mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2025-07-04 02:24:29 +00:00
Fix for PR2472. Use movss to set lower 32-bits of a zero XMM vector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53386 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
@ -2870,6 +2870,8 @@ def : Pat<(v4f32 (X86vzmovl (v4f32 (scalar_to_vector FR32:$src)))),
|
|||||||
(MOVLSS2PSrr (V_SET0), FR32:$src)>, Requires<[HasSSE2]>;
|
(MOVLSS2PSrr (V_SET0), FR32:$src)>, Requires<[HasSSE2]>;
|
||||||
def : Pat<(v4f32 (X86vzmovl (v4f32 VR128:$src))),
|
def : Pat<(v4f32 (X86vzmovl (v4f32 VR128:$src))),
|
||||||
(MOVLPSrr (V_SET0), VR128:$src)>, Requires<[HasSSE2]>;
|
(MOVLPSrr (V_SET0), VR128:$src)>, Requires<[HasSSE2]>;
|
||||||
|
def : Pat<(v4i32 (X86vzmovl (v4i32 VR128:$src))),
|
||||||
|
(MOVLPSrr (V_SET0), VR128:$src)>, Requires<[HasSSE2]>;
|
||||||
}
|
}
|
||||||
|
|
||||||
// Splat v2f64 / v2i64
|
// Splat v2f64 / v2i64
|
||||||
|
10
test/CodeGen/X86/vec_set-J.ll
Normal file
10
test/CodeGen/X86/vec_set-J.ll
Normal file
@ -0,0 +1,10 @@
|
|||||||
|
; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 | grep movss
|
||||||
|
; PR2472
|
||||||
|
|
||||||
|
define <4 x i32> @a(<4 x i32> %a) nounwind {
|
||||||
|
entry:
|
||||||
|
%vecext = extractelement <4 x i32> %a, i32 0
|
||||||
|
insertelement <4 x i32> zeroinitializer, i32 %vecext, i32 0
|
||||||
|
%add = add <4 x i32> %a, %0
|
||||||
|
ret <4 x i32> %add
|
||||||
|
}
|
Reference in New Issue
Block a user