diff options
Diffstat (limited to 'test/CodeGen/X86')
93 files changed, 1824 insertions, 285 deletions
diff --git a/test/CodeGen/X86/2006-05-22-FPSetEQ.ll b/test/CodeGen/X86/2006-05-22-FPSetEQ.ll index 35b0159..6c5a4fb 100644 --- a/test/CodeGen/X86/2006-05-22-FPSetEQ.ll +++ b/test/CodeGen/X86/2006-05-22-FPSetEQ.ll @@ -1,5 +1,5 @@ -; RUN: llc < %s -march=x86 | grep setnp -; RUN: llc < %s -march=x86 -enable-unsafe-fp-math -enable-no-nans-fp-math | \ +; RUN: llc < %s -march=x86 -mattr=-sse | grep setnp +; RUN: llc < %s -march=x86 -mattr=-sse -enable-unsafe-fp-math -enable-no-nans-fp-math | \ ; RUN: not grep setnp define i32 @test(float %f) { diff --git a/test/CodeGen/X86/2008-07-11-SpillerBug.ll b/test/CodeGen/X86/2008-07-11-SpillerBug.ll deleted file mode 100644 index dee7415..0000000 --- a/test/CodeGen/X86/2008-07-11-SpillerBug.ll +++ /dev/null @@ -1,52 +0,0 @@ -; RUN: llc < %s -march=x86 -relocation-model=static -disable-fp-elim -post-RA-scheduler=false -asm-verbose=0 | FileCheck %s -; PR2536 - -; CHECK: andl $65534, % -; CHECK-NEXT: movl % -; CHECK-NEXT: movzwl - -@g_5 = external global i16 ; <i16*> [#uses=2] -@g_107 = external global i16 ; <i16*> [#uses=1] -@g_229 = external global i32 ; <i32*> [#uses=1] -@g_227 = external global i16 ; <i16*> [#uses=1] - -define i32 @func_54(i32 %p_55, i16 zeroext %p_56) nounwind { -entry: - load i16* @g_5, align 2 ; <i16>:0 [#uses=1] - zext i16 %0 to i32 ; <i32>:1 [#uses=1] - %.mask = and i32 %1, 65534 ; <i32> [#uses=1] - icmp eq i32 %.mask, 0 ; <i1>:2 [#uses=1] - load i32* @g_229, align 4 ; <i32>:3 [#uses=1] - load i16* @g_227, align 2 ; <i16>:4 [#uses=1] - icmp eq i16 %4, 0 ; <i1>:5 [#uses=1] - load i16* @g_5, align 2 ; <i16>:6 [#uses=1] - br label %bb - -bb: ; preds = %bb7.preheader, %entry - %indvar4 = phi i32 [ 0, %entry ], [ %indvar.next5, %bb7.preheader ] ; <i32> [#uses=1] - %p_56_addr.1.reg2mem.0 = phi i16 [ %p_56, %entry ], [ %p_56_addr.0, %bb7.preheader ] ; <i16> [#uses=2] - br i1 %2, label %bb7.preheader, label %bb5 - -bb5: ; preds = %bb - store i16 %6, i16* @g_107, align 2 - br label %bb7.preheader - -bb7.preheader: ; preds = %bb5, %bb - icmp eq i16 %p_56_addr.1.reg2mem.0, 0 ; <i1>:7 [#uses=1] - %.0 = select i1 %7, i32 1, i32 %3 ; <i32> [#uses=1] - urem i32 1, %.0 ; <i32>:8 [#uses=1] - icmp eq i32 %8, 0 ; <i1>:9 [#uses=1] - %.not = xor i1 %9, true ; <i1> [#uses=1] - %.not1 = xor i1 %5, true ; <i1> [#uses=1] - %brmerge = or i1 %.not, %.not1 ; <i1> [#uses=1] - %iftmp.6.0 = select i1 %brmerge, i32 3, i32 0 ; <i32> [#uses=1] - mul i32 %iftmp.6.0, %3 ; <i32>:10 [#uses=1] - icmp eq i32 %10, 0 ; <i1>:11 [#uses=1] - %p_56_addr.0 = select i1 %11, i16 %p_56_addr.1.reg2mem.0, i16 1 ; <i16> [#uses=1] - %indvar.next5 = add i32 %indvar4, 1 ; <i32> [#uses=2] - %exitcond6 = icmp eq i32 %indvar.next5, 17 ; <i1> [#uses=1] - br i1 %exitcond6, label %bb25, label %bb - -bb25: ; preds = %bb7.preheader - ret i32 1 -} diff --git a/test/CodeGen/X86/2008-08-31-EH_RETURN32.ll b/test/CodeGen/X86/2008-08-31-EH_RETURN32.ll index b92c789..1d27fc5 100644 --- a/test/CodeGen/X86/2008-08-31-EH_RETURN32.ll +++ b/test/CodeGen/X86/2008-08-31-EH_RETURN32.ll @@ -1,5 +1,5 @@ ; Check that eh_return & unwind_init were properly lowered -; RUN: llc < %s | grep %ebp | count 7 +; RUN: llc < %s | grep %ebp | count 9 ; RUN: llc < %s | grep %ecx | count 5 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64" diff --git a/test/CodeGen/X86/2008-08-31-EH_RETURN64.ll b/test/CodeGen/X86/2008-08-31-EH_RETURN64.ll index 00ab735..d423bfc 100644 --- a/test/CodeGen/X86/2008-08-31-EH_RETURN64.ll +++ b/test/CodeGen/X86/2008-08-31-EH_RETURN64.ll @@ -1,5 +1,5 @@ ; Check that eh_return & unwind_init were properly lowered -; RUN: llc < %s | grep %rbp | count 5 +; RUN: llc < %s | grep %rbp | count 7 ; RUN: llc < %s | grep %rcx | count 3 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128" diff --git a/test/CodeGen/X86/2008-09-18-inline-asm-2.ll b/test/CodeGen/X86/2008-09-18-inline-asm-2.ll index 947a1f1..dfd165c 100644 --- a/test/CodeGen/X86/2008-09-18-inline-asm-2.ll +++ b/test/CodeGen/X86/2008-09-18-inline-asm-2.ll @@ -1,12 +1,32 @@ -; RUN: llc < %s -march=x86 -regalloc=linearscan | grep "#%ebp %edi %ebx 8(%esi) %eax %dl" -; RUN: llc < %s -march=x86 -regalloc=fast | grep "#%ebx %esi %edi 8(%ebp) %eax %dl" -; RUN: llc < %s -march=x86 -regalloc=basic | grep "#%ebp %esi %edx 8(%edi) %eax %bl" -; RUN: llc < %s -march=x86 -regalloc=greedy | grep "#%edx %edi %ebp 8(%esi) %eax %bl" +; RUN: llc < %s -march=x86 -regalloc=linearscan | FileCheck %s +; RUN: llc < %s -march=x86 -regalloc=fast | FileCheck %s +; RUN: llc < %s -march=x86 -regalloc=basic | FileCheck %s +; RUN: llc < %s -march=x86 -regalloc=greedy | FileCheck %s -; The 1st, 2nd, 3rd and 5th registers above must all be different. The registers +; The 1st, 2nd, 3rd and 5th registers must all be different. The registers ; referenced in the 4th and 6th operands must not be the same as the 1st or 5th -; operand. There are many combinations that work; this is what llc puts out now. -; ModuleID = '<stdin>' +; operand. +; +; CHECK: 1st=[[A1:%...]] +; CHECK-NOT: [[A1]] +; CHECK: 2nd=[[A2:%...]] +; CHECK-NOT: [[A1]] +; CHECK-NOT: [[A2]] +; CHECK: 3rd=[[A3:%...]] +; CHECK-NOT: [[A1]] +; CHECK-NOT: [[A2]] +; CHECK-NOT: [[A3]] +; CHECK: 5th=[[A5:%...]] +; CHECK-NOT: [[A1]] +; CHECK-NOT; [[A5]] +; CHECK: =4th + +; The 6th operand is an 8-bit register, and it mustn't alias the 1st and 5th. +; CHECK: 1%e[[S1:.]]x +; CHECK: 5%e[[S5:.]]x +; CHECK-NOT: %[[S1]] +; CHECK-NOT: %[[S5]] + target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i386-apple-darwin8" %struct.foo = type { i32, i32, i8* } @@ -19,7 +39,7 @@ entry: %3 = load i32* %0, align 4 ; <i32> [#uses=1] %4 = load i32* %1, align 4 ; <i32> [#uses=1] %5 = load i8* %state, align 1 ; <i8> [#uses=1] - %asmtmp = tail call { i32, i32, i32, i32 } asm sideeffect "#$0 $1 $2 $3 $4 $5", "=&r,=r,=r,=*m,=&q,=*imr,1,2,*m,5,~{dirflag},~{fpsr},~{flags},~{cx}"(i8** %2, i8* %state, i32 %3, i32 %4, i8** %2, i8 %5) nounwind ; <{ i32, i32, i32, i32 }> [#uses=3] + %asmtmp = tail call { i32, i32, i32, i32 } asm sideeffect "#1st=$0 $1 2nd=$1 $2 3rd=$2 $4 5th=$4 $3=4th 1$0 1%eXx 5$4 5%eXx 6th=$5", "=&r,=r,=r,=*m,=&q,=*imr,1,2,*m,5,~{dirflag},~{fpsr},~{flags},~{cx}"(i8** %2, i8* %state, i32 %3, i32 %4, i8** %2, i8 %5) nounwind ; <{ i32, i32, i32, i32 }> [#uses=3] %asmresult = extractvalue { i32, i32, i32, i32 } %asmtmp, 0 ; <i32> [#uses=1] %asmresult1 = extractvalue { i32, i32, i32, i32 } %asmtmp, 1 ; <i32> [#uses=1] store i32 %asmresult1, i32* %0 diff --git a/test/CodeGen/X86/2008-12-19-EarlyClobberBug.ll b/test/CodeGen/X86/2008-12-19-EarlyClobberBug.ll index 5eba9b9..75e0b8a 100644 --- a/test/CodeGen/X86/2008-12-19-EarlyClobberBug.ll +++ b/test/CodeGen/X86/2008-12-19-EarlyClobberBug.ll @@ -4,7 +4,7 @@ ; CHECK: ## InlineAsm End ; CHECK-NEXT: BB0_2: -; CHECK-NEXT: movl %esi, %eax +; CHECK-NEXT: {{movl %esi, %eax|addl %edi, %esi}} @"\01LC" = internal constant [7 x i8] c"n0=%d\0A\00" ; <[7 x i8]*> [#uses=1] diff --git a/test/CodeGen/X86/2009-03-13-PHIElimBug.ll b/test/CodeGen/X86/2009-03-13-PHIElimBug.ll index 2853930..45fc269 100644 --- a/test/CodeGen/X86/2009-03-13-PHIElimBug.ll +++ b/test/CodeGen/X86/2009-03-13-PHIElimBug.ll @@ -28,5 +28,5 @@ lpad: ; preds = %cont, %entry } ; CHECK: call{{.*}}f -; CHECK-NEXT: Ltmp0: -; CHECK-NEXT: movl %eax, %esi +; CHECK: movl %eax, %esi +; CHECK: call{{.*}}g diff --git a/test/CodeGen/X86/2010-05-25-DotDebugLoc.ll b/test/CodeGen/X86/2010-05-25-DotDebugLoc.ll index 848af82..2fceab6 100644 --- a/test/CodeGen/X86/2010-05-25-DotDebugLoc.ll +++ b/test/CodeGen/X86/2010-05-25-DotDebugLoc.ll @@ -1,8 +1,10 @@ -; RUN: llc -march=x86-64 -O2 < %s | FileCheck %s -; RUN: llc -march=x86-64 -O2 -regalloc=basic < %s | FileCheck %s +; RUN: llc -mtriple=x86_64-pc-linux -O2 < %s | FileCheck %s +; RUN: llc -mtriple=x86_64-pc-linux -O2 -regalloc=basic < %s | FileCheck %s ; Test to check .debug_loc support. This test case emits many debug_loc entries. ; CHECK: Loc expr size +; CHECK-NEXT: .short +; CHECK-NEXT: .Ltmp ; CHECK-NEXT: DW_OP_reg %0 = type { double } diff --git a/test/CodeGen/X86/2010-05-26-DotDebugLoc.ll b/test/CodeGen/X86/2010-05-26-DotDebugLoc.ll index 6600cc3..7909d27 100644 --- a/test/CodeGen/X86/2010-05-26-DotDebugLoc.ll +++ b/test/CodeGen/X86/2010-05-26-DotDebugLoc.ll @@ -68,9 +68,15 @@ declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone ; CHECK: Ldebug_loc0: ; CHECK-NEXT: .quad Lfunc_begin0 ; CHECK-NEXT: .quad [[LABEL]] -; CHECK-NEXT: .short 1 +; CHECK-NEXT: Lset{{.*}} = Ltmp{{.*}}-Ltmp{{.*}} ## Loc expr size +; CHECK-NEXT: .short Lset{{.*}} +; CHECK-NEXT: Ltmp{{.*}}: ; CHECK-NEXT: .byte 85 +; CHECK-NEXT: Ltmp{{.*}}: ; CHECK-NEXT: .quad [[LABEL]] ; CHECK-NEXT: .quad [[CLOBBER]] -; CHECK-NEXT: .short 1 +; CHECK-NEXT: Lset{{.*}} = Ltmp{{.*}}-Ltmp{{.*}} ## Loc expr size +; CHECK-NEXT: .short Lset{{.*}} +; CHECK-NEXT: Ltmp{{.*}}: ; CHECK-NEXT: .byte 83 +; CHECK-NEXT: Ltmp{{.*}}:
\ No newline at end of file diff --git a/test/CodeGen/X86/2010-06-25-CoalescerSubRegDefDead.ll b/test/CodeGen/X86/2010-06-25-CoalescerSubRegDefDead.ll index 6db3ce1..bb1db59 100644 --- a/test/CodeGen/X86/2010-06-25-CoalescerSubRegDefDead.ll +++ b/test/CodeGen/X86/2010-06-25-CoalescerSubRegDefDead.ll @@ -22,7 +22,7 @@ bb: ; it is. ; ; CHECK: # %bb -; CHECK: addq $64036, %rdi +; CHECK: leaq 64036(%rdx), %rdi ; CHECK: rep;stosl %tmp5 = bitcast i32* %tmp4 to i8* diff --git a/test/CodeGen/X86/2011-02-12-shuffle.ll b/test/CodeGen/X86/2011-02-12-shuffle.ll new file mode 100644 index 0000000..b4d56d1 --- /dev/null +++ b/test/CodeGen/X86/2011-02-12-shuffle.ll @@ -0,0 +1,32 @@ +; RUN: llc < %s +; PR9165 + +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f80:128:128-v64:64:64-v128:128:128-a0:0:64-f80:32:32-n8:16:32" +target triple = "i686-pc-win32" + +define void @m_387() nounwind { +entry: + br i1 undef, label %if.end, label %UnifiedReturnBlock + +if.end: ; preds = %entry + %tmp1067 = load <16 x i32> addrspace(1)* null, align 64 + %tmp1082 = shufflevector <16 x i32> <i32 0, i32 0, i32 0, i32 undef, i32 undef, i32 0, i32 0, i32 undef, i32 0, i32 0, i32 undef, i32 undef, i32 0, i32 undef, i32 undef, i32 undef>, + <16 x i32> %tmp1067, + <16 x i32> <i32 0, i32 1, i32 2, i32 undef, i32 26, i32 5, i32 6, i32 undef, i32 8, i32 9, i32 31, i32 30, i32 12, i32 undef, i32 undef, i32 undef> + + %tmp1100 = shufflevector <16 x i32> %tmp1082, + <16 x i32> %tmp1067, + <16 x i32> <i32 0, i32 1, i32 2, i32 undef, i32 4, i32 5, i32 6, i32 18, i32 8, i32 9, i32 10, i32 11, i32 12, i32 25, i32 undef, i32 17> + + %tmp1112 = shufflevector <16 x i32> %tmp1100, + <16 x i32> %tmp1067, + <16 x i32> <i32 0, i32 1, i32 2, i32 24, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 18, i32 15> + + store <16 x i32> %tmp1112, <16 x i32> addrspace(1)* undef, align 64 + + ret void + +UnifiedReturnBlock: ; preds = %entry + ret void +} + diff --git a/test/CodeGen/X86/2011-05-09-loaduse.ll b/test/CodeGen/X86/2011-05-09-loaduse.ll new file mode 100644 index 0000000..8673d74 --- /dev/null +++ b/test/CodeGen/X86/2011-05-09-loaduse.ll @@ -0,0 +1,13 @@ +; RUN: llc < %s -march=x86 -mcpu=corei7 | FileCheck %s + +;CHECK: test +;CHECK-not: pshufd +;CHECK: ret +define float @test(<4 x float>* %A) nounwind { +entry: + %T = load <4 x float>* %A + %R = extractelement <4 x float> %T, i32 3 + store <4 x float><float 0.0, float 0.0, float 0.0, float 0.0>, <4 x float>* %A + ret float %R +} + diff --git a/test/CodeGen/X86/2011-05-26-UnreachableBlockElim.ll b/test/CodeGen/X86/2011-05-26-UnreachableBlockElim.ll new file mode 100644 index 0000000..0f18f09 --- /dev/null +++ b/test/CodeGen/X86/2011-05-26-UnreachableBlockElim.ll @@ -0,0 +1,53 @@ +; RUN: llc < %s -verify-coalescing +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.0" + +%struct.attrib = type { i32, i32 } +%struct.dfa = type { [80 x i8], i32, %struct.state*, i32, i32, %struct.attrib*, i32, i32 } +%struct.state = type { i32, [4 x i32] } + +@aux_temp = external global %struct.dfa, align 8 + +declare i64 @llvm.objectsize.i64(i8*, i1) nounwind readnone + +declare void @__memset_chk() nounwind + +define void @dfa_add_string() nounwind uwtable ssp { +entry: + br label %if.end.i + +if.end.i: ; preds = %entry + %idxprom.i = add i64 0, 1 + br i1 undef, label %land.end.thread.i, label %land.end.i + +land.end.thread.i: ; preds = %if.end.i + %0 = call i64 @llvm.objectsize.i64(i8* undef, i1 false) nounwind + %cmp1710.i = icmp eq i64 %0, -1 + br i1 %cmp1710.i, label %cond.false156.i, label %cond.true138.i + +land.end.i: ; preds = %if.end.i + %1 = call i64 @llvm.objectsize.i64(i8* undef, i1 false) nounwind + %cmp17.i = icmp eq i64 %1, -1 + br i1 %cmp17.i, label %cond.false156.i, label %cond.true138.i + +cond.true138.i: ; preds = %for.end.i, %land.end.thread.i + call void @__memset_chk() nounwind + br label %cond.end166.i + +cond.false156.i: ; preds = %for.end.i, %land.end.thread.i + %idxprom1114.i = phi i64 [ undef, %land.end.thread.i ], [ %idxprom.i, %land.end.i ] + call void @__memset_chk() nounwind + br label %cond.end166.i + +cond.end166.i: ; preds = %cond.false156.i, %cond.true138.i + %idxprom1113.i = phi i64 [ %idxprom1114.i, %cond.false156.i ], [ undef, %cond.true138.i ] + %tmp235.i = load %struct.state** getelementptr inbounds (%struct.dfa* @aux_temp, i64 0, i32 2), align 8, !tbaa !0 + %att.i = getelementptr inbounds %struct.state* %tmp235.i, i64 %idxprom1113.i, i32 0 + store i32 0, i32* %att.i, align 4, !tbaa !3 + ret void +} + +!0 = metadata !{metadata !"any pointer", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA", null} +!3 = metadata !{metadata !"int", metadata !1} diff --git a/test/CodeGen/X86/2011-05-27-CrossClassCoalescing.ll b/test/CodeGen/X86/2011-05-27-CrossClassCoalescing.ll new file mode 100644 index 0000000..c595bba --- /dev/null +++ b/test/CodeGen/X86/2011-05-27-CrossClassCoalescing.ll @@ -0,0 +1,41 @@ +; RUN: llc < %s -verify-coalescing +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.0" + +@bit_count = external constant [256 x i32], align 16 + +define fastcc void @unate_intersect() nounwind uwtable ssp { +entry: + br label %for.body + +for.body: ; preds = %entry, %for.inc.i + br label %do.body.i + +do.body.i: ; preds = %do.body.i, %for.body + %exitcond149 = icmp eq i64 undef, undef + br i1 %exitcond149, label %land.lhs.true, label %do.body.i + +land.lhs.true: ; preds = %do.body.i + br label %for.body.i + +for.body.i: ; preds = %for.inc.i, %if.then + %tmp3524.i = phi i32 [ 0, %land.lhs.true ], [ %tmp351.i, %for.inc.i ] + %tmp6.i12 = load i32* undef, align 4 + br i1 undef, label %for.inc.i, label %if.then.i17 + +if.then.i17: ; preds = %for.body.i + %shr.i14 = lshr i32 %tmp6.i12, 8 + %and14.i = and i32 %shr.i14, 255 + %idxprom15.i = zext i32 %and14.i to i64 + %arrayidx16.i = getelementptr inbounds [256 x i32]* @bit_count, i64 0, i64 %idxprom15.i + %tmp17.i15 = load i32* %arrayidx16.i, align 4 + %add.i = add i32 0, %tmp3524.i + %add24.i = add i32 %add.i, %tmp17.i15 + %add31.i = add i32 %add24.i, 0 + %add33.i = add i32 %add31.i, 0 + br label %for.inc.i + +for.inc.i: ; preds = %if.then.i17, %for.body.i + %tmp351.i = phi i32 [ %add33.i, %if.then.i17 ], [ %tmp3524.i, %for.body.i ] + br label %for.body.i +} diff --git a/test/CodeGen/X86/2011-05-31-movmsk.ll b/test/CodeGen/X86/2011-05-31-movmsk.ll new file mode 100644 index 0000000..2b54d5c --- /dev/null +++ b/test/CodeGen/X86/2011-05-31-movmsk.ll @@ -0,0 +1,79 @@ +; RUN: llc -mcpu=core2 < %s | FileCheck %s +; ModuleID = '<stdin>' +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.6" + +%0 = type { double } +%union.anon = type { float } + +define i32 @double_signbit(double %d1) nounwind uwtable readnone ssp { +entry: + %__x.addr.i = alloca double, align 8 + %__u.i = alloca %0, align 8 + %0 = bitcast double* %__x.addr.i to i8* + %1 = bitcast %0* %__u.i to i8* + store double %d1, double* %__x.addr.i, align 8 + %__f.i = getelementptr inbounds %0* %__u.i, i64 0, i32 0 + store double %d1, double* %__f.i, align 8 + %tmp = bitcast double %d1 to i64 +; CHECK-NOT: shr +; CHECK: movmskpd +; CHECK-NEXT: and + %tmp1 = lshr i64 %tmp, 63 + %shr.i = trunc i64 %tmp1 to i32 + ret i32 %shr.i +} + +define i32 @double_add_signbit(double %d1, double %d2) nounwind uwtable readnone ssp { +entry: + %__x.addr.i = alloca double, align 8 + %__u.i = alloca %0, align 8 + %add = fadd double %d1, %d2 + %0 = bitcast double* %__x.addr.i to i8* + %1 = bitcast %0* %__u.i to i8* + store double %add, double* %__x.addr.i, align 8 + %__f.i = getelementptr inbounds %0* %__u.i, i64 0, i32 0 + store double %add, double* %__f.i, align 8 + %tmp = bitcast double %add to i64 +; CHECK-NOT: shr +; CHECK: movmskpd +; CHECK-NEXT: and + %tmp1 = lshr i64 %tmp, 63 + %shr.i = trunc i64 %tmp1 to i32 + ret i32 %shr.i +} + +define i32 @float_signbit(float %f1) nounwind uwtable readnone ssp { +entry: + %__x.addr.i = alloca float, align 4 + %__u.i = alloca %union.anon, align 4 + %0 = bitcast float* %__x.addr.i to i8* + %1 = bitcast %union.anon* %__u.i to i8* + store float %f1, float* %__x.addr.i, align 4 + %__f.i = getelementptr inbounds %union.anon* %__u.i, i64 0, i32 0 + store float %f1, float* %__f.i, align 4 + %2 = bitcast float %f1 to i32 +; CHECK-NOT: shr +; CHECK: movmskps +; CHECK-NEXT: and + %shr.i = lshr i32 %2, 31 + ret i32 %shr.i +} + +define i32 @float_add_signbit(float %f1, float %f2) nounwind uwtable readnone ssp { +entry: + %__x.addr.i = alloca float, align 4 + %__u.i = alloca %union.anon, align 4 + %add = fadd float %f1, %f2 + %0 = bitcast float* %__x.addr.i to i8* + %1 = bitcast %union.anon* %__u.i to i8* + store float %add, float* %__x.addr.i, align 4 + %__f.i = getelementptr inbounds %union.anon* %__u.i, i64 0, i32 0 + store float %add, float* %__f.i, align 4 + %2 = bitcast float %add to i32 +; CHECK-NOT: shr +; CHECK: movmskps +; CHECK-NEXT: and + %shr.i = lshr i32 %2, 31 + ret i32 %shr.i +} diff --git a/test/CodeGen/X86/2011-06-01-fildll.ll b/test/CodeGen/X86/2011-06-01-fildll.ll new file mode 100644 index 0000000..3a0b05f --- /dev/null +++ b/test/CodeGen/X86/2011-06-01-fildll.ll @@ -0,0 +1,15 @@ +; RUN: llc < %s -march=x86 | FileCheck %s +; ModuleID = '<stdin>' +target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32" +target triple = "i386-apple-macosx10.6.6" + +define float @f(i64* nocapture %x) nounwind readonly ssp { +entry: +; CHECK: movl +; CHECK-NOT: movl + %tmp1 = load i64* %x, align 4 +; CHECK: fildll + %conv = sitofp i64 %tmp1 to float + %add = fadd float %conv, 1.000000e+00 + ret float %add +} diff --git a/test/CodeGen/X86/2011-06-03-x87chain.ll b/test/CodeGen/X86/2011-06-03-x87chain.ll new file mode 100644 index 0000000..bf7f583 --- /dev/null +++ b/test/CodeGen/X86/2011-06-03-x87chain.ll @@ -0,0 +1,31 @@ +; RUN: llc < %s -march=x86 -mattr=+sse | FileCheck %s + +define float @chainfail1(i64* nocapture %a, i64* nocapture %b, i32 %x, i32 %y, float* nocapture %f) nounwind uwtable noinline ssp { +entry: + %tmp1 = load i64* %a, align 8 +; Insure x87 ops are properly chained, order preserved. +; CHECK: fildll + %conv = sitofp i64 %tmp1 to float +; CHECK: fstps + store float %conv, float* %f, align 4 +; CHECK: idivl + %div = sdiv i32 %x, %y + %conv5 = sext i32 %div to i64 + store i64 %conv5, i64* %b, align 8 + ret float %conv +} + +define float @chainfail2(i64* nocapture %a, i64* nocapture %b, i32 %x, i32 %y, float* nocapture %f) nounwind uwtable noinline ssp { +entry: +; CHECK: movl $0, + store i64 0, i64* %b, align 8 + %mul = mul nsw i32 %y, %x + %sub = add nsw i32 %mul, -1 + %idxprom = sext i32 %sub to i64 + %arrayidx = getelementptr inbounds i64* %a, i64 %idxprom + %tmp4 = load i64* %arrayidx, align 8 +; CHECK: fildll + %conv = sitofp i64 %tmp4 to float + store float %conv, float* %f, align 4 + ret float %conv +} diff --git a/test/CodeGen/X86/2011-06-06-fgetsign80bit.ll b/test/CodeGen/X86/2011-06-06-fgetsign80bit.ll new file mode 100644 index 0000000..d934148 --- /dev/null +++ b/test/CodeGen/X86/2011-06-06-fgetsign80bit.ll @@ -0,0 +1,8 @@ +; RUN: llc -march=x86-64 < %s +define i32 @signbitl(x86_fp80 %x) nounwind uwtable readnone { +entry: + %tmp4 = bitcast x86_fp80 %x to i80 + %tmp4.lobit = lshr i80 %tmp4, 79 + %tmp = trunc i80 %tmp4.lobit to i32 + ret i32 %tmp +} diff --git a/test/CodeGen/X86/2011-06-19-QuicksortCoalescerBug.ll b/test/CodeGen/X86/2011-06-19-QuicksortCoalescerBug.ll new file mode 100644 index 0000000..08178a3 --- /dev/null +++ b/test/CodeGen/X86/2011-06-19-QuicksortCoalescerBug.ll @@ -0,0 +1,31 @@ +; RUN: llc < %s -verify-coalescing +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.7.0" + +define void @Quicksort(i32* %a, i32 %l, i32 %r) nounwind ssp { +entry: + br label %tailrecurse + +tailrecurse: ; preds = %do.cond, %entry + %l.tr = phi i32 [ %l, %entry ], [ %i.1, %do.cond ] + %r.tr = phi i32 [ %r, %entry ], [ %l.tr, %do.cond ] + %idxprom12 = sext i32 %r.tr to i64 + %arrayidx14 = getelementptr inbounds i32* %a, i64 %idxprom12 + br label %do.body + +do.body: ; preds = %do.cond, %tailrecurse + %i.0 = phi i32 [ %l.tr, %tailrecurse ], [ %i.1, %do.cond ] + %add7 = add nsw i32 %i.0, 1 + %cmp = icmp sgt i32 %add7, %r.tr + br i1 %cmp, label %do.cond, label %if.then + +if.then: ; preds = %do.body + store i32 %add7, i32* %arrayidx14, align 4 + %add16 = add i32 %i.0, 2 + br label %do.cond + +do.cond: ; preds = %do.body, %if.then + %i.1 = phi i32 [ %add16, %if.then ], [ %add7, %do.body ] + %cmp19 = icmp sgt i32 %i.1, %r.tr + br i1 %cmp19, label %tailrecurse, label %do.body +} diff --git a/test/CodeGen/X86/9601.ll b/test/CodeGen/X86/9601.ll new file mode 100644 index 0000000..cd65a03 --- /dev/null +++ b/test/CodeGen/X86/9601.ll @@ -0,0 +1,12 @@ +; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu +; PR9601 +; Previously we'd crash trying to put a 32-bit float into a constraint +; for a normal 'r' register. +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-unknown-linux-gnu" + +define void @test() { +entry: + %0 = call float asm sideeffect "xchg $0, $1", "=r,*m,0,~{memory},~{dirflag},~{fpsr},~{flags}"(i32* undef, float 2.000000e+00) nounwind + unreachable +} diff --git a/test/CodeGen/X86/abi-isel.ll b/test/CodeGen/X86/abi-isel.ll index 7535e07..5068d29 100644 --- a/test/CodeGen/X86/abi-isel.ll +++ b/test/CodeGen/X86/abi-isel.ll @@ -12,17 +12,6 @@ ; RUN: llc < %s -asm-verbose=0 -mtriple=x86_64-apple-darwin -march=x86-64 -relocation-model=dynamic-no-pic -code-model=small | FileCheck %s -check-prefix=DARWIN-64-DYNAMIC ; RUN: llc < %s -asm-verbose=0 -mtriple=x86_64-apple-darwin -march=x86-64 -relocation-model=pic -code-model=small | FileCheck %s -check-prefix=DARWIN-64-PIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=i686-unknown-linux-gnu -march=x86 -relocation-model=static -code-model=small | FileCheck %s -check-prefix=LINUX-32-STATIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=i686-unknown-linux-gnu -march=x86 -relocation-model=static -code-model=small | FileCheck %s -check-prefix=LINUX-32-PIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=x86_64-unknown-linux-gnu -march=x86-64 -relocation-model=static -code-model=small | FileCheck %s -check-prefix=LINUX-64-STATIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=x86_64-unknown-linux-gnu -march=x86-64 -relocation-model=pic -code-model=small | FileCheck %s -check-prefix=LINUX-64-PIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=i686-apple-darwin -march=x86 -relocation-model=static -code-model=small | FileCheck %s -check-prefix=DARWIN-32-STATIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=i686-apple-darwin -march=x86 -relocation-model=dynamic-no-pic -code-model=small | FileCheck %s -check-prefix=DARWIN-32-DYNAMIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=i686-apple-darwin -march=x86 -relocation-model=pic -code-model=small | FileCheck %s -check-prefix=DARWIN-32-PIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=x86_64-apple-darwin -march=x86-64 -relocation-model=static -code-model=small | FileCheck %s -check-prefix=DARWIN-64-STATIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=x86_64-apple-darwin -march=x86-64 -relocation-model=dynamic-no-pic -code-model=small | FileCheck %s -check-prefix=DARWIN-64-DYNAMIC -; RUN: llc < %s -asm-verbose=0 -regalloc=basic -mtriple=x86_64-apple-darwin -march=x86-64 -relocation-model=pic -code-model=small | FileCheck %s -check-prefix=DARWIN-64-PIC - @src = external global [131072 x i32] @dst = external global [131072 x i32] @xsrc = external global [32 x i32] diff --git a/test/CodeGen/X86/add-of-carry.ll b/test/CodeGen/X86/add-of-carry.ll index f924ec8..a4abccb 100644 --- a/test/CodeGen/X86/add-of-carry.ll +++ b/test/CodeGen/X86/add-of-carry.ll @@ -4,9 +4,9 @@ define i32 @test1(i32 %sum, i32 %x) nounwind readnone ssp { entry: ; CHECK: test1: -; CHECK: sbbl %ecx, %ecx +; CHECK: cmpl %ecx, %eax ; CHECK-NOT: addl -; CHECK: subl %ecx, %eax +; CHECK: adcl $0, %eax %add4 = add i32 %x, %sum %cmp = icmp ult i32 %add4, %x %inc = zext i1 %cmp to i32 @@ -18,8 +18,7 @@ entry: ; CHECK: test2: ; CHECK: movl ; CHECK-NEXT: addl -; CHECK-NEXT: sbbl -; CHECK-NEXT: subl +; CHECK-NEXT: adcl $0 ; CHECK-NEXT: ret define i32 @test2(i32 %sum, i32 %x) nounwind readnone ssp { entry: diff --git a/test/CodeGen/X86/add.ll b/test/CodeGen/X86/add.ll index b95e5b5..7bf527a 100644 --- a/test/CodeGen/X86/add.ll +++ b/test/CodeGen/X86/add.ll @@ -1,6 +1,8 @@ ; RUN: llc < %s -march=x86 | FileCheck %s -check-prefix=X32 -; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s -check-prefix=X64 -; RUN: llc < %s -mtriple=x86_64-win32 | FileCheck %s -check-prefix=X64 +; RUN: llc < %s -mtriple=x86_64-linux -join-physregs | FileCheck %s -check-prefix=X64 +; RUN: llc < %s -mtriple=x86_64-win32 -join-physregs | FileCheck %s -check-prefix=X64 + +; Some of these tests depend on -join-physregs to commute instructions. ; The immediate can be encoded in a smaller way if the ; instruction is a sub instead of an add. diff --git a/test/CodeGen/X86/andimm8.ll b/test/CodeGen/X86/andimm8.ll index 640237d..a3dc85f 100644 --- a/test/CodeGen/X86/andimm8.ll +++ b/test/CodeGen/X86/andimm8.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-linux-gnu -show-mc-encoding | FileCheck %s +; RUN: llc < %s -march=x86-64 -mtriple=x86_64-pc-linux-gnu -show-mc-encoding -join-physregs | FileCheck %s ; PR8365 ; CHECK: andl $-64, %edi # encoding: [0x83,0xe7,0xc0] diff --git a/test/CodeGen/X86/basic-promote-integers.ll b/test/CodeGen/X86/basic-promote-integers.ll new file mode 100644 index 0000000..c80f2b0 --- /dev/null +++ b/test/CodeGen/X86/basic-promote-integers.ll @@ -0,0 +1,98 @@ +; Test that vectors are scalarized/lowered correctly +; (with both legalization methods). +; RUN: llc -march=x86 -promote-elements < %s +; RUN: llc -march=x86 < %s + +; A simple test to check copyToParts and copyFromParts. + +define <4 x i64> @test_param_0(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + ret <4 x i64> %A +} + +define <2 x i32> @test_param_1(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + ret <2 x i32> %B +} + +define <4 x i8> @test_param_2(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + ret <4 x i8> %C +} + +; Simple tests to check arithmetic and vector operations on types which need to +; be legalized (no loads/stores to/from memory here). + +define <4 x i64> @test_arith_0(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + %K = add <4 x i64> %A, <i64 0, i64 1, i64 3, i64 9> + ret <4 x i64> %K +} + +define <2 x i32> @test_arith_1(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + %K = add <2 x i32> %B, <i32 0, i32 1> + ret <2 x i32> %K +} + +define <4 x i8> @test_arith_2(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + %K = add <4 x i8> %C, <i8 0, i8 1, i8 3, i8 9> + ret <4 x i8> %K +} + +define i8 @test_arith_3(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + %K = add <4 x i8> %C, <i8 0, i8 1, i8 3, i8 9> + %Y = extractelement <4 x i8> %K, i32 1 + ret i8 %Y +} + +define <4 x i8> @test_arith_4(<4 x i64> %A, <2 x i32> %B, <4 x i8> %C) { + %Y = insertelement <4 x i8> %C, i8 1, i32 0 + ret <4 x i8> %Y +} + +define <4 x i32> @test_arith_5(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %Y = insertelement <4 x i32> %C, i32 1, i32 0 + ret <4 x i32> %Y +} + +define <4 x i32> @test_arith_6(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %F = extractelement <2 x i32> %B, i32 1 + %Y = insertelement <4 x i32> %C, i32 %F, i32 0 + ret <4 x i32> %Y +} + +define <4 x i64> @test_arith_7(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %F = extractelement <2 x i32> %B, i32 1 + %W = zext i32 %F to i64 + %Y = insertelement <4 x i64> %A, i64 %W, i32 0 + ret <4 x i64> %Y +} + +define i64 @test_arith_8(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %F = extractelement <2 x i32> %B, i32 1 + %W = zext i32 %F to i64 + %T = add i64 %W , 11 + ret i64 %T +} + +define <4 x i64> @test_arith_9(<4 x i64> %A, <2 x i32> %B, <4 x i16> %C) { + %T = add <4 x i16> %C, %C + %F0 = extractelement <4 x i16> %T, i32 0 + %F1 = extractelement <4 x i16> %T, i32 1 + %W0 = zext i16 %F0 to i64 + %W1 = zext i16 %F1 to i64 + %Y0 = insertelement <4 x i64> %A, i64 %W0, i32 0 + %Y1 = insertelement <4 x i64> %Y0, i64 %W1, i32 2 + ret <4 x i64> %Y1 +} + +define <4 x i16> @test_arith_10(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %F = bitcast <2 x i32> %B to <4 x i16> + %T = add <4 x i16> %F , <i16 0, i16 1, i16 2, i16 3> + ret <4 x i16> %T +} + + +; Simple tests to check saving/loading from memory +define <4 x i16> @test_mem_0(<4 x i64> %A, <2 x i32> %B, <4 x i32> %C) { + %F = bitcast <2 x i32> %B to <4 x i16> + %T = add <4 x i16> %F , <i16 0, i16 1, i16 2, i16 3> + ret <4 x i16> %T +} + diff --git a/test/CodeGen/X86/bool-zext.ll b/test/CodeGen/X86/bool-zext.ll index d2c30c6..3558376 100644 --- a/test/CodeGen/X86/bool-zext.ll +++ b/test/CodeGen/X86/bool-zext.ll @@ -1,8 +1,12 @@ -; RUN: llc < %s -march=x86-64 | FileCheck %s +; RUN: llc < %s -mtriple=x86_64-apple-darwin10 | FileCheck %s -check-prefix=X64 +; RUN: llc < %s -mtriple=x86_64-pc-win32 | FileCheck %s -check-prefix=WIN64 -; CHECK: @bar1 -; CHECK: movzbl -; CHECK: callq +; X64: @bar1 +; X64: movzbl +; X64: jmp +; WIN64: @bar1 +; WIN64: movzbl +; WIN64: callq define void @bar1(i1 zeroext %v1) nounwind ssp { entry: %conv = zext i1 %v1 to i32 @@ -10,9 +14,12 @@ entry: ret void } -; CHECK: @bar2 -; CHECK-NOT: movzbl -; CHECK: callq +; X64: @bar2 +; X64-NOT: movzbl +; X64: jmp +; WIN64: @bar2 +; WIN64-NOT: movzbl +; WIN64: callq define void @bar2(i8 zeroext %v1) nounwind ssp { entry: %conv = zext i8 %v1 to i32 @@ -20,11 +27,16 @@ entry: ret void } -; CHECK: @bar3 -; CHECK: callq -; CHECK-NOT: movzbl -; CHECK-NOT: and -; CHECK: ret +; X64: @bar3 +; X64: callq +; X64-NOT: movzbl +; X64-NOT: and +; X64: ret +; WIN64: @bar3 +; WIN64: callq +; WIN64-NOT: movzbl +; WIN64-NOT: and +; WIN64: ret define zeroext i1 @bar3() nounwind ssp { entry: %call = call i1 @foo2() nounwind diff --git a/test/CodeGen/X86/byval-align.ll b/test/CodeGen/X86/byval-align.ll new file mode 100644 index 0000000..c62a181 --- /dev/null +++ b/test/CodeGen/X86/byval-align.ll @@ -0,0 +1,59 @@ +; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s +%struct.S = type { i32} + +@.str = private constant [10 x i8] c"ptr = %p\0A\00", align 1 ; <[10 x i8]*> [#uses=1] +@.str1 = private constant [8 x i8] c"Failed \00", align 1 ; <[8 x i8]*> [#uses=1] +@.str2 = private constant [2 x i8] c"0\00", align 1 ; <[2 x i8]*> [#uses=1] +@.str3 = private constant [7 x i8] c"test.c\00", align 1 ; <[7 x i8]*> [#uses=1] +@__PRETTY_FUNCTION__.2067 = internal constant [13 x i8] c"aligned_func\00" ; <[13 x i8]*> [#uses=1] + +define void @aligned_func(%struct.S* byval align 64 %obj) nounwind { +entry: + %ptr = alloca i8* ; <i8**> [#uses=3] + %p = alloca i64 ; <i64*> [#uses=3] + %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] + %obj1 = bitcast %struct.S* %obj to i8* ; <i8*> [#uses=1] + store i8* %obj1, i8** %ptr, align 8 + %0 = load i8** %ptr, align 8 ; <i8*> [#uses=1] + %1 = ptrtoint i8* %0 to i64 ; <i64> [#uses=1] + store i64 %1, i64* %p, align 8 + %2 = load i8** %ptr, align 8 ; <i8*> [#uses=1] + %3 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([10 x i8]* @.str, i64 0, i64 0), i8* %2) nounwind ; <i32> [#uses=0] + %4 = load i64* %p, align 8 ; <i64> [#uses=1] + %5 = and i64 %4, 140737488355264 ; <i64> [#uses=1] + %6 = load i64* %p, align 8 ; <i64> [#uses=1] + %7 = icmp ne i64 %5, %6 ; <i1> [#uses=1] + br i1 %7, label %bb, label %bb2 + +bb: ; preds = %entry + %8 = call i32 @puts(i8* getelementptr inbounds ([8 x i8]* @.str1, i64 0, i64 0)) nounwind ; <i32> [#uses=0] + call void @__assert_fail(i8* getelementptr inbounds ([2 x i8]* @.str2, i64 0, i64 0), i8* getelementptr inbounds ([7 x i8]* @.str3, i64 0, i64 0), i32 18, i8* getelementptr inbounds ([13 x i8]* @__PRETTY_FUNCTION__.2067, i64 0, i64 0)) noreturn nounwind + unreachable + +bb2: ; preds = %entry + br label %return + +return: ; preds = %bb2 + ret void +} + +declare i32 @printf(i8*, ...) nounwind + +declare i32 @puts(i8*) + +declare void @__assert_fail(i8*, i8*, i32, i8*) noreturn nounwind + +define void @main() nounwind { +entry: +; CHECK: main +; CHECK: andq $-64, %rsp + %s1 = alloca %struct.S ; <%struct.S*> [#uses=4] + %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] + %0 = getelementptr inbounds %struct.S* %s1, i32 0, i32 0 ; <i32*> [#uses=1] + store i32 1, i32* %0, align 4 + call void @aligned_func(%struct.S* byval align 64 %s1) nounwind + br label %return + +return: ; preds = %entry + ret void +} diff --git a/test/CodeGen/X86/byval7.ll b/test/CodeGen/X86/byval7.ll index 686ed9c..98a26e4 100644 --- a/test/CodeGen/X86/byval7.ll +++ b/test/CodeGen/X86/byval7.ll @@ -9,7 +9,6 @@ entry: ; CHECK: main: ; CHECK: movl $1, (%esp) ; CHECK: leal 16(%esp), %edi -; CHECK: movl $36, %ecx ; CHECK: leal 160(%esp), %esi ; CHECK: rep;movsl %s = alloca %struct.S ; <%struct.S*> [#uses=2] diff --git a/test/CodeGen/X86/clz.ll b/test/CodeGen/X86/clz.ll index 623ac75..d76fab4 100644 --- a/test/CodeGen/X86/clz.ll +++ b/test/CodeGen/X86/clz.ll @@ -31,3 +31,18 @@ entry: } declare i16 @llvm.ctlz.i16(i16) nounwind readnone + +; Don't generate the cmovne when the source is known non-zero (and bsr would +; not set ZF). +; rdar://9490949 + +define i32 @t4(i32 %n) nounwind { +entry: +; CHECK: t4: +; CHECK: bsrl +; CHECK-NOT: cmov +; CHECK: ret + %or = or i32 %n, 1 + %tmp1 = tail call i32 @llvm.ctlz.i32(i32 %or) + ret i32 %tmp1 +} diff --git a/test/CodeGen/X86/coalescer-commute2.ll b/test/CodeGen/X86/coalescer-commute2.ll index 7306920..6e5c1cf 100644 --- a/test/CodeGen/X86/coalescer-commute2.ll +++ b/test/CodeGen/X86/coalescer-commute2.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s +; RUN: llc < %s -mtriple=x86_64-linux -join-physregs | FileCheck %s ; CHECK-NOT: mov ; CHECK: paddw ; CHECK-NOT: mov diff --git a/test/CodeGen/X86/dbg-const-int.ll b/test/CodeGen/X86/dbg-const-int.ll new file mode 100644 index 0000000..bfc96f1 --- /dev/null +++ b/test/CodeGen/X86/dbg-const-int.ll @@ -0,0 +1,29 @@ +; RUN: llc < %s - | FileCheck %s +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.7" +; Radar 9511391 + +;CHECK: .byte 4 ## DW_AT_const_value +define i32 @foo() nounwind uwtable readnone optsize ssp { +entry: + tail call void @llvm.dbg.value(metadata !8, i64 0, metadata !6), !dbg !9 + ret i32 42, !dbg !10 +} + +declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone + +!llvm.dbg.cu = !{!0} +!llvm.dbg.sp = !{!1} +!llvm.dbg.lv.foo = !{!6} + +!0 = metadata !{i32 589841, i32 0, i32 12, metadata !"a.c", metadata !"/private/tmp", metadata !"clang version 3.0 (trunk 132191)", i1 true, i1 true, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!1 = metadata !{i32 589870, i32 0, metadata !2, metadata !"foo", metadata !"foo", metadata !"", metadata !2, i32 1, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 0, i1 true, i32 ()* @foo, null, null} ; [ DW_TAG_subprogram ] +!2 = metadata !{i32 589865, metadata !"a.c", metadata !"/private/tmp", metadata !0} ; [ DW_TAG_file_type ] +!3 = metadata !{i32 589845, metadata !2, metadata !"", metadata !2, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{metadata !5} +!5 = metadata !{i32 589860, metadata !0, metadata !"int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] +!6 = metadata !{i32 590080, metadata !7, metadata !"i", metadata !2, i32 2, metadata !5, i32 0} ; [ DW_TAG_auto_variable ] +!7 = metadata !{i32 589835, metadata !1, i32 1, i32 11, metadata !2, i32 0} ; [ DW_TAG_lexical_block ] +!8 = metadata !{i32 42} +!9 = metadata !{i32 2, i32 12, metadata !7, null} +!10 = metadata !{i32 3, i32 2, metadata !7, null} diff --git a/test/CodeGen/X86/dbg-const.ll b/test/CodeGen/X86/dbg-const.ll new file mode 100644 index 0000000..5a51eb8 --- /dev/null +++ b/test/CodeGen/X86/dbg-const.ll @@ -0,0 +1,34 @@ +; RUN: llc < %s - | FileCheck %s +target triple = "x86_64-apple-darwin10.0.0" + +;CHECK: ## DW_OP_constu +;CHECK-NEXT: .byte 42 +define i32 @foobar() nounwind readonly noinline ssp { +entry: + %call = tail call i32 @bar(), !dbg !11 + tail call void @llvm.dbg.value(metadata !8, i64 0, metadata !6), !dbg !9 + %call2 = tail call i32 @bar(), !dbg !11 + tail call void @llvm.dbg.value(metadata !{i32 %call}, i64 0, metadata !6), !dbg !11 + %add = add nsw i32 %call2, %call, !dbg !12 + ret i32 %add, !dbg !10 +} + +declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone +declare i32 @bar() nounwind readnone + +!llvm.dbg.sp = !{!0} +!llvm.dbg.lv.foobar = !{!6} + +!0 = metadata !{i32 524334, i32 0, metadata !1, metadata !"foobar", metadata !"foobar", metadata !"foobar", metadata !1, i32 12, metadata !3, i1 false, i1 true, i32 0, i32 0, null, i1 false, i1 true, i32 ()* @foobar} +!1 = metadata !{i32 524329, metadata !"mu.c", metadata !"/private/tmp", metadata !2} +!2 = metadata !{i32 524305, i32 0, i32 12, metadata !"mu.c", metadata !"/private/tmp", metadata !"clang version 2.9 (trunk 114183)", i1 true, i1 true, metadata !"", i32 0} +!3 = metadata !{i32 524309, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i64 0, i32 0, null, metadata !4, i32 0, null} +!4 = metadata !{metadata !5} +!5 = metadata !{i32 524324, metadata !1, metadata !"int", metadata !1, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} +!6 = metadata !{i32 524544, metadata !7, metadata !"j", metadata !1, i32 15, metadata !5} +!7 = metadata !{i32 524299, metadata !0, i32 12, i32 52, metadata !1, i32 0} +!8 = metadata !{i32 42} +!9 = metadata !{i32 15, i32 12, metadata !7, null} +!10 = metadata !{i32 23, i32 3, metadata !7, null} +!11 = metadata !{i32 17, i32 3, metadata !7, null} +!12 = metadata !{i32 18, i32 3, metadata !7, null} diff --git a/test/CodeGen/X86/dbg-file-name.ll b/test/CodeGen/X86/dbg-file-name.ll index e7d5f92..3a849aa 100644 --- a/test/CodeGen/X86/dbg-file-name.ll +++ b/test/CodeGen/X86/dbg-file-name.ll @@ -1,7 +1,7 @@ ; RUN: llc -mtriple x86_64-apple-darwin10.0.0 < %s | FileCheck %s ; Radar 8884898 -; CHECK: file 1 "/Users/manav/one/two/simple.c" +; CHECK: file 1 "/Users/manav/one/two{{/|\\\\}}simple.c" declare i32 @printf(i8*, ...) nounwind diff --git a/test/CodeGen/X86/dbg-merge-loc-entry.ll b/test/CodeGen/X86/dbg-merge-loc-entry.ll index 76b93dd..afe1729 100644 --- a/test/CodeGen/X86/dbg-merge-loc-entry.ll +++ b/test/CodeGen/X86/dbg-merge-loc-entry.ll @@ -6,8 +6,11 @@ target triple = "x86_64-apple-darwin8" ;CHECK: Ldebug_loc0: ;CHECK-NEXT: .quad Lfunc_begin0 ;CHECK-NEXT: .quad L -;CHECK-NEXT: .short 1 ## Loc expr size +;CHECK-NEXT: Lset{{.*}} = Ltmp{{.*}}-Ltmp{{.*}} ## Loc expr size +;CHECK-NEXT: .short Lset +;CHECK-NEXT: Ltmp ;CHECK-NEXT: .byte 85 ## DW_OP_reg5 +;CHECK-NEXT: Ltmp7 ;CHECK-NEXT: .quad 0 ;CHECK-NEXT: .quad 0 diff --git a/test/CodeGen/X86/dbg-prolog-end.ll b/test/CodeGen/X86/dbg-prolog-end.ll new file mode 100644 index 0000000..81303bb --- /dev/null +++ b/test/CodeGen/X86/dbg-prolog-end.ll @@ -0,0 +1,55 @@ +; RUN: llc -O0 < %s | FileCheck %s +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.7" + +;CHECK: .loc 1 2 11 prologue_end +define i32 @foo(i32 %i) nounwind ssp { +entry: + %i.addr = alloca i32, align 4 + %j = alloca i32, align 4 + store i32 %i, i32* %i.addr, align 4 + call void @llvm.dbg.declare(metadata !{i32* %i.addr}, metadata !7), !dbg !8 + call void @llvm.dbg.declare(metadata !{i32* %j}, metadata !9), !dbg !11 + store i32 2, i32* %j, align 4, !dbg !12 + %tmp = load i32* %j, align 4, !dbg !13 + %inc = add nsw i32 %tmp, 1, !dbg !13 + store i32 %inc, i32* %j, align 4, !dbg !13 + %tmp1 = load i32* %j, align 4, !dbg !14 + %tmp2 = load i32* %i.addr, align 4, !dbg !14 + %add = add nsw i32 %tmp1, %tmp2, !dbg !14 + store i32 %add, i32* %j, align 4, !dbg !14 + %tmp3 = load i32* %j, align 4, !dbg !15 + ret i32 %tmp3, !dbg !15 +} + +declare void @llvm.dbg.declare(metadata, metadata) nounwind readnone + +define i32 @main() nounwind ssp { +entry: + %retval = alloca i32, align 4 + store i32 0, i32* %retval + %call = call i32 @foo(i32 21), !dbg !16 + ret i32 %call, !dbg !16 +} + +!llvm.dbg.cu = !{!0} +!llvm.dbg.sp = !{!1, !6} + +!0 = metadata !{i32 589841, i32 0, i32 12, metadata !"/tmp/a.c", metadata !"/private/tmp", metadata !"clang version 3.0 (trunk 131100)", i1 true, i1 false, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!1 = metadata !{i32 589870, i32 0, metadata !2, metadata !"foo", metadata !"foo", metadata !"", metadata !2, i32 1, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 256, i1 false, i32 (i32)* @foo, null, null} ; [ DW_TAG_subprogram ] +!2 = metadata !{i32 589865, metadata !"/tmp/a.c", metadata !"/private/tmp", metadata !0} ; [ DW_TAG_file_type ] +!3 = metadata !{i32 589845, metadata !2, metadata !"", metadata !2, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{metadata !5} +!5 = metadata !{i32 589860, metadata !0, metadata !"int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] +!6 = metadata !{i32 589870, i32 0, metadata !2, metadata !"main", metadata !"main", metadata !"", metadata !2, i32 7, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 0, i1 false, i32 ()* @main, null, null} ; [ DW_TAG_subprogram ] +!7 = metadata !{i32 590081, metadata !1, metadata !"i", metadata !2, i32 16777217, metadata !5, i32 0} ; [ DW_TAG_arg_variable ] +!8 = metadata !{i32 1, i32 13, metadata !1, null} +!9 = metadata !{i32 590080, metadata !10, metadata !"j", metadata !2, i32 2, metadata !5, i32 0} ; [ DW_TAG_auto_variable ] +!10 = metadata !{i32 589835, metadata !1, i32 1, i32 16, metadata !2, i32 0} ; [ DW_TAG_lexical_block ] +!11 = metadata !{i32 2, i32 6, metadata !10, null} +!12 = metadata !{i32 2, i32 11, metadata !10, null} +!13 = metadata !{i32 3, i32 2, metadata !10, null} +!14 = metadata !{i32 4, i32 2, metadata !10, null} +!15 = metadata !{i32 5, i32 2, metadata !10, null} +!16 = metadata !{i32 8, i32 2, metadata !17, null} +!17 = metadata !{i32 589835, metadata !6, i32 7, i32 12, metadata !2, i32 1} ; [ DW_TAG_lexical_block ] diff --git a/test/CodeGen/X86/dbg-value-dag-combine.ll b/test/CodeGen/X86/dbg-value-dag-combine.ll new file mode 100644 index 0000000..b115bf4 --- /dev/null +++ b/test/CodeGen/X86/dbg-value-dag-combine.ll @@ -0,0 +1,48 @@ +; RUN: llc < %s | FileCheck %s +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-darwin10.0.0" +; PR 9817 + + +declare <4 x i32> @__amdil_get_global_id_int() +declare void @llvm.dbg.value(metadata , i64 , metadata ) +define void @__OpenCL_test_kernel(i32 addrspace(1)* %ip) nounwind { +entry: + call void @llvm.dbg.value(metadata !{i32 addrspace(1)* %ip}, i64 0, metadata +!7), !dbg !8 + %0 = call <4 x i32> @__amdil_get_global_id_int() nounwind + %1 = extractelement <4 x i32> %0, i32 0 + call void @llvm.dbg.value(metadata !{i32 %1}, i64 0, metadata !9), !dbg !11 + call void @llvm.dbg.value(metadata !12, i64 0, metadata !13), !dbg !14 + %tmp2 = load i32 addrspace(1)* %ip, align 4, !dbg !15 + %tmp3 = add i32 0, %tmp2, !dbg !15 +; CHECK: ##DEBUG_VALUE: idx <- EAX+0 + call void @llvm.dbg.value(metadata !{i32 %tmp3}, i64 0, metadata !13), !dbg +!15 + %arrayidx = getelementptr i32 addrspace(1)* %ip, i32 %1, !dbg !16 + store i32 %tmp3, i32 addrspace(1)* %arrayidx, align 4, !dbg !16 + ret void, !dbg !17 +} +!llvm.dbg.sp = !{!0} + +!0 = metadata !{i32 589870, i32 0, metadata !1, metadata +!"__OpenCL_test_kernel", metadata !"__OpenCL_test_kernel", metadata +!"__OpenCL_test_kernel", metadata !1, i32 2, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 0, i1 false, null} ; [ DW_TAG_subprogram ] +!1 = metadata !{i32 589865, metadata !"OCL6368.tmp.cl", metadata !"E:\5CUsers\5Cmvillmow.AMD\5CAppData\5CLocal\5CTemp", metadata !2} ; [ DW_TAG_file_type ] +!2 = metadata !{i32 589841, i32 0, i32 1, metadata !"OCL6368.tmp.cl", metadata !"E:\5CUsers\5Cmvillmow.AMD\5CAppData\5CLocal\5CTemp", metadata !"clc", i1 true, i1 false, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!3 = metadata !{i32 589845, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{null, metadata !5} +!5 = metadata !{i32 589839, metadata !2, metadata !"", null, i32 0, i64 32, i64 32, i64 0, i32 0, metadata !6} ; [ DW_TAG_pointer_type ] +!6 = metadata !{i32 589860, metadata !2, metadata !"unsigned int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 7} ; [ DW_TAG_base_type ] +!7 = metadata !{i32 590081, metadata !0, metadata !"ip", metadata !1, i32 1, metadata !5, i32 0} ; [ DW_TAG_arg_variable ] +!8 = metadata !{i32 1, i32 42, metadata !0, null} +!9 = metadata !{i32 590080, metadata !10, metadata !"gid", metadata !1, i32 3, metadata !6, i32 0} ; [ DW_TAG_auto_variable ] +!10 = metadata !{i32 589835, metadata !0, i32 2, i32 1, metadata !1, i32 0} ; [ DW_TAG_lexical_block ] +!11 = metadata !{i32 3, i32 41, metadata !10, null} +!12 = metadata !{i32 0} +!13 = metadata !{i32 590080, metadata !10, metadata !"idx", metadata !1, i32 4, metadata !6, i32 0} ; [ DW_TAG_auto_variable ] +!14 = metadata !{i32 4, i32 20, metadata !10, null} +!15 = metadata !{i32 5, i32 15, metadata !10, null} +!16 = metadata !{i32 6, i32 18, metadata !10, null} +!17 = metadata !{i32 7, i32 1, metadata !0, null} + diff --git a/test/CodeGen/X86/dbg-value-isel.ll b/test/CodeGen/X86/dbg-value-isel.ll new file mode 100644 index 0000000..d1a9e57 --- /dev/null +++ b/test/CodeGen/X86/dbg-value-isel.ll @@ -0,0 +1,102 @@ +; RUN: llc < %s | FileCheck %s +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-darwin10.0.0" +; PR 9879 + +; CHECK: ##DEBUG_VALUE: tid <- +%0 = type { i8*, i8*, i8*, i8*, i32 } + +@sgv = internal addrspace(2) constant [1 x i8] zeroinitializer +@fgv = internal addrspace(2) constant [1 x i8] zeroinitializer +@lvgv = internal constant [0 x i8*] zeroinitializer +@llvm.global.annotations = appending global [1 x %0] [%0 { i8* bitcast (void (i32 addrspace(1)*)* @__OpenCL_nbt02_kernel to i8*), i8* bitcast ([1 x i8] addrspace(2)* @sgv to i8*), i8* bitcast ([1 x i8] addrspace(2)* @fgv to i8*), i8* bitcast ([0 x i8*]* @lvgv to i8*), i32 0 }], section "llvm.metadata" + +define void @__OpenCL_nbt02_kernel(i32 addrspace(1)* %ip) nounwind { +entry: + call void @llvm.dbg.value(metadata !{i32 addrspace(1)* %ip}, i64 0, metadata !8), !dbg !9 + %0 = call <4 x i32> @__amdil_get_local_id_int() nounwind + %1 = extractelement <4 x i32> %0, i32 0 + br label %2 + +; <label>:2 ; preds = %entry + %3 = phi i32 [ %1, %entry ] + br label %4 + +; <label>:4 ; preds = %2 + %5 = phi i32 [ %3, %2 ] + br label %get_local_id.exit + +get_local_id.exit: ; preds = %4 + %6 = phi i32 [ %5, %4 ] + call void @llvm.dbg.value(metadata !{i32 %6}, i64 0, metadata !10), !dbg !12 + %7 = call <4 x i32> @__amdil_get_global_id_int() nounwind + %8 = extractelement <4 x i32> %7, i32 0 + br label %9 + +; <label>:9 ; preds = %get_local_id.exit + %10 = phi i32 [ %8, %get_local_id.exit ] + br label %11 + +; <label>:11 ; preds = %9 + %12 = phi i32 [ %10, %9 ] + br label %get_global_id.exit + +get_global_id.exit: ; preds = %11 + %13 = phi i32 [ %12, %11 ] + call void @llvm.dbg.value(metadata !{i32 %13}, i64 0, metadata !13), !dbg !14 + %14 = call <4 x i32> @__amdil_get_local_size_int() nounwind + %15 = extractelement <4 x i32> %14, i32 0 + br label %16 + +; <label>:16 ; preds = %get_global_id.exit + %17 = phi i32 [ %15, %get_global_id.exit ] + br label %18 + +; <label>:18 ; preds = %16 + %19 = phi i32 [ %17, %16 ] + br label %get_local_size.exit + +get_local_size.exit: ; preds = %18 + %20 = phi i32 [ %19, %18 ] + call void @llvm.dbg.value(metadata !{i32 %20}, i64 0, metadata !15), !dbg !16 + %tmp5 = add i32 %6, %13, !dbg !17 + %tmp7 = add i32 %tmp5, %20, !dbg !17 + store i32 %tmp7, i32 addrspace(1)* %ip, align 4, !dbg !17 + br label %return, !dbg !17 + +return: ; preds = %get_local_size.exit + ret void, !dbg !18 +} + +declare void @llvm.dbg.declare(metadata, metadata) nounwind readnone + +declare <4 x i32> @__amdil_get_local_size_int() nounwind + +declare <4 x i32> @__amdil_get_local_id_int() nounwind + +declare <4 x i32> @__amdil_get_global_id_int() nounwind + +declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone + +!llvm.dbg.sp = !{!0} + +!0 = metadata !{i32 589870, i32 0, metadata !1, metadata !"__OpenCL_nbt02_kernel", metadata !"__OpenCL_nbt02_kernel", metadata !"__OpenCL_nbt02_kernel", metadata !1, i32 2, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 0, i1 false, null} ; [ DW_TAG_subprogram ] +!1 = metadata !{i32 589865, metadata !"OCLlLwTXZ.cl", metadata !"/tmp", metadata !2} ; [ DW_TAG_file_type ] +!2 = metadata !{i32 589841, i32 0, i32 1, metadata !"OCLlLwTXZ.cl", metadata !"/tmp", metadata !"clc", i1 true, i1 false, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!3 = metadata !{i32 589845, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{null, metadata !5} +!5 = metadata !{i32 589839, metadata !2, metadata !"", null, i32 0, i64 32, i64 32, i64 0, i32 0, metadata !6} ; [ DW_TAG_pointer_type ] +!6 = metadata !{i32 589846, metadata !2, metadata !"uint", metadata !1, i32 0, i64 0, i64 0, i64 0, i32 0, metadata !7} ; [ DW_TAG_typedef ] +!7 = metadata !{i32 589860, metadata !2, metadata !"unsigned int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 7} ; [ DW_TAG_base_type ] +!8 = metadata !{i32 590081, metadata !0, metadata !"ip", metadata !1, i32 1, metadata !5, i32 0} ; [ DW_TAG_arg_variable ] +!9 = metadata !{i32 1, i32 32, metadata !0, null} +!10 = metadata !{i32 590080, metadata !11, metadata !"tid", metadata !1, i32 3, metadata !6, i32 0} ; [ DW_TAG_auto_variable ] +!11 = metadata !{i32 589835, metadata !0, i32 2, i32 1, metadata !1, i32 1} ; [ DW_TAG_lexical_block ] +!12 = metadata !{i32 5, i32 24, metadata !11, null} +!13 = metadata !{i32 590080, metadata !11, metadata !"gid", metadata !1, i32 3, metadata !6, i32 0} ; [ DW_TAG_auto_variable ] +!14 = metadata !{i32 6, i32 25, metadata !11, null} +!15 = metadata !{i32 590080, metadata !11, metadata !"lsz", metadata !1, i32 3, metadata !6, i32 0} ; [ DW_TAG_auto_variable ] +!16 = metadata !{i32 7, i32 26, metadata !11, null} +!17 = metadata !{i32 9, i32 24, metadata !11, null} +!18 = metadata !{i32 10, i32 1, metadata !0, null} + diff --git a/test/CodeGen/X86/dbg-value-range.ll b/test/CodeGen/X86/dbg-value-range.ll index da49f2d..28d873b 100644 --- a/test/CodeGen/X86/dbg-value-range.ll +++ b/test/CodeGen/X86/dbg-value-range.ll @@ -1,5 +1,5 @@ ; RUN: llc -mtriple=x86_64-apple-darwin10 < %s | FileCheck %s -; RUN: llc -mtriple=x86_64-apple-darwin10 -regalloc=basic < %s | FileCheck %s +; RUN: llc -mtriple=x86_64-apple-darwin10 -regalloc=basic -join-physregs < %s | FileCheck %s %struct.a = type { i32 } @@ -53,7 +53,10 @@ declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone ;CHECK:Ldebug_loc0: ;CHECK-NEXT: .quad ;CHECK-NEXT: .quad [[CLOBBER]] -;CHECK-NEXT: .short 1 +;CHECK-NEXT: Lset{{.*}} = Ltmp{{.*}}-Ltmp{{.*}} +;CHECK-NEXT: .short Lset +;CHECK-NEXT: Ltmp ;CHECK-NEXT: .byte 85 +;CHECK-NEXT: Ltmp ;CHECK-NEXT: .quad 0 ;CHECK-NEXT: .quad 0 diff --git a/test/CodeGen/X86/div8.ll b/test/CodeGen/X86/div8.ll new file mode 100644 index 0000000..0825f79 --- /dev/null +++ b/test/CodeGen/X86/div8.ll @@ -0,0 +1,22 @@ +; RUN: llc < %s | FileCheck %s +; ModuleID = '8div.c' +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64" +target triple = "x86_64-apple-macosx10.6.6" + +define signext i8 @test_div(i8 %dividend, i8 %divisor) nounwind ssp { +entry: + %dividend.addr = alloca i8, align 2 + %divisor.addr = alloca i8, align 1 + %quotient = alloca i8, align 1 + store i8 %dividend, i8* %dividend.addr, align 2 + store i8 %divisor, i8* %divisor.addr, align 1 + %tmp = load i8* %dividend.addr, align 2 + %tmp1 = load i8* %divisor.addr, align 1 +; Insist on i8->i32 zero extension, even though divb demands only i16: +; CHECK: movzbl {{.*}}%eax +; CHECK: divb + %div = udiv i8 %tmp, %tmp1 + store i8 %div, i8* %quotient, align 1 + %tmp4 = load i8* %quotient, align 1 + ret i8 %tmp4 +} diff --git a/test/CodeGen/X86/eh_frame.ll b/test/CodeGen/X86/eh_frame.ll new file mode 100644 index 0000000..3b792b2 --- /dev/null +++ b/test/CodeGen/X86/eh_frame.ll @@ -0,0 +1,14 @@ +; RUN: llc < %s -mtriple x86_64-unknown-linux-gnu | FileCheck -check-prefix=STATIC %s +; RUN: llc < %s -mtriple x86_64-unknown-linux-gnu -relocation-model=pic | FileCheck -check-prefix=PIC %s + +@__FRAME_END__ = constant [1 x i32] zeroinitializer, section ".eh_frame" + +@foo = external global i32 +@bar1 = constant i8* bitcast (i32* @foo to i8*), section "my_bar1", align 8 + + +; STATIC: .section .eh_frame,"a",@progbits +; STATIC: .section my_bar1,"a",@progbits + +; PIC: .section .eh_frame,"a",@progbits +; PIC: .section my_bar1,"aw",@progbits diff --git a/test/CodeGen/X86/empty-functions.ll b/test/CodeGen/X86/empty-functions.ll index b303cd1..874c53a 100644 --- a/test/CodeGen/X86/empty-functions.ll +++ b/test/CodeGen/X86/empty-functions.ll @@ -6,10 +6,24 @@ entry: unreachable } ; CHECK-NO-FP: _func: -; CHECK-NO-FP-NOT: movq %rsp, %rbp +; CHECK-NO-FP-NEXT: : +; CHECK-NO-FP-NEXT: .cfi_startproc ; CHECK-NO-FP: nop +; CHECK-NO-FP-NEXT: : +; CHECK-NO-FP-NEXT: .cfi_endproc ; CHECK-FP: _func: -; CHECK-FP: movq %rsp, %rbp -; CHECK-FP-NEXT: Ltmp1: -; CHECK-FP: nop +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: .cfi_startproc +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: pushq %rbp +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: .cfi_def_cfa_offset 16 +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: .cfi_offset %rbp, -16 +; CHECK-FP-NEXT: movq %rsp, %rbp +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: .cfi_def_cfa_register %rbp +; CHECK-FP-NEXT: nop +; CHECK-FP-NEXT: : +; CHECK-FP-NEXT: .cfi_endproc diff --git a/test/CodeGen/X86/fast-isel-agg-constant.ll b/test/CodeGen/X86/fast-isel-agg-constant.ll new file mode 100644 index 0000000..ce0dff7 --- /dev/null +++ b/test/CodeGen/X86/fast-isel-agg-constant.ll @@ -0,0 +1,11 @@ +; RUN: llc < %s -march=x86-64 -O0 | FileCheck %s +; Make sure fast-isel doesn't screw up aggregate constants. +; (Failing out is okay, as long as we don't miscompile.) + +%bar = type { i32 } + +define i32 @foo() { + %tmp = extractvalue %bar { i32 3 }, 0 + ret i32 %tmp +; CHECK: movl $3, %eax +} diff --git a/test/CodeGen/X86/fast-isel-call.ll b/test/CodeGen/X86/fast-isel-call.ll index 5fcdbbb..2fbe4e2 100644 --- a/test/CodeGen/X86/fast-isel-call.ll +++ b/test/CodeGen/X86/fast-isel-call.ll @@ -1,6 +1,8 @@ -; RUN: llc < %s -fast-isel -march=x86 | grep and +; RUN: llc < %s -O0 -fast-isel-abort -march=x86 | FileCheck %s -define i32 @t() nounwind { +%struct.s = type {i32, i32, i32} + +define i32 @test1() nounwind { tak: %tmp = call i1 @foo() br i1 %tmp, label %BB1, label %BB2 @@ -8,6 +10,46 @@ BB1: ret i32 1 BB2: ret i32 0 +; CHECK: test1: +; CHECK: calll +; CHECK-NEXT: testb $1 } - declare i1 @foo() zeroext nounwind + +declare void @foo2(%struct.s* byval) + +define void @test2(%struct.s* %d) nounwind { + call void @foo2(%struct.s* %d byval) + ret void +; CHECK: test2: +; CHECK: movl (%eax) +; CHECK: movl {{.*}}, (%esp) +; CHECK: movl 4(%eax) +; CHECK: movl {{.*}}, 4(%esp) +; CHECK: movl 8(%eax) +; CHECK: movl {{.*}}, 8(%esp) +} + +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind + +define void @test3(i8* %a) { + call void @llvm.memset.p0i8.i32(i8* %a, i8 0, i32 100, i32 1, i1 false) + ret void +; CHECK: test3: +; CHECK: movl {{.*}}, (%esp) +; CHECK: movl $0, 4(%esp) +; CHECK: movl $100, 8(%esp) +; CHECK: calll {{.*}}memset +} + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind + +define void @test4(i8* %a, i8* %b) { + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %a, i8* %b, i32 100, i32 1, i1 false) + ret void +; CHECK: test4: +; CHECK: movl {{.*}}, (%esp) +; CHECK: movl {{.*}}, 4(%esp) +; CHECK: movl $100, 8(%esp) +; CHECK: calll {{.*}}memcpy +} diff --git a/test/CodeGen/X86/fast-isel-extract.ll b/test/CodeGen/X86/fast-isel-extract.ll new file mode 100644 index 0000000..f63396e --- /dev/null +++ b/test/CodeGen/X86/fast-isel-extract.ll @@ -0,0 +1,48 @@ +; RUN: llc < %s -mtriple x86_64-apple-darwin11 -O0 -fast-isel-abort | FileCheck %s + +%struct.x = type { i64, i64 } +%addovf = type { i32, i1 } +declare %struct.x @f() + +define void @test1(i64*) nounwind ssp { + %2 = tail call %struct.x @f() nounwind + %3 = extractvalue %struct.x %2, 0 + %4 = add i64 %3, 10 + store i64 %4, i64* %0 + ret void +; CHECK: test1: +; CHECK: callq _f +; CHECK-NEXT: addq $10, %rax +} + +define void @test2(i64*) nounwind ssp { + %2 = tail call %struct.x @f() nounwind + %3 = extractvalue %struct.x %2, 1 + %4 = add i64 %3, 10 + store i64 %4, i64* %0 + ret void +; CHECK: test2: +; CHECK: callq _f +; CHECK-NEXT: addq $10, %rdx +} + +declare %addovf @llvm.sadd.with.overflow.i32(i32, i32) nounwind readnone + +define void @test3(i32 %x, i32 %y, i32* %z) { + %r = call %addovf @llvm.sadd.with.overflow.i32(i32 %x, i32 %y) + %sum = extractvalue %addovf %r, 0 + %sum3 = mul i32 %sum, 3 + %bit = extractvalue %addovf %r, 1 + br i1 %bit, label %then, label %end + +then: + store i32 %sum3, i32* %z + br label %end + +end: + ret void +; CHECK: test3 +; CHECK: addl +; CHECK: seto %al +; CHECK: testb $1, %al +} diff --git a/test/CodeGen/X86/fast-isel-fneg.ll b/test/CodeGen/X86/fast-isel-fneg.ll index 5ffd48b..f42a4a2 100644 --- a/test/CodeGen/X86/fast-isel-fneg.ll +++ b/test/CodeGen/X86/fast-isel-fneg.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -fast-isel -fast-isel-abort -march=x86-64 | FileCheck %s +; RUN: llc < %s -fast-isel -fast-isel-abort -mtriple=x86_64-apple-darwin10 | FileCheck %s ; RUN: llc < %s -fast-isel -march=x86 -mattr=+sse2 | grep xor | count 2 ; CHECK: doo: diff --git a/test/CodeGen/X86/fast-isel-i1.ll b/test/CodeGen/X86/fast-isel-i1.ll index 5d572c1..bea18a1 100644 --- a/test/CodeGen/X86/fast-isel-i1.ll +++ b/test/CodeGen/X86/fast-isel-i1.ll @@ -1,14 +1,15 @@ -; RUN: llc < %s -march=x86 -fast-isel | FileCheck %s +; RUN: llc < %s -mtriple=i686-apple-darwin10 -fast-isel -fast-isel-abort | FileCheck %s +; RUN: llc < %s -mtriple=x86_64-apple-darwin10 -fast-isel -fast-isel-abort | FileCheck %s -declare i64 @test1a(i64) +declare i32 @test1a(i32) -define i32 @test1(i64 %x) nounwind { +define i32 @test1(i32 %x) nounwind { ; CHECK: test1: ; CHECK: andb $1, % - %y = add i64 %x, -3 - %t = call i64 @test1a(i64 %y) - %s = mul i64 %t, 77 - %z = trunc i64 %s to i1 + %y = add i32 %x, -3 + %t = call i32 @test1a(i32 %y) + %s = mul i32 %t, 77 + %z = trunc i32 %s to i1 br label %next next: ; preds = %0 diff --git a/test/CodeGen/X86/fast-isel-ret-ext.ll b/test/CodeGen/X86/fast-isel-ret-ext.ll new file mode 100644 index 0000000..fd768cb --- /dev/null +++ b/test/CodeGen/X86/fast-isel-ret-ext.ll @@ -0,0 +1,38 @@ +; RUN: llc < %s -O0 -fast-isel-abort -mtriple i686-apple-darwin10 | FileCheck %s +; RUN: llc < %s -O0 -fast-isel-abort -mtriple x86_64-apple-darwin10 | FileCheck %s + +define zeroext i8 @test1(i32 %y) nounwind { + %conv = trunc i32 %y to i8 + ret i8 %conv + ; CHECK: test1: + ; CHECK: movzbl {{.*}}, %eax +} + +define signext i8 @test2(i32 %y) nounwind { + %conv = trunc i32 %y to i8 + ret i8 %conv + ; CHECK: test2: + ; CHECK: movsbl {{.*}}, %eax +} + +define zeroext i16 @test3(i32 %y) nounwind { + %conv = trunc i32 %y to i16 + ret i16 %conv + ; CHECK: test3: + ; CHECK: movzwl {{.*}}, %eax +} + +define signext i16 @test4(i32 %y) nounwind { + %conv = trunc i32 %y to i16 + ret i16 %conv + ; CHECK: test4: + ; CHECK: movswl {{.*}}, %eax +} + +define zeroext i1 @test5(i32 %y) nounwind { + %conv = trunc i32 %y to i1 + ret i1 %conv + ; CHECK: test5: + ; CHECK: andb $1 + ; CHECK: movzbl {{.*}}, %eax +} diff --git a/test/CodeGen/X86/fast-isel.ll b/test/CodeGen/X86/fast-isel.ll index 5a1d213..8391860 100644 --- a/test/CodeGen/X86/fast-isel.ll +++ b/test/CodeGen/X86/fast-isel.ll @@ -1,5 +1,5 @@ ; RUN: llc < %s -fast-isel -fast-isel-abort -march=x86 -mattr=sse2 -; RUN: llc < %s -fast-isel -fast-isel-abort -march=x86-64 +; RUN: llc < %s -fast-isel -fast-isel-abort -mtriple=x86_64-apple-darwin10 ; This tests very minimal fast-isel functionality. @@ -27,7 +27,7 @@ exit: ret i32* %t8 } -define double @bar(double* %p, double* %q) nounwind { +define void @bar(double* %p, double* %q) nounwind { entry: %r = load double* %p %s = load double* %q @@ -41,7 +41,8 @@ fast: br label %exit exit: - ret double %t3 + store double %t3, double* %q + ret void } define i32 @cast() nounwind { @@ -68,24 +69,28 @@ define i8* @inttoptr_i32(i32 %p) nounwind { ret i8* %t } -define i8 @trunc_i32_i8(i32 %x) signext nounwind { +define void @trunc_i32_i8(i32 %x, i8* %p) nounwind { %tmp1 = trunc i32 %x to i8 - ret i8 %tmp1 + store i8 %tmp1, i8* %p + ret void } -define i8 @trunc_i16_i8(i16 signext %x) signext nounwind { +define void @trunc_i16_i8(i16 signext %x, i8* %p) nounwind { %tmp1 = trunc i16 %x to i8 - ret i8 %tmp1 + store i8 %tmp1, i8* %p + ret void } -define i8 @shl_i8(i8 %a, i8 %c) nounwind { - %tmp = shl i8 %a, %c - ret i8 %tmp +define void @shl_i8(i8 %a, i8 %c, i8* %p) nounwind { + %tmp = shl i8 %a, %c + store i8 %tmp, i8* %p + ret void } -define i8 @mul_i8(i8 %a) nounwind { - %tmp = mul i8 %a, 17 - ret i8 %tmp +define void @mul_i8(i8 %a, i8* %p) nounwind { + %tmp = mul i8 %a, 17 + store i8 %tmp, i8* %p + ret void } define void @load_store_i1(i1* %p, i1* %q) nounwind { diff --git a/test/CodeGen/X86/fold-xmm-zero.ll b/test/CodeGen/X86/fold-xmm-zero.ll new file mode 100644 index 0000000..b4eeb40 --- /dev/null +++ b/test/CodeGen/X86/fold-xmm-zero.ll @@ -0,0 +1,34 @@ +; RUN: llc < %s -mtriple=i386-apple-macosx10.6.7 -mattr=+sse2 | FileCheck %s + +; Simple test to make sure folding for special constants (like float zero) +; isn't completely broken. + +; CHECK: divss LCPI0 + +%0 = type { float, float, float, float, float, float, float, float } + +define void @f() nounwind ssp { +entry: + %0 = tail call %0 asm sideeffect "foo", "={xmm0},={xmm1},={xmm2},={xmm3},={xmm4},={xmm5},={xmm6},={xmm7},0,1,2,3,4,5,6,7,~{dirflag},~{fpsr},~{flags}"(float 1.000000e+00, float 2.000000e+00, float 3.000000e+00, float 4.000000e+00, float 5.000000e+00, float 6.000000e+00, float 7.000000e+00, float 8.000000e+00) nounwind + %asmresult = extractvalue %0 %0, 0 + %asmresult8 = extractvalue %0 %0, 1 + %asmresult9 = extractvalue %0 %0, 2 + %asmresult10 = extractvalue %0 %0, 3 + %asmresult11 = extractvalue %0 %0, 4 + %asmresult12 = extractvalue %0 %0, 5 + %asmresult13 = extractvalue %0 %0, 6 + %asmresult14 = extractvalue %0 %0, 7 + %div = fdiv float %asmresult, 0.000000e+00 + %1 = tail call %0 asm sideeffect "bar", "={xmm0},={xmm1},={xmm2},={xmm3},={xmm4},={xmm5},={xmm6},={xmm7},0,1,2,3,4,5,6,7,~{dirflag},~{fpsr},~{flags}"(float %div, float %asmresult8, float %asmresult9, float %asmresult10, float %asmresult11, float %asmresult12, float %asmresult13, float %asmresult14) nounwind + %asmresult24 = extractvalue %0 %1, 0 + %asmresult25 = extractvalue %0 %1, 1 + %asmresult26 = extractvalue %0 %1, 2 + %asmresult27 = extractvalue %0 %1, 3 + %asmresult28 = extractvalue %0 %1, 4 + %asmresult29 = extractvalue %0 %1, 5 + %asmresult30 = extractvalue %0 %1, 6 + %asmresult31 = extractvalue %0 %1, 7 + %div33 = fdiv float %asmresult24, 0.000000e+00 + %2 = tail call %0 asm sideeffect "baz", "={xmm0},={xmm1},={xmm2},={xmm3},={xmm4},={xmm5},={xmm6},={xmm7},0,1,2,3,4,5,6,7,~{dirflag},~{fpsr},~{flags}"(float %div33, float %asmresult25, float %asmresult26, float %asmresult27, float %asmresult28, float %asmresult29, float %asmresult30, float %asmresult31) nounwind + ret void +} diff --git a/test/CodeGen/X86/hidden-vis-pic.ll b/test/CodeGen/X86/hidden-vis-pic.ll index 217dba6..67be3d0 100644 --- a/test/CodeGen/X86/hidden-vis-pic.ll +++ b/test/CodeGen/X86/hidden-vis-pic.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -disable-cfi -mtriple=i386-apple-darwin9 -relocation-model=pic -disable-fp-elim -unwind-tables | FileCheck %s +; RUN: llc < %s -disable-cfi -mtriple=i386-apple-darwin9 -relocation-model=pic -disable-fp-elim | FileCheck %s @@ -26,7 +26,7 @@ entry: @.str = private constant [12 x i8] c"hello world\00", align 1 ; <[12 x i8]*> [#uses=1] -define hidden void @func() nounwind ssp { +define hidden void @func() nounwind ssp uwtable { entry: %0 = call i32 @puts(i8* getelementptr inbounds ([12 x i8]* @.str, i64 0, i64 0)) nounwind ; <i32> [#uses=0] br label %return @@ -37,7 +37,7 @@ return: ; preds = %entry declare i32 @puts(i8*) -define hidden i32 @main() nounwind ssp { +define hidden i32 @main() nounwind ssp uwtable { entry: %retval = alloca i32 ; <i32*> [#uses=1] %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] diff --git a/test/CodeGen/X86/hoist-common.ll b/test/CodeGen/X86/hoist-common.ll new file mode 100644 index 0000000..72e17c0 --- /dev/null +++ b/test/CodeGen/X86/hoist-common.ll @@ -0,0 +1,28 @@ +; RUN: llc < %s -mtriple=x86_64-apple-macosx | FileCheck %s + +; Common "xorb al, al" instruction in the two successor blocks should be +; moved to the entry block above the test + je. + +; rdar://9145558 + +define zeroext i1 @t(i32 %c) nounwind ssp { +entry: +; CHECK: t: +; CHECK: xorb %al, %al +; CHECK: test +; CHECK: je + %tobool = icmp eq i32 %c, 0 + br i1 %tobool, label %return, label %if.then + +if.then: +; CHECK: callq + %call = tail call zeroext i1 (...)* @foo() nounwind + br label %return + +return: +; CHECK: ret + %retval.0 = phi i1 [ %call, %if.then ], [ false, %entry ] + ret i1 %retval.0 +} + +declare zeroext i1 @foo(...) diff --git a/test/CodeGen/X86/inline-asm-error.ll b/test/CodeGen/X86/inline-asm-error.ll new file mode 100644 index 0000000..29c5ae5 --- /dev/null +++ b/test/CodeGen/X86/inline-asm-error.ll @@ -0,0 +1,17 @@ +; RUN: not llc -march x86 -regalloc=fast < %s 2> %t1 +; RUN: not llc -march x86 -regalloc=basic < %s 2> %t2 +; RUN: not llc -march x86 -regalloc=greedy < %s 2> %t3 +; RUN: FileCheck %s < %t1 +; RUN: FileCheck %s < %t2 +; RUN: FileCheck %s < %t3 + +; The register allocator must fail on this function, and it should print the +; inline asm in the diagnostic. +; CHECK: LLVM ERROR: Ran out of registers during register allocation! +; CHECK: INLINEASM <es:hello world> + +define void @f(i32 %x0, i32 %x1, i32 %x2, i32 %x3, i32 %x4, i32 %x5, i32 %x6, i32 %x7, i32 %x8, i32 %x9) nounwind ssp { +entry: + tail call void asm sideeffect "hello world", "r,r,r,r,r,r,r,r,r,r,~{dirflag},~{fpsr},~{flags}"(i32 %x0, i32 %x1, i32 %x2, i32 %x3, i32 %x4, i32 %x5, i32 %x6, i32 %x7, i32 %x8, i32 %x9) nounwind + ret void +} diff --git a/test/CodeGen/X86/isint.ll b/test/CodeGen/X86/isint.ll index 507a328..4a98e63 100644 --- a/test/CodeGen/X86/isint.ll +++ b/test/CodeGen/X86/isint.ll @@ -1,17 +1,15 @@ -; RUN: llc < %s -march=x86 -mattr=+sse2 > %t -; RUN: not grep cmp %t -; RUN: not grep xor %t -; RUN: grep jne %t | count 1 -; RUN: grep jp %t | count 1 -; RUN: grep setnp %t | count 1 -; RUN: grep sete %t | count 1 -; RUN: grep and %t | count 1 -; RUN: grep cvt %t | count 4 +; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s define i32 @isint_return(double %d) nounwind { +; CHECK-NOT: xor +; CHECK: cvt %i = fptosi double %d to i32 +; CHECK-NEXT: cvt %e = sitofp i32 %i to double +; CHECK: cmpeqsd %c = fcmp oeq double %d, %e +; CHECK-NEXT: movd +; CHECK-NEXT: andl %z = zext i1 %c to i32 ret i32 %z } @@ -19,9 +17,14 @@ define i32 @isint_return(double %d) nounwind { declare void @foo() define void @isint_branch(double %d) nounwind { +; CHECK: cvt %i = fptosi double %d to i32 +; CHECK-NEXT: cvt %e = sitofp i32 %i to double +; CHECK: ucomisd %c = fcmp oeq double %d, %e +; CHECK-NEXT: jne +; CHECK-NEXT: jp br i1 %c, label %true, label %false true: call void @foo() diff --git a/test/CodeGen/X86/lsr-overflow.ll b/test/CodeGen/X86/lsr-overflow.ll index 5bc4f7e..09c1c07 100644 --- a/test/CodeGen/X86/lsr-overflow.ll +++ b/test/CodeGen/X86/lsr-overflow.ll @@ -25,3 +25,21 @@ __ABContainsLabel.exit: %cmp = icmp eq i64 %indvar, 9223372036854775807 ret i1 %cmp } + +define void @func_37() noreturn nounwind readonly { +entry: + br label %for.body + +for.body: ; preds = %for.inc8, %entry + %indvar = phi i64 [ 0, %entry ], [ %indvar.next, %for.inc8 ] + %sub.i = add i64 undef, %indvar + %cmp.i = icmp eq i64 %sub.i, -9223372036854775808 + br i1 undef, label %for.inc8, label %for.cond4 + +for.cond4: ; preds = %for.cond4, %for.body + br label %for.cond4 + +for.inc8: ; preds = %for.body + %indvar.next = add i64 %indvar, 1 + br label %for.body +} diff --git a/test/CodeGen/X86/movntdq-no-avx.ll b/test/CodeGen/X86/movntdq-no-avx.ll new file mode 100644 index 0000000..8b7e6ef --- /dev/null +++ b/test/CodeGen/X86/movntdq-no-avx.ll @@ -0,0 +1,12 @@ +; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck %s + +; Test that we produce a movntdq, not a vmovntdq +; CHECK-NOT: vmovntdq + +define void @test(<2 x i64>* nocapture %a, <2 x i64> %b) nounwind optsize { +entry: + store <2 x i64> %b, <2 x i64>* %a, align 16, !nontemporal !0 + ret void +} + +!0 = metadata !{i32 1} diff --git a/test/CodeGen/X86/nontemporal.ll b/test/CodeGen/X86/nontemporal.ll new file mode 100644 index 0000000..1d09535 --- /dev/null +++ b/test/CodeGen/X86/nontemporal.ll @@ -0,0 +1,19 @@ +; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s + +define void @f(<4 x float> %A, i8* %B, <2 x double> %C, i32 %D, <2 x i64> %E) { +; CHECK: movntps + %cast = bitcast i8* %B to <4 x float>* + store <4 x float> %A, <4 x float>* %cast, align 16, !nontemporal !0 +; CHECK: movntdq + %cast1 = bitcast i8* %B to <2 x i64>* + store <2 x i64> %E, <2 x i64>* %cast1, align 16, !nontemporal !0 +; CHECK: movntpd + %cast2 = bitcast i8* %B to <2 x double>* + store <2 x double> %C, <2 x double>* %cast2, align 16, !nontemporal !0 +; CHECK: movnti + %cast3 = bitcast i8* %B to i32* + store i32 %D, i32* %cast3, align 16, !nontemporal !0 + ret void +} + +!0 = metadata !{i32 1} diff --git a/test/CodeGen/X86/optimize-max-3.ll b/test/CodeGen/X86/optimize-max-3.ll index e35eb70..e42aa9d 100644 --- a/test/CodeGen/X86/optimize-max-3.ll +++ b/test/CodeGen/X86/optimize-max-3.ll @@ -41,14 +41,13 @@ for.end: ; preds = %for.body, %entry ; CHECK: jle ; CHECK-NOT: cmov -; CHECK: xorl {{%edi, %edi|%ecx, %ecx}} +; CHECK: xorl {{%edi, %edi|%ecx, %ecx|%eax, %eax}} ; CHECK-NEXT: align ; CHECK-NEXT: BB1_2: -; CHECK-NEXT: callq +; CHECK: callq ; CHECK-NEXT: incl [[BX:%[a-z0-9]+]] ; CHECK-NEXT: cmpl [[R14:%[a-z0-9]+]], [[BX]] -; CHECK-NEXT: movq %rax, %r{{di|cx}} -; CHECK-NEXT: jl +; CHECK: jl define void @_Z18GenerateStatusPagei(i32 %jobs_to_display) nounwind { entry: diff --git a/test/CodeGen/X86/peep-setb.ll b/test/CodeGen/X86/peep-setb.ll new file mode 100644 index 0000000..0bab789 --- /dev/null +++ b/test/CodeGen/X86/peep-setb.ll @@ -0,0 +1,82 @@ +; RUN: llc -march=x86-64 < %s | FileCheck %s + +define i8 @test1(i8 %a, i8 %b) nounwind { + %cmp = icmp ult i8 %a, %b + %cond = zext i1 %cmp to i8 + %add = add i8 %cond, %b + ret i8 %add +; CHECK: test1: +; CHECK: adcb $0 +} + +define i32 @test2(i32 %a, i32 %b) nounwind { + %cmp = icmp ult i32 %a, %b + %cond = zext i1 %cmp to i32 + %add = add i32 %cond, %b + ret i32 %add +; CHECK: test2: +; CHECK: adcl $0 +} + +define i64 @test3(i64 %a, i64 %b) nounwind { + %cmp = icmp ult i64 %a, %b + %conv = zext i1 %cmp to i64 + %add = add i64 %conv, %b + ret i64 %add +; CHECK: test3: +; CHECK: adcq $0 +} + +define i8 @test4(i8 %a, i8 %b) nounwind { + %cmp = icmp ult i8 %a, %b + %cond = zext i1 %cmp to i8 + %sub = sub i8 %b, %cond + ret i8 %sub +; CHECK: test4: +; CHECK: sbbb $0 +} + +define i32 @test5(i32 %a, i32 %b) nounwind { + %cmp = icmp ult i32 %a, %b + %cond = zext i1 %cmp to i32 + %sub = sub i32 %b, %cond + ret i32 %sub +; CHECK: test5: +; CHECK: sbbl $0 +} + +define i64 @test6(i64 %a, i64 %b) nounwind { + %cmp = icmp ult i64 %a, %b + %conv = zext i1 %cmp to i64 + %sub = sub i64 %b, %conv + ret i64 %sub +; CHECK: test6: +; CHECK: sbbq $0 +} + +define i8 @test7(i8 %a, i8 %b) nounwind { + %cmp = icmp ult i8 %a, %b + %cond = sext i1 %cmp to i8 + %sub = sub i8 %b, %cond + ret i8 %sub +; CHECK: test7: +; CHECK: adcb $0 +} + +define i32 @test8(i32 %a, i32 %b) nounwind { + %cmp = icmp ult i32 %a, %b + %cond = sext i1 %cmp to i32 + %sub = sub i32 %b, %cond + ret i32 %sub +; CHECK: test8: +; CHECK: adcl $0 +} + +define i64 @test9(i64 %a, i64 %b) nounwind { + %cmp = icmp ult i64 %a, %b + %conv = sext i1 %cmp to i64 + %sub = sub i64 %b, %conv + ret i64 %sub +; CHECK: test9: +; CHECK: adcq $0 +} diff --git a/test/CodeGen/X86/phys_subreg_coalesce-2.ll b/test/CodeGen/X86/phys_subreg_coalesce-2.ll index 13e804d..02c519f 100644 --- a/test/CodeGen/X86/phys_subreg_coalesce-2.ll +++ b/test/CodeGen/X86/phys_subreg_coalesce-2.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86 | grep mov | count 4 +; RUN: llc < %s -march=x86 | FileCheck %s ; PR2659 define i32 @binomial(i32 %n, i32 %k) nounwind { @@ -12,7 +12,8 @@ forcond.preheader: ; preds = %entry ifthen: ; preds = %entry ret i32 0 - +; CHECK: forbody +; CHECK-NOT: mov forbody: ; preds = %forbody, %forcond.preheader %indvar = phi i32 [ 0, %forcond.preheader ], [ %divisor.02, %forbody ] ; <i32> [#uses=3] %accumulator.01 = phi i32 [ 1, %forcond.preheader ], [ %div, %forbody ] ; <i32> [#uses=1] diff --git a/test/CodeGen/X86/phys_subreg_coalesce-3.ll b/test/CodeGen/X86/phys_subreg_coalesce-3.ll index f23669e..4162015 100644 --- a/test/CodeGen/X86/phys_subreg_coalesce-3.ll +++ b/test/CodeGen/X86/phys_subreg_coalesce-3.ll @@ -1,6 +1,11 @@ -; RUN: llc < %s -mtriple=i386-apple-darwin | FileCheck %s +; RUN: llc < %s -mtriple=i386-apple-darwin -join-physregs | FileCheck %s ; rdar://5571034 +; This requires physreg joining, %vreg13 is live everywhere: +; 304L %CL<def> = COPY %vreg13:sub_8bit; GR32_ABCD:%vreg13 +; 320L %vreg15<def> = COPY %vreg19; GR32:%vreg15 GR32_NOSP:%vreg19 +; 336L %vreg15<def> = SAR32rCL %vreg15, %EFLAGS<imp-def,dead>, %CL<imp-use,kill>; GR32:%vreg15 + define void @foo(i32* nocapture %quadrant, i32* nocapture %ptr, i32 %bbSize, i32 %bbStart, i32 %shifts) nounwind ssp { ; CHECK: foo: entry: diff --git a/test/CodeGen/X86/pmul.ll b/test/CodeGen/X86/pmul.ll index bf5229a..d8ed4c0 100644 --- a/test/CodeGen/X86/pmul.ll +++ b/test/CodeGen/X86/pmul.ll @@ -1,7 +1,9 @@ -; RUN: llc < %s -march=x86 -mattr=sse41 -stack-alignment=16 > %t +; RUN: llc < %s -march=x86 -mattr=sse41 -stack-alignment=16 -join-physregs > %t ; RUN: grep pmul %t | count 12 ; RUN: grep mov %t | count 11 +; The f() arguments in %xmm0 and %xmm1 cause an extra movdqa without -join-physregs. + define <4 x i32> @a(<4 x i32> %i) nounwind { %A = mul <4 x i32> %i, < i32 117, i32 117, i32 117, i32 117 > ret <4 x i32> %A diff --git a/test/CodeGen/X86/pr10068.ll b/test/CodeGen/X86/pr10068.ll new file mode 100644 index 0000000..8829c5d --- /dev/null +++ b/test/CodeGen/X86/pr10068.ll @@ -0,0 +1,22 @@ +; RUN: llc < %s -march=x86 + +define void @foobar() { +entry: + %sub.i = trunc i64 undef to i32 + %shr80.i = ashr i32 %sub.i, 16 + %add82.i = add nsw i32 %shr80.i, 1 + %notlhs.i = icmp slt i32 %shr80.i, undef + %notrhs.i = icmp sgt i32 %add82.i, -1 + %or.cond.not.i = and i1 %notrhs.i, %notlhs.i + %cmp154.i = icmp slt i32 0, undef + %or.cond406.i = and i1 %or.cond.not.i, %cmp154.i + %or.cond406.not.i = xor i1 %or.cond406.i, true + %or.cond407.i = or i1 undef, %or.cond406.not.i + br i1 %or.cond407.i, label %if.then158.i, label %if.end163.i + +if.then158.i: + ret void + +if.end163.i: ; preds = %if.end67.i + ret void +} diff --git a/test/CodeGen/X86/pr2659.ll b/test/CodeGen/X86/pr2659.ll index ef0f9ea..5dab5c9 100644 --- a/test/CodeGen/X86/pr2659.ll +++ b/test/CodeGen/X86/pr2659.ll @@ -1,5 +1,4 @@ -; RUN: llc < %s -march=x86 -mtriple=i686-apple-darwin9.4.0 | grep movl | count 4 -; RUN: llc < %s -march=x86 -mtriple=i686-apple-darwin9.4.0 | FileCheck %s +; RUN: llc < %s -march=x86 -mtriple=i686-apple-darwin9.4.0 -disable-branch-fold | FileCheck %s ; PR2659 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" @@ -19,7 +18,11 @@ forcond.preheader: ; preds = %entry ; CHECK-NOT: xorl ; CHECK-NOT: movl ; CHECK-NOT: LBB -; CHECK: je +; CHECK: jne + +; There should be no moves required in the for loop body. +; CHECK: %forbody +; CHECK-NOT: mov ifthen: ; preds = %entry ret i32 0 diff --git a/test/CodeGen/X86/pr9127.ll b/test/CodeGen/X86/pr9127.ll index 9b251f5..ba92c77 100644 --- a/test/CodeGen/X86/pr9127.ll +++ b/test/CodeGen/X86/pr9127.ll @@ -10,4 +10,4 @@ entry: } ; test that the load is folded. -; CHECK: ucomisd (%{{rdi|rdx}}), %xmm0 +; CHECK: cmpeqsd (%{{rdi|rdx}}), %xmm0 diff --git a/test/CodeGen/X86/pr9743.ll b/test/CodeGen/X86/pr9743.ll index 8feccd9..6597c23 100644 --- a/test/CodeGen/X86/pr9743.ll +++ b/test/CodeGen/X86/pr9743.ll @@ -9,9 +9,9 @@ define void @f() { ; CHECK-NEXT: : ; CHECK-NEXT: .cfi_def_cfa_offset 16 ; CHECK-NEXT: : -; CHECK-NEXT: .cfi_offset 6, -16 +; CHECK-NEXT: .cfi_offset %rbp, -16 ; CHECK-NEXT: movq %rsp, %rbp ; CHECK-NEXT: : -; CHECK-NEXT: .cfi_def_cfa_register 6 +; CHECK-NEXT: .cfi_def_cfa_register %rbp ; CHECK-NEXT: popq %rbp ; CHECK-NEXT: ret diff --git a/test/CodeGen/X86/ret-mmx.ll b/test/CodeGen/X86/ret-mmx.ll index 04b57dd..865e147 100644 --- a/test/CodeGen/X86/ret-mmx.ll +++ b/test/CodeGen/X86/ret-mmx.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86-64 -mattr=+mmx,+sse2 +; RUN: llc < %s -mtriple=x86_64-apple-darwin11 -mattr=+mmx,+sse2 | FileCheck %s ; rdar://6602459 @g_v1di = external global <1 x i64> @@ -8,19 +8,32 @@ entry: %call = call <1 x i64> @return_v1di() ; <<1 x i64>> [#uses=0] store <1 x i64> %call, <1 x i64>* @g_v1di ret void +; CHECK: t1: +; CHECK: callq +; CHECK-NEXT: movq _g_v1di +; CHECK-NEXT: movq %rax, } declare <1 x i64> @return_v1di() define <1 x i64> @t2() nounwind { ret <1 x i64> <i64 1> +; CHECK: t2: +; CHECK: movl $1 +; CHECK-NEXT: ret } define <2 x i32> @t3() nounwind { ret <2 x i32> <i32 1, i32 0> +; CHECK: t3: +; CHECK: movl $1 +; CHECK: movd {{.*}}, %xmm0 } define double @t4() nounwind { ret double bitcast (<2 x i32> <i32 1, i32 0> to double) +; CHECK: t4: +; CHECK: movl $1 +; CHECK: movd {{.*}}, %xmm0 } diff --git a/test/CodeGen/X86/setoeq.ll b/test/CodeGen/X86/setoeq.ll index 4a9c1ba..aa2f0af 100644 --- a/test/CodeGen/X86/setoeq.ll +++ b/test/CodeGen/X86/setoeq.ll @@ -1,5 +1,4 @@ -; RUN: llc < %s -march=x86 | grep set | count 2 -; RUN: llc < %s -march=x86 | grep and +; RUN: llc < %s -march=x86 -mattr=+sse2 | FileCheck %s define zeroext i8 @t(double %x) nounwind readnone { entry: @@ -7,5 +6,16 @@ entry: %1 = sitofp i32 %0 to double ; <double> [#uses=1] %2 = fcmp oeq double %1, %x ; <i1> [#uses=1] %retval12 = zext i1 %2 to i8 ; <i8> [#uses=1] +; CHECK: cmpeqsd + ret i8 %retval12 +} + +define zeroext i8 @u(double %x) nounwind readnone { +entry: + %0 = fptosi double %x to i32 ; <i32> [#uses=1] + %1 = sitofp i32 %0 to double ; <double> [#uses=1] + %2 = fcmp une double %1, %x ; <i1> [#uses=1] + %retval12 = zext i1 %2 to i8 ; <i8> [#uses=1] +; CHECK: cmpneqsd ret i8 %retval12 } diff --git a/test/CodeGen/X86/shift-pair.ll b/test/CodeGen/X86/shift-pair.ll new file mode 100644 index 0000000..24ba1fc --- /dev/null +++ b/test/CodeGen/X86/shift-pair.ll @@ -0,0 +1,11 @@ +; RUN: llc < %s -march=x86-64 | FileCheck %s + +define i64 @test(i64 %A) { +; CHECK: @test +; CHECK: shrq $54 +; CHECK: andq $1020 +; CHECK: ret + %B = lshr i64 %A, 56 + %C = shl i64 %B, 2 + ret i64 %C +} diff --git a/test/CodeGen/X86/sibcall.ll b/test/CodeGen/X86/sibcall.ll index de2a81e8..4a98efb 100644 --- a/test/CodeGen/X86/sibcall.ll +++ b/test/CodeGen/X86/sibcall.ll @@ -198,7 +198,7 @@ declare i32 @foo6(i32, i32, %struct.t* byval align 4) ; rdar://r7717598 %struct.ns = type { i32, i32 } -%struct.cp = type { float, float } +%struct.cp = type { float, float, float, float, float } define %struct.ns* @t13(%struct.cp* %yy) nounwind ssp { ; 32: t13: @@ -229,7 +229,7 @@ entry: ; 64: t14: ; 64: movq 32(%rdi) ; 64-NOT: movq 16(%rdi) -; 64: jmpq *16(%rdi) +; 64: jmpq *16({{%rdi|%rax}}) %0 = getelementptr inbounds %struct.__block_literal_2* %.block_descriptor, i64 0, i32 5 ; <void ()**> [#uses=1] %1 = load void ()** %0, align 8 ; <void ()*> [#uses=2] %2 = bitcast void ()* %1 to %struct.__block_literal_1* ; <%struct.__block_literal_1*> [#uses=1] diff --git a/test/CodeGen/X86/smul-with-overflow-2.ll b/test/CodeGen/X86/smul-with-overflow-2.ll deleted file mode 100644 index 7c23adb..0000000 --- a/test/CodeGen/X86/smul-with-overflow-2.ll +++ /dev/null @@ -1,20 +0,0 @@ -; RUN: llc < %s -march=x86 | grep mul | count 1 -; RUN: llc < %s -march=x86 | grep add | count 3 - -define i32 @t1(i32 %a, i32 %b) nounwind readnone { -entry: - %tmp0 = add i32 %b, %a - %tmp1 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 %tmp0, i32 2) - %tmp2 = extractvalue { i32, i1 } %tmp1, 0 - ret i32 %tmp2 -} - -define i32 @t2(i32 %a, i32 %b) nounwind readnone { -entry: - %tmp0 = add i32 %b, %a - %tmp1 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 %tmp0, i32 4) - %tmp2 = extractvalue { i32, i1 } %tmp1, 0 - ret i32 %tmp2 -} - -declare { i32, i1 } @llvm.smul.with.overflow.i32(i32, i32) nounwind diff --git a/test/CodeGen/X86/smul-with-overflow-3.ll b/test/CodeGen/X86/smul-with-overflow-3.ll deleted file mode 100644 index 49c31f5..0000000 --- a/test/CodeGen/X86/smul-with-overflow-3.ll +++ /dev/null @@ -1,23 +0,0 @@ -; RUN: llc < %s -march=x86 | grep {jno} | count 1 - -@ok = internal constant [4 x i8] c"%d\0A\00" -@no = internal constant [4 x i8] c"no\0A\00" - -define i1 @func1(i32 %v1, i32 %v2) nounwind { -entry: - %t = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %v1, i32 %v2) - %sum = extractvalue {i32, i1} %t, 0 - %obit = extractvalue {i32, i1} %t, 1 - br i1 %obit, label %overflow, label %normal - -overflow: - %t2 = tail call i32 (i8*, ...)* @printf( i8* getelementptr ([4 x i8]* @no, i32 0, i32 0) ) nounwind - ret i1 false - -normal: - %t1 = tail call i32 (i8*, ...)* @printf( i8* getelementptr ([4 x i8]* @ok, i32 0, i32 0), i32 %sum ) nounwind - ret i1 true -} - -declare i32 @printf(i8*, ...) nounwind -declare {i32, i1} @llvm.smul.with.overflow.i32(i32, i32) diff --git a/test/CodeGen/X86/smul-with-overflow.ll b/test/CodeGen/X86/smul-with-overflow.ll index 6d125e4..7c2e247 100644 --- a/test/CodeGen/X86/smul-with-overflow.ll +++ b/test/CodeGen/X86/smul-with-overflow.ll @@ -1,9 +1,9 @@ -; RUN: llc < %s -march=x86 | grep {jo} | count 1 +; RUN: llc < %s -march=x86 | FileCheck %s @ok = internal constant [4 x i8] c"%d\0A\00" @no = internal constant [4 x i8] c"no\0A\00" -define i1 @func1(i32 %v1, i32 %v2) nounwind { +define i1 @test1(i32 %v1, i32 %v2) nounwind { entry: %t = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %v1, i32 %v2) %sum = extractvalue {i32, i1} %t, 0 @@ -17,7 +17,53 @@ normal: overflow: %t2 = tail call i32 (i8*, ...)* @printf( i8* getelementptr ([4 x i8]* @no, i32 0, i32 0) ) nounwind ret i1 false +; CHECK: test1: +; CHECK: imull +; CHECK-NEXT: jo +} + +define i1 @test2(i32 %v1, i32 %v2) nounwind { +entry: + %t = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %v1, i32 %v2) + %sum = extractvalue {i32, i1} %t, 0 + %obit = extractvalue {i32, i1} %t, 1 + br i1 %obit, label %overflow, label %normal + +overflow: + %t2 = tail call i32 (i8*, ...)* @printf( i8* getelementptr ([4 x i8]* @no, i32 0, i32 0) ) nounwind + ret i1 false + +normal: + %t1 = tail call i32 (i8*, ...)* @printf( i8* getelementptr ([4 x i8]* @ok, i32 0, i32 0), i32 %sum ) nounwind + ret i1 true +; CHECK: test2: +; CHECK: imull +; CHECK-NEXT: jno } declare i32 @printf(i8*, ...) nounwind declare {i32, i1} @llvm.smul.with.overflow.i32(i32, i32) + +define i32 @test3(i32 %a, i32 %b) nounwind readnone { +entry: + %tmp0 = add i32 %b, %a + %tmp1 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 %tmp0, i32 2) + %tmp2 = extractvalue { i32, i1 } %tmp1, 0 + ret i32 %tmp2 +; CHECK: test3: +; CHECK: addl +; CHECK-NEXT: addl +; CHECK-NEXT: ret +} + +define i32 @test4(i32 %a, i32 %b) nounwind readnone { +entry: + %tmp0 = add i32 %b, %a + %tmp1 = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 %tmp0, i32 4) + %tmp2 = extractvalue { i32, i1 } %tmp1, 0 + ret i32 %tmp2 +; CHECK: test4: +; CHECK: addl +; CHECK: mull +; CHECK-NEXT: ret +} diff --git a/test/CodeGen/X86/sse-minmax.ll b/test/CodeGen/X86/sse-minmax.ll index 348121a..ff0af25 100644 --- a/test/CodeGen/X86/sse-minmax.ll +++ b/test/CodeGen/X86/sse-minmax.ll @@ -1,6 +1,6 @@ -; RUN: llc < %s -march=x86-64 -asm-verbose=false | FileCheck %s -; RUN: llc < %s -march=x86-64 -asm-verbose=false -enable-unsafe-fp-math -enable-no-nans-fp-math | FileCheck -check-prefix=UNSAFE %s -; RUN: llc < %s -march=x86-64 -asm-verbose=false -enable-no-nans-fp-math | FileCheck -check-prefix=FINITE %s +; RUN: llc < %s -march=x86-64 -asm-verbose=false -join-physregs | FileCheck %s +; RUN: llc < %s -march=x86-64 -asm-verbose=false -join-physregs -enable-unsafe-fp-math -enable-no-nans-fp-math | FileCheck -check-prefix=UNSAFE %s +; RUN: llc < %s -march=x86-64 -asm-verbose=false -join-physregs -enable-no-nans-fp-math | FileCheck -check-prefix=FINITE %s ; Some of these patterns can be matched as SSE min or max. Some of ; then can be matched provided that the operands are swapped. @@ -12,6 +12,9 @@ ; y_ : use -0.0 instead of %y ; _inverse : swap the arms of the select. +; Some of these tests depend on -join-physregs commuting instructions to +; eliminate copies. + ; CHECK: ogt: ; CHECK-NEXT: maxsd %xmm1, %xmm0 ; CHECK-NEXT: ret diff --git a/test/CodeGen/X86/sse3.ll b/test/CodeGen/X86/sse3.ll index 8e72f13..8c2e58d 100644 --- a/test/CodeGen/X86/sse3.ll +++ b/test/CodeGen/X86/sse3.ll @@ -62,11 +62,10 @@ define <8 x i16> @t4(<8 x i16> %A, <8 x i16> %B) nounwind { %tmp = shufflevector <8 x i16> %A, <8 x i16> %B, <8 x i32> < i32 0, i32 7, i32 2, i32 3, i32 1, i32 5, i32 6, i32 5 > ret <8 x i16> %tmp ; X64: t4: -; X64: pextrw $7, %xmm0, %eax -; X64: pshufhw $100, %xmm0, %xmm1 -; X64: pinsrw $1, %eax, %xmm1 -; X64: pextrw $1, %xmm0, %eax -; X64: movdqa %xmm1, %xmm0 +; X64: pextrw $7, [[XMM0:%xmm[0-9]+]], %eax +; X64: pshufhw $100, [[XMM0]], [[XMM1:%xmm[0-9]+]] +; X64: pinsrw $1, %eax, [[XMM1]] +; X64: pextrw $1, [[XMM0]], %eax ; X64: pinsrw $4, %eax, %xmm0 ; X64: ret } @@ -251,13 +250,13 @@ entry: %tmp9 = shufflevector <16 x i8> %tmp8, <16 x i8> %T0, <16 x i32> < i32 0, i32 1, i32 2, i32 17, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef , i32 undef > ret <16 x i8> %tmp9 ; X64: t16: -; X64: pinsrw $0, %eax, %xmm1 -; X64: pextrw $8, %xmm0, %eax -; X64: pinsrw $1, %eax, %xmm1 -; X64: pextrw $1, %xmm1, %ecx -; X64: movd %xmm1, %edx -; X64: pinsrw $0, %edx, %xmm1 -; X64: pinsrw $1, %eax, %xmm0 +; X64: pinsrw $0, %eax, [[X1:%xmm[0-9]+]] +; X64: pextrw $8, [[X0:%xmm[0-9]+]], %eax +; X64: pinsrw $1, %eax, [[X1]] +; X64: pextrw $1, [[X1]], %ecx +; X64: movd [[X1]], %edx +; X64: pinsrw $0, %edx, %xmm +; X64: pinsrw $1, %eax, %xmm ; X64: ret } diff --git a/test/CodeGen/X86/sse42.ll b/test/CodeGen/X86/sse42.ll index 1723909..c787523 100644 --- a/test/CodeGen/X86/sse42.ll +++ b/test/CodeGen/X86/sse42.ll @@ -1,38 +1,39 @@ ; RUN: llc < %s -mtriple=i686-apple-darwin9 -mattr=sse42 | FileCheck %s -check-prefix=X32 ; RUN: llc < %s -mtriple=x86_64-apple-darwin9 -mattr=sse42 | FileCheck %s -check-prefix=X64 -declare i32 @llvm.x86.sse42.crc32.8(i32, i8) nounwind -declare i32 @llvm.x86.sse42.crc32.16(i32, i16) nounwind -declare i32 @llvm.x86.sse42.crc32.32(i32, i32) nounwind +declare i32 @llvm.x86.sse42.crc32.32.8(i32, i8) nounwind +declare i32 @llvm.x86.sse42.crc32.32.16(i32, i16) nounwind +declare i32 @llvm.x86.sse42.crc32.32.32(i32, i32) nounwind -define i32 @crc32_8(i32 %a, i8 %b) nounwind { - %tmp = call i32 @llvm.x86.sse42.crc32.8(i32 %a, i8 %b) +define i32 @crc32_32_8(i32 %a, i8 %b) nounwind { + %tmp = call i32 @llvm.x86.sse42.crc32.32.8(i32 %a, i8 %b) ret i32 %tmp -; X32: _crc32_8: +; X32: _crc32_32_8: ; X32: crc32b 8(%esp), %eax -; X64: _crc32_8: -; X64: crc32b %sil, %eax +; X64: _crc32_32_8: +; X64: crc32b %sil, } -define i32 @crc32_16(i32 %a, i16 %b) nounwind { - %tmp = call i32 @llvm.x86.sse42.crc32.16(i32 %a, i16 %b) +define i32 @crc32_32_16(i32 %a, i16 %b) nounwind { + %tmp = call i32 @llvm.x86.sse42.crc32.32.16(i32 %a, i16 %b) ret i32 %tmp -; X32: _crc32_16: +; X32: _crc32_32_16: ; X32: crc32w 8(%esp), %eax -; X64: _crc32_16: -; X64: crc32w %si, %eax +; X64: _crc32_32_16: +; X64: crc32w %si, } -define i32 @crc32_32(i32 %a, i32 %b) nounwind { - %tmp = call i32 @llvm.x86.sse42.crc32.32(i32 %a, i32 %b) +define i32 @crc32_32_32(i32 %a, i32 %b) nounwind { + %tmp = call i32 @llvm.x86.sse42.crc32.32.32(i32 %a, i32 %b) ret i32 %tmp -; X32: _crc32_32: +; X32: _crc32_32_32: ; X32: crc32l 8(%esp), %eax -; X64: _crc32_32: -; X64: crc32l %esi, %eax +; X64: _crc32_32_32: +; X64: crc32l %esi, } + diff --git a/test/CodeGen/X86/sse42_64.ll b/test/CodeGen/X86/sse42_64.ll new file mode 100644 index 0000000..8b3a69b --- /dev/null +++ b/test/CodeGen/X86/sse42_64.ll @@ -0,0 +1,21 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin9 -mattr=sse42 | FileCheck %s -check-prefix=X64 + +declare i64 @llvm.x86.sse42.crc32.64.8(i64, i8) nounwind +declare i64 @llvm.x86.sse42.crc32.64.64(i64, i64) nounwind + +define i64 @crc32_64_8(i64 %a, i8 %b) nounwind { + %tmp = call i64 @llvm.x86.sse42.crc32.64.8(i64 %a, i8 %b) + ret i64 %tmp + +; X64: _crc32_64_8: +; X64: crc32b %sil, +} + +define i64 @crc32_64_64(i64 %a, i64 %b) nounwind { + %tmp = call i64 @llvm.x86.sse42.crc32.64.64(i64 %a, i64 %b) + ret i64 %tmp + +; X64: _crc32_64_64: +; X64: crc32q %rsi, +} + diff --git a/test/CodeGen/X86/sse_reload_fold.ll b/test/CodeGen/X86/sse_reload_fold.ll index 02399c4..a57fa58 100644 --- a/test/CodeGen/X86/sse_reload_fold.ll +++ b/test/CodeGen/X86/sse_reload_fold.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -mtriple=x86_64-linux -mattr=+64bit,+sse3 -print-failed-fuse-candidates |& FileCheck %s +; RUN: llc < %s -mtriple=x86_64-linux -mattr=+64bit,+sse3 -print-failed-fuse-candidates -regalloc=basic |& FileCheck %s ; CHECK: fail ; CHECK-NOT: fail @@ -117,7 +117,16 @@ define <2 x double> @d8(<2 x double> %f) { ret <2 x double> %t } -; This one should fail to fuse. +; This one should fail to fuse, but -regalloc=greedy isn't even trying. Instead +; it produces: +; callq test_vd +; movapd (%rsp), %xmm1 # 16-byte Reload +; hsubpd %xmm0, %xmm1 +; movapd %xmm1, %xmm0 +; addq $24, %rsp +; ret +; RABasic still tries to fold this one. + define <2 x double> @z0(<2 x double> %f) { %y = call <2 x double> @test_vd(<2 x double> %f) %t = call <2 x double> @llvm.x86.sse3.hsub.pd(<2 x double> %f, <2 x double> %y) diff --git a/test/CodeGen/X86/tail-opts.ll b/test/CodeGen/X86/tail-opts.ll index 77710ad..d6c16ca 100644 --- a/test/CodeGen/X86/tail-opts.ll +++ b/test/CodeGen/X86/tail-opts.ll @@ -412,9 +412,9 @@ return: ; can fall-through into the ret and the other side has to branch anyway. ; CHECK: TESTE: -; CHECK: imulq -; CHECK-NEXT: LBB8_2: -; CHECK-NEXT: ret +; CHECK: ret +; CHECK-NOT: ret +; CHECK: size TESTE define i64 @TESTE(i64 %parami, i64 %paraml) nounwind readnone { entry: diff --git a/test/CodeGen/X86/tailcallstack64.ll b/test/CodeGen/X86/tailcallstack64.ll index 060ce0f..c18c7aa 100644 --- a/test/CodeGen/X86/tailcallstack64.ll +++ b/test/CodeGen/X86/tailcallstack64.ll @@ -2,7 +2,7 @@ ; RUN: llc < %s -tailcallopt -mtriple=x86_64-win32 -post-RA-scheduler=true | FileCheck %s ; FIXME: Redundant unused stack allocation could be eliminated. -; CHECK: subq ${{24|72}}, %rsp +; CHECK: subq ${{24|72|80}}, %rsp ; Check that lowered arguments on the stack do not overwrite each other. ; Add %in1 %p1 to a different temporary register (%eax). diff --git a/test/CodeGen/X86/umul-with-overflow.ll b/test/CodeGen/X86/umul-with-overflow.ll index c997661..84fcbc7 100644 --- a/test/CodeGen/X86/umul-with-overflow.ll +++ b/test/CodeGen/X86/umul-with-overflow.ll @@ -12,3 +12,27 @@ define i1 @a(i32 %x) zeroext nounwind { ; CHECK: movzbl %al, %eax ; CHECK: ret } + +define i32 @test2(i32 %a, i32 %b) nounwind readnone { +entry: + %tmp0 = add i32 %b, %a + %tmp1 = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %tmp0, i32 2) + %tmp2 = extractvalue { i32, i1 } %tmp1, 0 + ret i32 %tmp2 +; CHECK: test2: +; CHECK: addl +; CHECK-NEXT: addl +; CHECK-NEXT: ret +} + +define i32 @test3(i32 %a, i32 %b) nounwind readnone { +entry: + %tmp0 = add i32 %b, %a + %tmp1 = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 %tmp0, i32 4) + %tmp2 = extractvalue { i32, i1 } %tmp1, 0 + ret i32 %tmp2 +; CHECK: test3: +; CHECK: addl +; CHECK: mull +; CHECK-NEXT: ret +} diff --git a/test/CodeGen/X86/use-add-flags.ll b/test/CodeGen/X86/use-add-flags.ll index 8fbbd39..a0448ec 100644 --- a/test/CodeGen/X86/use-add-flags.ll +++ b/test/CodeGen/X86/use-add-flags.ll @@ -7,10 +7,10 @@ ; Use the flags on the add. ; CHECK: test1: -; CHECK: addl (%r[[A0:di|cx]]), {{%esi|%edx}} -; CHECK-NEXT: movl {{%edx|%r8d}}, %eax -; CHECK-NEXT: cmovnsl {{%ecx|%r9d}}, %eax -; CHECK-NEXT: ret +; CHECK: addl +; CHECK-NOT: test +; CHECK: cmovnsl +; CHECK: ret define i32 @test1(i32* %x, i32 %y, i32 %a, i32 %b) nounwind { %tmp2 = load i32* %x, align 4 ; <i32> [#uses=1] @@ -42,7 +42,7 @@ false: ; Do use the flags result of the and here, since the and has another use. ; CHECK: test3: -; CHECK: andl $16, %e[[A0]] +; CHECK: andl $16, %e ; CHECK-NEXT: jne define void @test3(i32 %x) nounwind { diff --git a/test/CodeGen/X86/vararg_tailcall.ll b/test/CodeGen/X86/vararg_tailcall.ll new file mode 100644 index 0000000..73d80eb --- /dev/null +++ b/test/CodeGen/X86/vararg_tailcall.ll @@ -0,0 +1,98 @@ +; RUN: llc < %s -mtriple=x86_64-apple-darwin10 | FileCheck %s -check-prefix=X64 +; RUN: llc < %s -mtriple=x86_64-pc-win32 | FileCheck %s -check-prefix=WIN64 + +@.str = private unnamed_addr constant [5 x i8] c"%ld\0A\00" +@sel = external global i8* +@sel3 = external global i8* +@sel4 = external global i8* +@sel5 = external global i8* +@sel6 = external global i8* +@sel7 = external global i8* + +; X64: @foo +; X64: jmp +; WIN64: @foo +; WIN64: callq +define void @foo(i64 %arg) nounwind optsize ssp noredzone { +entry: + %call = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %arg) nounwind optsize noredzone + ret void +} + +declare i32 @printf(i8*, ...) optsize noredzone + +; X64: @bar +; X64: jmp +; WIN64: @bar +; WIN64: jmp +define void @bar(i64 %arg) nounwind optsize ssp noredzone { +entry: + tail call void @bar2(i8* getelementptr inbounds ([5 x i8]* @.str, i64 0, i64 0), i64 %arg) nounwind optsize noredzone + ret void +} + +declare void @bar2(i8*, i64) optsize noredzone + +; X64: @foo2 +; X64: jmp +; WIN64: @foo2 +; WIN64: callq +define i8* @foo2(i8* %arg) nounwind optsize ssp noredzone { +entry: + %tmp1 = load i8** @sel, align 8, !tbaa !0 + %call = tail call i8* (i8*, i8*, ...)* @x2(i8* %arg, i8* %tmp1) nounwind optsize noredzone + ret i8* %call +} + +declare i8* @x2(i8*, i8*, ...) optsize noredzone + +; X64: @foo6 +; X64: jmp +; WIN64: @foo6 +; WIN64: callq +define i8* @foo6(i8* %arg1, i8* %arg2) nounwind optsize ssp noredzone { +entry: + %tmp2 = load i8** @sel3, align 8, !tbaa !0 + %tmp3 = load i8** @sel4, align 8, !tbaa !0 + %tmp4 = load i8** @sel5, align 8, !tbaa !0 + %tmp5 = load i8** @sel6, align 8, !tbaa !0 + %call = tail call i8* (i8*, i8*, i8*, ...)* @x3(i8* %arg1, i8* %arg2, i8* %tmp2, i8* %tmp3, i8* %tmp4, i8* %tmp5) nounwind optsize noredzone + ret i8* %call +} + +declare i8* @x3(i8*, i8*, i8*, ...) optsize noredzone + +; X64: @foo7 +; X64: callq +; WIN64: @foo7 +; WIN64: callq +define i8* @foo7(i8* %arg1, i8* %arg2) nounwind optsize ssp noredzone { +entry: + %tmp2 = load i8** @sel3, align 8, !tbaa !0 + %tmp3 = load i8** @sel4, align 8, !tbaa !0 + %tmp4 = load i8** @sel5, align 8, !tbaa !0 + %tmp5 = load i8** @sel6, align 8, !tbaa !0 + %tmp6 = load i8** @sel7, align 8, !tbaa !0 + %call = tail call i8* (i8*, i8*, i8*, i8*, i8*, i8*, i8*, ...)* @x7(i8* %arg1, i8* %arg2, i8* %tmp2, i8* %tmp3, i8* %tmp4, i8* %tmp5, i8* %tmp6) nounwind optsize noredzone + ret i8* %call +} + +declare i8* @x7(i8*, i8*, i8*, i8*, i8*, i8*, i8*, ...) optsize noredzone + +; X64: @foo8 +; X64: callq +; WIN64: @foo8 +; WIN64: callq +define i8* @foo8(i8* %arg1, i8* %arg2) nounwind optsize ssp noredzone { +entry: + %tmp2 = load i8** @sel3, align 8, !tbaa !0 + %tmp3 = load i8** @sel4, align 8, !tbaa !0 + %tmp4 = load i8** @sel5, align 8, !tbaa !0 + %tmp5 = load i8** @sel6, align 8, !tbaa !0 + %call = tail call i8* (i8*, i8*, i8*, ...)* @x3(i8* %arg1, i8* %arg2, i8* %tmp2, i8* %tmp3, i8* %tmp4, i8* %tmp5, i32 48879, i32 48879) nounwind optsize noredzone + ret i8* %call +} + +!0 = metadata !{metadata !"any pointer", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA", null} diff --git a/test/CodeGen/X86/vec_extract-sse4.ll b/test/CodeGen/X86/vec_extract-sse4.ll index dab5dd1..f487654 100644 --- a/test/CodeGen/X86/vec_extract-sse4.ll +++ b/test/CodeGen/X86/vec_extract-sse4.ll @@ -1,8 +1,8 @@ -; RUN: llc < %s -march=x86 -mattr=+sse41 -o %t -; RUN: grep extractps %t | count 1 -; RUN: grep pextrd %t | count 1 +; RUN: llc < %s -mcpu=corei7 -march=x86 -mattr=+sse41 -o %t +; RUN: not grep extractps %t +; RUN: not grep pextrd %t ; RUN: not grep pshufd %t -; RUN: not grep movss %t +; RUN: grep movss %t | count 2 define void @t1(float* %R, <4 x float>* %P1) nounwind { %X = load <4 x float>* %P1 diff --git a/test/CodeGen/X86/vec_extract.ll b/test/CodeGen/X86/vec_extract.ll index b013730..2c8796b 100644 --- a/test/CodeGen/X86/vec_extract.ll +++ b/test/CodeGen/X86/vec_extract.ll @@ -1,7 +1,7 @@ -; RUN: llc < %s -march=x86 -mattr=+sse2,-sse41 -o %t -; RUN: grep movss %t | count 3 +; RUN: llc < %s -mcpu=corei7 -march=x86 -mattr=+sse2,-sse41 -o %t +; RUN: grep movss %t | count 4 ; RUN: grep movhlps %t | count 1 -; RUN: grep pshufd %t | count 1 +; RUN: not grep pshufd %t ; RUN: grep unpckhpd %t | count 1 define void @test1(<4 x float>* %F, float* %f) nounwind { diff --git a/test/CodeGen/X86/vec_shuffle-16.ll b/test/CodeGen/X86/vec_shuffle-16.ll index 2ee87fe..06f38ed 100644 --- a/test/CodeGen/X86/vec_shuffle-16.ll +++ b/test/CodeGen/X86/vec_shuffle-16.ll @@ -1,8 +1,9 @@ ; RUN: llc < %s -march=x86 -mattr=+sse,-sse2 -mtriple=i386-apple-darwin | FileCheck %s -check-prefix=sse ; RUN: llc < %s -march=x86 -mattr=+sse2 -mtriple=i386-apple-darwin | FileCheck %s -check-prefix=sse2 +; sse: t1: +; sse2: t1: define <4 x float> @t1(<4 x float> %a, <4 x float> %b) nounwind { -; sse: movaps ; sse: shufps ; sse2: pshufd ; sse2-NEXT: ret @@ -10,6 +11,8 @@ define <4 x float> @t1(<4 x float> %a, <4 x float> %b) nounwind { ret <4 x float> %tmp1 } +; sse: t2: +; sse2: t2: define <4 x float> @t2(<4 x float> %A, <4 x float> %B) nounwind { ; sse: shufps ; sse2: pshufd @@ -18,8 +21,9 @@ define <4 x float> @t2(<4 x float> %A, <4 x float> %B) nounwind { ret <4 x float> %tmp } +; sse: t3: +; sse2: t3: define <4 x float> @t3(<4 x float> %A, <4 x float> %B) nounwind { -; sse: movaps ; sse: shufps ; sse2: pshufd ; sse2-NEXT: ret @@ -27,7 +31,10 @@ define <4 x float> @t3(<4 x float> %A, <4 x float> %B) nounwind { ret <4 x float> %tmp } +; sse: t4: +; sse2: t4: define <4 x float> @t4(<4 x float> %A, <4 x float> %B) nounwind { + ; sse: shufps ; sse2: pshufd ; sse2-NEXT: ret diff --git a/test/CodeGen/X86/vec_uint_to_fp.ll b/test/CodeGen/X86/vec_uint_to_fp.ll index 39e7d71..fe7fa2f 100644 --- a/test/CodeGen/X86/vec_uint_to_fp.ll +++ b/test/CodeGen/X86/vec_uint_to_fp.ll @@ -1,4 +1,4 @@ -; RUN: llc < %s -march=x86 -mcpu=sandybridge | FileCheck %s +; RUN: llc < %s -march=x86 -mcpu=corei7-avx | FileCheck %s ; Test that we are not lowering uinttofp to scalars define <4 x float> @test1(<4 x i32> %A) nounwind { diff --git a/test/CodeGen/X86/visibility2.ll b/test/CodeGen/X86/visibility2.ll new file mode 100644 index 0000000..72ea733 --- /dev/null +++ b/test/CodeGen/X86/visibility2.ll @@ -0,0 +1,18 @@ +; This test case ensures that when the visibility of a global declaration is +; emitted they are not treated as definitions. Test case for r132825. +; Fixes <rdar://problem/9429892>. +; +; RUN: llc -mtriple=x86_64-apple-darwin %s -o - | FileCheck %s + +@foo_private_extern_str = external hidden global i8* + +define void @foo1() nounwind ssp { +entry: + %tmp = load i8** @foo_private_extern_str, align 8 + call void @foo3(i8* %tmp) + ret void +} + +declare void @foo3(i8*) + +; CHECK-NOT: .private_extern diff --git a/test/CodeGen/X86/widen_load-0.ll b/test/CodeGen/X86/widen_load-0.ll index 82c8252..c91627c 100644 --- a/test/CodeGen/X86/widen_load-0.ll +++ b/test/CodeGen/X86/widen_load-0.ll @@ -4,15 +4,15 @@ ; Both loads should happen before either store. -; CHECK: movl (%rdi), %eax -; CHECK: movl (%rsi), %ecx -; CHECK: movl %ecx, (%rdi) -; CHECK: movl %eax, (%rsi) +; CHECK: movl (%rdi), %[[R1:...]] +; CHECK: movl (%rsi), %[[R2:...]] +; CHECK: movl %[[R2]], (%rdi) +; CHECK: movl %[[R1]], (%rsi) -; WIN64: movl (%rcx), %eax -; WIN64: movl (%rdx), %esi -; WIN64: movl %esi, (%rcx) -; WIN64: movl %eax, (%rdx) +; WIN64: movl (%rcx), %[[R1:...]] +; WIN64: movl (%rdx), %[[R2:...]] +; WIN64: movl %[[R2]], (%rcx) +; WIN64: movl %[[R1]], (%rdx) define void @short2_int_swap(<2 x i16>* nocapture %b, i32* nocapture %c) nounwind { entry: diff --git a/test/CodeGen/X86/win64_alloca_dynalloca.ll b/test/CodeGen/X86/win64_alloca_dynalloca.ll index cbd38da..e39d007 100644 --- a/test/CodeGen/X86/win64_alloca_dynalloca.ll +++ b/test/CodeGen/X86/win64_alloca_dynalloca.ll @@ -1,9 +1,12 @@ -; RUN: llc < %s -mtriple=x86_64-mingw32 | FileCheck %s -check-prefix=M64 -; RUN: llc < %s -mtriple=x86_64-win32 | FileCheck %s -check-prefix=W64 -; RUN: llc < %s -mtriple=x86_64-win32-macho | FileCheck %s -check-prefix=EFI +; RUN: llc < %s -join-physregs -mtriple=x86_64-mingw32 | FileCheck %s -check-prefix=M64 +; RUN: llc < %s -join-physregs -mtriple=x86_64-win32 | FileCheck %s -check-prefix=W64 +; RUN: llc < %s -join-physregs -mtriple=x86_64-win32-macho | FileCheck %s -check-prefix=EFI ; PR8777 ; PR8778 +; Passing the same value in two registers creates a false interference that +; only -join-physregs resolves. It could also be handled by a parallel copy. + define i64 @foo(i64 %n, i64 %x) nounwind { entry: @@ -40,9 +43,9 @@ entry: ; W64: subq %rax, %rsp ; W64: movq %rsp, %rax -; EFI: leaq 15(%rcx), [[R1:%r..]] +; EFI: leaq 15(%rcx), [[R1:%r.*]] ; EFI: andq $-16, [[R1]] -; EFI: movq %rsp, [[R64:%r..]] +; EFI: movq %rsp, [[R64:%r.*]] ; EFI: subq [[R1]], [[R64]] ; EFI: movq [[R64]], %rsp diff --git a/test/CodeGen/X86/x86-64-and-mask.ll b/test/CodeGen/X86/x86-64-and-mask.ll index 2465f23..07ccb23 100644 --- a/test/CodeGen/X86/x86-64-and-mask.ll +++ b/test/CodeGen/X86/x86-64-and-mask.ll @@ -39,7 +39,7 @@ define void @ccc(i64 %x) nounwind { ; This requires a mov and a 64-bit and. ; CHECK: ddd: -; CHECK: movabsq $4294967296, %rax +; CHECK: movabsq $4294967296, %r ; CHECK: andq %rax, %rdi define void @ddd(i64 %x) nounwind { diff --git a/test/CodeGen/X86/x86-64-extend-shift.ll b/test/CodeGen/X86/x86-64-extend-shift.ll index 6852785..6ebaeee 100644 --- a/test/CodeGen/X86/x86-64-extend-shift.ll +++ b/test/CodeGen/X86/x86-64-extend-shift.ll @@ -2,7 +2,7 @@ ; Formerly there were two shifts. define i64 @baz(i32 %A) nounwind { -; CHECK: shlq $49, %rax +; CHECK: shlq $49, %r %tmp1 = shl i32 %A, 17 %tmp2 = zext i32 %tmp1 to i64 %tmp3 = shl i64 %tmp2, 32 diff --git a/test/CodeGen/X86/x86-shifts.ll b/test/CodeGen/X86/x86-shifts.ll new file mode 100644 index 0000000..fdf68f9 --- /dev/null +++ b/test/CodeGen/X86/x86-shifts.ll @@ -0,0 +1,142 @@ +; RUN: llc < %s -march=x86-64 -mcpu=corei7 | FileCheck %s + +; Splat patterns below + + +define <4 x i32> @shl4(<4 x i32> %A) nounwind { +entry: +; CHECK: shl4 +; CHECK: pslld +; CHECK-NEXT: pslld + %B = shl <4 x i32> %A, < i32 2, i32 2, i32 2, i32 2> + %C = shl <4 x i32> %A, < i32 1, i32 1, i32 1, i32 1> + %K = xor <4 x i32> %B, %C + ret <4 x i32> %K +} + +define <4 x i32> @shr4(<4 x i32> %A) nounwind { +entry: +; CHECK: shr4 +; CHECK: psrld +; CHECK-NEXT: psrld + %B = lshr <4 x i32> %A, < i32 2, i32 2, i32 2, i32 2> + %C = lshr <4 x i32> %A, < i32 1, i32 1, i32 1, i32 1> + %K = xor <4 x i32> %B, %C + ret <4 x i32> %K +} + +define <4 x i32> @sra4(<4 x i32> %A) nounwind { +entry: +; CHECK: sra4 +; CHECK: psrad +; CHECK-NEXT: psrad + %B = ashr <4 x i32> %A, < i32 2, i32 2, i32 2, i32 2> + %C = ashr <4 x i32> %A, < i32 1, i32 1, i32 1, i32 1> + %K = xor <4 x i32> %B, %C + ret <4 x i32> %K +} + +define <2 x i64> @shl2(<2 x i64> %A) nounwind { +entry: +; CHECK: shl2 +; CHECK: psllq +; CHECK-NEXT: psllq + %B = shl <2 x i64> %A, < i64 2, i64 2> + %C = shl <2 x i64> %A, < i64 9, i64 9> + %K = xor <2 x i64> %B, %C + ret <2 x i64> %K +} + +define <2 x i64> @shr2(<2 x i64> %A) nounwind { +entry: +; CHECK: shr2 +; CHECK: psrlq +; CHECK-NEXT: psrlq + %B = lshr <2 x i64> %A, < i64 8, i64 8> + %C = lshr <2 x i64> %A, < i64 1, i64 1> + %K = xor <2 x i64> %B, %C + ret <2 x i64> %K +} + + +define <8 x i16> @shl8(<8 x i16> %A) nounwind { +entry: +; CHECK: shl8 +; CHECK: psllw +; CHECK-NEXT: psllw + %B = shl <8 x i16> %A, < i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2> + %C = shl <8 x i16> %A, < i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> + %K = xor <8 x i16> %B, %C + ret <8 x i16> %K +} + +define <8 x i16> @shr8(<8 x i16> %A) nounwind { +entry: +; CHECK: shr8 +; CHECK: psrlw +; CHECK-NEXT: psrlw + %B = lshr <8 x i16> %A, < i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2> + %C = lshr <8 x i16> %A, < i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> + %K = xor <8 x i16> %B, %C + ret <8 x i16> %K +} + +define <8 x i16> @sra8(<8 x i16> %A) nounwind { +entry: +; CHECK: sra8 +; CHECK: psraw +; CHECK-NEXT: psraw + %B = ashr <8 x i16> %A, < i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2> + %C = ashr <8 x i16> %A, < i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> + %K = xor <8 x i16> %B, %C + ret <8 x i16> %K +} + +; non splat test + + +define <8 x i16> @sll8_nosplat(<8 x i16> %A) nounwind { +entry: +; CHECK: sll8_nosplat +; CHECK-NOT: psll +; CHECK-NOT: psll + %B = shl <8 x i16> %A, < i16 1, i16 2, i16 3, i16 6, i16 2, i16 2, i16 2, i16 2> + %C = shl <8 x i16> %A, < i16 9, i16 7, i16 5, i16 1, i16 4, i16 1, i16 1, i16 1> + %K = xor <8 x i16> %B, %C + ret <8 x i16> %K +} + + +define <2 x i64> @shr2_nosplat(<2 x i64> %A) nounwind { +entry: +; CHECK: shr2_nosplat +; CHECK-NOT: psrlq +; CHECK-NOT: psrlq + %B = lshr <2 x i64> %A, < i64 8, i64 1> + %C = lshr <2 x i64> %A, < i64 1, i64 0> + %K = xor <2 x i64> %B, %C + ret <2 x i64> %K +} + + +; Other shifts + +define <2 x i32> @shl2_other(<2 x i32> %A) nounwind { +entry: +; CHECK: shl2_other +; CHECK-not: psllq + %B = shl <2 x i32> %A, < i32 2, i32 2> + %C = shl <2 x i32> %A, < i32 9, i32 9> + %K = xor <2 x i32> %B, %C + ret <2 x i32> %K +} + +define <2 x i32> @shr2_other(<2 x i32> %A) nounwind { +entry: +; CHECK: shr2_other +; CHECK-NOT: psrlq + %B = lshr <2 x i32> %A, < i32 8, i32 8> + %C = lshr <2 x i32> %A, < i32 1, i32 1> + %K = xor <2 x i32> %B, %C + ret <2 x i32> %K +} diff --git a/test/CodeGen/X86/xor.ll b/test/CodeGen/X86/xor.ll index b90d81a..178c59d 100644 --- a/test/CodeGen/X86/xor.ll +++ b/test/CodeGen/X86/xor.ll @@ -29,9 +29,8 @@ entry: ret i32 %tmp4 ; X64: test3: -; X64: notl [[A1:%esi|%edx]] -; X64: andl [[A0:%edi|%ecx]], [[A1]] -; X64: movl [[A1]], %eax +; X64: notl +; X64: andl ; X64: shrl %eax ; X64: ret @@ -139,7 +138,7 @@ entry: %t2 = add i32 %t1, -1 ret i32 %t2 ; X64: test8: -; X64: notl %eax +; X64: notl {{%eax|%edi|%ecx}} ; X32: test8: ; X32: notl %eax } |
