iris: don't synchronize BO for batch decoding
We don't need to go to the kernel to synchronize the BO we want to
decode with INTEL_DEBUG=bat, mostly because we'll decode what was
written by the driver in the batch.
This also works around an issue in the simulation environment.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 9ac81f1890
("iris: decoder fixes")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15517>
This commit is contained in:
parent
729f95a110
commit
a55cc061fd
|
@ -143,7 +143,7 @@ decode_get_bo(void *v_batch, bool ppgtt, uint64_t address)
|
|||
return (struct intel_batch_decode_bo) {
|
||||
.addr = bo_address,
|
||||
.size = bo->size,
|
||||
.map = iris_bo_map(batch->dbg, bo, MAP_READ),
|
||||
.map = iris_bo_map(batch->dbg, bo, MAP_READ | MAP_ASYNC),
|
||||
};
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in New Issue