FreeResource() keeps clientTable[cid].elements up to date with the
number of resources allocated to the client. The other free
resource functions (FreeResourceByType(),
FreeClientNeverRetainResources() and FreeClientResources()) don't
maintain this invariant.
Typically, the only consequence is that the element count is too high
and we end up allocating the hash table bigger than necessary. However,
FreeResource() also relies on the element count to restart the search if
the list of resources has been changed during a resource destruction
callback. Since FreeResourceByType() doesn't update the count, if we call
that from a resource destruction callback from FreeResource(), the
loop isn't restarted and we end up following an invalid next pointer.
Furthermore, LookupClientResourceComplex() and
FreeClientNeverRetainResources() don't use the element count to detect
if a callback deleted a resource and may end up following an invalid
next pointer if the resource system is called into recursively.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
GetPairedDevice() may not always return the keyboard, resulting in a
null-pointer dereference when accessing the XKB state.
For floating devices, the GetMaster() returns the device itself.
X.Org Bug 27573 <http://bugs.freedesktop.org/show_bug.cgi?id=27573>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Tested-by: Ben Hutchings <ben@decadent.org.uk>
Currently, SyncComputeBracketValues reuses old values of bracket_greater
and bracket_less when startOver = FALSE. This can result in incorrect bracket
values. To fix this issue, the startOver parameter is removed, and we do not
reuse old values of bracket_greater and bracket_less.
X.Org Bug 27023 <http://bugs.freedesktop.org/show_bug.cgi?id=27023>
Signed-off-by: David James <davidjames@google.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
miPolyPoint ought to leave the GC unchanged even if it fails. ajax says:
> We have a new winner for the oldest-bug competition! It's actually
> been like that since X11R1:
>
> -rw-r--r--. 1 ajax ajax 2817 1987-09-12 01:20 ddx/mi/mipolypnt.c
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
This does make wide ellipses slower, by a factor of twoish:
946000.0 521000.0 ( 0.55) 10-pixel wide ellipse
98300.0 49900.0 ( 0.51) 100-pixel wide ellipse
17700.0 9310.0 ( 0.53) 500-pixel wide ellipse
16900.0 7980.0 ( 0.47) 100-pixel wide dashed ellipse
16100.0 5370.0 ( 0.33) 100-pixel wide double-dashed ellipse
But no one cares, and it's a modest size win:
text data bss dec hex filename
1773824 69552 59288 1902664 1d0848 hw/vfb/Xvfb.before
1773112 69552 58648 1901312 1d0300 hw/vfb/Xvfb
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
I believe this is what was intended.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Same goes for VERIFY_ALPHA, VERIFY_XIN_PICTURE, and VERIFY_XIN_ALPHA.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Presumably no implications, especially security-wise.
Signed-off-by: Matthias Hopf <mhopf@suse.de>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
With Xinerama enabled, event coordinates are relative to Screen 0, so
they can be negative. The new DeviceEvent's coordinates are of type
uint16_t, making screens above and to the left of Screen 0 unusable.
X.Org Bug 24986 <https://bugs.freedesktop.org/show_bug.cgi?id=24986>
Signed-off-by: Chris Humbert <freedesktop@mahadri.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Properties allocated through XIGetKnownProperty() aren't reset on the second
server generation but keep the old value. As a result, wrong Atoms are
supplied to the driver, resulting in potential data corruption or weird
error message.
Reproducible by running "xlsatom | grep FLOAT" twice on a plain X server.
The second X server generation won't have the FLOAT atom defined anymore,
despite the users of this atom not noticing any errors.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
commit 206531f75c added localization files
for ar, add them to the Makefile.
Signed-off-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
This uses the same hack that dix uses for the rootCursor -- allocate
a resource ID for the invisible cursor so that it gets freed at reset
time. This also allows us to unconditionally create it during
extension initialization; necessary as the privates layout may well be
different on subsequent generations.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
A few cursor value assignments weren't getting correctly ref counted,
causing leaks of cursor objects.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
this shut up some warnings.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
In the process, fixes a memory leak in CloseDevice, and an unchecked
memory allocation in InitializePredictableAccelerationProperties.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Needed as the privates stuff uses the callbacks before they would have
been defined in dix.h
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
The functions exported by the devPrivates code were poorly documented. I tried
to spruce it up a little.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
We call NIDR on all devices that make it through the config backend.
Including some that have no driver assigned to them (/dev/input/mouse0 for
example). Those ones then simply get ignored by NIDR, but this should not be
noted as an error in the log file.
X_INFO is sufficient, and it may just prevent some bugreports.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Acked-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
A rewrite of the XWin DDX AIGLX code to actually make it do something useful again
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Colin Harrison <colin.harrison@virgin.net>
Running "make check" will lead to build problems in
scratchbox. Building the first test that wraps dixLookupWindow fails
because symbol __real_dixLookupWindow can't be resolved. Defining
wrapping options as linker options instead of compiler options makes
everything build nicely in scratchbox.
Signed-off-by: Rami Ylimäki <ext-rami.ylimaki@nokia.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Banked framebuffers are so 1990. As of 7.4 the only drivers remaining
that used this were chips, neomagic, trident, and vesa. vesa only used
it when not using shadowfb, which is broadly undesirable anyway, and no
longer uses it at all as of 2.3.0. neomagic never used it by default,
and support for it is gone in git master. The other two effectively
only ever used it for ISA chips; since ISA support is now gone from
the server, they have been modified to only compile mibank support when
ISA support is available.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Most DGA requests allow at most one client to be using DGA on each
screen. Instead of keeping track of the current client in a
MAXSCREEN-sized array, track it in a per-screen private.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
If a -screen option specifies a screen number higher than any previously
specified, reallocate the vfb-private array of screen-info structs.
If built with a DIX that still has a MAXSCREENS limit, asking for too
many screens won't be detected until InitOutput calls AddScreen.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Inspired by Jon Turney's "Xwin: Simplify screen option processing"
patch, which does something similar for the Xwin server.
Besides making the code more readable, this eliminates most of Xvfb's
references to MAXSCREENS.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
The screen number can be inferred from the position in the vfbScreens
array, and it was only used in two places, so it was hardly important.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Several DDXes allow each screen to have at most one (or in some cases,
exactly one) installed colormap. These all use the same pattern: Declare
a global-lifetime array of MAXSCREENS ColormapPtrs, and index it by
screen number. This patch converts most of those to use screen privates
instead.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Acked-by: Tiago Vignatti <tiago.vignatti@nokia.com>
The structure of the adaptor-matching algorithm was obscured by trying
to use "continue" and "break" as the only control-flow primitives. It's
a lot more clear if you add "return" to that set.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
A call to miPointerUpdateSprite for the XTEST keyboard may result in a
NULL pointer dereference in miDCPutUpCursor() when the save buffer is NULL.
XTS test case: Xlib 11 KeymapNotify
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
_XkbCopyGeom did not copy all of the data from the source geometry. This
resulted in failures when trying to obtain the keymap from a server
where the default geometry has not been replaced by a custom
configuration.
Signed-off-by: Dirk Wallenstein <halsmit@t-online.de>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
