The authorative source of the request frame size is client->req_len,
especially with big requests larger than 2^18 bytes.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Since we're now using xcb for upstream X11 connection, it's cleaner to
use it's type for the window IDs.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Now that we completely ported from Xlib to XCB, we can finally stop
importing Xlib :)
FIN.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Now that no Xlib operations (besides opening and closing connection)
aren't used anymore, we can move over the last pieces and use XCB
instead of Xlib for connecting the upstream Xserver.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Now that no X11 calls are being done via Xlib anymore, we're free to
also move over event receiving, leaving Xlib pretty much unused.
Also need to add a simple event queue mechanism, because we've go a
screen operation (see xnestBitBlitHelper) that needs to collect up
certain events for it's return value.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
xnestWindowExposures() is a micro-optimization for the specific case that
a newly created window receives exposure events (from our upstream server)
inside the region we're already exposing on our own (miWindowExposures()):
it peeks the Xlib event queue for all expose events, checks whether their
areas are inside our exposure region and requeue's those that aren't.
Unfortunately, this depends on Xlib's internal queue mechamism, thus standing
in the way of moving to XCB (which doesn't have that).
Removing this doens't seem to make any practical difference, even with
demanding applications like GIMP. The only cost is potentially having some
initial window content painted twice, *if* the application really draws
something complicated right after creating the window.
*If* there'll really be a demand for such an optimization some day, it can
be reimplemented without any message queue: just redirecting all expose events
into recording them in a region, which is flushed out later. But for now,
there really doesn't seem to be any practical need for that.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Use the visuals lookup table introduced by previous commit for
looking up local vs upstream visuals and their colormaps.
Replacing the the old Xlib visuals table.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Record the associations between host's and our visuals as well their
corresponding cmaps in a global table, which's used later for lookups.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This is a temporary measure, until xcbproto / libxcb is fixed:
keep an own copy of the fixed xcb_xkb_get_kbd_by_name(), renamed
as xcb_xkb_get_kbd_by_name_1().
Once xcbproto/libxcb is fixed (and new xcb release is out), this
commit can be reverted.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Now that no Xlib drawing functions used anymore, we can finally switch over
to using GC XID's directly, instead of Xlib's GC struct.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Replace XDrawImageString() by xcb_image_text_8(), as well as their 16-bit
counterparts.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Replace XDrawString8() by xcb_poly_text_8(), as well as XDrawString16()
by xcb_poly_text_16(). Some care needs to be taken to prepend the xTextElt
header before sending the request out.
GC operation handlers don't need to care about poly-strings or length
above 254, as this is already handled by their caller, doPolyText().
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Instead of going through mi machinery, just pass the ClearArea
request to the upstream window.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Use xcb_create_colormap() and XFreeColormap() instead of XCreateColormap()
and XFreeColormap().
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Using xcb_shape_rectangles() and xcb_shape_mask() instead of Xlib's
XShapeCombineRegion() and XShapeCombineMask().
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Upcoming patches will need to retieve GC's XIDs on the upstream connection.
Moving this out into separate .c file, in order to not creating more
dependencies on Xlib headers, which we wanna get rid of.
For now, looking at the Xlib GC structure, attached to our DDX GCs.
When all users of the Xlib GC have gone (ie. moved all consumers to xcb),
we'll create the GC via xcb directly, thus replacing the Xlib GC struct
by XID - the interface of this helper will remain the same.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Fetching the setup data from xcb instead of Xlib, storing in our own struct,
holding all information needed for one particular upstream connection.
For now, there's only one, but future multi-upstream implementation will
change this to an array (and storing pointers to particular upstream in
various places).
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
* unexport functions from dixgrab.h, that aren't used by any driver/module.
* add paremeter names to prototypes
* add doxygen-style documentation for all the prototypes
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This reverts commit 58a2fb8b6f.
Needed by xf86-video-intel driver. Didn't get noticed, because we don't
have this driver in our CI yet.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Since we're not indirectly writing via FILE anymore, this option has
become meaningless: it meant flushing out our in-process buffer to
the kernel, but we're now doing direct write() calls anyways.
xf86 still accepts the "flush" config file flag for backwards compatibility,
but it hasn't any practical meaning anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's just a wrapper around LogVMessageVerb() and no external module
using it, so can easily be optimized-away.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's nothing but a wrapper, doing the same as LogMessageVerb(X_NONE, ...),
and no external module / driver needs it, so can be easily optimized away.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
These aren't used by any modules/drivers, so no need to keep them exported.
Also drop the return value, which isn't used by any caller.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Direct calls to ScreenRec->DestroyPixmap() blocks cleaning up the wrapping
jungle, so use the proper dix function instead.
See: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1754
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Direct calls to ScreenRec->DestroyPixmap() blocks cleaning up the wrapping
jungle, so use the proper dix function instead.
See: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1754
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Direct calls to ScreenRec->DestroyPixmap() blocks cleaning up the wrapping
jungle, so use the proper dix function instead.
See: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1754
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Direct calls to ScreenRec->DestroyPixmap() blocks cleaning up the wrapping
jungle, so use the proper dix function instead.
See: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1754
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Direct calls to ScreenRec->DestroyPixmap() blocks cleaning up the wrapping
jungle, so use the proper dix function instead.
See: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1754
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Xwin's DestroyPixmap proc just free()s the PixmapRec directly, w/o catering
for devPrivate's, so leaving a memleak. The correct DIX function for this
is FreePixmap().
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Even if the manual redirect didn't result in the surface window we
wanted, it might be different again without the redirect.
I don't know of any specific scenario hitting this though, it's more of
a theoretical safeguard.
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1637>
using EGL (e.g., eglQueryString, epoxy_has_egl_extension)
before establishing this connection
enables the GBM/EGL implementation to potentially consume the
WAYLAND_SOCKET fd, which, if closed, will cause the compositor
to kill this xserver
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1646>
$ echo "#foo\nfoo" > custom_config $ X -config custom_config
will trigger the double free because the contents of xf86_lex_val.str
have been realloc()ed aready when free is called in read.c:209.
This copies the lex token and adds all the necessary free() calls to
avoid leaking it
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1176>
The functions declared here aren't used by any driver, so no need to keep
them in the public driver API. Since the whole file isn't included by anybody
outside the xserver tree itself, it doesn't need to be installed at all,
so making it internal and move it to Xi directory.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1325>
These aren't used by any drivers/modules, so no need to keep them exported.
As already touching them, give them a proper name prefix for clarity.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1581>
This header includes libxfont2.h, but the dependency isn't stated anywhere,
causing some drivers to FTBS (when libxont2.h is in non-standard location).
Since this header doesn't seem to need including libxfont2.h at all, just
stop including it, instead of adding yet another dependency to server SDK.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1610>
Most of the support for such OS'es was removed in 2010 for
xorg-server-1.10.0, but a few bits lingered on, and a few
comments were left out-of-date.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1667>
Tidy it up a bit and split the actual request handler from the request
reading loop, making upcoming reworks easier to review (smaller diffs).
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1661>
Make the code easier to read/understand by scoping variables to where
exactly they're needed. Also a preparation for subsequent reworks
(keep the diff smaller and thus easier to review)
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1661>
This header makes use of types and functions defined in gcstruct.h and
privates.h. It should include them, instead of demanding it's consumers
having done that before including this file.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1659>
It's safer to zero-out the cursor-private memory on allocation,
instead of relying on being cleared initialized somewhere later.
Fixes: 3f3ff971ec - Replace X-allocation functions with their C89 counterparts
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1652>
We alread have several of these calls, that aren't interested in result value,
explicitly casting to void. Fixing this up for the remaining ones.
This is helpful for the human reader as well as quality analysis tools.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1648>
The vidmode extension emulation in Xwayland reports the modeline based
on the current mode.
To do so, it searches for the mode using `xwl_output_find_mode(-1, -1)`
which is supposed to return the current mode, whatever that mode is.
With XRandR emulation, in rootless mode, the default value is the mode
at index 0. That assumption, however is not true when running rootful.
That means that the vidmode extension will always return the highest
mode available, which is 5120x2880, with Xwayland running rootful:
$ xwayland-run -geometry 1024x768 -- xvidtune -show
"5120x2880" 1276.50 5120 5560 6128 7136 2880 2883 2888 2982 -hsync +vsync
Luckily, when Xwayland is running rootful, we have the current mode size
conveniently stored in dedicated fields of the xwayland output struct,
so we can use that to search for the right mode being used and report
that through the vidmode extension:
$ xwayland-run -geometry 1024x768 -- xvidtune -show
"1024x768" 63.50 1024 1064 1176 1328 768 771 775 798 -hsync +vsync
That fixes legacy games using the vidmode extension and rendering at the
wrong size when running within Xwayland rootful.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1641>
The code here assumed a `leave` event always occurs between two `enter`
events. On Sway (and presumably other compositors) this happens even if
the client has destroyed the `wl_surface`, but the client gets a null
`surface` here. (Which presumably on on the wire is the id of the
destroyed surface.)
This seems like a bad thing to rely on, and is easy to avoid. But if
this is correct to assume, the Wayland protocol should be explicit about
this.
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1527>
Now that we won't enable DRI3 if <sys/eventfd.h> is not available, there
is not point in trying to include that header without DRI3.
That allows to build Xwayland with GLAMOR enabled (without DRI3) on
platforms which do not support eventfd.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1523>
Currently, we would start dequeuing events as soon as a device is
resumed, regardless of its capabilities.
If the capabilities are not available, we would just fallback to the
regular XTEST code path and not use input emulation.
As a result, it is very likely that we shall lose the first events until
the compositor resumes first a device with the requested capabilities.
To avoid that issue, start emulating only once we have the requested
capabilities, if they match the seat capabilities.
Closes: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1732
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1631>
This is a small code refactoring to help with clarity, simply move the
code from the switch case for device resume to a dedicated function.
No functional change.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1631>
Enrico Weigelt, metux IT consult
Michel Dänzer
Mike Blumenkrantz
Olivier Fourdan
Matthieu Herrb
Alan Coopersmith
Ian Douglas Scott