( 2.09) 1990 ( 2.07) Copy 500x500
So it's a modest performance hit, but correctness demands it, and it's
probably worth keeping the 2x speedup from having the fast path in the
first place.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Otherwise sys_pitch will be stale when a system memory copy is allocated.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=38322 and a crash when
unlocking the screen with xscreensaver, reported by Janne Huttunen.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Janne Huttunen <jahuttun@gmail.com>
Tested-by: Jan Kriho <Erbureth@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Fixes assertion failure when calling dixSetPrivate
Debian bug#632549 <http://bugs.debian.org/632549>
Reported-and-tested-by: Mohammed Sameer <msameer@foolab.org>
Signed-off-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
The composite extension spec says that window background painting
should be inhibited when the subwindow redirection mode is set to
manual.
This eliminates the ugly flashing effect when compiz unredirects a
fullscreen window.
Signed-off-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Owen Taylor <otaylor@fishsoup.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This reverts commit db8840600e.
It was an optimization for the resize case, but 193ecc8b45 made
it so that no backfilling takes place on resize if left in.
Signed-off-by: Pierre-Loup A. Griffais <pgriffais@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Conflicts:
composite/compalloc.c
Signed-off-by: Keith Packard <keithp@keithp.com>
RRPostPendingProperties tries to compare the pending and current
property values to decide whether they're actually changing. However,
it does this using a memcmp that passes in pending_value->size as the
number of bytes. This is actually the number of elements, where each
element is (pending_value->format / 8) bytes long. This causes the
pending value to not be propagated if the first pending_value->size
bytes are the same and only the end of it is changing.
Fix this by computing the total number of bytes to compare in the
memcmp.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Instead of just closing the log when everything is done, put one more
message in stating that we're actually terminating. Users or scripts that
look at the Xorg.log will then know that a) the server has terminated
properly and b) why the server terminated (to some degree, given that most
real-world errors will be caused by AbortServer()).
Acked-by: Gaetan Nadon <memsize@videotron.ca>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If config/udev was enabled, this would default to base, which means that
after regen the devices would get the wrong rules, and hilarity would
ensue.
It's probably safe to default to evdev unconditionally on Linux by now.
Reported-by: Bastian Blank <waldi@debian.org>
Signed-off-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
For hotplugged devices, xf86AllocateInput does that for us but the xorg.conf
path is different. Since not all drivers reset the fd during PreInit but may
still call close(pInfo->fd) in all cases, this can terminate the logging
early.
Reproducible: add a wacom driver InputDevice section with no Option Device.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
xf86ConfigLayout.inputs contains the information from the xorg.conf
file. Passing this into xf86NewInputDevice means the device will get
cleaned up on exit and the pointers in xf86ConfigLayout.inputs are left
dangling. In the second server generation, this results in a server
crash.
Also, rename pDev to pInfo. pDev is pretty much reserved for DeviceIntPtr
types.
Reproducible: AutoAddDevices off and xorg.conf input sections, trigger
server regeneration.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Devices that succeeded during PreInit and DEVICE_INIT but failed in
DEVICE_ON would be deleted through xf86DeleteInput but not removed from the
list of input devices (and not turned off). The result was a double free on
server shutdown.
Fix this by calling RemoveDevice if EnableDevice fails.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
The nature of hotplug is that a device we enumerated might already be
gone by the time we look at it, so don't assume otherwise.
Signed-off-by: Lennart Poettering <lennart@poettering.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Don't enumerate/monitor all devices of the system (since that can be
quite a few), but limit our search to devices from the "input"
subsystem, as well as the "tty" subsystem (to cover Wacom tablets).
This should make X start up a bit faster and reduce the number of
unnecessary wake-ups of the X server.
Signed-off-by: Lennart Poettering <lennart@poettering.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
udev gives no guarantee that before each "changed" event for a device
there's an "add" event, or that before each "remove" is an "add", or
that before each "add" there was no "add" already and so on. Users can
trigger these events at any time with "udevadm trigger", and netlink is
a lossy transport, hence the events can come in unexpected ordering.
With other words: regardless which event is generated, the X server must
not choke on it and make the best of it, hence make sure that if we get
an "add" event for an existing device we don't add the device a second
time.
Signed-off-by: Lennart Poettering <lennart@poettering.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This preference allows users to override the related option in Mac OS X's
Mouse/Trackpad preferences. This effectively lets the user determine
which "context" all of X11 fits into for context-based scrolling until
such API exists within X11 itself to pass along to X11 clients.
This is applicable to Mav OS X 10.7+
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
This occurred to me in hindsight after the last commit. If the
original developer had done this, we would have noticed the
problem sooner.
(cherry picked from commit aa0a57996f3e7d16238f69976958c2526821388b)
xp_destroy_surface was called with a surface id of 0, due to some
premature cleanup that set it to 0. This means the surfaces
weren't being destroyed until the window was.
The code that did that was: pDRIDrawablePriv->sid = 0;
In long running applications this leak may or may not have been
harmful. With the old libGL the surfaces weren't destroyed until
the context was destroyed or a new context created. In the new
libGL they are reference counted, and released much sooner, so we
ran into a resource leak more noticeably with some tests.
Make the Apple DRI code dispatch events to the client(s) for
destroyed surfaces, when a resource is destroyed. This seems to
work in my tests, however this clearly wasn't working for a while,
so bugs may result in the future if it enables some new (unexpected)
side effects.
Also add a few helpful comments to aid in understanding the code
in the future.
Tested with the test suite, Pymol, and various Mesa demos.
(cherry picked from commit bede83eb19a1629396fcd5a46441f8476a8fcd1b)
Also add some comments that weren't merged in from server-1.4-apple's
99babae1326485c27eb9253db83afdd6aef9e362
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Peter Hutterer
Lennart Poettering
Daniel Kurtz
Adam Jackson
Pelle Johansson
Michel Dänzer
Julien Cristau
Keith Packard
Ville Syrjälä
Pierre-Loup A. Griffais
Jeremy Huddleston
Aaron Plattner
Jon TURNEY
George Staplin