r600g: Fix segfault in r600_compute_global_transfer_map()
authorTom Stellard <thomas.stellard@amd.com>
Tue, 16 Oct 2012 14:21:40 +0000 (14:21 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Tue, 16 Oct 2012 14:39:16 +0000 (14:39 +0000)
This segfault was caused by commit
369e46888904c6d379b8b477d9242cff1608e30e, however it is my fault for not
testing the patch while it was on the list.

src/gallium/drivers/r600/evergreen_compute.c

index 655cf756186e90896d9796b1a9550d7f8a535d13..557159496467493a2696eda1a5ef87a11e25e356 100644 (file)
@@ -851,7 +851,7 @@ void *r600_compute_global_transfer_map(
        struct compute_memory_pool *pool = rctx->screen->global_pool;
        struct pipe_transfer *transfer = util_slab_alloc(&rctx->pool_transfers);
        struct r600_resource_global* buffer =
-               (struct r600_resource_global*)transfer->resource;
+               (struct r600_resource_global*)resource;
        uint32_t* map;
 
        compute_memory_finalize_pending(pool, ctx_);