Skip to content
Snippets Groups Projects
  1. Jan 16, 2019
  2. Jan 11, 2019
  3. Jan 10, 2019
  4. Jan 09, 2019
    • Olivier Fourdan's avatar
      xwayland: handle case without any crtc · e8295c50
      Olivier Fourdan authored
      
      Xwayland creates and destroys the CRTC along with the Wayland outputs,
      so there is possibly a case where the number of CRTC drops to 0.
      
      However, `xwl_present_get_crtc()` always return `crtcs[0]` which is
      invalid when `numCrtcs` is 0.
      
      That leads to crash if a client queries the Present capabilities when
      there is no CRTC, the backtrace looks like:
      
        #0  raise() from libc.so
        #1  abort() from libc.so
        #2  OsAbort() at utils.c:1350
        #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  main_arena() from libc.so
        #9  proc_present_query_capabilities() at present_request.c:236
        #10 Dispatch() at dispatch.c:478
        #11 dix_main() at main.c:276
      
      To avoid returning an invalid pointer (`crtcs[0]`) in that case, simply
      check for `numCrtcs` being 0 and return `NULL` in that case.
      
      Thanks to Michel Dänzer <michel.daenzer@amd.com> for pointing this as a
      possible cause of the crash.
      
      Signed-off-by: default avatarOlivier Fourdan <ofourdan@redhat.com>
      Reviewed-by: default avatarMichel Dänzer <michel.daenzer@amd.com>
      Bugzilla: https://bugzilla.redhat.com/1609181
      e8295c50
  5. Jan 07, 2019
  6. Jan 02, 2019
  7. Dec 21, 2018
  8. Dec 20, 2018
  9. Dec 17, 2018
  10. Dec 14, 2018
  11. Dec 12, 2018
  12. Dec 11, 2018
  13. Nov 29, 2018
    • Adam Jackson's avatar
    • Lionel Landwerlin's avatar
      present: fix compile warning with debug traces · a425eee6
      Lionel Landwerlin authored and Adam Jackson's avatar Adam Jackson committed
      
      Signed-off-by: default avatarLionel Landwerlin <lionel.g.landwerlin@intel.com>
      a425eee6
    • Lyude Paul's avatar
      modesetting: Actually disable CRTCs in legacy mode · 7a44e8d4
      Lyude Paul authored and Adam Jackson's avatar Adam Jackson committed
      
      Believe it or not, somehow we've never done this in legacy mode! We
      currently simply change the DPMS property on the CRTC's output's
      respective DRM connector, but this means that we're just setting the
      CRTC as inactive-not disabled. From the perspective of the kernel, this
      means that any shared resources used by the CRTC are still in use.
      
      This can cause problems for drivers that are not yet fully atomic,
      despite using the atomic helpers internally. For instance: if CRTC-1 and
      CRTC-2 are still enabled and use shared resources within the kernel (an
      MST topology, for example), and then userspace tries to go enable CRTC-3
      on the same topology this might suddenly fail if CRTC-3 needs the shared
      resources CRTC-1 and CRTC-2 are using. While I don't know of any
      situations in the mainline kernel that actually trigger this, future
      plans for reworking the atomic check of MST drivers are absolutely
      going to make this into a real issue (they already are in my WIP
      branches for the kernel).
      
      So: actually do the right thing here and disable CRTCs when they're not
      going to be used anymore, even in legacy mode.
      
      Signed-off-by: Lyude Paul's avatarLyude Paul <lyude@redhat.com>
      7a44e8d4
  14. Nov 25, 2018
  15. Nov 19, 2018
    • Michel Dänzer's avatar
      xwayland: Don't take buffer release queue into account for frame timer · e6cd1c9b
      Michel Dänzer authored and Adam Jackson's avatar Adam Jackson committed
      The buffer release queue has two kinds of entries:
      
      * Pending async flips.
      * Completed flips waiting for their buffer to be released by the Wayland
        compositor.
      
      xwl_present_timer_callback neither completes async flips nor releases
      buffers, so the timer isn't needed for the buffer release queue.
      e6cd1c9b
    • Michel Dänzer's avatar
      xwayland: Don't need xwl_window anymore in xwl_present_queue_vblank · f5416153
      Michel Dänzer authored and Adam Jackson's avatar Adam Jackson committed
      Fixes issue #12. Presumably the problem was that Present operations on
      unmapped windows were executed immediately instead of only when reaching
      the target MSC.
      f5416153
    • Michel Dänzer's avatar
      xwayland: Add xwl_present_unrealize_window · 8c953857
      Michel Dänzer authored and Adam Jackson's avatar Adam Jackson committed
      When a window is unrealized, a pending frame callback may never be
      called, which could result in repeatedly freezing until the frame timer
      fires after a second.
      
      Fixes these symptoms when switching from fullscreen to windowed mode in
      sauerbraten.
      8c953857
    • Michel Dänzer's avatar
      xwayland: Replace xwl_window::present_window with ::present_flipped · 6b016d58
      Michel Dänzer authored and Adam Jackson's avatar Adam Jackson committed
      There's no need to keep track of the window which last performed a
      Present flip. This fixes crashes due to the assertion in
      xwl_present_flips_stop failing. Fixes issue #10.
      
      The damage generated by a flip only needs to be ignored once, then
      xwl_window::present_flipped can be cleared. This may fix freezing in
      the (hypothetical) scenario where Present flips are performed on a
      window, followed by other drawing requests using the window as the
      destination, but nothing triggering xwl_present_flips_stop. The damage
      from the latter drawing requests would continue being ignored.
      6b016d58
    • Ray Strode's avatar
      dix: ensure work queues are cleared on reset · 8738ce85
      Ray Strode authored and Adam Jackson's avatar Adam Jackson committed
      If the server resets, most client workqueues are cleaned up as the
      clients are killed.
      
      The one exception is the server's client, which is exempt from
      the killing spree.
      
      If that client has a queued work procedure active, it won't get
      cleared on reset.
      
      This commit ensures it gets cleared too.
      8738ce85
    • Samuel Thibault's avatar
      dix: do not send focus event when grab actually does not change · 364d6498
      Samuel Thibault authored and Adam Jackson's avatar Adam Jackson committed
      
      c67f2eac ("dix: always send focus event on grab change") made dix
      always sent events when it's a NotifyGrab or NotifyUngrab, even if
      from == to, because 'from' can just come from a previous XSetInputFocus
      call.
      
      However, when an application calls XGrabKeyboard several times on
      the same window, we are now sending spurious FocusOut+FocusIn with
      NotifyGrab, even if the grab does not actually change. This makes screen
      readers for blind people spuriously emit activity events which disturb
      screen reading workflow when e.g. switching between menus.
      
      This commit avoids calling DoFocusEvents in that precise case, i.e. when
      oldWin is a previous grab and the new grab is the same window.
      
      Signed-off-by: default avatarSamuel Thibault <samuel.thibault@ens-lyon.org>
      Reviewed-by: Adam Jackson's avatarAdam Jackson <ajax@redhat.com>
      364d6498
  16. Nov 14, 2018
Loading