[ARM][AARCH64][NEON]: Wrong return type of NEON intrinsic vqrshrunh_n_s16, vqrshruns_n_s32, and vqrshrund_n_s64 in arm_neon.h #80819
Conversation
|
Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write If you have received no comments on your PR for a week, you can request a review If you have further questions, they may be answered by the LLVM GitHub User Guide. You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums. |
llvmbot
added
clang
|
@llvm/pr-subscribers-clang Author: None (hlivin01) Changes
Full diff: https://github.com/llvm/llvm-project/pull/80819.diff 2 Files Affected:
diff --git a/clang/include/clang/Basic/arm_neon.td b/clang/include/clang/Basic/arm_neon.td
index 9cb7e0981384b0..f16de97f4e6bda 100644
--- a/clang/include/clang/Basic/arm_neon.td
+++ b/clang/include/clang/Basic/arm_neon.td
@@ -1354,9 +1354,9 @@ let isScalarNarrowShift = 1 in {
// Signed/Unsigned Saturating Rounded Shift Right Narrow (Immediate)
def SCALAR_SQRSHRN_N: SInst<"vqrshrn_n", "(1<)1I", "SsSiSlSUsSUiSUl">;
// Signed Saturating Shift Right Unsigned Narrow (Immediate)
- def SCALAR_SQSHRUN_N: SInst<"vqshrun_n", "(1<)1I", "SsSiSl">;
+ def SCALAR_SQSHRUN_N: SInst<"vqshrun_n", "(1<U)1I", "SsSiSl">;
// Signed Saturating Rounded Shift Right Unsigned Narrow (Immediate)
- def SCALAR_SQRSHRUN_N: SInst<"vqrshrun_n", "(1<)1I", "SsSiSl">;
+ def SCALAR_SQRSHRUN_N: SInst<"vqrshrun_n", "(1<U)1I", "SsSiSl">;
}
////////////////////////////////////////////////////////////////////////////////
diff --git a/clang/test/CodeGen/aarch64-neon-intrinsics.c b/clang/test/CodeGen/aarch64-neon-intrinsics.c
index 7c53b9b0af6bbc..eeb50d095a5c97 100644
--- a/clang/test/CodeGen/aarch64-neon-intrinsics.c
+++ b/clang/test/CodeGen/aarch64-neon-intrinsics.c
@@ -14132,8 +14132,8 @@ int32_t test_vqshrund_n_s64(int64_t a) {
// CHECK: [[VQRSHRUNH_N_S16:%.*]] = call <8 x i8> @llvm.aarch64.neon.sqrshrun.v8i8(<8 x i16> [[TMP0]], i32 8)
// CHECK: [[TMP1:%.*]] = extractelement <8 x i8> [[VQRSHRUNH_N_S16]], i64 0
// CHECK: ret i8 [[TMP1]]
-int8_t test_vqrshrunh_n_s16(int16_t a) {
- return (int8_t)vqrshrunh_n_s16(a, 8);
+uint8_t test_vqrshrunh_n_s16(int16_t a) {
+ return (uint8_t)vqrshrunh_n_s16(a, 8);
}
// CHECK-LABEL: @test_vqrshruns_n_s32(
@@ -14141,15 +14141,15 @@ int8_t test_vqrshrunh_n_s16(int16_t a) {
// CHECK: [[VQRSHRUNS_N_S32:%.*]] = call <4 x i16> @llvm.aarch64.neon.sqrshrun.v4i16(<4 x i32> [[TMP0]], i32 16)
// CHECK: [[TMP1:%.*]] = extractelement <4 x i16> [[VQRSHRUNS_N_S32]], i64 0
// CHECK: ret i16 [[TMP1]]
-int16_t test_vqrshruns_n_s32(int32_t a) {
- return (int16_t)vqrshruns_n_s32(a, 16);
+uint16_t test_vqrshruns_n_s32(int32_t a) {
+ return (uint16_t)vqrshruns_n_s32(a, 16);
}
// CHECK-LABEL: @test_vqrshrund_n_s64(
// CHECK: [[VQRSHRUND_N_S64:%.*]] = call i32 @llvm.aarch64.neon.sqrshrun.i32(i64 %a, i32 32)
// CHECK: ret i32 [[VQRSHRUND_N_S64]]
-int32_t test_vqrshrund_n_s64(int64_t a) {
- return (int32_t)vqrshrund_n_s64(a, 32);
+uint32_t test_vqrshrund_n_s64(int64_t a) {
+ return (uint32_t)vqrshrund_n_s64(a, 32);
}
// CHECK-LABEL: @test_vcvts_n_f32_s32(
|
vqrshrunh_n_s16, vqrshruns_n_s32, and vqrshrund_n_s64 in arm_neon.h * fixes llvm#71751 * changed return types in the table gen file responsible for generation of the problematic intrinsics * this is to ensure that the return type for the functions is the same as specified in the Arm Developer Documentation and avoid casting bugs (https://developer.arm.com/architectures/instruction-sets/intrinsics/vqrshrunh_n_s16) * updated lit tests to reflect the change in return type, worth noting that LLVM does not seems to differentiate signed and unsigned ints in the IR, hence the change in type cannot be checked in IR as far as I am aware
|
The author (Gleb) doesn't yet have llvm commit access, so has asked me to merge it for him. |
|
@hlivin01 Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested Please check whether problems have been caused by your change specifically, as How to do this, and the rest of the post-merge process, is covered in detail here. If your change does cause a problem, it may be reverted, or you can revert it yourself. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! |
of the problematic intrinsics
as specified in the Arm Developer Documentation and avoid casting
bugs
(https://developer.arm.com/architectures/instruction-sets/intrinsics/vqrshrunh_n_s16)
that LLVM does not seems to differentiate signed and unsigned ints in
the IR, hence the change in type cannot be checked in IR as far as I am
aware