Handle -displayfd and an explicit display number sensibly, e.g. use the
explicitly specified display number, and write it to the displayfd
v2: displayfd might be 0, so use -1 as invalid value
v3: Rebase for addition of NoListenAll flag
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
miOpqStipDrawable resets the stipple after painting. When that stipple
was NULL, ChangeGC needs to handle that and not crash.
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-off-by: Eric Anholt <eric@anholt.net>
Whenever the master changes, push the locked modifier state to the attached
slave devices, then update the indicators. This way, when NumLock or CapsLock
are hit on any device, the LED will light up on all devices. Likewise, a new
keyboard attached to a master device will light up with the correct
indicators.
The indicators are handled per-keyboard, depending on the layout, i.e. if one
keyboard has grp_led:num set, the NumLock LED won't light up on that keyboard.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
If there is a selection left over from a previous execution of the
main loop, and that selection has privates allocated for it, the X
server will crash. This is because dixResetPrivates() resets the
privates refcounts to zero without accounting for the reference held
by the selection object. When the selection is then deleted in
InitSelections() after the call to dixResetPrivates(), the refcount
for its privates type goes negative and bad things happen.
To fix this, we should delete any existing selections before calling
dixResetPrivates(). This will properly release the selection's
privates and avoid the crash.
A more thorough description of the problem and a test case to
reproduce the crash is available at a previous mail:
"Negative Selection devPrivates refcount?"
By Andrew Eikum to xorg-devel on 10 Dec 2013
http://lists.freedesktop.org/archives/xorg-devel/2013-December/039492.html
Signed-off-by: Andrew Eikum <aeikum@codeweavers.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
dispatch.c: In function 'SetVendorString':
dispatch.c:481:29: warning: declaration of 'string' shadows a global declaration [-Wshadow]
SetVendorString(const char *string)
^
dispatch.c:135:21: warning: shadowed declaration is here [-Wshadow]
typedef const char *string;
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
dev->button->down is a bitmask, not a normal array. Use the helper function to
check, we technically allow the mapping to change after the physical button
has been pressed (but not yet processed yet), so only check BUTTON_PROCESSED.
From XSetPointerMapping(3):
"If any of the buttons to be altered are logically in the down state,
XSetPointerMapping returns MappingBusy, and the mapping is not changed."
Reported-by: Ilja Van Sprundel <ivansprundel@ioactive.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
The server internally relies on arrays with a MAX_BUTTONS maximum size (which
is the max the core protocol can transport). Make sure a driver adheres to
that.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Fallout from fecc7eb1cf, and reverts most of the
rest of that patch.
The device name is allocated and may even change during PreInit. The const
warnings came from the test codes, the correct fix here is to fix the test
code.
touch.c: In function ‘touch_init’:
touch.c:254:14: warning: assignment discards ‘const’ qualifier from pointer target type [enabled by default]
dev.name = "test device";
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Introduced in fecc7eb1cf and reverts most of
that but it's helpfully mixed with other stuff.
InputAttributes are not const, they're strdup'd everywhere but the test code
and freed properly. Revert the const char changes and fix the test up instead.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Checked against randrproto.txt & randr.h
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Flagged by cppcheck 1.62:
[dix/dixfonts.c:1792]: (error) Common realloc mistake:
'font_path_string' nulled but not freed upon failure
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
A client which is ready, but hasn't run for a while, should receive
the same benefit as one which has simply been idle for a while. Use
the smart_stop_tick to see how long it has been since a client has
run instead of smart_check_tick, which got reset each time a client
was ready, even if it didn't get to run.
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
This just removes the comment markers from around the formals in
several function prototypes near where pointer -> void * changes were
made. There are plenty more of these to fix.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
This lets us stop using the 'pointer' typedef in Xdefs.h as 'pointer'
is used throughout the X server for other things, and having duplicate
names generates compiler warnings.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Lots more const char stuff.
Remove duplicate defs of CoreKeyboardProc and CorePointerProc from
test/xi2/protocol-common.c
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
As usual, mostly const char changes. However, filter_device_events had
a potentially uninitialized value, 'raw', which I added a bunch of
checks for. I suspect most of those are 'can't happen', but it's hard
to see that inside the function.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
CheckPassiveGrabsOnWindow() calls AllocGrab() which can fail and return NULL.
This return value is not checked, and can cause NULL pointer dereferences.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
GrabDevice() calls AllocGrab() which can fail and return NULL.
This return value is not checked, and can cause NULL pointer dereferences.
Reported-by: Ilja Van Sprundel <ivansprundel@ioactive.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
If either the initial calloc or the xi2mask_new fails, grab is NULL,
but if a src grab is passed in, it was always being written to by
CopyGrab (and if that failed, dereferenced again in teardown).
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Since we're using RedirectAutomatic to do this, we don't actually
preserve contents when unmapped.
v2: Don't say WhenMapped if Composite didn't initialize [vsyrjala]
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
A default timeslice of 20ms means a pathological client can ruin up to
two frames per scheduler tick. And a fifth of a second is just insane.
Pick two different numbers out of the hat. A 5ms slice means you can
probably keep up with two or three abusive clients, and letting it burst
to 15ms should give you about all the timeslice you need for a
fullscreen game (that's doing server-side rendering for some reason).
If you're running on a system with a 10ms granularity on SIGALRM, then
this effectively changes the intervals to 10ms and 30ms. Which is still
better, just not as better.
I suspect this is about as good as we can do without actually going
preemptive, which is an entire other nightmare.
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
This allocates a new region structure and copies a source region into
it in a single API rather than forcing the caller to do both steps themselves.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
The time between the idle reset and the IdleTimeWakeupHandler to be called is
indeterminate. Clients with an PositiveTransition or NegativeTransition alarm
on a low threshold may miss an alarm.
Work around this by keeping a reset flag for each device. When the
WakeupHandler triggers and the reset flag is set, we force a re-calculation of
everything and pretend the current idle time is zero. Immediately after is the
next calculation with the real idle time.
Relatively reproducible test case: Set up a XSyncNegativeTransition alarm for
a threshold of 1 ms. May trigger, may not.
X.Org Bug 70476 <http://bugs.freedesktop.org/show_bug.cgi?id=70476>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
And now that we have the accessors, localize it. No functional changes, just
preparing for a future change.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Save a pointer to the passed in closure structure before copying it
and overwriting the *c pointer to point to our copy instead of the
original. If we hit an error, once we free(c), reset c to point to
the original structure before jumping to the cleanup code that
references *c.
Since one of the errors being checked for is whether the server was
able to malloc(c->nChars * itemSize), the client can potentially pass
a number of characters chosen to cause the malloc to fail and the
error path to be taken, resulting in the read from freed memory.
Since the memory is accessed almost immediately afterwards, and the
X server is mostly single threaded, the odds of the free memory having
invalid contents are low with most malloc implementations when not using
memory debugging features, but some allocators will definitely overwrite
the memory there, leading to a likely crash.
Reported-by: Pedro Ribeiro <pedrib@gmail.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Use the grabtype to determine which type of event to send - all other events
are pointless and may result in erroneous events being delivered.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
For an active grab, grab->eventMask can be either the core or the XI1 mask.
With the overlap of event filters, calling DeliverOneGrabbedEvent(XI1) for a
ProximityOut event will trigger if the client has selected for enter events -
the filter is the same for both.
Thus, we end up delivering a proximity event to a client not expecting one.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If a client calls XAllowEvents(SyncPointer) it expects events as normal until
the next button press or release event - that freezes the device. An e.g.
proximity event must thus not freeze the pointer.
As per the spec, only button and key events may do so, so narrow it to these
cases.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
There's no reason to do this as (nmasks + 2) callocs, and it's a
surprisingly hot path. Turns out you hit this ~once per passive grab,
and you do a few bajillion passive grab changes every time you enter or
leave the overview in gnome-shell. According to a callgrind of Xorg
with gnome-shell-perf-tool run against it:
Ir before: 721437275
Ir after: 454227086
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
There's no reason not to, and it simplifies quite a few callers.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Adam Jackson <ajax@redhat.com>
It's already not optional at configure time, this just makes it so at
build time too.
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Adam Jackson <ajax@redhat.com>
CreateCursor (Xlib call XCreatePixmapCursor) with a non-bitmap
source pixmap and a None mask is supposed to error out with BadMatch,
but didn't.
From der Mouse <mouse@Rodents-Montreal.ORG>, changed following
comments by Alan Coopersmith <alan.coopersmith@oracle.com>.
Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
if the grab type isn't XI2, grab->xi2mask is random. That random data may
have the enter/leave mask set, leading to events sent to the client that the
client can't handler.
Source of these errors:
_xgeWireToEvent: Unknown extension 131, this should never happen.
Simplest reproducer:
Start Xephyr, press button inside window, move out. As the pointer leaves
the Xephyr window, the errors appear.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
If we have a client which has registered for a DeviceButton grab
be sure to pass this to CheckDeviceGrabAndHintWindow(). Since the
order of clients is arbitrary there is no guarantee that the last
client in the list is the one that belongs to this class.
Signed-off-by: Egbert Eich <eich@freedesktop.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
XI 1.x only allows for first + num valuators, so if a device sends data for
valuators 0 and 2+ only (i.e. valuator 1 is missing) we still need to get
the data for that from somewhere.
XI 1.x uses the hack of an unset valuator mask to get the right coordinates,
i.e. we set the value but don't set the mask for it so XI2 events have the
right mask.
For an absolute device in relative mode, this broke in b28a1af55c, the
value was now always 0. This wasn't visible on the cursor, only in an XI 1.x
client. The GIMP e.g. sees jumps to x/0 every few events.
Drop the condition introduced in b28a1af55c, data in valuators is always
absolute, regardless of the mode.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
grab->type is only non-zero for passive grabs. We're checking an active grab
here, so we need to check if the touch mask is set on the grab.
Test case: grab the device, then start two simultaneous touches. The
grabbing client won't see the second touchpoints because grab->type is 0
and the second touch is not an emulating pointer.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
XQuartz already conditionally renames main() as dix_main() so it can provide
it's own main(). This isn't the ideal way of doing this, as it prevents libdix
built this way from being useful with any other DDX.
So instead, always name that function dix_main(), and also provide a stub main()
which just calls dix_main(), which can be overriden in the DDX.
Add a main() to XWin (XQuartz already has one, of course).
It's no longer neccessary to link XWin and XQuartz with libmain.
v2: Remove unneeded stub main hw/xwin/InitOutput.c
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
The peculiar way we handle coordinates results in relative coordinates on
absolute devices being added to the last value, then that value is mapped to
the screen (taking the device dimensions into account). From that mapped
value we get the final coordinates, both screen and device coordinates.
To avoid uneven scaling on relative coordinates, they are pre-scaled by
screen ratio:resolution:device ratio factor before being mapped. This
ensures that a circle drawn on the device is a circle on the screen.
Previously, we used the ratio to scale x up. Synaptics already does its own
scaling based on the resolution and that is done by scaling y down by the
ratio. So we can remove the code from the driver and get approximately the
same behaviour here.
Minor ABI bump, so we can remove this from synaptics.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Emmanuel Benisty <benisty.e@gmail.com>
A constant deceleration of x simply means (delta * 1/x). We limited that to
values >= 1.0f for obvious reasons, but can also allow values from 0-1.
That means that ConstantDeceleration is actually a ConstantAcceleration, but
hey, if someone needs it...
X.Org Bug 66134 <http://bugs.freedesktop.org/show_bug.cgi?id=66134>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Ungrabbing a device during an active touch grab rejects the grab. Ungrabbing
a device during an active pointer grab accepts the grab.
Rejection is not really an option for a pointer-emulated grab, if a client
has a button mask on the window it would get a ButtonPress emulated after
UngrabDevice. That is against the core grab behaviour.
X.Org Bug 66720 <http://bugs.freedesktop.org/show_bug.cgi?id=66720>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
This shouldn't have been in the patch
Reported-by: Colin Harrison <colin.harrison@virgin.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
defmin/defmax are screen coords and thus use a min-inclusive, max-exclusive
range. device axes ranges are inclusive, so bump the max up by one to get the
scaling right.
This fixes off-by-one coordinate errors if the coordinate matrix is used to
bind the device to a fraction of the screen. It introduces an off-by-one
scaling error in the device coordinate range, but since most devices have a
higher resolution than the screen (e.g. a Wacom I4 has 5080 dpi) the effect
of this should be limited.
This error manifests when we have numScreens > 1, as the scaling from
desktop size back to screen size drops one device unit.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
commit 61a99aff9d
dix: pre-scale relative events from abs devices to desktop ratio (#31636)
added pre-scaling of relative coordinates coming from absolute devices to
undo uneven scaling based on the screen dimensions.
Devices have their own device width/height ratio as well (in a specific
resolution) and this must be applied for relative devices as well to avoid
scaling of the relative events into the device's ratio.
e.g. a Wacom Intuos4 6x9 is in 16:10 format with equal horiz/vert
resolution (dpi). A movement by 1000/1000 coordinates is a perfect diagonal
on the tablet and must be reflected as such on the screen.
However, we map the relative device-coordinate events to absolute screen
coordinates based on the axis ranges. This results in an effective scaling
of 1000/(1000 * 1.6) and thus an uneven x/y axis movement - the y
axis is always faster.
So we need to pre-scale not only by the desktop dimenstions but also by the
device width/height ratio _and_ the resolution ratio.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Too many callers relied on the refcnt being handled correctly. Use a simple
wrapper to handle that case.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
ProcessTouchEvents() calls UDS for all touch events, but if the event type
was switched to TouchUpdate(pending end) UDS is a noop.
Daniel Drake found this can cause stuck buttons if a touch grab is
activated, rejected and the touch event is passed to a regular listener.
This sequence causes the TouchEnd to be changed to TouchUpdate(pending end).
The actual TouchEnd event is later generated by the server once it is passed
to the next listener. UDS is never called for this event, thus the button
remains logically down.
A previous patch suggested for UDS to handle TouchUpdate events [1], however
this would release the button when the first TouchEvent is processed, not
when the last grab has been released (as is the case for sync pointer
grabs). A client may thus have the grab on the device, receive a ButtonPress
but see the button logically up in an XQueryPointer request.
This patch adds a call to UDS to TouchEmitTouchEnd(). The device state must
be updated once a TouchEnd event was sent to the last grabbing listener and
the number of grabs on the touchpoint is 0.
[1] http://patchwork.freedesktop.org/patch/13464/
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The cursor is referenced during CopyGrab(), thus doesn't need to be handled
manually anymore. It does need to be refcounted for temp grabs though.
The oldGrab handling in ProcGrabPointer is a leftover from the cursor in the
grab being refcounted, but the grab itself being a static struct in the
DeviceIntRec. Now that all grabs are copied, this lead to a double-free of
the cursor (Reproduced in Thunderbird, dragging an email twice (or more
often) causes a crash).
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
A client may call XIGrabDevice twice, overwriting the existing grab. Thus,
make sure we free the old copy after we copied it. Free it last, to make
sure our refcounts don't run to 0 and inadvertantly free something on the
way.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
TouchListenerGone cleans up if a client disappears. Having this in
FreeGrab() triggers cyclic removal of grabs, emitting wrong events. In
particular, it would clean up a passive grab record while that grab is
active.
Move it to CloseDownClient() instead, cleaning up before we go.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Introduced in xorg-server-1.13.99.901-2-g9ad0fdb. Storing the grab pointer
in the listener turns out to be a bad idea. If the grab is not an active
grab or an implicit grab, the pointer stored is the one to the grab attached
on the window. This grab may be removed if the client calls UngrabButton or
similar while the touch is still active, leaving a dangling pointer.
To avoid this, copy the grab wherever we need to reference it later. This
is also what we do for pointer/keyboard grabs, where we copy the grab as
soon as it becomes active.
Reported-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Obsolete since 4bc2761ad5. This struct
existed so copying a passive grab could be simply done by
activeGrab = *grab
and thus have a copy of the GrabPtr we'd get from various sources but still
be able to check device->grab for NULL.
Since 4bc2761 activeGrab is a pointer itself and points to the same memory
as grabinfo->grab, leaving us with the potential of dangling pointers if
either calls FreeGrab() and doesn't reset the other one.
There is no reader of activeGrab anyway, so simply removing it is
sufficient.
Note: field is merely renamed to keep the ABI. Should be removed in the
future.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Change the single if condition in the loop body to a
if (!foo) continue;
and re-indent the rest.
No functional changes.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
A sync grab is the owner once it gets events. If it doesn't replay the
event it will get all events from this touch, equivalent to accepting it.
If the touch has ended before XAllowEvents() is called, we also now need to
send the TouchEnd event and clean-up since we won't see anything more from
this touch.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If a device is frozen in results to a grab, we need to enqueue the events.
This makes things complicated, and hard to follow since touch events are now
replayed in the history, pushed into EnqueueEvent, then replayed later
during PlayReleasedEvents in response to an XAllowEvents.
While the device is frozen, no touch events are processed, so if there is a
touch client with ownership mask _below_ the grab this will delay the
delivery and potentially screw gesture recognition. However, this is the
behaviour we have already anyway if the top-most client is a sync pgrab or
there is a sync grab active on the device when the TouchBegin was generated.
(also note, such a client would only reliably work in case of ReplayPointer
anyway)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If the device is currently grabbed as the result of a passive grab
activating, do not prepend that grab to the listeners (unlike active grabs).
Otherwise, a client with a passive pointer grab will prevent touch grabs
from activating higher up in the window stack.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If drivers supply incorrect values don't just quietly return False, spew to
the log so we can detect what's going on. All these cases are driver bugs
and should be fixed immediately.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
InitPointerClassDeviceStruct/InitKeyboardDeviceStruct allocate a
proximity/focus class, respectively. If a driver calls
InitFocusClassDeviceStruct or InitProximityClassDeviceStruct beforehand,
the previously allocated class is overwritten, leaking the memory.
Neither takes a parameter other than the device, so we can simply skip
initialising it if we already have one.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
==21860== 24 bytes in 1 blocks are still reachable in loss record 85 of 397
==21860== at 0x4C2B3F8: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==21860== by 0x61ED93: AllocateOutputBuffer (io.c:1037)
==21860== by 0x61E15A: WriteToClient (io.c:764)
==21860== by 0x457B30: ProcQueryExtension (extension.c:275)
==21860== by 0x43596B: Dispatch (dispatch.c:432)
==21860== by 0x425DAB: main (main.c:295)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
DeleteCallbackManager() introduced for better symmetry in the caller, they
do the same thing.
==20085== 24 bytes in 1 blocks are still reachable in loss record 11 of 103
==20085== at 0x4C2A4CD: malloc (vg_replace_malloc.c:236)
==20085== by 0x43A097: CreateCallbackList (dixutils.c:837)
==20085== by 0x43A1D3: AddCallback (dixutils.c:869)
==20085== by 0x4B1736: GEExtensionInit (geext.c:209)
==20085== by 0x41C8A8: InitExtensions (miinitext.c:389)
==20085== by 0x5AC918: main (main.c:208)
==2042== 8 bytes in 1 blocks are still reachable in loss record 2 of 97
==2042== at 0x4C2A4CD: malloc (vg_replace_malloc.c:236)
==2042== by 0x4C2A657: realloc (vg_replace_malloc.c:525)
==2042== by 0x4802F5: XNFrealloc (utils.c:1095)
==2042== by 0x43A17A: CreateCallbackList (dixutils.c:855)
==2042== by 0x43A1EF: AddCallback (dixutils.c:870)
==2042== by 0x4B1752: GEExtensionInit (geext.c:209)
==2042== by 0x41C8A8: InitExtensions (miinitext.c:389)
==2042== by 0x5AC9E4: main (main.c:208)
==2042==
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Heaps of these:
==2042== 15,360 bytes in 120 blocks are still reachable in loss record 94 of
97
==2042== at 0x4C2A4CD: malloc (vg_replace_malloc.c:236)
==2042== by 0x4C2A657: realloc (vg_replace_malloc.c:525)
==2042== by 0x45FB91: double_size (registry.c:65)
==2042== by 0x45FC97: RegisterRequestName (registry.c:85)
==2042== by 0x460095: RegisterExtensionNames (registry.c:179)
==2042== by 0x460729: dixResetRegistry (registry.c:334)
==2042== by 0x5AC992: main (main.c:201)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
==2547== 1 bytes in 1 blocks are still reachable in loss record 1 of 111
==2547== at 0x4C2A4CD: malloc (vg_replace_malloc.c:236)
==2547== by 0x64D1551: strdup (strdup.c:43)
==2547== by 0x4802FB: Xstrdup (utils.c:1113)
==2547== by 0x585B6C: XkbSetRulesUsed (xkbInit.c:219)
==2547== by 0x58700F: InitKeyboardDeviceStruct (xkbInit.c:595)
==2547== by 0x419FA3: vfbKeybdProc (InitInput.c:74)
==2547== by 0x425A3D: ActivateDevice (devices.c:540)
==2547== by 0x425F65: InitAndStartDevices (devices.c:713)
==2547== by 0x5ACA57: main (main.c:259)
and a few more of the above.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
I didn't think we needed this before, but after doing some more
work with reverse optimus it seems like it should be called.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
this allows the pixmap dirty helper to be used for reverse optimus,
where the GPU wants to copy from the shared pixmap to its VRAM copy.
[airlied: slave_dst is wrong name now but pointless ABI churn at this point]
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
We should have no problem allowing output/offload from the same slave,
I asserted here, but in order to implement reverse optimus this makes
perfect sense. (reverse optimus is intel outputting to nvidia).
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
==15562== 1,800 bytes in 1 blocks are definitely lost in loss record 298 of 330
==15562== at 0x4A06B6F: calloc (vg_replace_malloc.c:593)
==15562== by 0x4312C7: InitTouchClassDeviceStruct (devices.c:1644)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Slave devices don't need these and the matching code in CloseDevice() has a
IsMaster() condition on freeing these, causing a leak.
==16111== 384 bytes in 4 blocks are definitely lost in loss record 72 of 105
==16111== at 0x4C28BB4: calloc (vg_replace_malloc.c:467)
==16111== by 0x42AEE2: AllocDevicePair (devices.c:2707)
==16111== by 0x4BAA27: AllocXTestDevice (xtest.c:617)
==16111== by 0x4BA89A: InitXTestDevices (xtest.c:570)
==16111== by 0x425F5E: InitCoreDevices (devices.c:690)
==16111== by 0x5ACB2D: main (main.c:257)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Fixes reading random memory read beyond the end of original event.
sizeof device_event: 424
sizeof internal_event: 2800
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Unlike pointer/keyboard events, the flags field for ET_Touch* is a set of
server-internal defines that we need to convert to XI protocol defines.
Currently only two of those defines actually translate to the protocol, so
make sure we don't send internal garbage down the wire.
No effect to current clients since they shouldn't look at undefined bits
anyway.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
e02f864fdf "Suppress cursor display until the first XDefineCursor() request"
disabled cursor display a priori unless -retro is given.
On a plain server, caling XFixesHideCursor() and XFixesShowCursor() would
show the default root cursor, despite no client actually defining a cursor.
Change the logic, disable CursorVisible by default and only enable it from
the window's CWCursor logic. If no window ever defines a cursor, said cursor
stays invisible.
X.Org Bug 58398 <http://bugs.freedesktop.org/show_bug.cgi?id=58398>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Tested-by: Bastien Nocera <hadess@hadess.net>
Reviewed-by: Daniel Martin <consume.noise@gmail.com>
==5712== 6 bytes in 1 blocks are still reachable in loss record 17 of 585
==5712== at 0x4A074CD: malloc (vg_replace_malloc.c:236)
==5712== by 0x3D1DE885B1: strndup (strndup.c:46)
==5712== by 0x41CB71: MakeAtom (atom.c:121)
==5712== by 0x55AE3E: XIGetKnownProperty (xiproperty.c:401)
==5712== by 0x4251C9: AddInputDevice (devices.c:312)
==5712== by 0x42AC0C: AllocDevicePair (devices.c:2657)
==5712== by 0x425E6E: InitCoreDevices (devices.c:677)
==5712== by 0x5ACA05: main (main.c:257)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
If we're about to abort, we're already in the signal handler and cannot call
down to the default device cleanup routines (which reset, free, alloc, and
do a bunch of other things).
Add a new DEVICE_ABORT mode to signal a driver's DeviceProc that it must
reset the hardware if needed but do nothing else. An actual HW reset is only
required for some drivers dealing with the HW directly.
This is largely backwards-compatible, hence the input ABI minor bump only.
Drivers we care about either return BadValue on a mode that's not
DEVICE_{INIT|ON|OFF|CLOSE} or print an error and return BadValue. Exception
here is vmmouse, which currently ignores it and would not reset anything.
This should be fixed if the reset is required.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
The transformation matrix we previously stored was a scaled matrix based on
the axis ranges of the device. For relative movements, the scaling is not
required (or desired).
Store two separate matrices, one as requested by the client, one as the
product of [scale . matrix . inv_scale]. Depending on the type of movement,
apply the respective matrix.
For relative movements, also drop the translation component since it doesn't
really make sense to use that bit.
Input ABI 19
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Set on DeviceEnterLeaveEvent() the xXIEnterEvent->focus field
similarly to how the CoreEnterLeaveEvent() function above does
for core events.
This fixes bug https://bugzilla.gnome.org/show_bug.cgi?id=677329
reported to GTK+, where focus handling on window managers with
sloppy focus or no window manager present was broken due to this
field being always set to FALSE.
Signed-off-by: Carlos Garnacho <carlosg@gnome.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
XTest devices are the first ones in the list, being initialised together
with the master devices. If we disable the devices in-order and a device has
a button down when being disabled, the XTest device is checked for a
required button release (xkbAccessX.c's ProcessPointerEvent). This fails if
the device is already NULL.
Instead of putting the check there, disable the devices in the reverse order
they are initialised. Disable physical slaves first, then xtest devices,
then the master devices.
Testcase: shut down server with a button still held down on a physical
device
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Absolute devices may send relative events depending on the mode (synaptics
by default, wacom per option). The relative events are added to the previous
position, converted into device coordinates and then scaled into desktop
coordinates for pointer movement.
Because the device range must be mapped into the desktop coordinate range,
this results in uneven scaling depending dimensions, e.g. on a setup with
width == 2 * height, a relative movement of 10/10 in device coordinates
results in a cursor movement of 20/10 (+ acceleration)
Other commonly user-visible results:
* the touchpad changing acceleration once an external monitor as added.
* drawing a circle on a wacom tablet in relative mode gives an ellipsis in
the same ratio as the desktop dimensions.
Solution: pre-scale the incoming relative x/y coordinates by width/height
ratio of the total desktop size. Then add them to the previous
coordinates and scale back with the previous mapping, which will undo the
pre-scaling and give us the right movement.
X.Org Bug 31636 <http://bugs.freedesktop.org/show_bug.cgi?id=31636>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
If we're already using our own custom macro, might as well use it properly.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
This is mostly sigsafe code, so use sigsave printf. And update some fields
to double that used to be int.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
For absolute events, if the client specifies a screen number offset the
coordinates by that. And add a new flag so we know when _not_ to add the
screen offset in GPE.
Without this offset and the flag, GPE would simply add the offset of the
current screen if POINTER_SCREEN is set.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
