Index: lib/Target/ARM/ARMCallingConv.td
===================================================================
--- lib/Target/ARM/ARMCallingConv.td
+++ lib/Target/ARM/ARMCallingConv.td
@@ -30,8 +30,8 @@
CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>,
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
// f64 and v2f64 are passed in adjacent GPRs, possibly split onto the stack
CCIfType<[f64, v2f64], CCCustom<"CC_ARM_APCS_Custom_f64">>,
@@ -56,8 +56,8 @@
CCIfSwiftError<CCIfType<[i32], CCAssignToReg<[R8]>>>,
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
CCIfType<[f64, v2f64], CCCustom<"RetCC_ARM_APCS_Custom_f64">>,
@@ -71,8 +71,8 @@
let Entry = 1 in
def FastCC_ARM_APCS : CallingConv<[
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>,
CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>,
@@ -91,8 +91,8 @@
let Entry = 1 in
def RetFastCC_ARM_APCS : CallingConv<[
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
CCIfType<[v2f64], CCAssignToReg<[Q0, Q1, Q2, Q3]>>,
CCIfType<[f64], CCAssignToReg<[D0, D1, D2, D3, D4, D5, D6, D7]>>,
@@ -108,8 +108,8 @@
let Entry = 1 in
def CC_ARM_APCS_GHC : CallingConv<[
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
CCIfType<[v2f64], CCAssignToReg<[Q4, Q5]>>,
CCIfType<[f64], CCAssignToReg<[D8, D9, D10, D11]>>,
@@ -165,8 +165,8 @@
CCIfNest<CCAssignToReg<[R12]>>,
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
// Pass SwiftSelf in a callee saved register.
CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>,
@@ -182,8 +182,8 @@
let Entry = 1 in
def RetCC_ARM_AAPCS : CallingConv<[
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v8f16,v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v8f16,v16i8, v4f32], CCBitConvertToType<v2f64>>,
// Pass SwiftSelf in a callee saved register.
CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>,
@@ -208,8 +208,8 @@
CCIfByVal<CCPassByVal<4, 4>>,
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
// Pass SwiftSelf in a callee saved register.
CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>,
@@ -230,8 +230,8 @@
let Entry = 1 in
def RetCC_ARM_AAPCS_VFP : CallingConv<[
// Handle all vector types as either f64 or v2f64.
- CCIfType<[v1i64, v2i32, v4i16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
- CCIfType<[v2i64, v4i32, v8i16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+ CCIfType<[v1i64, v2i32, v4i16, v4f16, v4f16, v8i8, v2f32], CCBitConvertToType<f64>>,
+ CCIfType<[v2i64, v4i32, v8i16, v8f16, v8f16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
// Pass SwiftSelf in a callee saved register.
CCIfSwiftSelf<CCIfType<[i32], CCAssignToReg<[R10]>>>,
Index: test/CodeGen/ARM/fp16-vector-argument.ll
===================================================================
--- /dev/null
+++ test/CodeGen/ARM/fp16-vector-argument.ll
@@ -0,0 +1,56 @@
+; RUN: llc -mtriple=armv8.2a -mattr=+armv8.2-a,+fullfp16,+neon -target-abi=apcs-gnu %s | FileCheck %s --ckeck-prefix=EL
+; RUN: llc -mtriple=armeb -mattr=+armv8.2-a,+fullfp16,+neon -target-abi=apcs-gnu %s | FileCheck %s --check-prefix=EB
+declare <4 x half> @llvm.fabs.v4f16(<4 x half>)
+declare <8 x half> @llvm.fabs.v8f16(<8 x half>)
+
+define <4 x half> @test_vabs_f16(<4 x half> %a) {
+; EL-LABEL: test_vabs_f16:
+; EL: .fnstart
+; EL-NEXT: @ %bb.0:
+; EL-NEXT: vmov d16, r0, r1
+; EL-NEXT: vabs.f16 d16, d16
+; EL-NEXT: vmov r0, r1, d16
+; EL-NEXT: bx lr
+
+; EB-LABEL: test_vabs_f16:
+; EB: .fnstart
+; EB-NEXT: @ %bb.0:
+; EB-NEXT: vmov d16, r1, r0
+; EB-NEXT: vrev64.16 d16, d16
+; EB-NEXT: vabs.f16 d16, d16
+; EB-NEXT: vrev64.16 d16, d16
+; EB-NEXT: vmov r1, r0, d16
+; EB-NEXT: bx lr
+
+entry:
+ %vabs1.i = tail call <4 x half> @llvm.fabs.v4f16(<4 x half> %a)
+ ret <4 x half> %vabs1.i
+}
+
+define <8 x half> @test2_vabs_f16(<8 x half> %a) {
+; EL-LABEL: test2_vabs_f16:
+; EL: .fnstart
+; EL-NEXT: @ %bb.0:
+; EL-NEXT: vmov d17, r2, r3
+; EL-NEXT: vmov d16, r0, r1
+; EL-NEXT: vabs.f16 q8, q8
+; EL-NEXT: vmov r0, r1, d16
+; EL-NEXT: vmov r2, r3, d17
+; EL-NEXT: bx lr
+
+; EB-NEXT: test2_vabs_f16:
+; EB-NEXT: .fnstart
+; EB-NEXT: @ %bb.0:
+; EB-NEXT: vmov d17, r3, r2
+; EB-NEXT: vmov d16, r1, r0
+; EB-NEXT: vrev64.16 q8, q8
+; EB-NEXT: vabs.f16 q8, q8
+; EB-NEXT: vrev64.16 q8, q8
+; EB-NEXT: vmov r1, r0, d16
+; EB-NEXT: vmov r3, r2, d17
+; EB-NEXT: bx lr
+entry:
+ %vabs1.i = tail call <8 x half> @llvm.fabs.v8f16(<8 x half> %a)
+ ret <8 x half> %vabs1.i
+}
+