gbm: fix get_back_bo() failure with gbm_surface and implicit modifiers

Before 361f362258 ("dri: Unify createImage and
createImageWithModifiers"), gbm_surface_create_with_modifiers() would
fail with ENOSYS on drivers missing explicit modifiers support. After
that commit, it succeeds and fails later when it tries to allocate a
new back buffer.

Restore the previous behavior.

Signed-off-by: Simon Ser <contact@emersion.fr>
Fixes: 105fcb9cfd ("dri: revert INVALID modifier special-casing")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12283
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32726>

src/gbm/backends/dri/gbm_dri.c

@@ -1078,11 +1078,17 @@ gbm_dri_surface_create(struct gbm_device *gbm,
 		       uint32_t format, uint32_t flags,
                        const uint64_t *modifiers, const unsigned count)
 {
+   struct gbm_dri_device *dri = gbm_dri_device(gbm);
    struct gbm_dri_surface *surf;
 
    if (count)
       assert(modifiers);
 
+   if (count > 0 && !dri->screen->base.screen->resource_create_with_modifiers) {
+      errno = ENOSYS;
+      return NULL;
+   }
+
    /* It's acceptable to create an image with INVALID modifier in the list,
     * but it cannot be on the only modifier (since it will certainly fail
     * later). While we could easily catch this after modifier creation, doing