Since group_info and width are used for the key actions allocations,
when modifying them we need to take care of reallocation key actions if
needed.

Some devices may not have keyboard information.

Fixes #574

Set a linear gamma ramp.  This avoids the xrandr command always warning
'Failed to get size of gamma for output default'
(perhaps we should be using GDI GetDeviceGammaRamp(), if possible?)

Make CRTC report non-zero physical dimensions initially

The rrGetInfo hook is not called for all RANDR requests (e.g.
RRGetOutputInfo), so we must always keep the fake mode information up to
date, rather than doing it lazily in the rrGetInfo hook)

Because we are so bad, most GTK+3 versions treat the output name 'default'
specially, and don't try to use RANDR with it.  But versions 3.21.6 to
3.22.24, don't do this, and get badly confused by a CRTC with size 0x0.

See:
https://bugzilla.gnome.org/show_bug.cgi?id=771033
https://bugzilla.gnome.org/show_bug.cgi?id=780101

Future work: Rather than reporting a single fake CRTC with a mode matching
the entire virtual display, the fake CRTCs we report should match our
'pseudo-xinerama' monitors

Make QueryMonitor() slightly less insane, making it return TRUE if the
specified monitor exists, rather than always returning TRUE (which we
are uselessly checking, and then also checking if the specified monitor
exists)

(Note that EnumDisplayMonitors() doesn't seem to have meaningful way to
return errors, see 5940580f)

Also: Spamming the long UseMsg() after "Invalid monitor number" isn't very
helpful.

Also: If we are exiting in ddxProcessArgument() due to an error in
options, use a non-zero exit status.

The PresentOptionSuboptimal support code is not optional (once you've
enabled building Present at all), so require a protocol package version
that defines what we need.

Fixes: xorg/xserver#821

Adam Jackson authored 5 years ago and Adam Jackson committed 5 years ago

Suppose you're in a Hyper-V guest and are trying to use PCI passthrough.
The ID_PATH that udev will construct for that looks something like
"acpi-VMBUS:00-pci-b8c8:00:00.0", and obviously looking for "pci-" in
the first four characters of that is going to not work.

Instead, strstr. I suppose it's possible you could have _multiple_ PCI
buses in the path, in which case you'd want strrstr, if that were a
thing.

Olivier Fourdan authored 5 years ago and Adam Jackson committed 5 years ago

Using the existing command line option "-listen" for passing file
descriptors between the Wayland compositor and Xwayland is misleading,
Xwayland should add is own command line option for that specific use.

As XWayland is spawned by the Wayland compositor, we cannot just change
the option, as that would break all existing Wayland compositors using
Xwayland, so we add a new options "-listenfd" and mark the previous one
as deprecated and log a warning, but it still works for backward
compatibility.

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
!214

Olivier Fourdan authored 5 years ago and Adam Jackson committed 5 years ago

Xwayland uses the command line option “-listen” to pass file descriptors
from the Wayland compositor.

That breaks the traditional, documented behavior of the “-listen”
command line option which is to enable a transport type.

Checks if the given option starts with a digit, otherwise treat it as a
regular transport type.

Closes: #817


Suggested-by: Rodrigo Exterckötter Tjäder
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>

a2rgb10 configs would end up with channel masks corresponding to
argb8888. This would confuse the GLX core code into matching an a2rgb10
config to the root window visual, and that would make things look wrong
and bad.

Fix this by handling more cases. We're still not fully general here, and
this could still be wrong on big-endian. The XXX comment about doing
something less ugly still applies, ideally we would get this information
out of EGL instead of making lucky guesses. Still, better than it was.

Fixes: #824

Add an option to turn on the use of the X window's alpha channel in
multiwindow mode, i.e. this uses the X window's alpha channel for
compositing into the native desktop.

This works on W7/Vista (using DwmEnableBlurBehindWindow()), and Windows
10 (using the undocumented SetWindowCompositionAttribute()), but not on
Windows 8/8.1

-compositewm must be enabled for this to be useful, as we only have a
pixmap with an alpha channel for the X window in that case.  The
framebuffer/root window doesn't have one (unless perhaps you are using
the rootless extension, maybe...).

v2:
Update meson.build

Future work:

A window property to control use of alpha?
Option to turn off blur on W7/Vista
Implement _NET_WM_WINDOW_OPACITY

Set convenience variables in winTopLevelWindowProc() for WM_CREATE as
well.

I think that a major cost in the current implementation is doing a
CreateDIBSection()/DestroyObject() on every refresh.  So provide our own
CreatePixmap() instead, which does the CreateDIBSection(), once.

Testcase: glxgears or foobillard with direct swrast
Testcase: scrolling in a full-screen xterm

v2:
Fix handling of RENDER Scratch Pixmaps
(A problem easily shown with gitk or emacs)

v3:
Note that we don't own screen pixmap to release in DestroyPixmap
Log if unimplemented slow-path ever gets hit

Fill the area outside the current window size with black, rather than
leaking framebuffer contents or leaving it undrawn.

Rather than drawing the window contents from the shadow framebuffer, use
Composite extension redirection to cause the server to maintain a bitmap
image of each top-level X window, and draw the window contents from
that, so that window contents which are occluded in the framebuffer show
correctly in the task bar and task switcher previews.

v2:
Fix incorrect use of memset() found by gcc5

hw/xwin/winshadgdi.c: In function ‘winBltExposedWindowRegionShadowGDI’:
hw/xwin/winshadgdi.c:861:9: warning: ‘memset’ used with constant zero length parameter; this could be due to transposed parameters [-Wmemset-transposed-args]

v3:
Turn on -compositewm by default

v4:
Ignore -swcursor if -compositewm

-swcursor is not compatible with -compositewm (because the window
contents are drawn from an off-screen pixmap, not from the screen
pixmap, where the software cursor will be drawn).

v5:
Update meson.build also
Add -compositewm option to help output
Update CI to install prerequisites

Push the multiwindow wndproc WM_PAINT handling down into the drawing
engine.  Only the GDI engine is supported in multiwindow mode currently,
so we only need to do this in the GDI engine.

Make winBltExposedRegionsShadowGDI() do the same stuff that
winTopLevelWindowProc()'s WM_PAINT handler does.

Note that winBltExposedRegionsShadowGDI() is currently used 1) in
windowed mode when the GDI engine is selected, and 2) in multiwindow
mode when "Hide Root Window" is off.

Peter Hutterer authored 5 years ago and Adam Jackson committed 5 years ago

We can't have high keycodes because everything in XKB relies on 8 bits. XI2's
API allows for 32-bit keycodes so we have to take those but nothing in the
server is really ready for this. The effect of this right now is that any high
keycode grab is clipped to 255 and thus ends up grabbing a different key
instead.

https://bugzilla.redhat.com/show_bug.cgi?id=1697804



Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

Carlos Garnacho authored 5 years ago and Olivier Fourdan committed 5 years ago

Hiding the tablet tool cursor results in it being hidden forever after.
This is due to the stale frame callback that will neither be disposed
or replaced. This can be reproduced in krita (X11) as the pointer
cursor is hidden while over the canvas.

Clearing the frame callback ensures the correct behavior in future
xwl_tablet_tool_set_cursor() calls (i.e. a new cursor surface being
displayed, and a new frame callback created), and is 1:1
with xwl_seat_set_cursor() for pointers.

Signed-off-by: Carlos Garnacho <carlosg@gnome.org>

Olivier Fourdan authored 5 years ago and Adam Jackson committed 5 years ago

If a pixmap is not exportable, `glamor_gbm_bo_from_pixmap()` would fail
and the modesettings driver would consequently fail to do its page flip,
which both prevents Present from working and also fill up the logs with
error messages such as:

  (EE) modeset(0): Failed to get GBM bo for flip to new front.
  (EE) modeset(0): present flip failed

Refactor the code so that `glamor_gbm_bo_from_pixmap()` takes care of
making the pixmap exportable.

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Signed-off-by: Yuxuan Shui <yshui@hadean.com>
See-also: !131
Closes: #68
Fixes: 86b2d874 "glamor: Reallocate pixmap storage without modifiers
       if necessary"

Olivier Fourdan authored 6 years ago and Adam Jackson committed 5 years ago

With `glamor_set_pixmap_texture()` returning its status, remove the hack
and use the return value.

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>

Olivier Fourdan authored 6 years ago and Adam Jackson committed 5 years ago

Chnage the API for `glamor_set_pixmap_texture()` to return a status,
so that the caller can know whether it succeeded or not.

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>

Olivier Fourdan authored 6 years ago and Adam Jackson committed 5 years ago

The current code in `xwl_glamor_gbm_create_pixmap_for_bo()` may fail in
several cases that are not checked for:

 - `eglCreateImageKHR()` may have failed to create the image,
 - `glEGLImageTargetTexture2DOES()` may fail and set an error,
 - `glamor_set_pixmap_texture()` may fail for very large pixmaps
    because the corresponding FBO could not be created.

Trying to upload content to a pixmap with no texture will crash Mesa,
glamor and Xwayland, e.g.:

  XXX fail to create fbo.
  (EE)
  (EE) Backtrace:
  (EE) 0: Xwayland (OsSigHandler+0x29)
  (EE) 1: libpthread.so.0 (funlockfile+0x50)
  (EE) 2: libc.so.6 (__memmove_avx_unaligned_erms+0x215)
  (EE) 3: dri/i965_dri.so (_mesa_format_convert+0xab3)
  (EE) 4: dri/i965_dri.so (_mesa_texstore+0x205)
  (EE) 5: dri/i965_dri.so (store_texsubimage+0x28c)
  (EE) 6: dri/i965_dri.so (intel_upload_tex+0x13b)
  (EE) 7: dri/i965_dri.so (texture_sub_image+0x134)
  (EE) 8: dri/i965_dri.so (texsubimage_err+0x150)
  (EE) 9: dri/i965_dri.so (_mesa_TexSubImage2D+0x48)
  (EE) 10: Xwayland (glamor_upload_boxes+0x246)
  (EE) 11: Xwayland (glamor_copy+0x4d1)
  (EE) 12: Xwayland (miCopyRegion+0x96)
  (EE) 13: Xwayland (miDoCopy+0x43c)
  (EE) 14: Xwayland (glamor_copy_area+0x24)
  (EE) 15: Xwayland (damageCopyArea+0xba)
  (EE) 16: Xwayland (compCopyWindow+0x31c)
  (EE) 17: Xwayland (damageCopyWindow+0xd3)
  (EE) 18: Xwayland (miResizeWindow+0x7b7)
  (EE) 19: Xwayland (compResizeWindow+0x3a)
  (EE) 20: Xwayland (ConfigureWindow+0xa96)
  (EE) 21: Xwayland (ProcConfigureWindow+0x7d)
  (EE) 22: Xwayland (Dispatch+0x320)
  (EE) 23: Xwayland (dix_main+0x366)
  (EE) 24: libc.so.6 (__libc_start_main+0xf3)
  (EE) 25: Xwayland (_start+0x2e)
  (EE)
  Fatal server error:
  (EE) Caught signal 11 (Segmentation fault). Server aborting
  (EE)

Check for the possible cases of failure above and fallback to the
regular glamor pixmap creation when an error is detected.

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Closes: #661

Adam Jackson authored 5 years ago and Adam Jackson committed 5 years ago

Desktop GL can handle arbitrary rops here, GLES can't. The switch
statement attempts to optimize some cases that GLES can still handle if
we precompute the right pixel value, but we were then throwing that
pixel value away and using gc->fgPixel anyway. Fix this, and now
xts-render passes against glamor+gles.

Without this we're using driswrast to set up GLX visuals. This is
unfortunate because llvmpipe does not expose multisample configs, so
various apps that expect them will fail. With this we just query the
capabilities of the EGL that's backing glamor, and reflect that to the
GLX clients. This also paves the way for xserver to stop being a DRI
driver loader, which is nice.

Fixes: #640
Fixes: #643
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=98272


Reviewed-by: Olivier Fourdan <ofourdan@redhat.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>

Commit 56c0a71f incremented the
GLXSERVER_VENDOR_ABI_MINOR_VERSION define, but this define was not actually
being used to set glxServer.minorVersion.

Update the initializer for glxServer to use the correct version numbers.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>

Olivier Fourdan authored 5 years ago and Olivier Fourdan committed 5 years ago

On pointer enter notification, Xwayland checks for an existing pointer
warp with a `NULL` sprite.

In turn, `xwl_pointer_warp_emulator_maybe_lock()` checks for an existing
grab and the destination window using `XYToWindow()` which does not
check for the actual sprite not being `NULL`.

So, in some cases, when the pointer enters the surface and there is an
existing X11 grab which is not an ownerEvents grab, Xwayland would crash
trying to dereference the `NULL` sprite pointer:

  #0  __GI_raise ()
  #1  __GI_abort () at abort.c:79
  #2  OsAbort () at utils.c:1351
  #3  AbortServer () at log.c:879
  #4  FatalError () at log.c:1017
  #5  OsSigHandler () at osinit.c:156
  #6  OsSigHandler () at osinit.c:110
  #7  <signal handler called>
  #8  XYToWindow (pSprite=0x0, x=0, y=0) at events.c:2880
  #9  xwl_pointer_warp_emulator_maybe_lock () at xwayland-input.c:2673
  #10 pointer_handle_enter () at xwayland-input.c:434

Avoid the crash by simply checking for the sprite being not `NULL` in
`xwl_pointer_warp_emulator_maybe_lock()`

Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Bugzilla: https://bugzilla.redhat.com/1708119

I don't know how I feel about the promise made in 2e7f790b :)

Only _sigset_t is defined by MinGW's signal.h

MinGW defines SIG_BLOCK, but doesn't have signal masks, so rather than
checking for SIG_BLOCK, add a configure check for sigprocmask.

v2:
Also add check to meson.build

Include winsock.h for definition of fd_set type and select()

Future work: Maybe this also needs some error checking for fd > FD_SETSIZE?

I don't think an input thread can ever be useful on Windows.

There is a pthread emulation, so having the thread itself isn't much of
a problem.

However, there is no device to wait on for Windows events, and even if
we were to replace select() with WFMO, Windows wants to send events for
a window to the thread which created that window.

So, disable input thread by default for MinGW

v2:
Also add similar to meson.build

Kyle Brenneman authored 5 years ago and Aaron Plattner committed 5 years ago

Add a new function, GlxServerExports::setClientScreenVendor, which will change
the vendor that handles GLX requests for a screen, but only for requests from
a specific client.

v2: Increment the GLXVND minor version number.
v3: Note the GLXVND version requirement for setClientScreenVendor.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>

Kyle Brenneman authored 5 years ago and Aaron Plattner committed 5 years ago

When GlxGetXIDMap looks up an unknown XID, it will now look up a vendor based
on the screen number for the XID and the client that sent the current request.

In GlxGetXIDMap, if the XID is for a regular X window, then it won't be in the
(XID -> vendor) mapping, so we have to look up a vendor by screen number.

With this change, GlxGetXIDMap will use the (screen -> vendor) map for
whichever client sent the current request, instead of using the global
(screen -> vendor) map.

Since GlxGetXIDMap doesn't take a ClientPtr argument, GlxDispatchRequest will
store the client for the current request in a global variable. That way, the
ABI for GLXVND doesn't need to change.

v2: Fix an error check in GlxDispatchRequest.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>

Kyle Brenneman authored 7 years ago and Aaron Plattner committed 5 years ago

Each client now has its own (screen, vendor) mapping.

Currently, it's just a copy of the global mapping, but later changes will allow
it to change.

Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>

glxc->drawPriv will be NULL if the context is direct, or if it is
current but without a bound drawable. Mesa's libGL won't normally emit
protocol for direct contexts for these calls, but a malign client could
still crash the server.