| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
A request, like input device grabs, may check a request timestamp
against currentTime. It is possible for currentTime to lag a previously
sent event timestamp. If the client makes a request based on such an
event timestamp, the request may fail the validity check against
currentTime unless we always update the time before processing the
request.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
If a touch is physically active, the pointer core state should reflect
that the first button is pressed. Currently, this only occurs when there
are active listeners of the touch sequence. By moving the device state
updating to the beginning of touch processing we ensure it is updated
according to the processed physical state no matter what.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
The current code checks the core event mask as though it were an XI
mask. This change fixes the checks so the proper client and event masks
are used.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
As a special case, if a still physically active pointer emulated touch
has no listeners and the device is explicitly grabbed for pointer
events, create a new dix touch record for the grab only.
This allows for clients to "hand off" grabs. For example, when dragging
a window under compiz the window decorator sees the button press and
then ungrabs the implicit grab. It then tells compiz to grab the device,
and compiz then moves the window with the pointer motion. This is racy,
but is allowed by the input protocol for pointer events when there are
no other clients with a grab on the device.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
The function will be used for building a sprite for pointer emulation
after an explicit device grab. This commit refactors the code so that
TouchBuildSprite will function with any event type and moves the checks
to the caller.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Pointer passive grabs may be changed by the grabbing client. This allows
for a selecting client to change an implicit grab to an active grab,
which is the mechanism used for pop-up windows like application menus.
We need to do the same thing with touches. If the grabbing client is the
owner of a touch sequence, change the listener record to reflect the new
grab. If the grabbing client is not the owner, nothing changes for the
touch.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
See UpdateCurrentTime() for reference. I don't know what bug this might
trigger, but it wouldn't hurt to fix this.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Explicit pointer grabs are placed at the head of the touch listener
array for pointer emulated touches. If the grab is deactivated, we must
remove it from all touches for the device.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
This will be used for accepting and rejecting touches in the future.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
QueryPointer is part of the core protocol. As such, it knows nothing
about touch devices. Touches are converted to button 1 press, pointer
motion, and button 1 release for core clients, so we should ensure the
pointer state mask has button 1 set when XQueryPointer is used.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Events from button-only devices still need coordinates, and they get them
from scale_to_desktop(). Therefore, a dev without valuators is not a bug.
However, a dev with valuators, but less than two of them still is a bug.
This was noticed when unplugging a "Creative Technology SB Arena Headset",
which has some BTNs and some KEYs, but no REL or ABS valuators.
It emits [BTN_3] = 0 on unplug, which would trigger the BUG_WARN.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
|
|
Previously, we only had one idle alarm that was triggered for all devices,
whenever the user used any device, came back from suspend, etc.
Add system SyncCounters for each device (named "DEVICEIDLETIME x", with x
being the device id) that trigger on that device only. This allows for
enabling/disabling devices based on interaction with other devices.
Popular use-case: disable the touchpad when the keyboard just above the
touchpad stops being idle.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: James Jones <jajones@nvidia.com>
|
|
Preparation work for per-device idle counters.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: James Jones <jajones@nvidia.com>
|
|
|
|
For driver debugging, it is helpful to know whether the driver has actually
submitted an event to the server. dtrace hooks can help here.
Note that GetPointerEvents and friends may also be triggered by the server
for other emulated devices, some care must be taken when analysing the
results.
Additional difficulty: proximity events have a run-time assigned type, so
this may make automatic detection a tad harder. If in doubt, go for any
event > 64 since the only two that can have that value are ProximityIn and
ProximityOut.
An example systemtap script is below:
# Compile+run with
# stap -g xorg.stp /usr/bin/Xorg
#
function print_valuators:string(nvaluators:long, mask_in:long, valuators_in:long) %{
int i;
unsigned char *mask = (unsigned char*)THIS->mask_in;
double *valuators = (double*)THIS->valuators_in;
char str[128] = {0};
char *s = str;
#define BitIsSet(ptr, bit) (((unsigned char*)(ptr))[(bit)>>3] & (1 << ((bit) & 7)))
s += sprintf(s, "nval: %d ::", (int)THIS->nvaluators);
for (i = 0; i < THIS->nvaluators; i++)
{
s += sprintf(s, " %d: ", i);
if (BitIsSet(mask, i))
s += sprintf(s, "%d", (int)valuators[i]);
}
sprintf(THIS->__retvalue, "%s", str);
%}
probe process(@1).mark("input__event")
{
deviceid = $arg1
type = $arg2
detail = $arg3
flags = $arg4
nvaluators = $arg5
str = print_valuators(nvaluators, $arg6, $arg7)
printf("Event: device %d type %d detail %d flags %#x %s\n",
deviceid, type, detail, flags, str);
}
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Jeremy Huddleston <jeremyhu@apple.com>
|
|
Regression introduced in 4e52cc0ef48145134cd58d357fb7289e6f8bb709
Raw event values are values as-is from the driver, modified only be
transformation or acceleration. 4e52cc caused the mask to be updated from
relative to absolute coordinates which then got written into the raw events.
Move the raw event update into the respective branches for absolute/relative
events.
X.Org Bug 46976 <http://bugs.freedesktop.org/show_bug.cgi?id=46976>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Tested-by: Sven Arvidsson <sa@whiz.se>
Reviewed-by: Simon Thum <simon.thum@gmx.de>
|
|
getevents.c: In function 'updateSlaveDeviceCoords':
getevents.c:326:15: warning: unused variable 'scr' [-Wunused-variable]
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
ended
All touches of an indirect device, such as a trackpad, are sent to the
same window set. When there are no active touches, a new window set is
created; otherwise, the window set of an existing touch is copied.
The current code checks for any logically active touches. This includes
touches that have physically ended but are still logically active due to
unhandled touch grabs. Instead, we want a new window set whenever there
are no physically active touches.
This change skips over logically active but pending end touches, which
are touches that have physically ended.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
There are a few subtle bugs during startup where IsFloating() returns true
if the device is a master device that is not yet paired with its keyboard
device.
Force IsFloating() to always return FALSE for master devices, that was the
intent after all and any code that relies on the other behaviour should be
fixed instead.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jon TURNEY <jon.turney@dronecode.org.uk>
|
|
master->last.valuators[] is in desktop dimensions, so use those as
rescale axis ranges, not the screen. Otherwise, a rescale on any screen
not the top-left will cause out-of-bounds coordinates which will always
map to the bottom-right screen, causing the device to be stuck on that
screen.
X.Org Bug 46657 <http://bugs.freedesktop.org/show_bug.cgi?id=46657>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
|
|
If the typedef wasn't perfect, indent would get confused and change:
foo = (SomePointlessTypedef *) &stuff[1];
to:
foo = (SomePointlessTypedef *) & stuff[1];
Fix this up with a really naïve sed script, plus some hand-editing to
change some false positives in XKB back.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
|
|
This is strictly the application of the script 'x-indent-all.sh'
from util/modular. Compared to the patch that Daniel posted in
January, I've added a few indent flags:
-bap
-psl
-T PrivatePtr
-T pmWait
-T _XFUNCPROTOBEGIN
-T _XFUNCPROTOEND
-T _X_EXPORT
The typedefs were needed to make the output of sdksyms.sh match the
previous output, otherwise, the code is formatted badly enough that
sdksyms.sh generates incorrect output.
The generated code was compared with the previous version and found to
be essentially identical -- "assert" line numbers and BUILD_TIME were
the only differences found.
The comparison was done with this script:
dir1=$1
dir2=$2
for dir in $dir1 $dir2; do
(cd $dir && find . -name '*.o' | while read file; do
dir=`dirname $file`
base=`basename $file .o`
dump=$dir/$base.dump
objdump -d $file > $dump
done)
done
find $dir1 -name '*.dump' | while read dump; do
otherdump=`echo $dump | sed "s;$dir1;$dir2;"`
diff -u $dump $otherdump
done
Signed-off-by: Keith Packard <keithp@keithp.com>
Acked-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Alan Coopersmith <alan.coopersmith@oracle.com>
|
|
And for such devices simply take the last.valuators[] which must be valid at
all times anyway. UpdateSlaveDeviceCoords takes care of that.
X.Org Bug 38313 <http://bugs.freedesktop.org/show_bug.cgi?id=38313>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
|
|
This reverts commit 2bfb802839688ecf328119c4c6979390fc60348d.
This commit caused a regression.
See: http://xquartz.macosforge.org/trac/ticket/517#comment:10
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
|
|
last.scroll remained on the last-submitted scrolling value but last.valuator
was changed whenever the slave device changed. The first scrolling delta
after a switch was then calculated as (last.scroll - new abs value), causing
erroneous scrolling events.
Test case:
- synaptics with a scrolling method enabled, other device with 3+ axes (e.g.
wacom)
- scroll on touchpad
- use other device
- scroll on touchpad
The second scroll caused erroneous button press/release events.
X.Org Bug 45611 <http://bugs.freedesktop.org/show_bug.cgi?id=45611>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
|
|
|
|
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
This doesn't really implement early accept as it should. Ideally, the
server should send end events to all subsequent touch clients as soon as
an early accept comes in. However, this implementation is still protocol
compliant. We can always improve it later.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
This call was supposed to have no functional changes but in some cases
DeliverDeviceEvents() was called with a uninitialised win variable.
Revert, safer than trying to sort this out otherwise.
This reverts commit 6eff14a789341d366b3013c5aa020e959c954651.
Reported-by: Mathieu Taillefumier <mathieu.taillefumier@free.fr>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
|
|
Rename functions/macros from list_* to xorg_list_*
Rename struct from struct list to struct xorg_list.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
In-sed-I-trust: Peter Hutterer <peter.hutterer@who-t.net>
|
|
For expediency, it made sense to always have the X and Y axes set for
direct touch device event propagation. The last X and Y values are
stored internally. However, indirect device touch event propagation
does not depend on the touch's X and Y values. Thus, we don't need to
set the values for every indirect touch event.
On top of this, the previous X and Y values aren't stored for indirect
touches, so without this change the axes get erroneously set to 0.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
The kill_client argument to UngrabAllClients specifies if we want to
kill the client holding the grab or just deactivate the grab.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reported-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
Indirect touch devices provide valuator values in pure device
coordinates. They also don't need to be fixed up for screen crossings.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
When {XI,X,}AllowEvents is called, the timestamp is compared against the
grab time to ensure that the request pertains to the current grab in the
server. While many clients may use CurrentTime (client-side), the
timestamp of the event causing the grab is also valid.
This change ensures that the server's notion of the grab time is the
time of the event that activated the grab rather than the time that the
grab is actually activated.
This bug was exposed through nested touch then pointer grabs.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
When the screen is restructured, the pointer limits need to be reset for
floating slave devices as well, not just for master pointers. Only skip
devices that don't have a cursor (attached slaves and keyboard)
Bug reproducer: float an absolute slave device, rotate the screen - the
device is now confined to a section of the screen only.
X.Org Bug 43635 <http://bugs.freedesktop.org/show_bug.cgi?id=43635>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
|
|
This reverts commit 78fa121f4097d29458e5453c13473595df06e26e.
ABI change pended for 1.13
|
|
|
|
This reverts commit f04fe06ae244b851b38be824b1a80f2f8a030591.
dixLookupWindow no longer returns BadMatch. No other caller was checking
for it, so this problem is now fixed in the utility function.
Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
dixLookupWindow uses dixLookupDrawable internally, which returns
BadMatch when the XID matches a non-Window drawable. Users
of dixLookupWindow don't care about this, just that it's not
a valid Window.
This is a generalised version of the fix for X.Org Bug 23562,
where GetProperty was incorrectly returning BadMatch. Auditing other
window requests, all that I checked would incorrectly return BadMatch
in these circumstances. An incomplete list of calls that could
incorrectly return BadMatch is: ListProperties, SetSelectionOwner,
{Destroy,Map,Unmap}{,Sub}Window.
None of the callers of dixLookupWindow, except for GetProperty, check
for BadMatch
Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
|
|
The vast vast vast majority of resource lookups are successful. Move some
work to the error paths so we don't punish success.
Before:
40000000 trep @ 0.0009 msec (1109091.3/sec): PutImage 10x10 square
60000000 trep @ 0.0005 msec (2072652.2/sec): ShmPutImage 10x10 square
After:
40000000 trep @ 0.0009 msec (1148346.9/sec): PutImage 10x10 square
60000000 trep @ 0.0005 msec (2091666.1/sec): ShmPutImage 10x10 square
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
Don't call LookupMajorName if the hooks aren't active, it's quite expensive.
Before:
40000000 trep @ 0.0009 msec (1087458.5/sec): PutImage 10x10 square
60000000 trep @ 0.0005 msec (2012238.6/sec): ShmPutImage 10x10 square
After:
40000000 trep @ 0.0009 msec (1109091.3/sec): PutImage 10x10 square
60000000 trep @ 0.0005 msec (2072652.2/sec): ShmPutImage 10x10 square
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
Forwarding proxies like sshd will appear to be local, even though they
aren't really. This leads to weird behaviour for extensions that truly
require running under the same OS services as the client, like MIT-SHM
and DRI2.
Add two new legal values for the initial connection's byteOrder field,
'r' and 'R'. These act like 'l' and 'B' respectively, but have the side
effect of forcing the client to be treated as non-local. Forwarding
proxies should attempt to munge the first packet of the connection
accordingly; older servers will reject connections thusly munged, so the
proxy should fall back to passthrough if the munged connection attempt
fails.
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
