[Mesa-dev] [PATCH] radeonsi: fix an assertion failure: assert(!res->b.is_shared)

[Marek Olšák]

From: Marek Olšák <marek.olsak at amd.com>

This only appears to happen on Raven2.

Possible way to reproduce:

resource_get_handle(WINSYS_HANDLE_TYPE_KMS) --> sets is_shared = true
resource_get_handle(WINSYS_HANDLE_TYPE_DMABUF) --> fail

Cc: 19.1 19.2 <mesa-stable at lists.freedesktop.org>
---
 src/gallium/drivers/radeonsi/si_texture.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/src/gallium/drivers/radeonsi/si_texture.c b/src/gallium/drivers/radeonsi/si_texture.c
index 904a39b6fed..be614e66440 100644
--- a/src/gallium/drivers/radeonsi/si_texture.c
+++ b/src/gallium/drivers/radeonsi/si_texture.c
@@ -912,22 +912,21 @@ static bool si_texture_get_handle(struct pipe_screen* screen,
 		/* This is not supported now, but it might be required for OpenCL
 		 * interop in the future.
 		 */
 		if (resource->nr_samples > 1 || tex->is_depth)
 			return false;
 
 		/* Move a suballocated texture into a non-suballocated allocation. */
 		if (sscreen->ws->buffer_is_suballocated(res->buf) ||
 		    tex->surface.tile_swizzle ||
 		    (tex->buffer.flags & RADEON_FLAG_NO_INTERPROCESS_SHARING &&
-		     sscreen->info.has_local_buffers &&
-		     whandle->type != WINSYS_HANDLE_TYPE_KMS)) {
+		     sscreen->info.has_local_buffers)) {
 			assert(!res->b.is_shared);
 			si_reallocate_texture_inplace(sctx, tex,
 							PIPE_BIND_SHARED, false);
 			flush = true;
 			assert(res->b.b.bind & PIPE_BIND_SHARED);
 			assert(res->flags & RADEON_FLAG_NO_SUBALLOC);
 			assert(!(res->flags & RADEON_FLAG_NO_INTERPROCESS_SHARING));
 			assert(tex->surface.tile_swizzle == 0);
 		}
 
-- 
2.17.1