[clang] [Headers][X86] Allow SSE2/AVX2/AVX512F/AVX512BW/AVX512DQ integer arithmetic intrinsics to be used in constexpr (PR #157582)

[Simon Pilgrim via cfe-commits]

================
@@ -4174,40 +4174,46 @@ __m512i test_mm512_rolv_epi32(__m512i __A, __m512i __B) {
   // CHECK: @llvm.fshl.v16i32
   return _mm512_rolv_epi32(__A, __B); 
 }
+TEST_CONSTEXPR(match_v16si(_mm512_rolv_epi32((__m512i)(__v16si){ -1, -2, 3, -4, -5, -6, 7, 8, 9, -10, -11, -12, -13, 14, 15, -16}, (__m512i)(__v16si){ 16, 15, -14, 13, -12, -11, 10, -9, 8, -7, 6, 5, 4, -3, 2, -1}), -1, -32769, 786432, -24577, -4194305, -10485761, 7168, 67108864, 2304, -301989889, -641, -353, -193, -1073741823, 60, 2147483640));
----------------
RKSimon wrote:

it might be better if you rebase against trunk latest as this is from a recent commit that the merge seems to have gotten confused by

https://github.com/llvm/llvm-project/pull/157582