ac/llvm: support shifts on 16 bit vec2
authorKarol Herbst <kherbst@redhat.com>
Thu, 20 Apr 2023 10:58:19 +0000 (12:58 +0200)
committerKarol Herbst <kherbst@redhat.com>
Mon, 24 Apr 2023 11:17:05 +0000 (13:17 +0200)
commit9d7ba3801396bac2b01cd4fa9486c6f20a55d0a0
tree8dc81d8ed14c4ea53af8f068bde83506c4788e14
parent037b56bf0fa589bdd9887937c089f4b50a75b351
ac/llvm: support shifts on 16 bit vec2

In OpenCL we can actually end up with those.

Fixes `basic astype` and those `integer_ops` OpenCL CTS tests:
integer_hadd
integer_rhadd
integer_upsample
quick_short_shift
quick_ushort_shift

Cc: mesa-stable
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22597>
src/amd/llvm/ac_nir_to_llvm.c