diff options
author | Bastien Montagne <montagne29@wanadoo.fr> | 2017-09-25 10:56:02 +0300 |
---|---|---|
committer | Bastien Montagne <montagne29@wanadoo.fr> | 2017-09-25 10:56:02 +0300 |
commit | 01a3c6b204bfbe149dc2cc2278dd58f206af7ae2 (patch) | |
tree | d997f65227966b031083bf50f60718513a6b0a36 /source/blender/blenkernel/intern/editderivedmesh.c | |
parent | cb6f07f59e4045f4bc3d4c3f8e2bba681cc6bb79 (diff) |
Tweak to fix for thread concurency in looptri generation.
Even if pointer assignment may be atomic, it does not prevent reordering
and other nifty compiler tricks, we need a memory barrier to ensure not
only that transferring pointer from wip array to final one is atomic,
but also that all previous writing to memory are “flushed” to
(visible by) all CPUs...
Thanks @sergey for finding the potential (though quite unlikely) issue.
Diffstat (limited to 'source/blender/blenkernel/intern/editderivedmesh.c')
-rw-r--r-- | source/blender/blenkernel/intern/editderivedmesh.c | 5 |
1 files changed, 4 insertions, 1 deletions
diff --git a/source/blender/blenkernel/intern/editderivedmesh.c b/source/blender/blenkernel/intern/editderivedmesh.c index 9f688432988..357420179dd 100644 --- a/source/blender/blenkernel/intern/editderivedmesh.c +++ b/source/blender/blenkernel/intern/editderivedmesh.c @@ -41,6 +41,8 @@ * is likely to be a little slow. */ +#include "atomic_ops.h" + #include "BLI_math.h" #include "BLI_jitter.h" #include "BLI_bitmap.h" @@ -662,7 +664,8 @@ static void emDM_recalcLoopTri(DerivedMesh *dm) } BLI_assert(dm->looptris.array == NULL); - SWAP(MLoopTri *, dm->looptris.array, dm->looptris.array_wip); + atomic_cas_z((size_t *)&dm->looptris.array, *(size_t *)&dm->looptris.array, *(size_t *)&dm->looptris.array_wip); + dm->looptris.array_wip = NULL; } static void emDM_foreachMappedVert( |