Previously the parent constrained borderClip was copied over
when compRedirectWindow() is called. That is insufficient eg. in
case the window was already redirected, but not yet realized. So
copy the borderClip over in compAllocPixmap() instead.
Example:
Window 1 is below an automatically redirect window 2. Window 2 is
unmapped and moved outside the extents of window 1. Window 2 is
then mapped again, and MarkOverlappedWindows() uses the up to
date borderSize of window 2 to mark windows, which leaves
window 1 unmarked. Then exposures are calculated using the stale
borderClip of window 2, which causes the window below window 2,
to be exposed through an apparent hole in window 1.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=22566
Signed-off-by: Ville Syrjala <syrjala@sci.fi>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Spaces or tabs do not affect the text output layout.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Keith Packard <keithp@keithp.com>
It's fairly common to have multiple, identical monitors plugged in. In
that case, it's preferable to run the monitor's preferred mode on each
output, rather than just matching the width & height and end up with
different timings or refresh rates.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than perform an intermediate copy and expand the strip and the
fan into a triangle list (thereby tripling the number of edges that the
driver needs to process), allow the backend to hook directly into the
appropriate Composite function.
In order to extend the PictureScreen, without needlessly bumping the
ABI, we move the existing copy implementations to mipict.c and assign
those by default. To notify the ddx that the new entry points are
available, we introduce PICTURE_SCREEN_VERSION.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Immutable in randr means that clients are not able to alter the
property itself, they are only allowed to alter the property value.
This logically means that the property then should not be deleted
by the client either.
Signed-off-by: Luc Verhaegen <libv@skynet.be>
Reviewed-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Use fchmod() to change permissions of the lock file instead
of chmod(), thus avoid the race that can be exploited to set
a symbolic link to any file or directory in the system.
Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
use O_NOFOLLOW to open the existing lock file, so symbolic links
aren't followed, thus avoid revealing if it point to an existing
file.
Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
The function does not initialize the module so it has no business
loading it. If some user of DuplicateModule expects a module actually
loaded they should use LoadModule.
Signed-off-by: Michal Suchanek <hramrach@centrum.cz>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Document option introduced in commit 8976e97.
Signed-off-by: Michal Suchanek <hramrach@centrum.cz>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Bug introduced by 9dca441670
xfree86: add a hook to replace the new console handler.
console_handler was not being set, making the server eat up CPU spinning
in WaitForSomething selecting consoleFd over and over again, every time
trying to unregister drain_console without success due to
console_handler being NULL.
Let's just fix the unregistration in xf86SetConsoleHandler() and use that.
But wait, there could be a catch: If some driver replaced the handler using
xf86SetConsoleHandler(), the unregistration in xf86CloseConsole will unregister
that one. I don't understand Xorg well enough to know whether this poses a
problem (could mess up driver deinit somehow or something like that). As it is,
xf86SetConsoleHandler() doesn't offer any way to prevent this (i.e. check which
handler is currently registered).
I had been using it for two days on my machine that previously hit 100% CPU
several times a day. That has now gone away without any new problems appearing.
Signed-off-by: Tomas Trnka <tomastrnka@gmx.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The protocol requires that the emulated event is marked as such. So if a
driver with smooth scrolling axis sends legacy button events, the motion
event must be marked as emulated.
Pass the real type to emulate_scroll_button_events and create the events
accordingly. For real button press or relase events, only that event must be
generated since a release event will follow or a press event has already
occured, respectively. (This fixes a bug where we'd get two release events
for each legacy button event)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Using this call simplifies callers that don't know if the mask bit is set.
Before:
if (valuator_mask_isset(mask, valnum))
value = valuator_mask_get_double(mask, valnum));
else
value = someothervalue;
Now:
if (!valuator_mask_fetch_double(mask, valnum, &value))
value = someothervalue;
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>
The ABI changed in the previous series of changes, so bump the ABI version for
the next release.
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Every module building against xorg-server does not *Require* pixman nor
libpciaccess. If such modules need pixman or pciaccess, they should be
depending on them directly rather than inheriting a dependency from
xorg-server. To do this, they should use PKG_CHECK_MODULES in configure.ac
to check for pixman-1 or pciaccess and include the apporpriate _LIBS variable
to the appropriate _LDFLAGS variable in Makefile.am
This also moves pixman-1 to Requires.private, so CPPFLAGS is set right to
to satisfy include dependencies but avoid linking needlessly.
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Gaetan Nadon <memsize@videotron.ca>
This removes the need to pass the following to configure on darwin:
--disable-pciaccess -disable-int10-module --disable-vbe --disable-vgahw --disable-libdrm
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
This makes a difference on darwin (and apparently nowhere else)
https://www.gnu.org/s/libtool/manual/libtool.html#Modules-for-libltdl
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Gaetan Nadon <memsize@videotron.ca>
--disable-pciaccess, used together with --disable-module-int10, can be used to
disable all pci code inside the server.
Note that XSERVER_LIBPCIACCESS was previously used only in the driver side and
now it defines also whether the library is used inside the server. Also,
XORG_BUS_PCI automake variable is introduced to track PCI code needs.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
No semantical changes. Just moved code around, grouping PCI related stuff in a
single chunk.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
People that don't want VGA arbiter active can go to the library and enable the
stubs there.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Automake:
"Be careful when selecting library components conditionally. Because building
an empty library is not portable, you should ensure that any library
always contains at least one object."
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Gaetan Nadon <memsize@videotron.ca>
This API is apparently semi-deprecated even by XFree86 standards, and
there are only four drivers left using it. Let's start chopping it off.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
This is slightly draconian, but that API is just awful. In all but
one case in the callers it's used to get a map of some legacy VGA
memory, and it would be cleaner for the caller to just call
pci_device_map_legacy.
The sole exception is in the vesa driver, which uses it to avoid having
to look up which device the BAR belongs to. That's similarly trivial to
fix.
Having done that, Linux's PCI layer is now very small indeed.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
... instead of rolling our own, badly.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
It is kept around to help drivers through the API transition and will be
removed at some point in the future.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
pciaccess handles this now.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
If you haven't ported 2.6 to your architecture in the intervening seven
years, you can keep running older servers.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Per-domain I/O is now something drivers must manually request, and must
keep track of within their own state rather than in the ScrnInfoRec.
It's not really possible to split that into two steps without an
additional intermediate ABI break, so don't even try. Drivers that want
source compatibility should ifdef on the presence of xf86UnmapLegacyIO.
As a fringe benefit, domain-aware I/O is now OS-independent, relying
only on support in pciaccess. Simplify OS PCI setup to reflect this.
The IOADDRESS type is kept around to help drivers through the API
transition and will be removed at some point in the future.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
In fact, don't default to anything; drivers must explicitly say which
kind they want, and they are strongly encouraged to do MMIO if possible.
This is an ABI change in that drivers that don't will crash, but drivers
that are explicit will work with both old and new servers.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
This is really a vga-specific hack anyway. The only modern driver that
uses it is trident, but it's already loaded vgahw by the time it would
call xf86GetClocks.
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
The test was memsetting the wrong thing, this fixes make check in my tinderbox.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Co-authored-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
miPointerSetPosition traditionally took coordinates on a per-screen basis,
triggering a screen switch when these went out-of-bounds. For absolute
devices, this prevented screen crossing in the negative x/y direction.
This patch changes the event generation patch to handle screen coordinates
in a desktop range (i.e. all screens together). Screen switches are
triggered when these coordinates are not on the current screen.
This unifies the pointer behaviour of single ScreenRec multihead and
multiple ScreenRecs multihead in that the cursor by default moves about the
whole screen rather than be confined to one single screen. The
transformation matrix may then be used to actually confine the cursor to the
screen again.
Note: fill_pointer_events has to deal with several different coordinate
systems. Make sure you read the comment before trying to understand the code.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
