Fix implicit vector conversion

[MartinWehking]

Previously, the unsigned NEON intrinsic variants of 'vqshrun_high_n' and 'vqrshrun_high_n' were using signed integer types for their first argument and return values.
These should be unsigned according to developer.arm.com, however.

Adjust the test cases accordingly.

[llvmbot]

@llvm/pr-subscribers-clang

Author: Martin Wehking (MartinWehking)

Changes

Previously, the unsigned NEON intrinsic variants of 'vqshrun_high_n' and 'vqrshrun_high_n' were using signed integer types for their first argument and return values.
These should be unsigned according to developer.arm.com, however.

Adjust the test cases accordingly.

---

Full diff: https://github.com/llvm/llvm-project/pull/149970.diff

2 Files Affected:

• (modified) clang/include/clang/Basic/arm_neon.td (+2-2)
• (modified) clang/test/CodeGen/AArch64/neon-intrinsics.c (+6-6)

diff --git a/clang/include/clang/Basic/arm_neon.td b/clang/include/clang/Basic/arm_neon.td
index 0daef4ad1701f..ef196103035e8 100644
--- a/clang/include/clang/Basic/arm_neon.td
+++ b/clang/include/clang/Basic/arm_neon.td
@@ -964,11 +964,11 @@ def SLI_N : WInst<"vsli_n", "...I", "PlQPl", [ImmCheck<2, ImmCheckShiftLeft, 0>]
 // Right shift narrow high
 def SHRN_HIGH_N    : IOpInst<"vshrn_high_n", "<(<q).I",
                              "HsHiHlHUsHUiHUl", OP_NARROW_HI>;
-def QSHRUN_HIGH_N  : SOpInst<"vqshrun_high_n", "<(<q).I",
+def QSHRUN_HIGH_N  : SOpInst<"vqshrun_high_n", "(<U)(<Uq).I",
                              "HsHiHl", OP_NARROW_HI>;
 def RSHRN_HIGH_N   : IOpInst<"vrshrn_high_n", "<(<q).I",
                              "HsHiHlHUsHUiHUl", OP_NARROW_HI>;
-def QRSHRUN_HIGH_N : SOpInst<"vqrshrun_high_n", "<(<q).I",
+def QRSHRUN_HIGH_N : SOpInst<"vqrshrun_high_n", "(<U)(<Uq).I",
                              "HsHiHl", OP_NARROW_HI>;
 def QSHRN_HIGH_N   : SOpInst<"vqshrn_high_n", "<(<q).I",
                              "HsHiHlHUsHUiHUl", OP_NARROW_HI>;
diff --git a/clang/test/CodeGen/AArch64/neon-intrinsics.c b/clang/test/CodeGen/AArch64/neon-intrinsics.c
index 6304245614342..035e1ca1b45e8 100644
--- a/clang/test/CodeGen/AArch64/neon-intrinsics.c
+++ b/clang/test/CodeGen/AArch64/neon-intrinsics.c
@@ -8585,7 +8585,7 @@ uint32x2_t test_vqshrun_n_s64(int64x2_t a) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> [[VQSHRUN_N3]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 // CHECK-NEXT:    ret <16 x i8> [[SHUFFLE_I]]
 //
-int8x16_t test_vqshrun_high_n_s16(int8x8_t a, int16x8_t b) {
+uint8x16_t test_vqshrun_high_n_s16(uint8x8_t a, int16x8_t b) {
   return vqshrun_high_n_s16(a, b, 3);
 }
 
@@ -8598,7 +8598,7 @@ int8x16_t test_vqshrun_high_n_s16(int8x8_t a, int16x8_t b) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <4 x i16> [[A]], <4 x i16> [[VQSHRUN_N3]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 // CHECK-NEXT:    ret <8 x i16> [[SHUFFLE_I]]
 //
-int16x8_t test_vqshrun_high_n_s32(int16x4_t a, int32x4_t b) {
+uint16x8_t test_vqshrun_high_n_s32(uint16x4_t a, int32x4_t b) {
   return vqshrun_high_n_s32(a, b, 9);
 }
 
@@ -8611,7 +8611,7 @@ int16x8_t test_vqshrun_high_n_s32(int16x4_t a, int32x4_t b) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <2 x i32> [[A]], <2 x i32> [[VQSHRUN_N3]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 // CHECK-NEXT:    ret <4 x i32> [[SHUFFLE_I]]
 //
-int32x4_t test_vqshrun_high_n_s64(int32x2_t a, int64x2_t b) {
+uint32x4_t test_vqshrun_high_n_s64(uint32x2_t a, int64x2_t b) {
   return vqshrun_high_n_s64(a, b, 19);
 }
 
@@ -8810,7 +8810,7 @@ uint32x2_t test_vqrshrun_n_s64(int64x2_t a) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <8 x i8> [[A]], <8 x i8> [[VQRSHRUN_N3]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 // CHECK-NEXT:    ret <16 x i8> [[SHUFFLE_I]]
 //
-int8x16_t test_vqrshrun_high_n_s16(int8x8_t a, int16x8_t b) {
+uint8x16_t test_vqrshrun_high_n_s16(uint8x8_t a, int16x8_t b) {
   return vqrshrun_high_n_s16(a, b, 3);
 }
 
@@ -8823,7 +8823,7 @@ int8x16_t test_vqrshrun_high_n_s16(int8x8_t a, int16x8_t b) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <4 x i16> [[A]], <4 x i16> [[VQRSHRUN_N3]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 // CHECK-NEXT:    ret <8 x i16> [[SHUFFLE_I]]
 //
-int16x8_t test_vqrshrun_high_n_s32(int16x4_t a, int32x4_t b) {
+uint16x8_t test_vqrshrun_high_n_s32(uint16x4_t a, int32x4_t b) {
   return vqrshrun_high_n_s32(a, b, 9);
 }
 
@@ -8836,7 +8836,7 @@ int16x8_t test_vqrshrun_high_n_s32(int16x4_t a, int32x4_t b) {
 // CHECK-NEXT:    [[SHUFFLE_I:%.*]] = shufflevector <2 x i32> [[A]], <2 x i32> [[VQRSHRUN_N3]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 // CHECK-NEXT:    ret <4 x i32> [[SHUFFLE_I]]
 //
-int32x4_t test_vqrshrun_high_n_s64(int32x2_t a, int64x2_t b) {
+uint32x4_t test_vqrshrun_high_n_s64(uint32x2_t a, int64x2_t b) {
   return vqrshrun_high_n_s64(a, b, 19);
 }
 

[Lukacma]

LGTM

[FantasqueX]

Hi, thans for fixing! I wonder whether this could be backport to llvm-21?