gallium/u_threaded: implement asynchronous flushes

This requires out-of-band creation of fences, and will be signaled to
the pipe_context::flush implementation by a special TC_FLUSH_ASYNC flag.

v2:
- remove an incorrect assertion
- handle fence_server_sync for unsubmitted fences by
  relying on the improved cs_add_fence_dependency
- only implement asynchronous flushes on amdgpu

Reviewed-by: Marek Olšák <[email protected]>

1 files changed, 2 insertions, 0 deletions

diff --git a/src/gallium/drivers/radeonsi/si_pipe.h b/src/gallium/drivers/radeonsi/si_pipe.h
index 84e8a4c8673..de089e351ce 100644
--- a/src/gallium/drivers/radeonsi/si_pipe.h
+++ b/src/gallium/drivers/radeonsi/si_pipe.h
@@ -604,6 +604,8 @@ void si_init_dma_functions(struct si_context *sctx);
 /* si_fence.c */
 void si_init_fence_functions(struct si_context *ctx);
 void si_init_screen_fence_functions(struct si_screen *screen);
+struct pipe_fence_handle *si_create_fence(struct pipe_context *ctx,
+					  struct tc_unflushed_batch_token *tc_token);
 
 /* si_hw_context.c */
 void si_destroy_saved_cs(struct si_saved_cs *scs);