Fix a stackmap bug introduced in r220710.

[oota-llvm.git] / test / CodeGen / X86 / vec_ss_load_fold.ll

diff --git a/test/CodeGen/X86/vec_ss_load_fold.ll b/test/CodeGen/X86/vec_ss_load_fold.ll
index 42831f00225be12527ee624304a58997325521dc..80f12a2dec2cf62b5a19aae88954d090775b98f5 100644 (file)
--- a/test/CodeGen/X86/vec_ss_load_fold.ll
+++ b/test/CodeGen/X86/vec_ss_load_fold.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=x86 -mattr=+sse,+sse2 | FileCheck %s
+; RUN: llc < %s -march=x86 -mattr=+sse,+sse2,+sse4.1 | FileCheck %s
 
 target datalayout = "e-p:32:32"
 target triple = "i686-apple-darwin8.7.2"
@@ -15,10 +15,10 @@ define i16 @test1(float %f) nounwind {
        %tmp.upgrd.1 = tail call i32 @llvm.x86.sse.cvttss2si( <4 x float> %tmp59 )              ; <i32> [#uses=1]
        %tmp69 = trunc i32 %tmp.upgrd.1 to i16          ; <i16> [#uses=1]
        ret i16 %tmp69
-; CHECK: test1:
-; CHECK: subss LCPI1_
-; CHECK: mulss LCPI1_
-; CHECK: minss LCPI1_
+; CHECK-LABEL: test1:
+; CHECK: subss LCPI0_
+; CHECK: mulss LCPI0_
+; CHECK: minss LCPI0_
 }
 
 define i16 @test2(float %f) nounwind {
@@ -30,10 +30,10 @@ define i16 @test2(float %f) nounwind {
        %tmp = tail call i32 @llvm.x86.sse.cvttss2si( <4 x float> %tmp59 )              ; <i32> [#uses=1]
        %tmp69 = trunc i32 %tmp to i16          ; <i16> [#uses=1]
        ret i16 %tmp69
-; CHECK: test2:
-; CHECK: addss LCPI2_
-; CHECK: mulss LCPI2_
-; CHECK: minss LCPI2_
+; CHECK-LABEL: test2:
+; CHECK: addss LCPI1_
+; CHECK: mulss LCPI1_
+; CHECK: minss LCPI1_
 }
 
 declare <4 x float> @llvm.x86.sse.sub.ss(<4 x float>, <4 x float>)
@@ -45,3 +45,42 @@ declare <4 x float> @llvm.x86.sse.min.ss(<4 x float>, <4 x float>)
 declare <4 x float> @llvm.x86.sse.max.ss(<4 x float>, <4 x float>)
 
 declare i32 @llvm.x86.sse.cvttss2si(<4 x float>)
+
+
+declare <4 x float> @llvm.x86.sse41.round.ss(<4 x float>, <4 x float>, i32)
+declare <4 x float> @f()
+
+define <4 x float> @test3(<4 x float> %A, float *%b, i32 %C) nounwind {
+  %a = load float *%b
+  %B = insertelement <4 x float> undef, float %a, i32 0
+  %X = call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %A, <4 x float> %B, i32 4)
+  ret <4 x float> %X
+; CHECK-LABEL: test3:
+; CHECK: roundss       $4, (%eax), %xmm0
+}
+
+define <4 x float> @test4(<4 x float> %A, float *%b, i32 %C) nounwind {
+  %a = load float *%b
+  %B = insertelement <4 x float> undef, float %a, i32 0
+  %q = call <4 x float> @f()
+  %X = call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %q, <4 x float> %B, i32 4)
+  ret <4 x float> %X
+; CHECK-LABEL: test4:
+; CHECK: movss (%eax), %xmm
+; CHECK: call
+; CHECK: roundss $4, %xmm{{.*}}, %xmm0
+}
+
+; PR13576 
+define  <2 x double> @test5() nounwind uwtable readnone noinline {
+entry:
+  %0 = tail call <2 x double> @llvm.x86.sse2.cvtsi2sd(<2 x double> <double
+4.569870e+02, double 1.233210e+02>, i32 128) nounwind readnone
+  ret <2 x double> %0
+; CHECK-LABEL: test5:
+; CHECK: mov
+; CHECK: mov
+; CHECK: cvtsi2sd
+}
+
+declare <2 x double> @llvm.x86.sse2.cvtsi2sd(<2 x double>, i32) nounwind readnone