gv100/ir: Use system wide atomics
authorAlistair Popple <apopple@nvidia.com>
Mon, 7 Dec 2020 01:23:17 +0000 (12:23 +1100)
committerMarge Bot <eric+marge@anholt.net>
Thu, 7 Jan 2021 08:59:10 +0000 (08:59 +0000)
commit7f9a084e7e10769c1a297d4aaca569ca99ef0ba4
tree5264cdfb3988f86378167ad809b392224ef65430
parentb02e3053ea478e5d1f4b96d107e1a5d12233e055
gv100/ir: Use system wide atomics

Increase the scope of atomic operations from GPU to system. This is
required for support of SVM to ensure atomic access is maintained for
memory buffers that are not local to the current GPU.

Signed-off-by: Alistair Popple <apopple@nvidia.com>
Reviewed-by: Ben Skeggs <bskeggs@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7955>
src/gallium/drivers/nouveau/codegen/nv50_ir_emit_gv100.cpp