If we end up reading all pending X events in the course of other server
execution, then our notify FD callback won't get invoked and we won't
process them. Fix this by noting that there are queued events in the
block handler, setting the poll timeout to zero and queuing a work
proc to clear the event queue.
v2: use a work proc to clear the event queue rather than doing it in
the block handler directly.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Under glamor, we need to re-create the screen pixmap at the new size
so that we can ask glamor for the associated texture. Fortunately, we
can simply use ephyr_glamor_create_screen_resources to create the new
pixmap.
Because this is being done after the server has started, we need to
walk the window heirarchy and reset any windows pointing at the old
pixmap. I could easily be convinced that this TraverseTree should be
moved to miSetScreenPixmap.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
The screen block and wakeup handlers are the only ones which provide a
well known ordering between the wrapping layers; placing these as
close as possible to the server blocking provides a way for the driver
to control the flow of execution correctly.
Switch the shadow code to run in the screen block handler so that it
now occurrs just before the server goes to sleep.
Switch glamor to call down to the driver after it has executed its own
block handler piece, in case the driver needs to perform additional
flushing work after glamor has called glFlush.
These changes ensure that the following modules update the screen in
the correct order:
animated cursors (uses RegisterBlockAndWakeupHandlers dynamically)
composite (dynamic wrapping)
misprite (dynamic wrapping)
shadow (static wrapping)
glamor (static wrapping)
driver (static wrapping)
It looks like there's still a bit of confusion between composite and
misprite; if composite updates after misprite, then it's possible
you'd exit the block handler chain with the cursor left hidden. To fix
that, misprite should be wrapping during ScreenInit time and not
unwrapping. And composite might as well join in that fun, just to make
things consistent.
[v2] Unwrap BlockHandler in shadowCloseScreen (ajax)
[v3] ephyr: Use screen block handler for flushing changes
ephyr needs to make sure it calls glXSwapBuffers after glamor finishes
its rendering. As the screen block handler is now called last, we have
to use that instead of a registered block/wakeup handler to make sure
the GL rendering is done before we copy it to the front buffer.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
When multiple processes are using GL with indirect rendering a race
condition can make drawables refcount never drop to zero.
This situation could happen when there are many X clients using indirect
GLX:
1 - client1: calls glXMakeCurrent
2 - client2: calls glXMakeCurrent
This is the first context switch for this client. So old_context_tag=0
3 - client1: calls glXRender
For the client, its context is already current.
For the server side lastGLContext points to client2's context.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Guilherme Quentel Melo <gqmelo@gmail.com>
When there aren't any devices, the input thread is going to be pretty
lonely, so don't bother to even start it.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The call to 'InitButtonClassDeviceStruct' which initializes the pointer
buttons only results in the first three buttons being created due to a
hardcoded '3'. In order to expose all the buttons defined in the
btn_labels array, we subtitute 'NBUTTONS' in its place.
Signed-off-by: Jason Gerecke <jason.gerecke@wacom.com>
Reviewed-by: Olivier Fourdan <ofourdan@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Substitute a few errant tab characters with eight spaces to conform to the
prevailing style.
Signed-off-by: Jason Gerecke <jason.gerecke@wacom.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
[fix copied from 40191d82370e in xf86-video-ati]
Without this, we end up setting rotated CRTCs back to their previous
framebuffer right after we perform a rotation. Reproducer:
- Have two monitors connected at the same resolution
- Rotate one monitor from normal straight to inverted
- Watch as the monitor you didn't rotate either freezes or shows intense
flickering
Signed-off-by: Lyude <cpaul@redhat.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
If we're doing reverse-prime; or doing rotation the main fb is not used,
and there is no reason to add it in this case.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
drmmode_set_mode_major() is the only user of drmmode->fb_id and will
create it if necessary.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
This ensures the fb gets re-added when a shared pixmap is re-used for
a second drmmode_set_scanout_pixmap_cpu call.
Note currently the xserver never re-uses a shared pixmap in this way,
so this is mostly a sanity fix.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
drmmode_set_scanout_pixmap_gpu(pix) adds drmmod->fb_id through a call
to drmmode_xf86crtc_resize(), but on a subsequent
drmmode_set_scanout_pixmap_gpu(NULL) it would not remove the fb.
This keeps the crtc marked as busy, which causes the dgpu to not
being able to runtime suspend, after an output attached to the dgpu
has been used once. Which causes burning through an additional 10W
of power and the laptop to run quite hot.
This commit adds the missing remove fb call, allowing the dgpu to runtime
suspend after an external monitor has been plugged into the laptop.
Note this also makes drmmode_set_scanout_pixmap_gpu(NULL) match the
behavior of drmmode_set_scanout_pixmap_cpu(NULL) which was already
removing the fb.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
In prime configurations master's last set time may not be latest
and greatest, adjust it with slaves last set time, pick up greatest
one. Otherwise xserver may end with events which has
lastSetTime < lastConfigTime even if that's not
the case and confuse xrandr client.
[hdegoede: rebase on "xrandrprovider: Do not use separate lists for unbound /
source / offload slaves"]
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
A single provider can be both a offload and source slave at the same time,
the use of seperate lists breaks in this case e.g. :
xrandr --listproviders
Providers: number : 2
Provider 0: id: 0x7b cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 3 outputs: 2 associated providers: 0 name:modesetting
Provider 1: id: 0x46 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 2 outputs: 5 associated providers: 0 name:modesetting
xrandr --setprovideroutputsource 1 0x7b
xrandr --listproviders
Providers: number : 2
Provider 0: id: 0x7b cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 3 outputs: 2 associated providers: 1 name:modesetting
Provider 1: id: 0x46 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 2 outputs: 5 associated providers: 1 name:modesetting
xrandr --setprovideroffloadsink 1 0x7b
xrandr --listproviders
Providers: number : 3
Provider 0: id: 0x7b cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 3 outputs: 2 associated providers: 2 name:modesetting
Provider 1: id: 0x46 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 2 outputs: 5 associated providers: 2 name:modesetting
Provider 2: id: 0x46 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 2 outputs: 5 associated providers: 2 name:modesetting
Not good. The problem is that the provider with id 0x46 now is on both
the output_slave_list and the offload_slave_list of the master screen.
This commit fixes this by unifying all 3 lists into a single slaves list.
Note that this does change the struct _Screen definition, so this is an ABI
break. I do not expect any of the drivers to actually use the removed / changed
fields so a recompile should suffice.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
The flip queue currently only holds events submitted to the driver for
flipping, awaiting the completion notifier. It is short. We therefore
can speed up interrupt processing by keeping the small number of events
ready to be flipped on the end of the flip queue. By appending the
events to the flip_queue in the order that they become ready, we also
resolve one issue causing Present to display frames out of order.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-and-tested-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
With large numbers of queued vblank, the list iteration on every
interupt dominates processing time. If we reorder the list to be in
ascending event order, then not only is also likely to be in order for
notification queries (i.e. the notification will be near the start of
the list), we can also stop iterating when past the target event_id.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-and-tested-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Fixes freedesktop.org bug https://bugs.freedesktop.org/show_bug.cgi?id=67484
If t->bottom is close to MIN_INT, removing top can wraparound, so do the check properly.
A similar fix should also be applied to pixman.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Nice of FreePicture to take a void * instead of a PicturPtr so that
this error wasn't caught by the compiler.
Noticed when resetting the X server left a dangling pixmap around.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
On cursor unrealize, the associated pixmap is destroyed, make sure we
clear the pointer from the private resource and check for the value
being non-null when setting or destroying the cursor.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96246
RRScreenSizeNotify() will update the connection information block, but
if this occurs during initialization before ConnectionInfo is even
initialized, this will lead to a crash.
Simply check for ConnectionInfo prior to update it to avoid the crash.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=95337
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
This partially revert commit 984be78
The rountrip in Xwayland's InitInput() is unlikely the culprit for the
crash reported in bug 95337, even though it's triggered from
InitInput().
Startup goes like this:
xwl_screen_init()
xwl_output_create()
wl_display_roundtrip()
InitInput()
wl_display_roundtrip()
ConnectionInfo initialized
What happens in bug 95337 is that some output data is already available
when we reach InitInput()'s wl_display_roundtrip() and therefore we end
up trying to update the ConnectionInfo's data from RR routines before
ConnectionInfo is actually initialized.
Removing the wl_display_roundtrip() from InitInput() will not fix the
issue (although it would make it less lileky to happen), because
xwl_screen_init() also does a wl_display_roundtrip() after creating the
output, so the race that led to bug 95337 remains.
However, re-setting the xwl_screen->expecting_event to 0 again in
InitInput() still doesn't seem right. so this part is not restored
(thus a partial revert).
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=95337
There are no in-tree consumers of the audit hooks, and they are in any
case redundant with the dtrace dispatch hooks. Neither is there any
in-tree user of the core request dispatch hook. The extension hook is
only used for non-default security cases, but in the absence of LTO we
always have to take the function call into XaceHookDispatch to find out
that there's no callback registered.
Cc: Eamon Walsh <ewalsh@tycho.nsa.gov>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
As documented in xorg.conf(5), a value of ConstantDeceleration between 0
and 1 will speed up the pointer. However, values less than 1 actually
had no effect. Fix this.
Note that this bug only affected "ConstantDeceleration" as configured
through xorg.conf, not "Device Accel Constant Deceleration" as configured
through xinput. The property handler AccelSetDecelProperty() also did
not need to be changed, as it did not limit the values of the property.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=92766
Signed-off-by: Eric Biggers <ebiggers3@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
We want to notice that it's set, but still pass it through to dix.
Return 0 to indicate this.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Can't find any reference of pointer_limbo_window in the code, let's
remove it.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
A new --with-fallback-input-driver=foo option allows selecting a
fallback driver for the server if the driver configured for the device
is not found. Note that this only applies when the device has a driver
assigned and that module fails to load, devices without a driver are
ignored as usual.
This avoids the situation where a configuration assigns e.g. the
synaptics driver but that driver is not available on the system,
resulting in a dead device. A fallback driver can at least provides some
functionality.
This becomes more important as we move towards making other driver true
leaf nodes that can be installed/uninstalled as requested. Specifically,
wacom and synaptics, a config that assigns either driver should be
viable even when the driver itself is not (yet) installed on the system.
It is up to the distributions to make sure that the fallback driver is
always installed. The fallback driver can be disabled with
--without-fallback-input-driver and is disabled by default on non-Linux
systems because we don't have generic drivers on those platforms.
Default driver on Linux is libinput, evdev is the only other serious
candidate here.
Sample log output:
[ 3274.421] (II) config/udev: Adding input device SynPS/2 Synaptics TouchPad (/dev/input/event4)
[ 3274.421] (**) SynPS/2 Synaptics TouchPad: Applying InputClass "touchpad weird driver"
[ 3274.421] (II) LoadModule: "banana"
[ 3274.422] (WW) Warning, couldn't open module banana
[ 3274.422] (II) UnloadModule: "banana"
[ 3274.422] (II) Unloading banana
[ 3274.422] (EE) Failed to load module "banana" (module does not exist, 0)
[ 3274.422] (EE) No input driver matching `banana'
[ 3274.422] (II) Falling back to input driver `libinput'
.. server proceeds to assign libinput, init the device, world peace and rainbows
everywhere, truly what a sight. Shame about the banana though.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This is an ABI break, in that we now pass NULL to a function that hasn't
accepted it before.
Alex Goins had a different patch for this but it wasn't symmetrical, it
freed something in a very different place than it allocated it, this
attempts to retain symmetry in the releasing of the backing bo.
v2: use a new toplevel API, though it still passes NULL to something
that wasn't expecting it.
v3: pass -1 instead of 0.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Alex Goins <agoins at nvidia.com>
Not visible in the patch, but the same stanza is repeated below inside
the #ifdef GLXEXT. There's no reason to bother with checking it if we
built without GLXEXT so remove the unconditional one.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Current Mesa Git master checks that the EGL display actually supports
the API passed to eglBindAPI, which can only succeed after
eglInitialize.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Current Mesa Git master checks that the EGL display actually supports
the API passed to eglBindAPI, which can only succeed after
eglInitialize.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96344
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
The doc text is wrong at this point, input processing isn't going to
vary based on this, so we shouldn't say it does. The only thing this
_does_ get used for is DRI1 SwapBuffers (on everything but savage), and
if you disable it you're not going to get DRI1 at all, so we really
shouldn't even mention it.
Still, leave the option wired up to the parser so we don't break any
DRI1-driver-using setup relying on it being disabled, and so we don't
complain about unused options elsewhere.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Acked-by: Keith Packard <keithp@keithp.com>
This code was broken anyway. Note that DEVICE_OFF would make dmx think
_no_ devices were using SIGIO anymore, which means 'xinput disable' on
your mouse would probably do weird things to your keyboard too. Rather
than try to repair that and keep SIGIO working on this one niche DDX,
just rip it out and use the thread model like everyone else.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Acked-by: Keith Packard <keithp@keithp.com>
As the man page for the latter states:
The effects of signal() in a multithreaded process are unspecified.
We already have an interface to call sigaction() instead, use it.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Timer processing can happen on either the main thread or the input
thread. As a result, it must be done under the input lock.
Signals are unrelated to timers now that SIGIO isn't used for input
processing, so stop blocking signals while processing timers.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Read and dispatch pending Wayland events to make sure we do not miss a
possible reply from the compositor prior to discard a key repeat.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Key repeat is handled by the X server, but input events need to be
processed and forwarded by the Wayland compositor first.
Make sure the Wayland compositor is actually processing events, to
avoid repeating keys in Xwayland while the Wayland compositor cannot
deal with input events for whatever reason, thus not dispatching key
release events, leading to repeated keys while the user has already
released the key.
Bugzilla: https://bugzilla.gnome.org/show_bug.cgi?id=762618
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
The xserver generates the key repeat by itself.
But when used with another server processing inputs first (e.g. a
Wayland compositor), the other server may be busy dealing with some
other things and not queue up key release events in time.
Add a vfunc in XkbSrvInfo to possibly add a check before re-emitting a
keypress event in the AccessX timer handler, so that the key repeat has
a chance to be denied if the server processing the input is not ready.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
In patch 137ac094e7, Adam moved an
expensive call to UpdateCurrentTime out of the main dispatch
loop. That's a good change as the original fix from Chase was a bit
expensive. However, it breaks grab processing and so a couple of the
calls to UpdateCurrenTime need to be removed.
Input event processing can generate a stream of events; a button press
that activates a grab will send a press followed by a sequence of
enter/leave events. All of these should have the same time stamp on
the wire as they occur at the 'same' time.
More importantly, the grab time recorded in the device is pulled from
currentTime after all of the events are delivered, so if currentTime
doesn't match the time in the device event, then future grab
modifications will fail as the time marked in the device will be
'later' than the grab time known to the client (which is defined as
the timestamp from the activating input event).
A bit of history here -- it used to be that currentTime was driven
*entirely* by input events; those timestamps didn't even have to be
related to the system time in any way. Then we started doing ICCCM
stuff and people got confused when PropertyNotify events would have
the same timestamp even when delivered minutes apart because no input
events were delivered.
We added code in the server to go update the time, but only if no
input events were pending (so that the clock "wouldn't" go
backwards). The only places where this is necessary is in request
processing which may generate an event with a timestamp, and there
only at the very top of the request processing code so that the whole
request would be processed at the 'same time', just like events.
cc: Chase Douglas <chase.douglas@canonical.com>
cc: Peter Hutterer <peter.hutterer@who-t.net>
cc: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
If the server calls AbortServer during the first-time initialization
(which can happen if you start the server on an already using
DISPLAY), then the dbus code will shut down and call the notify fd
interface. If the notify fd list hasn't been initialized, the server
will crash.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
The list of input devices may be changed by hotplugging while the
server is active, and those changes may come from either the main
thread or the input thread. That means the list of input devices needs
to be protected by a mutex.
This prevents input drivers from receiving I/O ready callbacks after
removing a device.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
By default the X server will try CLOCK_MONOTONIC_COARSE before
CLOCK_MONOTONIC, while A Wayland compositor may only support getting
their timestamps from the CLOCK_MONOTONIC clock. This causes various
issues since it may happen that a timestamp from CLOCK_MONOTONIC
retrieved before a sending an X request will still be "later" than the
timestamp the X server than gets after receiving the request, due to the
fact that CLOCK_MONOTONIC_COARSE has a lower resolution.
To avoid these issues, make Xwayland always use CLOCK_MONOTONIC, so
that it becomes possible for Wayland compositor only supporting
CLOCK_MONOTONIC and X server to use the same clock.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Tested-by: Olivier Fourdan <ofourdan@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
No longer needed now that xf86CursorResetCursor is getting called for
each CRTC configuration change.
v2: Keep xf86_reload_cursors as a deprecated empty inline function
until all drivers stop calling it. (Adam Jackson)
Reviewed-by: Adam Jackson <ajax@redhat.com>
