From 7e31fe7e20d75a0629092edab0844c45ed79eb82 Mon Sep 17 00:00:00 2001 From: Quentin Colombet Date: Fri, 29 May 2015 01:40:00 +0000 Subject: [PATCH] Add a test for the MachineCopyPropagation change landed in r238518. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@238537 91177308-0d34-0410-b5e6-96231b3b80d8 --- test/CodeGen/X86/machine-cp.ll | 55 ++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+) diff --git a/test/CodeGen/X86/machine-cp.ll b/test/CodeGen/X86/machine-cp.ll index 0006b6ea713..aaed0f0a23d 100644 --- a/test/CodeGen/X86/machine-cp.ll +++ b/test/CodeGen/X86/machine-cp.ll @@ -58,3 +58,58 @@ while.end: ; preds = %while.body, %entry %t = trunc i64 %a.addr.0.lcssa to i32 ret i32 %t } + +; Check that copy propagation does not kill thing like: +; dst = copy src <-- do not kill that. +; ... = op1 dst +; ... = op2 dst <-- this is used here. +; +; CHECK-LABEL: foo: +; CHECK: psllw $7, +; CHECK: psllw $7, +; CHECK-NEXT: pand +; CHECK-NEXT: pcmpgtb +; CHECK-NEXT: pand %xmm{{[0-9]+}}, [[SRC:%xmm[0-9]+]] +; Machine propagation used to delete the first copy as the +; first few uses were . +; CHECK-NEXT: movdqa [[SRC]], [[CPY1:%xmm[0-9]+]] +; CHECK-NEXT: movdqa [[SRC]], [[CPY2:%xmm[0-9]+]] +; CHECK-NEXT: punpckhbw [[SRC]], +; Check that CPY1 is not redefined. +; CHECK-NOT: , [[CPY1]] +; undef use, we do not care. +; CHECK: punpcklwd [[CPY1]], +; Check that CPY1 is not redefined. +; CHECK-NOT: , [[CPY1]] +; CHECK: punpcklbw [[CPY2]], [[CPY2]] +; CHECK-NEXT: punpckhwd [[CPY2]], [[CPY2]] +; CHECK-NEXT pslld $31, [[CPY2]] +; Check that CPY1 is not redefined. +; CHECK-NOT: , [[CPY1]] +; CHECK: punpcklbw [[CPY1]], [[CPY1]] +; CHECK-NEXT: punpcklwd [[CPY1]], [[CPY1]] +; CHECK-NEXT pslld $31, [[CPY1]] +define <16 x float> @foo(<16 x float> %x) { +bb: + %v3 = icmp slt <16 x i32> undef, zeroinitializer + %v14 = zext <16 x i1> %v3 to <16 x i32> + %v16 = fcmp olt <16 x float> %x, zeroinitializer + %v17 = sext <16 x i1> %v16 to <16 x i32> + %v18 = zext <16 x i1> %v16 to <16 x i32> + %v19 = xor <16 x i32> %v14, %v18 + %v20 = or <16 x i32> %v17, undef + %v21 = fptosi <16 x float> %x to <16 x i32> + %v22 = sitofp <16 x i32> %v21 to <16 x float> + %v69 = fcmp ogt <16 x float> %v22, zeroinitializer + %v75 = and <16 x i1> %v69, %v3 + %v77 = bitcast <16 x float> %v22 to <16 x i32> + %v79 = sext <16 x i1> %v75 to <16 x i32> + %v80 = and <16 x i32> undef, %v79 + %v81 = xor <16 x i32> %v77, %v80 + %v82 = and <16 x i32> undef, %v81 + %v83 = xor <16 x i32> %v19, %v82 + %v84 = and <16 x i32> %v83, %v20 + %v85 = xor <16 x i32> %v19, %v84 + %v86 = bitcast <16 x i32> %v85 to <16 x float> + ret <16 x float> %v86 +} -- 2.34.1