Commit Graph

4555 Commits

Author SHA1 Message Date
Tudor Brindus 7964a313e8 xwayland/selection: use one X11 window per incoming transfer
This commit introduces logic for using a new X11 window for each
incoming transfer, rather than having a global window for each selection
source.

This eliminates a whole class of bugs involving multiple concurrent
incoming transfers.

For now, we retain the outgoing transfer queue, and the selection
source-specific windows to support it. Source-specific windows are no
longer used in the incoming path, and will be removed in a future PR.

Refs #1497.
2021-02-04 13:33:59 +01:00
Brandon Dowdy 0977633457 examples/egl_common: make attribs const
Make (config, context)_attribs const just to be on the safe side.
2021-02-04 09:38:27 +01:00
Simon Ser 8e27418dd3 backend/wayland: flush remote display
It turns out wl_event_source_check is not enough to guarantee that the
remote wl_display will be flushed after we queue requests. We need to
explicitly flush, just like we do in our X11 code.

References: https://gitlab.freedesktop.org/wayland/wayland/-/issues/187
2021-02-03 15:22:36 +01:00
Simon Ser a39dc1f7a8 region: remove resource_list arg from region_create
This is never used.
2021-02-02 23:42:53 +01:00
Simon Ser 45f992b27b region: make wlr_region_create private
This function should only be called from the handler for
wl_compositor.create_region requests.
2021-02-02 23:42:53 +01:00
Simon Ser 3d7aa73867 render/dmabuf: make src arg const in wlr_dmabuf_attributes_copy 2021-02-02 19:53:16 +01:00
Simon Ser 7ac2ce25e3 render/dmabuf: cleanup on wlr_dmabuf_attributes_copy error 2021-02-02 19:52:20 +01:00
Quantum 975d14b799 render/wlr_texture: clamp texture coordinates to edge by default
Clamping texture coordinates prevents OpenGL from blending the left and
right edge (or top and bottom edge) when scaling textures with GL_LINEAR
filtering. This prevents visual artifacts like swaywm/sway#5809.

Per discussion on IRC, this behaviour is made default. Compositors that want
the wrapping behaviour (e.g. for tiled patterns) can override this by doing:

    struct wlr_gles2_texture_attribs attribs;
    wlr_gles2_texture_get_attribs(texture, &attribs);

    glBindTexture(attribs.target, attribs.tex);
    glTexParameteri(attribs.target, GL_TEXTURE_WRAP_S, GL_REPEAT);
    glTexParameteri(attribs.target, GL_TEXTURE_WRAP_T, GL_REPEAT);
    glBindTexture(attribs.target, 0);
2021-02-01 21:19:17 +01:00
Simon Ser 01d21cdd9f render/egl: log whether DMA-BUF modifiers are supported 2021-02-01 16:16:37 +01:00
Brandon Dowdy f1d37c54c8 render/egl: add check for EGL_KHR_surfaceless_context
As surfaces are no longer going to be used for wlr_egl, I may as well just go and add this check as it is needed for safety whenever surface-less rendering is being used.
2021-02-01 09:31:44 +01:00
Tudor Brindus dd4c8aa45e xwayland/selection: make xwm_selection_init take a wlr_xwm_selection *
This makes it consistent with xwm_selection_finish.
2021-01-31 19:17:04 +01:00
Tudor Brindus b3d782f818 xwayland/selection: introduce `xwm_selection_transfer_init`
Currently, all this does is initialize `wl_client_fd` to -1, so that
comparisons with 0 are meaningful.
2021-01-31 19:17:04 +01:00
Tudor Brindus aa86a022fa xwayland/selection: make xwm_selection_finish take a wlr_xwm_selection *
Previously it took a wlr_xwm *, which was a bit surprising in that it
freed members of wlr_xwm *, not just its respective selections.
2021-01-31 19:17:04 +01:00
Tudor Brindus b6ba595862 xwayland/selection: destroy all selections on Xwayland restart
Previously, Xwayland could restart, and we'd get events for transfers
pointing to the previous (now freed) xwm instance. This led to
use-after-free segfaults.

Closes #2565.
2021-01-31 10:24:59 +01:00
Tudor Brindus 3417fc0cca xwayland/selection: don't leak Wayland fd if ConvertSelection fails
If our ConvertSelection failed, we would previously leak the pending
Wayland client fd.

Refs swaywm/sway#5946.
2021-01-31 10:24:53 +01:00
Tudor Brindus e0dfc14983 xwayland/selection: don't request another selection while one is pending
This will hopefully be fixed in the future by having separate windows
for each X11-to-Wayland transfer, but until then, let's avoid a
compositor crash.
2021-01-31 10:24:47 +01:00
Brandon Dowdy c89dba9435 examples: remove "major" and "minor" from egl_common.c
Remove "major" and "minor" from egl_common.c as they are not used by the examples that use egl_common.c.
2021-01-30 00:31:20 +01:00
Simon Ser 7b50f5d324 backend/wayland: remove unnecessary cast 2021-01-29 12:12:47 +01:00
Tudor Brindus 211c1e23be xwayland/selection: end incr transfer on empty prop, not next selection
Previously, `transfer->incr` was being cleared on the next selection.
However, if the next selection was *also* incremental, it's possible
that `xwm_handle_selection_property_notify` would route us to
`xwm_get_incr_chunk` instead of `xwm_selection_get_data`.
2021-01-29 10:18:03 +01:00
Tudor Brindus 703c17ae41 xwayland/selection: refactor remaining incremental transfer code 2021-01-29 10:18:03 +01:00
Tudor Brindus 23148d283f xwayland/selection: extract out property requests
Apart from reducing duplication, this has the positive side-effect of
allowing all deallocs to use
`xwm_selection_transfer_destroy_property_reply`, as opposed to the
latter and a mix of ad-hoc `free`s.
2021-01-29 10:18:03 +01:00
Tudor Brindus dea94f2bad xwayland/selection: simplify incremental transfer control flow
Previously, if the Wayland client died before an incremental transfer
was complete, the logs would be spammed by "write error to target fd" as
wlroots entered some control flow wherein it'd continually try
scheduling further writes to the already-dead pipe.

This commit contains no behavioral changes, but introduces explicit
handling for draining the X11 selection in case of Wayland client death.
2021-01-29 10:18:03 +01:00
Tudor Brindus 10a2d57055 xwayland/selection: explicitly bail if first write to Wayland fd fails
If `xwm_data_source_write` failed, it's failed permanently. In fact, a
failing `xwm_data_source_write` sets `transfer->property_reply` to
null as part of its error handling.

Instead of relying on an indirect check (whether
`transfer->property_reply` is still non-null), explicitly use the return
value from `xwm_data_source_write`.
2021-01-29 10:18:03 +01:00
Tudor Brindus 40b2e7669a xwayland/selection: make `xwm_data_source_write` return 0 on failure
The `fd` is marked `O_NONBLOCK`, so `write` will never spuriously return
`EINTR`. Therefore, `write` failing is permanent, and we can return 0 to
make the return value meaningful.
2021-01-29 10:18:03 +01:00
Brandon Dowdy 8aa38fe73e render/egl: remove *config_attribs and wlr_egl->config
Breaking changes:

Both "EGLint *config_attribs" and "wlr_egl->config" no longer exist.
2021-01-29 10:03:24 +01:00
Brandon Dowdy 705b3da7cb render/egl: remove wlr_egl_{create,destroy}_surface
Breaking changes:

wlr_egl_create_surface and wlr_egl_destroy_surface have been
removed and no longer exist.
2021-01-29 10:03:24 +01:00
Brandon Dowdy 34e7f69d69 examples: remove dependency on wlr_egl from clients
The specified clients in this commit used to rely on wlr_egl and
some of its related functions in order to render surfaces.

This is no longer the case as of this commit.
2021-01-29 10:03:24 +01:00
Simon Ser 50b9921642
backend/x11: remove output_set_refresh
The X11 backend uses the Present extension to schedule frames. The
refresh rate is unused.
2021-01-28 18:47:47 +01:00
Simon Ser f8a66072e7 xwayland: fix extraneous NET_WM_STATE log messages
wlroots would log "Unhandled NET_WM_STATE property change" log
messages for atoms we know about. Simplify the code structure
and remove these extra messages.
2021-01-28 12:03:50 +01:00
Simon Ser a406f19479 render/egl: fix NULL dereference when using llvmpipe
Fixes: 6becc69ec9 ("render/egl: fail to create wlr_egl on software renderer")
2021-01-27 21:12:19 +01:00
Simon Zeni 6becc69ec9 render/egl: fail to create wlr_egl on software renderer
The creation of `wlr_egl` will fail is the device extension
EGL_MESA_device_software is defined. The creation process is allowed to
continue only if the environment variable `WLR_RENDERER_ALLOW_SOFTWARE`
is defined to the value 1.
2021-01-27 15:50:19 +01:00
Simon Ser 5a2ef794dc xwayland: drop unused enum net_wm_state_action 2021-01-27 10:47:16 +01:00
Tudor Brindus e75f483aeb xwayland/selection: rename Wayland-facing data and helpers
Previously, wlr_xwm_selection_transfer.source_fd meant:

- the source of data in a Wayland -> X11 copy (good)
- the destination of data in a X11 -> Wayland copy (confusing)

This made reading through xwayland/selection/incoming.c difficult: in
many places, "source" actually means "destination".
2021-01-25 21:02:55 +01:00
fwsmit 1b8330d1f8 examples/foreign-toplevel: Add documenation and output selection 2021-01-25 10:14:49 +01:00
Tudor Brindus 0db191d3bf xwayland/selection: prevent fd leak on unsupported MIME type
Since we never end up calling xcb_convert_selection, the file descriptor
ends up getting leaked (i.e., not cleaned up within
xwm_data_source_write).
2021-01-25 09:46:20 +01:00
Tudor Brindus abb56152ff xwayland: use wlr_log_errno instead of %m
Previously, any error would be masked by an internal isatty call:

  24:31:48.174 [DEBUG] [wlr] [xwayland/selection/incoming.c:386] XCB_SELECTION_NOTIFY (selection=277, property=278, target=256)
  24:31:48.174 [ERROR] [wlr] [xwayland/selection/incoming.c:30] write error to target fd: Inappropriate ioctl for device
2021-01-25 09:22:04 +01:00
Simon Ser 73ffab70b0 backend/drm: don't blit in drm_fb_import
Instead blit in drm_plane_lock_surface. This makes drm_fb_import simpler
and better fits its name.
2021-01-24 18:33:56 +01:00
Simon Ser 4af85f4c19 backend/drm: simplify drm_fb_lock_surface
Make it take a plane instead, and rename to drm_plane_lock_surface.
2021-01-24 18:22:03 +01:00
Simon Ser 44a3d6e74d backend/drm: document wlr_drm_plane surface fields 2021-01-24 18:21:38 +01:00
Simon Ser 4f06ce2550 render/gbm_allocator: avoid drmPrimeHandleToFD
Some drivers (e.g. nouveau) need to do some book-keeping when exporting
a BO handle to a DMA-BUF [1]. Using drmPrimeHandleToFD directly hides
the export from the user-space drivers and causes some ref'counting
issues.

So we can't properly handle BOs with a different handle for each plane
until [2] lands.

[1]: https://gitlab.freedesktop.org/drm/nouveau/-/issues/43#note_772661
[2]: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5442

Suggested-by: Ilia Mirkin <imirkin@alum.mit.edu>
2021-01-23 16:41:54 +01:00
Ilia Mirkin 7bc8dbb991 backend/x11: keep track of exposed rects, add them to damage regions
When we receive an Expose event, that means that we must redraw that
region of the X11 window. Keep track of these regions with pixman
regions, and merge them with the additional output damaged regions.

Fixes #2670
2021-01-21 15:14:54 +01:00
Ilia Mirkin 922b7f415d backend/x11: skip events we don't care about
These are ones I see log messages about in my setup.
2021-01-21 15:14:54 +01:00
Ilia Mirkin bb92fd4c90 backend/x11: add support for scanout mode
This makes full-screen with weston-terminal work.
2021-01-21 12:35:52 +01:00
Simon Ser 7c995b78b2 Revert "render: add wlr_egl_create_from_drm_fd function"
This reverts commit ee31be167b.
2021-01-20 21:32:50 +01:00
Simon Ser 54e5ef39c0 Revert "render/egl: remove unused gbm_device member"
This reverts commit 306cf11d87.
2021-01-20 21:32:50 +01:00
Simon Zeni 306cf11d87 render/egl: remove unused gbm_device member 2021-01-20 15:29:00 +01:00
Simon Zeni ee31be167b render: add wlr_egl_create_from_drm_fd function 2021-01-20 15:29:00 +01:00
Ilia Mirkin 62f37ee319 backend/x11: make sure output transform matrix is initialized
The transform matrix was all 0's, which meant that effectively nothing
got rendered other than the clear color.
2021-01-20 14:55:27 +01:00
BrassyPanache d6649a8a4b Expose ICCCM input status
In certain situations windows can have their input field set to false
but still expect to receive input focus by passively listening to key
presses via a parent window. The ICCCM specification outlines how focus
should be given to clients.

Further reading: https://tronche.com/gui/x/icccm/sec-4.html#s-4.1.7

Relates to #2604
2021-01-20 10:38:58 +01:00
Isaac Freund f6fe439718 xdg-output: destroy outputs before manager
Since output_destroy() calls wl_list_remove() on the output's link,
the manager must still be valid. This is the same bug fixed in bf926e3
but with a different interface.
2021-01-19 11:56:31 +01:00