Summary:
This fixes failure when trying to select
insertelement <4 x half> undef, half %a, i64 0
which gets transformed to a scalar_to_vector node.
The accompanying v4 and v8 tests fail instruction selection without this
patch.
Reviewers: ab, jmolloy
Subscribers: srhines, llvm-commits
Differential Revision: http://reviews.llvm.org/D15322
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@255072
91177308-0d34-0410-b5e6-
96231b3b80d8
(v2i64 (INSERT_SUBREG (v2i64 (IMPLICIT_DEF)),
(i64 FPR64:$Rn), dsub))>;
+def : Pat<(v4f16 (scalar_to_vector (f16 FPR16:$Rn))),
+ (INSERT_SUBREG (v4f16 (IMPLICIT_DEF)), FPR16:$Rn, hsub)>;
+def : Pat<(v8f16 (scalar_to_vector (f16 FPR16:$Rn))),
+ (INSERT_SUBREG (v8f16 (IMPLICIT_DEF)), FPR16:$Rn, hsub)>;
+
def : Pat<(v4f32 (scalar_to_vector (f32 FPR32:$Rn))),
(INSERT_SUBREG (v4f32 (IMPLICIT_DEF)), FPR32:$Rn, ssub)>;
def : Pat<(v2f32 (scalar_to_vector (f32 FPR32:$Rn))),
ret <4 x half> %1
}
+define void @test_insert_at_zero(half %a, <4 x half>* %b) #0 {
+; CHECK-LABEL: test_insert_at_zero:
+; CHECK-NEXT: str d0, [x0]
+; CHECK-NEXT: ret
+ %1 = insertelement <4 x half> undef, half %a, i64 0
+ store <4 x half> %1, <4 x half>* %b, align 4
+ ret void
+}
+
attributes #0 = { nounwind }
ret <8 x half> %1
}
+define void @test_insert_at_zero(half %a, <8 x half>* %b) #0 {
+; CHECK-LABEL: test_insert_at_zero:
+; CHECK-NEXT: str q0, [x0]
+; CHECK-NEXT: ret
+ %1 = insertelement <8 x half> undef, half %a, i64 0
+ store <8 x half> %1, <8 x half>* %b, align 4
+ ret void
+}
+
attributes #0 = { nounwind }