egl/wayland: Use per-surface event queues

During display initialisation, we need a separate event queue to handle
the registry events, which is correctly handled. But we also need
separate per-surface event queues to handle swapchain-related events,
such as surface frame events and buffer release events. This avoids two
surfaces from the same EGLDisplay, both current on separate threads,
dispatching each other's events.

Create separate per-surface event queues, create wl_surface and wl_drm
proxy wrapper objects per surface, so we eliminate the race around
sending events to the wrong queue. swrast buffers do not need a
dedicated proxy wrapper, as the wl_shm_pool used to create the
wl_buffers, being transient, can itself be assigned to a queue.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 36b9976e1f99 ("egl/wayland: Avoid race conditions when on non-main thread")
Cc: mesa-stable@lists.freedesktop.org
diff --git a/src/egl/drivers/dri2/egl_dri2.h b/src/egl/drivers/dri2/egl_dri2.h
index b1e2242..b3d4848 100644
--- a/src/egl/drivers/dri2/egl_dri2.h
+++ b/src/egl/drivers/dri2/egl_dri2.h
@@ -256,6 +256,10 @@
    struct wl_egl_window  *wl_win;
    int                    dx;
    int                    dy;
+   struct wl_event_queue *wl_queue;
+   struct wl_surface     *wl_surface_wrapper;
+   struct wl_display     *wl_dpy_wrapper;
+   struct wl_drm         *wl_drm_wrapper;
    struct wl_callback    *throttle_callback;
    int                    format;
 #endif
diff --git a/src/egl/drivers/dri2/platform_wayland.c b/src/egl/drivers/dri2/platform_wayland.c
index 561609d..52ab9f7 100644
--- a/src/egl/drivers/dri2/platform_wayland.c
+++ b/src/egl/drivers/dri2/platform_wayland.c
@@ -146,6 +146,37 @@
    }
 
    dri2_surf->wl_win = window;
+   dri2_surf->wl_queue = wl_display_create_queue(dri2_dpy->wl_dpy);
+   if (!window) {
+      _eglError(EGL_BAD_ALLOC, "dri2_create_surface");
+      goto cleanup_surf;
+   }
+
+   if (dri2_dpy->wl_drm) {
+      dri2_surf->wl_drm_wrapper = wl_proxy_create_wrapper(dri2_dpy->wl_drm);
+      if (!dri2_surf->wl_drm_wrapper) {
+         _eglError(EGL_BAD_ALLOC, "dri2_create_surface");
+         goto cleanup_queue;
+      }
+      wl_proxy_set_queue((struct wl_proxy *)dri2_surf->wl_drm_wrapper,
+                         dri2_surf->wl_queue);
+   }
+
+   dri2_surf->wl_dpy_wrapper = wl_proxy_create_wrapper(dri2_dpy->wl_dpy);
+   if (!dri2_surf->wl_dpy_wrapper) {
+      _eglError(EGL_BAD_ALLOC, "dri2_create_surface");
+      goto cleanup_drm;
+   }
+   wl_proxy_set_queue((struct wl_proxy *)dri2_surf->wl_dpy_wrapper,
+                      dri2_surf->wl_queue);
+
+   dri2_surf->wl_surface_wrapper = wl_proxy_create_wrapper(window->surface);
+   if (!dri2_surf->wl_surface_wrapper) {
+      _eglError(EGL_BAD_ALLOC, "dri2_create_surface");
+      goto cleanup_drm;
+   }
+   wl_proxy_set_queue((struct wl_proxy *)dri2_surf->wl_surface_wrapper,
+                      dri2_surf->wl_queue);
 
    dri2_surf->wl_win->private = dri2_surf;
    dri2_surf->wl_win->destroy_window_callback = destroy_window_callback;
@@ -176,6 +207,11 @@
 
    return &dri2_surf->base;
 
+ cleanup_drm:
+   if (dri2_surf->wl_drm_wrapper)
+      wl_proxy_wrapper_destroy(dri2_surf->wl_drm_wrapper);
+ cleanup_queue:
+   wl_event_queue_destroy(dri2_surf->wl_queue);
  cleanup_surf:
    free(dri2_surf);
 
@@ -241,6 +277,12 @@
       dri2_surf->wl_win->destroy_window_callback = NULL;
    }
 
+   if (dri2_surf->wl_drm_wrapper)
+      wl_proxy_wrapper_destroy(dri2_surf->wl_drm_wrapper);
+   wl_proxy_wrapper_destroy(dri2_surf->wl_surface_wrapper);
+   wl_proxy_wrapper_destroy(dri2_surf->wl_dpy_wrapper);
+   wl_event_queue_destroy(dri2_surf->wl_queue);
+
    free(surf);
 
    return EGL_TRUE;
@@ -309,7 +351,7 @@
    }
 
    /* There might be a buffer release already queued that wasn't processed */
-   wl_display_dispatch_queue_pending(dri2_dpy->wl_dpy, dri2_dpy->wl_queue);
+   wl_display_dispatch_queue_pending(dri2_dpy->wl_dpy, dri2_surf->wl_queue);
 
    if (dri2_surf->back == NULL) {
       for (i = 0; i < ARRAY_SIZE(dri2_surf->color_buffers); i++) {
@@ -610,7 +652,7 @@
       dri2_dpy->image->queryImage(image, __DRI_IMAGE_ATTRIB_STRIDE, &stride);
 
       dri2_surf->current->wl_buffer =
-         wl_drm_create_prime_buffer(dri2_dpy->wl_drm,
+         wl_drm_create_prime_buffer(dri2_surf->wl_drm_wrapper,
                                     fd,
                                     dri2_surf->base.Width,
                                     dri2_surf->base.Height,
@@ -624,7 +666,7 @@
       dri2_dpy->image->queryImage(image, __DRI_IMAGE_ATTRIB_STRIDE, &stride);
 
       dri2_surf->current->wl_buffer =
-         wl_drm_create_buffer(dri2_dpy->wl_drm,
+         wl_drm_create_buffer(dri2_surf->wl_drm_wrapper,
                               name,
                               dri2_surf->base.Width,
                               dri2_surf->base.Height,
@@ -632,8 +674,6 @@
                               dri2_surf->format);
    }
 
-   wl_proxy_set_queue((struct wl_proxy *) dri2_surf->current->wl_buffer,
-                      dri2_dpy->wl_queue);
    wl_buffer_add_listener(dri2_surf->current->wl_buffer,
                           &wl_buffer_listener, dri2_surf);
 }
@@ -645,14 +685,14 @@
 {
    int i;
 
-   if (wl_proxy_get_version((struct wl_proxy *) dri2_surf->wl_win->surface)
+   if (wl_proxy_get_version((struct wl_proxy *) dri2_surf->wl_surface_wrapper)
        < WL_SURFACE_DAMAGE_BUFFER_SINCE_VERSION)
       return EGL_FALSE;
 
    for (i = 0; i < n_rects; i++) {
       const int *rect = &rects[i * 4];
 
-      wl_surface_damage_buffer(dri2_surf->wl_win->surface,
+      wl_surface_damage_buffer(dri2_surf->wl_surface_wrapper,
                                rect[0],
                                dri2_surf->base.Height - rect[1] - rect[3],
                                rect[2], rect[3]);
@@ -675,7 +715,7 @@
 
    while (dri2_surf->throttle_callback != NULL)
       if (wl_display_dispatch_queue(dri2_dpy->wl_dpy,
-                                    dri2_dpy->wl_queue) == -1)
+                                    dri2_surf->wl_queue) == -1)
          return -1;
 
    for (i = 0; i < ARRAY_SIZE(dri2_surf->color_buffers); i++)
@@ -691,11 +731,9 @@
 
    if (draw->SwapInterval > 0) {
       dri2_surf->throttle_callback =
-         wl_surface_frame(dri2_surf->wl_win->surface);
+         wl_surface_frame(dri2_surf->wl_surface_wrapper);
       wl_callback_add_listener(dri2_surf->throttle_callback,
                                &throttle_listener, dri2_surf);
-      wl_proxy_set_queue((struct wl_proxy *) dri2_surf->throttle_callback,
-                         dri2_dpy->wl_queue);
    }
 
    dri2_surf->back->age = 1;
@@ -704,7 +742,7 @@
 
    create_wl_buffer(dri2_surf);
 
-   wl_surface_attach(dri2_surf->wl_win->surface,
+   wl_surface_attach(dri2_surf->wl_surface_wrapper,
                      dri2_surf->current->wl_buffer,
                      dri2_surf->dx, dri2_surf->dy);
 
@@ -718,7 +756,7 @@
     * ignore the damage region and post maximum damage, due to
     * https://bugs.freedesktop.org/78190 */
    if (!n_rects || !try_damage_buffer(dri2_surf, rects, n_rects))
-      wl_surface_damage(dri2_surf->wl_win->surface,
+      wl_surface_damage(dri2_surf->wl_surface_wrapper,
                         0, 0, INT32_MAX, INT32_MAX);
 
    if (dri2_dpy->is_different_gpu) {
@@ -736,14 +774,14 @@
    dri2_flush_drawable_for_swapbuffers(disp, draw);
    dri2_dpy->flush->invalidate(dri2_surf->dri_drawable);
 
-   wl_surface_commit(dri2_surf->wl_win->surface);
+   wl_surface_commit(dri2_surf->wl_surface_wrapper);
 
    /* If we're not waiting for a frame callback then we'll at least throttle
     * to a sync callback so that we always give a chance for the compositor to
     * handle the commit and send a release event before checking for a free
     * buffer */
    if (dri2_surf->throttle_callback == NULL) {
-      dri2_surf->throttle_callback = wl_display_sync(dri2_dpy->wl_dpy_wrapper);
+      dri2_surf->throttle_callback = wl_display_sync(dri2_surf->wl_dpy_wrapper);
       wl_callback_add_listener(dri2_surf->throttle_callback,
                                &throttle_listener, dri2_surf);
    }
@@ -1386,11 +1424,13 @@
 
 
 static EGLBoolean
-dri2_wl_swrast_allocate_buffer(struct dri2_egl_display *dri2_dpy,
+dri2_wl_swrast_allocate_buffer(struct dri2_egl_surface *dri2_surf,
                                int format, int w, int h,
                                void **data, int *size,
                                struct wl_buffer **buffer)
 {
+   struct dri2_egl_display *dri2_dpy =
+      dri2_egl_display(dri2_surf->base.Resource.Display);
    struct wl_shm_pool *pool;
    int fd, stride, size_map;
    void *data_map;
@@ -1411,6 +1451,7 @@
 
    /* Share it in a wl_buffer */
    pool = wl_shm_create_pool(dri2_dpy->wl_shm, fd, size_map);
+   wl_proxy_set_queue((struct wl_proxy *)pool, dri2_surf->wl_queue);
    *buffer = wl_shm_pool_create_buffer(pool, 0, w, h, stride, format);
    wl_shm_pool_destroy(pool);
    close(fd);
@@ -1446,7 +1487,7 @@
    /* find back buffer */
 
    /* There might be a buffer release already queued that wasn't processed */
-   wl_display_dispatch_queue_pending(dri2_dpy->wl_dpy, dri2_dpy->wl_queue);
+   wl_display_dispatch_queue_pending(dri2_dpy->wl_dpy, dri2_surf->wl_queue);
 
    /* try get free buffer already created */
    for (i = 0; i < ARRAY_SIZE(dri2_surf->color_buffers); i++) {
@@ -1462,7 +1503,7 @@
       for (i = 0; i < ARRAY_SIZE(dri2_surf->color_buffers); i++) {
          if (!dri2_surf->color_buffers[i].locked) {
              dri2_surf->back = &dri2_surf->color_buffers[i];
-             if (!dri2_wl_swrast_allocate_buffer(dri2_dpy,
+             if (!dri2_wl_swrast_allocate_buffer(dri2_surf,
                                                  dri2_surf->format,
                                                  dri2_surf->base.Width,
                                                  dri2_surf->base.Height,
@@ -1472,8 +1513,6 @@
                 _eglError(EGL_BAD_ALLOC, "failed to allocate color buffer");
                  return -1;
              }
-             wl_proxy_set_queue((struct wl_proxy *) dri2_surf->back->wl_buffer,
-                                dri2_dpy->wl_queue);
              wl_buffer_add_listener(dri2_surf->back->wl_buffer,
                                     &wl_buffer_listener, dri2_surf);
              break;
@@ -1529,22 +1568,20 @@
 
    while (dri2_surf->throttle_callback != NULL)
       if (wl_display_dispatch_queue(dri2_dpy->wl_dpy,
-                                    dri2_dpy->wl_queue) == -1)
+                                    dri2_surf->wl_queue) == -1)
          return;
 
    if (dri2_surf->base.SwapInterval > 0) {
       dri2_surf->throttle_callback =
-         wl_surface_frame(dri2_surf->wl_win->surface);
+         wl_surface_frame(dri2_surf->wl_surface_wrapper);
       wl_callback_add_listener(dri2_surf->throttle_callback,
                                &throttle_listener, dri2_surf);
-      wl_proxy_set_queue((struct wl_proxy *) dri2_surf->throttle_callback,
-                         dri2_dpy->wl_queue);
    }
 
    dri2_surf->current = dri2_surf->back;
    dri2_surf->back = NULL;
 
-   wl_surface_attach(dri2_surf->wl_win->surface,
+   wl_surface_attach(dri2_surf->wl_surface_wrapper,
                      dri2_surf->current->wl_buffer,
                      dri2_surf->dx, dri2_surf->dy);
 
@@ -1554,9 +1591,9 @@
    dri2_surf->dx = 0;
    dri2_surf->dy = 0;
 
-   wl_surface_damage(dri2_surf->wl_win->surface,
+   wl_surface_damage(dri2_surf->wl_surface_wrapper,
                      0, 0, INT32_MAX, INT32_MAX);
-   wl_surface_commit(dri2_surf->wl_win->surface);
+   wl_surface_commit(dri2_surf->wl_surface_wrapper);
 
    /* If we're not waiting for a frame callback then we'll at least throttle
     * to a sync callback so that we always give a chance for the compositor to