; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc %s -o - -enable-shrink-wrap=true -pass-remarks-output=%t | FileCheck %s --check-prefix=ENABLE ; RUN: cat %t | FileCheck %s --check-prefix=REMARKS ; RUN: llc %s -o - -enable-shrink-wrap=false | FileCheck %s --check-prefix=DISABLE ; ; Note: Lots of tests use inline asm instead of regular calls. ; This allows to have a better control on what the allocation will do. ; Otherwise, we may have spill right in the entry block, defeating ; shrink-wrapping. Moreover, some of the inline asm statement (nop) ; are here to ensure that the related paths do not end up as critical ; edges. target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128" target triple = "x86_64-apple-macosx" ; Initial motivating example: Simple diamond with a call just on one side. define i32 @foo(i32 %a, i32 %b) { ; ENABLE-LABEL: foo: ; ENABLE: ## %bb.0: ; ENABLE-NEXT: movl %edi, %eax ; ENABLE-NEXT: cmpl %esi, %edi ; ENABLE-NEXT: jge LBB0_2 ; ENABLE-NEXT: ## %bb.1: ## %true ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: movl %eax, {{[0-9]+}}(%rsp) ; ENABLE-NEXT: leaq {{[0-9]+}}(%rsp), %rsi ; ENABLE-NEXT: xorl %edi, %edi ; ENABLE-NEXT: callq _doSomething ; ENABLE-NEXT: addq $8, %rsp ; ENABLE-NEXT: LBB0_2: ## %false ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: foo: ; DISABLE: ## %bb.0: ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: movl %edi, %eax ; DISABLE-NEXT: cmpl %esi, %edi ; DISABLE-NEXT: jge LBB0_2 ; DISABLE-NEXT: ## %bb.1: ## %true ; DISABLE-NEXT: movl %eax, {{[0-9]+}}(%rsp) ; DISABLE-NEXT: leaq {{[0-9]+}}(%rsp), %rsi ; DISABLE-NEXT: xorl %edi, %edi ; DISABLE-NEXT: callq _doSomething ; DISABLE-NEXT: LBB0_2: ## %false ; DISABLE-NEXT: popq %rcx ; DISABLE-NEXT: retq %tmp = alloca i32, align 4 %tmp2 = icmp slt i32 %a, %b br i1 %tmp2, label %true, label %false true: store i32 %a, ptr %tmp, align 4 %tmp4 = call i32 @doSomething(i32 0, ptr %tmp) br label %false false: %tmp.0 = phi i32 [ %tmp4, %true ], [ %a, %0 ] ret i32 %tmp.0 } ; Function Attrs: optsize declare i32 @doSomething(i32, ptr) ; Check that we do not perform the restore inside the loop whereas the save ; is outside. define i32 @freqSaveAndRestoreOutsideLoop(i32 %cond, i32 %N) { ; ENABLE-LABEL: freqSaveAndRestoreOutsideLoop: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: testl %edi, %edi ; ENABLE-NEXT: je LBB1_4 ; ENABLE-NEXT: ## %bb.1: ## %for.preheader ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbx, -16 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: movl $10, %ecx ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB1_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: addl %edx, %eax ; ENABLE-NEXT: decl %ecx ; ENABLE-NEXT: jne LBB1_2 ; ENABLE-NEXT: ## %bb.3: ## %for.end ; ENABLE-NEXT: shll $3, %eax ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB1_4: ## %if.else ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: addl %esi, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: freqSaveAndRestoreOutsideLoop: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbx, -16 ; DISABLE-NEXT: testl %edi, %edi ; DISABLE-NEXT: je LBB1_4 ; DISABLE-NEXT: ## %bb.1: ## %for.preheader ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: movl $10, %ecx ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB1_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: addl %edx, %eax ; DISABLE-NEXT: decl %ecx ; DISABLE-NEXT: jne LBB1_2 ; DISABLE-NEXT: ## %bb.3: ## %for.end ; DISABLE-NEXT: shll $3, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB1_4: ## %if.else ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: addl %esi, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: %tobool = icmp eq i32 %cond, 0 br i1 %tobool, label %if.else, label %for.preheader for.preheader: tail call void asm "nop", ""() br label %for.body for.body: ; preds = %entry, %for.body %i.05 = phi i32 [ %inc, %for.body ], [ 0, %for.preheader ] %sum.04 = phi i32 [ %add, %for.body ], [ 0, %for.preheader ] %call = tail call i32 asm sideeffect "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.04 %inc = add nuw nsw i32 %i.05, 1 %exitcond = icmp eq i32 %inc, 10 br i1 %exitcond, label %for.end, label %for.body for.end: ; preds = %for.body %shl = shl i32 %add, 3 br label %if.end if.else: ; preds = %entry %mul = shl nsw i32 %N, 1 br label %if.end if.end: ; preds = %if.else, %for.end %sum.1 = phi i32 [ %shl, %for.end ], [ %mul, %if.else ] ret i32 %sum.1 } declare i32 @something(...) ; Check that we do not perform the shrink-wrapping inside the loop even ; though that would be legal. The cost model must prevent that. define i32 @freqSaveAndRestoreOutsideLoop2(i32 %cond) { ; ENABLE-LABEL: freqSaveAndRestoreOutsideLoop2: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbx, -16 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: movl $10, %ecx ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB2_1: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: addl %edx, %eax ; ENABLE-NEXT: decl %ecx ; ENABLE-NEXT: jne LBB2_1 ; ENABLE-NEXT: ## %bb.2: ## %for.exit ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: freqSaveAndRestoreOutsideLoop2: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbx, -16 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: movl $10, %ecx ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB2_1: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: addl %edx, %eax ; DISABLE-NEXT: decl %ecx ; DISABLE-NEXT: jne LBB2_1 ; DISABLE-NEXT: ## %bb.2: ## %for.exit ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: br label %for.preheader for.preheader: tail call void asm "nop", ""() br label %for.body for.body: ; preds = %for.body, %entry %i.04 = phi i32 [ 0, %for.preheader ], [ %inc, %for.body ] %sum.03 = phi i32 [ 0, %for.preheader ], [ %add, %for.body ] %call = tail call i32 asm sideeffect "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.03 %inc = add nuw nsw i32 %i.04, 1 %exitcond = icmp eq i32 %inc, 10 br i1 %exitcond, label %for.exit, label %for.body for.exit: tail call void asm "nop", ""() br label %for.end for.end: ; preds = %for.body ret i32 %add } ; Check with a more complex case that we do not have save within the loop and ; restore outside. define i32 @loopInfoSaveOutsideLoop(i32 %cond, i32 %N) { ; ENABLE-LABEL: loopInfoSaveOutsideLoop: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: testl %edi, %edi ; ENABLE-NEXT: je LBB3_4 ; ENABLE-NEXT: ## %bb.1: ## %for.preheader ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbx, -16 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: movl $10, %ecx ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB3_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: addl %edx, %eax ; ENABLE-NEXT: decl %ecx ; ENABLE-NEXT: jne LBB3_2 ; ENABLE-NEXT: ## %bb.3: ## %for.end ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: shll $3, %eax ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB3_4: ## %if.else ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: addl %esi, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: loopInfoSaveOutsideLoop: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbx, -16 ; DISABLE-NEXT: testl %edi, %edi ; DISABLE-NEXT: je LBB3_4 ; DISABLE-NEXT: ## %bb.1: ## %for.preheader ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: movl $10, %ecx ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB3_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: addl %edx, %eax ; DISABLE-NEXT: decl %ecx ; DISABLE-NEXT: jne LBB3_2 ; DISABLE-NEXT: ## %bb.3: ## %for.end ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: shll $3, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB3_4: ## %if.else ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: addl %esi, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: %tobool = icmp eq i32 %cond, 0 br i1 %tobool, label %if.else, label %for.preheader for.preheader: tail call void asm "nop", ""() br label %for.body for.body: ; preds = %entry, %for.body %i.05 = phi i32 [ %inc, %for.body ], [ 0, %for.preheader ] %sum.04 = phi i32 [ %add, %for.body ], [ 0, %for.preheader ] %call = tail call i32 asm sideeffect "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.04 %inc = add nuw nsw i32 %i.05, 1 %exitcond = icmp eq i32 %inc, 10 br i1 %exitcond, label %for.end, label %for.body for.end: ; preds = %for.body tail call void asm "nop", "~{ebx}"() %shl = shl i32 %add, 3 br label %if.end if.else: ; preds = %entry %mul = shl nsw i32 %N, 1 br label %if.end if.end: ; preds = %if.else, %for.end %sum.1 = phi i32 [ %shl, %for.end ], [ %mul, %if.else ] ret i32 %sum.1 } ; Check with a more complex case that we do not have restore within the loop and ; save outside. define i32 @loopInfoRestoreOutsideLoop(i32 %cond, i32 %N) nounwind { ; ENABLE-LABEL: loopInfoRestoreOutsideLoop: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: testl %edi, %edi ; ENABLE-NEXT: je LBB4_4 ; ENABLE-NEXT: ## %bb.1: ## %if.then ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: movl $10, %ecx ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB4_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: addl %edx, %eax ; ENABLE-NEXT: decl %ecx ; ENABLE-NEXT: jne LBB4_2 ; ENABLE-NEXT: ## %bb.3: ## %for.end ; ENABLE-NEXT: shll $3, %eax ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB4_4: ## %if.else ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: addl %esi, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: loopInfoRestoreOutsideLoop: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: testl %edi, %edi ; DISABLE-NEXT: je LBB4_4 ; DISABLE-NEXT: ## %bb.1: ## %if.then ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: movl $10, %ecx ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB4_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: addl %edx, %eax ; DISABLE-NEXT: decl %ecx ; DISABLE-NEXT: jne LBB4_2 ; DISABLE-NEXT: ## %bb.3: ## %for.end ; DISABLE-NEXT: shll $3, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB4_4: ## %if.else ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: addl %esi, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: %tobool = icmp eq i32 %cond, 0 br i1 %tobool, label %if.else, label %if.then if.then: ; preds = %entry tail call void asm "nop", "~{ebx}"() br label %for.body for.body: ; preds = %for.body, %if.then %i.05 = phi i32 [ 0, %if.then ], [ %inc, %for.body ] %sum.04 = phi i32 [ 0, %if.then ], [ %add, %for.body ] %call = tail call i32 asm sideeffect "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.04 %inc = add nuw nsw i32 %i.05, 1 %exitcond = icmp eq i32 %inc, 10 br i1 %exitcond, label %for.end, label %for.body for.end: ; preds = %for.body %shl = shl i32 %add, 3 br label %if.end if.else: ; preds = %entry %mul = shl nsw i32 %N, 1 br label %if.end if.end: ; preds = %if.else, %for.end %sum.1 = phi i32 [ %shl, %for.end ], [ %mul, %if.else ] ret i32 %sum.1 } ; Check that we handle function with no frame information correctly. define i32 @emptyFrame() { ; ENABLE-LABEL: emptyFrame: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: emptyFrame: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: retq entry: ret i32 0 } ; Check that we handle inline asm correctly. define i32 @inlineAsm(i32 %cond, i32 %N) { ; ENABLE-LABEL: inlineAsm: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: testl %edi, %edi ; ENABLE-NEXT: je LBB6_4 ; ENABLE-NEXT: ## %bb.1: ## %for.preheader ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbx, -16 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: movl $10, %eax ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB6_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: addl $1, %ebx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: decl %eax ; ENABLE-NEXT: jne LBB6_2 ; ENABLE-NEXT: ## %bb.3: ## %for.exit ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB6_4: ## %if.else ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: addl %esi, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: inlineAsm: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbx, -16 ; DISABLE-NEXT: testl %edi, %edi ; DISABLE-NEXT: je LBB6_4 ; DISABLE-NEXT: ## %bb.1: ## %for.preheader ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: movl $10, %eax ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB6_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: addl $1, %ebx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: decl %eax ; DISABLE-NEXT: jne LBB6_2 ; DISABLE-NEXT: ## %bb.3: ## %for.exit ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB6_4: ## %if.else ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: addl %esi, %eax ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: %tobool = icmp eq i32 %cond, 0 br i1 %tobool, label %if.else, label %for.preheader for.preheader: tail call void asm "nop", ""() br label %for.body for.body: ; preds = %entry, %for.body %i.03 = phi i32 [ %inc, %for.body ], [ 0, %for.preheader ] tail call void asm "addl $$1, %ebx", "~{ebx}"() %inc = add nuw nsw i32 %i.03, 1 %exitcond = icmp eq i32 %inc, 10 br i1 %exitcond, label %for.exit, label %for.body for.exit: tail call void asm "nop", ""() br label %if.end if.else: ; preds = %entry %mul = shl nsw i32 %N, 1 br label %if.end if.end: ; preds = %for.body, %if.else %sum.0 = phi i32 [ %mul, %if.else ], [ 0, %for.exit ] ret i32 %sum.0 } ; Check that we handle calls to variadic functions correctly. define i32 @callVariadicFunc(i32 %cond, i32 %N) { ; ENABLE-LABEL: callVariadicFunc: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: testl %edi, %edi ; ENABLE-NEXT: je LBB7_2 ; ENABLE-NEXT: ## %bb.1: ## %if.then ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: movl %eax, (%rsp) ; ENABLE-NEXT: movl %eax, %edi ; ENABLE-NEXT: movl %eax, %esi ; ENABLE-NEXT: movl %eax, %edx ; ENABLE-NEXT: movl %eax, %ecx ; ENABLE-NEXT: movl %eax, %r8d ; ENABLE-NEXT: movl %eax, %r9d ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: callq _someVariadicFunc ; ENABLE-NEXT: shll $3, %eax ; ENABLE-NEXT: addq $8, %rsp ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB7_2: ## %if.else ; ENABLE-NEXT: addl %eax, %eax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: callVariadicFunc: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: testl %edi, %edi ; DISABLE-NEXT: je LBB7_2 ; DISABLE-NEXT: ## %bb.1: ## %if.then ; DISABLE-NEXT: movl %eax, (%rsp) ; DISABLE-NEXT: movl %eax, %edi ; DISABLE-NEXT: movl %eax, %esi ; DISABLE-NEXT: movl %eax, %edx ; DISABLE-NEXT: movl %eax, %ecx ; DISABLE-NEXT: movl %eax, %r8d ; DISABLE-NEXT: movl %eax, %r9d ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: callq _someVariadicFunc ; DISABLE-NEXT: shll $3, %eax ; DISABLE-NEXT: popq %rcx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB7_2: ## %if.else ; DISABLE-NEXT: addl %eax, %eax ; DISABLE-NEXT: popq %rcx ; DISABLE-NEXT: retq entry: %tobool = icmp eq i32 %cond, 0 br i1 %tobool, label %if.else, label %if.then if.then: ; preds = %entry %call = tail call i32 (i32, ...) @someVariadicFunc(i32 %N, i32 %N, i32 %N, i32 %N, i32 %N, i32 %N, i32 %N) %shl = shl i32 %call, 3 br label %if.end if.else: ; preds = %entry %mul = shl nsw i32 %N, 1 br label %if.end if.end: ; preds = %if.else, %if.then %sum.0 = phi i32 [ %shl, %if.then ], [ %mul, %if.else ] ret i32 %sum.0 } declare i32 @someVariadicFunc(i32, ...) ; Check that we use LEA not to clobber EFLAGS. %struct.temp_slot = type { ptr, ptr, ptr, i32, i64, ptr, ptr, i8, i8, i32, i32, i64, i64 } %union.tree_node = type { %struct.tree_decl } %struct.tree_decl = type { %struct.tree_common, ptr, i32, i32, ptr, i48, %union.anon, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, %union.anon.1, ptr, ptr, ptr, i64, ptr } %struct.tree_common = type { ptr, ptr, i32 } %union.anon = type { i64 } %union.anon.1 = type { ptr } %struct.function = type { ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, i32, i32, i32, i32, ptr, %struct.ix86_args, ptr, ptr, ptr, ptr, i32, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, ptr, i64, ptr, ptr, ptr, ptr, i32, ptr, ptr, i32, i32, i32, ptr, i32, i32, ptr, ptr, ptr, i32, i32, ptr, i32, i32, ptr, ptr, i24 } %struct.eh_status = type opaque %struct.stmt_status = type opaque %struct.expr_status = type { i32, i32, i32, ptr, ptr, ptr, ptr } %struct.emit_status = type { i32, i32, ptr, ptr, ptr, ptr, i32, i32, ptr, i32, ptr, ptr, ptr } %struct.sequence_stack = type { ptr, ptr, ptr, ptr } %struct.varasm_status = type opaque %struct.ix86_args = type { i32, i32, i32, i32, i32, i32, i32 } %struct.initial_value_struct = type opaque %struct.var_refs_queue = type { ptr, i32, i32, ptr } %struct.machine_function = type opaque %struct.language_function = type opaque %struct.lang_decl = type opaque %struct.rtx_def = type { i32, [1 x %union.rtunion_def] } %union.rtunion_def = type { i64 } declare hidden fastcc ptr @find_temp_slot_from_address(ptr readonly) define void @useLEA(ptr readonly %x) { ; ENABLE-LABEL: useLEA: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: testq %rdi, %rdi ; ENABLE-NEXT: je LBB8_7 ; ENABLE-NEXT: ## %bb.1: ## %if.end ; ENABLE-NEXT: cmpw $66, (%rdi) ; ENABLE-NEXT: jne LBB8_7 ; ENABLE-NEXT: ## %bb.2: ## %lor.lhs.false ; ENABLE-NEXT: movq 8(%rdi), %rdi ; ENABLE-NEXT: movzwl (%rdi), %eax ; ENABLE-NEXT: leal -54(%rax), %ecx ; ENABLE-NEXT: cmpl $14, %ecx ; ENABLE-NEXT: ja LBB8_3 ; ENABLE-NEXT: ## %bb.8: ## %lor.lhs.false ; ENABLE-NEXT: movl $24599, %edx ## imm = 0x6017 ; ENABLE-NEXT: btl %ecx, %edx ; ENABLE-NEXT: jae LBB8_3 ; ENABLE-NEXT: LBB8_7: ## %cleanup ; ENABLE-NEXT: popq %rax ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB8_3: ## %lor.lhs.false ; ENABLE-NEXT: cmpl $134, %eax ; ENABLE-NEXT: je LBB8_7 ; ENABLE-NEXT: ## %bb.4: ## %lor.lhs.false ; ENABLE-NEXT: cmpl $140, %eax ; ENABLE-NEXT: je LBB8_7 ; ENABLE-NEXT: ## %bb.5: ## %if.end.55 ; ENABLE-NEXT: callq _find_temp_slot_from_address ; ENABLE-NEXT: testq %rax, %rax ; ENABLE-NEXT: je LBB8_7 ; ENABLE-NEXT: ## %bb.6: ## %if.then.60 ; ENABLE-NEXT: movb $1, 57(%rax) ; ENABLE-NEXT: popq %rax ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: useLEA: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: testq %rdi, %rdi ; DISABLE-NEXT: je LBB8_7 ; DISABLE-NEXT: ## %bb.1: ## %if.end ; DISABLE-NEXT: cmpw $66, (%rdi) ; DISABLE-NEXT: jne LBB8_7 ; DISABLE-NEXT: ## %bb.2: ## %lor.lhs.false ; DISABLE-NEXT: movq 8(%rdi), %rdi ; DISABLE-NEXT: movzwl (%rdi), %eax ; DISABLE-NEXT: leal -54(%rax), %ecx ; DISABLE-NEXT: cmpl $14, %ecx ; DISABLE-NEXT: ja LBB8_3 ; DISABLE-NEXT: ## %bb.8: ## %lor.lhs.false ; DISABLE-NEXT: movl $24599, %edx ## imm = 0x6017 ; DISABLE-NEXT: btl %ecx, %edx ; DISABLE-NEXT: jae LBB8_3 ; DISABLE-NEXT: LBB8_7: ## %cleanup ; DISABLE-NEXT: popq %rax ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB8_3: ## %lor.lhs.false ; DISABLE-NEXT: cmpl $134, %eax ; DISABLE-NEXT: je LBB8_7 ; DISABLE-NEXT: ## %bb.4: ## %lor.lhs.false ; DISABLE-NEXT: cmpl $140, %eax ; DISABLE-NEXT: je LBB8_7 ; DISABLE-NEXT: ## %bb.5: ## %if.end.55 ; DISABLE-NEXT: callq _find_temp_slot_from_address ; DISABLE-NEXT: testq %rax, %rax ; DISABLE-NEXT: je LBB8_7 ; DISABLE-NEXT: ## %bb.6: ## %if.then.60 ; DISABLE-NEXT: movb $1, 57(%rax) ; DISABLE-NEXT: popq %rax ; DISABLE-NEXT: retq entry: %cmp = icmp eq ptr %x, null br i1 %cmp, label %cleanup, label %if.end if.end: ; preds = %entry %bf.load = load i32, ptr %x, align 8 %bf.clear = and i32 %bf.load, 65535 %cmp1 = icmp eq i32 %bf.clear, 66 br i1 %cmp1, label %lor.lhs.false, label %cleanup lor.lhs.false: ; preds = %if.end %arrayidx = getelementptr inbounds %struct.rtx_def, ptr %x, i64 0, i32 1, i64 0 %tmp1 = load ptr, ptr %arrayidx, align 8 %bf.load2 = load i32, ptr %tmp1, align 8 %bf.clear3 = and i32 %bf.load2, 65535 switch i32 %bf.clear3, label %if.end.55 [ i32 67, label %cleanup i32 68, label %cleanup i32 54, label %cleanup i32 55, label %cleanup i32 58, label %cleanup i32 134, label %cleanup i32 56, label %cleanup i32 140, label %cleanup ] if.end.55: ; preds = %lor.lhs.false %call = tail call fastcc ptr @find_temp_slot_from_address(ptr %tmp1) #2 %cmp59 = icmp eq ptr %call, null br i1 %cmp59, label %cleanup, label %if.then.60 if.then.60: ; preds = %if.end.55 %addr_taken = getelementptr inbounds %struct.temp_slot, ptr %call, i64 0, i32 8 store i8 1, ptr %addr_taken, align 1 br label %cleanup cleanup: ; preds = %if.then.60, %if.end.55, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %lor.lhs.false, %if.end, %entry ret void } ; Make sure we do not insert unreachable code after noreturn function. ; Although this is not incorrect to insert such code, it is useless ; and it hurts the binary size. define i32 @noreturn(i8 signext %bad_thing) { ; ENABLE-LABEL: noreturn: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: testb %dil, %dil ; ENABLE-NEXT: jne LBB9_2 ; ENABLE-NEXT: ## %bb.1: ## %if.end ; ENABLE-NEXT: movl $42, %eax ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB9_2: ## %if.abort ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: callq _abort ; ; DISABLE-LABEL: noreturn: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: testb %dil, %dil ; DISABLE-NEXT: jne LBB9_2 ; DISABLE-NEXT: ## %bb.1: ## %if.end ; DISABLE-NEXT: movl $42, %eax ; DISABLE-NEXT: popq %rcx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB9_2: ## %if.abort ; DISABLE-NEXT: callq _abort entry: %tobool = icmp eq i8 %bad_thing, 0 br i1 %tobool, label %if.end, label %if.abort if.abort: tail call void @abort() #0 unreachable if.end: ret i32 42 } declare void @abort() #0 attributes #0 = { noreturn nounwind } ; Make sure that we handle infinite loops properly When checking that the Save ; and Restore blocks are control flow equivalent, the loop searches for the ; immediate (post) dominator for the (restore) save blocks. When either the Save ; or Restore block is located in an infinite loop the only immediate (post) ; dominator is itself. In this case, we cannot perform shrink wrapping, but we ; should return gracefully and continue compilation. ; The only condition for this test is the compilation finishes correctly. ; define void @infiniteloop() { ; ENABLE-LABEL: infiniteloop: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rbp ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbp, -16 ; ENABLE-NEXT: movq %rsp, %rbp ; ENABLE-NEXT: .cfi_def_cfa_register %rbp ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_offset %rbx, -24 ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: testb %al, %al ; ENABLE-NEXT: jne LBB10_3 ; ENABLE-NEXT: ## %bb.1: ## %if.then ; ENABLE-NEXT: movq %rsp, %rcx ; ENABLE-NEXT: addq $-16, %rcx ; ENABLE-NEXT: movq %rcx, %rsp ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB10_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: addl %edx, %eax ; ENABLE-NEXT: movl %eax, (%rcx) ; ENABLE-NEXT: jmp LBB10_2 ; ENABLE-NEXT: LBB10_3: ## %if.end ; ENABLE-NEXT: leaq -8(%rbp), %rsp ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: popq %rbp ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: infiniteloop: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbp ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbp, -16 ; DISABLE-NEXT: movq %rsp, %rbp ; DISABLE-NEXT: .cfi_def_cfa_register %rbp ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_offset %rbx, -24 ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: testb %al, %al ; DISABLE-NEXT: jne LBB10_3 ; DISABLE-NEXT: ## %bb.1: ## %if.then ; DISABLE-NEXT: movq %rsp, %rcx ; DISABLE-NEXT: addq $-16, %rcx ; DISABLE-NEXT: movq %rcx, %rsp ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB10_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: addl %edx, %eax ; DISABLE-NEXT: movl %eax, (%rcx) ; DISABLE-NEXT: jmp LBB10_2 ; DISABLE-NEXT: LBB10_3: ## %if.end ; DISABLE-NEXT: leaq -8(%rbp), %rsp ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: popq %rbp ; DISABLE-NEXT: retq entry: br i1 undef, label %if.then, label %if.end if.then: %ptr = alloca i32, i32 4 br label %for.body for.body: ; preds = %for.body, %entry %sum.03 = phi i32 [ 0, %if.then ], [ %add, %for.body ] %call = tail call i32 asm "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.03 store i32 %add, ptr %ptr br label %for.body if.end: ret void } ; Another infinite loop test this time with a body bigger than just one block. define void @infiniteloop2() { ; ENABLE-LABEL: infiniteloop2: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rbp ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbp, -16 ; ENABLE-NEXT: movq %rsp, %rbp ; ENABLE-NEXT: .cfi_def_cfa_register %rbp ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_offset %rbx, -24 ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: testb %al, %al ; ENABLE-NEXT: jne LBB11_5 ; ENABLE-NEXT: ## %bb.1: ## %if.then ; ENABLE-NEXT: movq %rsp, %rcx ; ENABLE-NEXT: addq $-16, %rcx ; ENABLE-NEXT: movq %rcx, %rsp ; ENABLE-NEXT: xorl %edx, %edx ; ENABLE-NEXT: jmp LBB11_2 ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB11_4: ## %body2 ; ENABLE-NEXT: ## in Loop: Header=BB11_2 Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: LBB11_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: movl %edx, %esi ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: movl $1, %edx ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: addl %esi, %edx ; ENABLE-NEXT: movl %edx, (%rcx) ; ENABLE-NEXT: testb %al, %al ; ENABLE-NEXT: jne LBB11_4 ; ENABLE-NEXT: ## %bb.3: ## %body1 ; ENABLE-NEXT: ## in Loop: Header=BB11_2 Depth=1 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: jmp LBB11_2 ; ENABLE-NEXT: LBB11_5: ## %if.end ; ENABLE-NEXT: leaq -8(%rbp), %rsp ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: popq %rbp ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: infiniteloop2: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbp ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbp, -16 ; DISABLE-NEXT: movq %rsp, %rbp ; DISABLE-NEXT: .cfi_def_cfa_register %rbp ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_offset %rbx, -24 ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: testb %al, %al ; DISABLE-NEXT: jne LBB11_5 ; DISABLE-NEXT: ## %bb.1: ## %if.then ; DISABLE-NEXT: movq %rsp, %rcx ; DISABLE-NEXT: addq $-16, %rcx ; DISABLE-NEXT: movq %rcx, %rsp ; DISABLE-NEXT: xorl %edx, %edx ; DISABLE-NEXT: jmp LBB11_2 ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB11_4: ## %body2 ; DISABLE-NEXT: ## in Loop: Header=BB11_2 Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: LBB11_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: movl %edx, %esi ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: movl $1, %edx ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: addl %esi, %edx ; DISABLE-NEXT: movl %edx, (%rcx) ; DISABLE-NEXT: testb %al, %al ; DISABLE-NEXT: jne LBB11_4 ; DISABLE-NEXT: ## %bb.3: ## %body1 ; DISABLE-NEXT: ## in Loop: Header=BB11_2 Depth=1 ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: jmp LBB11_2 ; DISABLE-NEXT: LBB11_5: ## %if.end ; DISABLE-NEXT: leaq -8(%rbp), %rsp ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: popq %rbp ; DISABLE-NEXT: retq entry: br i1 undef, label %if.then, label %if.end if.then: %ptr = alloca i32, i32 4 br label %for.body for.body: ; preds = %for.body, %entry %sum.03 = phi i32 [ 0, %if.then ], [ %add, %body1 ], [ 1, %body2] %call = tail call i32 asm "movl $$1, $0", "=r,~{ebx}"() %add = add nsw i32 %call, %sum.03 store i32 %add, ptr %ptr br i1 undef, label %body1, label %body2 body1: tail call void asm sideeffect "nop", "~{ebx}"() br label %for.body body2: tail call void asm sideeffect "nop", "~{ebx}"() br label %for.body if.end: ret void } ; Another infinite loop test this time with two nested infinite loop. define void @infiniteloop3() { ; ENABLE-LABEL: infiniteloop3: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: testb %al, %al ; ENABLE-NEXT: jne LBB12_2 ; ENABLE-NEXT: ## %bb.1: ## %body ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: testb %al, %al ; ENABLE-NEXT: jne LBB12_7 ; ENABLE-NEXT: LBB12_2: ## %loop2a.preheader ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: xorl %ecx, %ecx ; ENABLE-NEXT: movq %rax, %rsi ; ENABLE-NEXT: jmp LBB12_4 ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB12_3: ## %loop2b ; ENABLE-NEXT: ## in Loop: Header=BB12_4 Depth=1 ; ENABLE-NEXT: movq %rdx, (%rsi) ; ENABLE-NEXT: movq %rdx, %rsi ; ENABLE-NEXT: LBB12_4: ## %loop1 ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: movq %rcx, %rdx ; ENABLE-NEXT: testq %rax, %rax ; ENABLE-NEXT: movq (%rax), %rcx ; ENABLE-NEXT: jne LBB12_3 ; ENABLE-NEXT: ## %bb.5: ## in Loop: Header=BB12_4 Depth=1 ; ENABLE-NEXT: movq %rdx, %rax ; ENABLE-NEXT: movq %rdx, %rsi ; ENABLE-NEXT: jmp LBB12_4 ; ENABLE-NEXT: LBB12_7: ## %end ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: infiniteloop3: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: testb %al, %al ; DISABLE-NEXT: jne LBB12_2 ; DISABLE-NEXT: ## %bb.1: ## %body ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: testb %al, %al ; DISABLE-NEXT: jne LBB12_7 ; DISABLE-NEXT: LBB12_2: ## %loop2a.preheader ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: xorl %ecx, %ecx ; DISABLE-NEXT: movq %rax, %rsi ; DISABLE-NEXT: jmp LBB12_4 ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB12_3: ## %loop2b ; DISABLE-NEXT: ## in Loop: Header=BB12_4 Depth=1 ; DISABLE-NEXT: movq %rdx, (%rsi) ; DISABLE-NEXT: movq %rdx, %rsi ; DISABLE-NEXT: LBB12_4: ## %loop1 ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: movq %rcx, %rdx ; DISABLE-NEXT: testq %rax, %rax ; DISABLE-NEXT: movq (%rax), %rcx ; DISABLE-NEXT: jne LBB12_3 ; DISABLE-NEXT: ## %bb.5: ## in Loop: Header=BB12_4 Depth=1 ; DISABLE-NEXT: movq %rdx, %rax ; DISABLE-NEXT: movq %rdx, %rsi ; DISABLE-NEXT: jmp LBB12_4 ; DISABLE-NEXT: LBB12_7: ## %end ; DISABLE-NEXT: retq entry: br i1 undef, label %loop2a, label %body body: ; preds = %entry br i1 undef, label %loop2a, label %end loop1: ; preds = %loop2a, %loop2b %var.phi = phi ptr [ %next.phi, %loop2b ], [ %var, %loop2a ] %next.phi = phi ptr [ %next.load, %loop2b ], [ %next.var, %loop2a ] %0 = icmp eq ptr %var, null %next.load = load ptr, ptr undef br i1 %0, label %loop2a, label %loop2b loop2a: ; preds = %loop1, %body, %entry %var = phi ptr [ null, %body ], [ null, %entry ], [ %next.phi, %loop1 ] %next.var = phi ptr [ undef, %body ], [ null, %entry ], [ %next.load, %loop1 ] br label %loop1 loop2b: ; preds = %loop1 store ptr %next.phi, ptr %var.phi br label %loop1 end: ret void } ; Check that we just don't bail out on RegMask. ; In this case, the RegMask does not touch a CSR so we are good to go! define i32 @regmask(i32 %a, i32 %b, ptr %addr) { ; ENABLE-LABEL: regmask: ; ENABLE: ## %bb.0: ; ENABLE-NEXT: cmpl %esi, %edi ; ENABLE-NEXT: jge LBB13_2 ; ENABLE-NEXT: ## %bb.1: ## %true ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbx, -16 ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: xorl %edi, %edi ; ENABLE-NEXT: movq %rdx, %rsi ; ENABLE-NEXT: callq _doSomething ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB13_2: ## %false ; ENABLE-NEXT: movl $6, %edi ; ENABLE-NEXT: movq %rdx, %rsi ; ENABLE-NEXT: jmp _doSomething ## TAILCALL ; ; DISABLE-LABEL: regmask: ; DISABLE: ## %bb.0: ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbx, -16 ; DISABLE-NEXT: cmpl %esi, %edi ; DISABLE-NEXT: jge LBB13_2 ; DISABLE-NEXT: ## %bb.1: ## %true ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: xorl %edi, %edi ; DISABLE-NEXT: movq %rdx, %rsi ; DISABLE-NEXT: callq _doSomething ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB13_2: ## %false ; DISABLE-NEXT: movl $6, %edi ; DISABLE-NEXT: movq %rdx, %rsi ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: jmp _doSomething ## TAILCALL %tmp2 = icmp slt i32 %a, %b br i1 %tmp2, label %true, label %false true: ; Clobber a CSR so that we check something on the regmask ; of the tail call. tail call void asm sideeffect "nop", "~{ebx}"() %tmp4 = call i32 @doSomething(i32 0, ptr %addr) br label %end false: %tmp5 = tail call i32 @doSomething(i32 6, ptr %addr) br label %end end: %tmp.0 = phi i32 [ %tmp4, %true ], [ %tmp5, %false ] ret i32 %tmp.0 } @b = internal unnamed_addr global i1 false @c = internal unnamed_addr global i8 0, align 1 @a = common global i32 0, align 4 ; Make sure the prologue does not clobber the EFLAGS when ; it is live accross. ; PR25629. ; Note: The registers may change in the following patterns, but ; because they imply register hierarchy (e.g., eax, al) this is ; tricky to write robust patterns. define i32 @useLEAForPrologue(i32 %d, i32 %a, i8 %c) #3 { ; ENABLE-LABEL: useLEAForPrologue: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: subq $16, %rsp ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: cmpb $0, _b(%rip) ; ENABLE-NEXT: movl $48, %ecx ; ENABLE-NEXT: cmovnel %eax, %ecx ; ENABLE-NEXT: movb %cl, _c(%rip) ; ENABLE-NEXT: je LBB14_4 ; ENABLE-NEXT: ## %bb.1: ## %for.body.lr.ph ; ENABLE-NEXT: ## InlineAsm Start ; ENABLE-NEXT: nop ; ENABLE-NEXT: ## InlineAsm End ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB14_2: ## %for.body ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: cmpl %esi, %edi ; ENABLE-NEXT: setl %al ; ENABLE-NEXT: xorl %esi, %esi ; ENABLE-NEXT: movb %al, %sil ; ENABLE-NEXT: incb %dl ; ENABLE-NEXT: cmpb $45, %dl ; ENABLE-NEXT: jl LBB14_2 ; ENABLE-NEXT: ## %bb.3: ## %for.cond.for.end_crit_edge ; ENABLE-NEXT: movq _a@GOTPCREL(%rip), %rax ; ENABLE-NEXT: movl %esi, (%rax) ; ENABLE-NEXT: LBB14_4: ## %for.end ; ENABLE-NEXT: xorl %edi, %edi ; ENABLE-NEXT: callq _varfunc ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: addq $16, %rsp ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: useLEAForPrologue: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: subq $16, %rsp ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: cmpb $0, _b(%rip) ; DISABLE-NEXT: movl $48, %ecx ; DISABLE-NEXT: cmovnel %eax, %ecx ; DISABLE-NEXT: movb %cl, _c(%rip) ; DISABLE-NEXT: je LBB14_4 ; DISABLE-NEXT: ## %bb.1: ## %for.body.lr.ph ; DISABLE-NEXT: ## InlineAsm Start ; DISABLE-NEXT: nop ; DISABLE-NEXT: ## InlineAsm End ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB14_2: ## %for.body ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: cmpl %esi, %edi ; DISABLE-NEXT: setl %al ; DISABLE-NEXT: xorl %esi, %esi ; DISABLE-NEXT: movb %al, %sil ; DISABLE-NEXT: incb %dl ; DISABLE-NEXT: cmpb $45, %dl ; DISABLE-NEXT: jl LBB14_2 ; DISABLE-NEXT: ## %bb.3: ## %for.cond.for.end_crit_edge ; DISABLE-NEXT: movq _a@GOTPCREL(%rip), %rax ; DISABLE-NEXT: movl %esi, (%rax) ; DISABLE-NEXT: LBB14_4: ## %for.end ; DISABLE-NEXT: xorl %edi, %edi ; DISABLE-NEXT: callq _varfunc ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: addq $16, %rsp ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: retq entry: %tmp = alloca i3 %.b = load i1, ptr @b, align 1 %bool = select i1 %.b, i8 0, i8 48 store i8 %bool, ptr @c, align 1 br i1 %.b, label %for.body.lr.ph, label %for.end for.body.lr.ph: ; preds = %entry tail call void asm sideeffect "nop", "~{ebx}"() br label %for.body for.body: ; preds = %for.body.lr.ph, %for.body %inc6 = phi i8 [ %c, %for.body.lr.ph ], [ %inc, %for.body ] %cond5 = phi i32 [ %a, %for.body.lr.ph ], [ %conv3, %for.body ] %cmp2 = icmp slt i32 %d, %cond5 %conv3 = zext i1 %cmp2 to i32 %inc = add i8 %inc6, 1 %cmp = icmp slt i8 %inc, 45 br i1 %cmp, label %for.body, label %for.cond.for.end_crit_edge for.cond.for.end_crit_edge: ; preds = %for.body store i32 %conv3, ptr @a, align 4 br label %for.end for.end: ; preds = %for.cond.for.end_crit_edge, %entry %call = tail call i32 (ptr) @varfunc(ptr null) ret i32 0 } declare i32 @varfunc(ptr nocapture readonly) @sum1 = external hidden thread_local global i32, align 4 ; Function Attrs: nounwind ; Make sure the TLS call used to access @sum1 happens after the prologue ; and before the epilogue. ; TLS calls used to be wrongly model and shrink-wrapping would have inserted ; the prologue and epilogue just around the call to doSomething. ; PR25820. define i32 @tlsCall(i1 %bool1, i32 %arg, ptr readonly dereferenceable(4) %sum1) #3 { ; ENABLE-LABEL: tlsCall: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: testb $1, %dil ; ENABLE-NEXT: je LBB15_2 ; ENABLE-NEXT: ## %bb.1: ## %main ; ENABLE-NEXT: movl (%rdx), %ecx ; ENABLE-NEXT: movq _sum1@TLVP(%rip), %rdi ; ENABLE-NEXT: callq *(%rdi) ; ENABLE-NEXT: movl %ecx, (%rax) ; ENABLE-NEXT: jmp LBB15_3 ; ENABLE-NEXT: LBB15_2: ## %else ; ENABLE-NEXT: xorl %edi, %edi ; ENABLE-NEXT: xorl %esi, %esi ; ENABLE-NEXT: callq _doSomething ; ENABLE-NEXT: movl %eax, %esi ; ENABLE-NEXT: LBB15_3: ## %exit ; ENABLE-NEXT: movl %esi, %eax ; ENABLE-NEXT: popq %rcx ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: tlsCall: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: testb $1, %dil ; DISABLE-NEXT: je LBB15_2 ; DISABLE-NEXT: ## %bb.1: ## %main ; DISABLE-NEXT: movl (%rdx), %ecx ; DISABLE-NEXT: movq _sum1@TLVP(%rip), %rdi ; DISABLE-NEXT: callq *(%rdi) ; DISABLE-NEXT: movl %ecx, (%rax) ; DISABLE-NEXT: jmp LBB15_3 ; DISABLE-NEXT: LBB15_2: ## %else ; DISABLE-NEXT: xorl %edi, %edi ; DISABLE-NEXT: xorl %esi, %esi ; DISABLE-NEXT: callq _doSomething ; DISABLE-NEXT: movl %eax, %esi ; DISABLE-NEXT: LBB15_3: ## %exit ; DISABLE-NEXT: movl %esi, %eax ; DISABLE-NEXT: popq %rcx ; DISABLE-NEXT: retq entry: br i1 %bool1, label %main, label %else main: %tmp1 = load i32, ptr %sum1, align 4 store i32 %tmp1, ptr @sum1, align 4 br label %exit else: %call = call i32 @doSomething(i32 0, ptr null) br label %exit exit: %res = phi i32 [ %arg, %main], [ %call, %else ] ret i32 %res } attributes #3 = { nounwind } @irreducibleCFGa = common global i32 0, align 4 @irreducibleCFGf = common global i8 0, align 1 @irreducibleCFGb = common global i32 0, align 4 ; Check that we do not run shrink-wrapping on irreducible CFGs until ; it is actually supported. ; At the moment, on those CFGs the loop information may be incorrect ; and since we use that information to do the placement, we may end up ; inserting the prologue/epilogue at incorrect places. ; PR25988. ; Make sure we emit missed optimization remarks for this. ; REMARKS: Pass: shrink-wrap ; REMARKS-NEXT: Name: UnsupportedIrreducibleCFG ; REMARKS-NEXT: Function: irreducibleCFG ; REMARKS-NEXT: Args: ; REMARKS-NEXT: - String: Irreducible CFGs are not supported yet define i32 @irreducibleCFG() #4 { ; ENABLE-LABEL: irreducibleCFG: ; ENABLE: ## %bb.0: ## %entry ; ENABLE-NEXT: pushq %rbp ; ENABLE-NEXT: .cfi_def_cfa_offset 16 ; ENABLE-NEXT: .cfi_offset %rbp, -16 ; ENABLE-NEXT: movq %rsp, %rbp ; ENABLE-NEXT: .cfi_def_cfa_register %rbp ; ENABLE-NEXT: pushq %rbx ; ENABLE-NEXT: pushq %rax ; ENABLE-NEXT: .cfi_offset %rbx, -24 ; ENABLE-NEXT: movq _irreducibleCFGf@GOTPCREL(%rip), %rax ; ENABLE-NEXT: cmpb $0, (%rax) ; ENABLE-NEXT: je LBB16_2 ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB16_1: ## %preheader ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: jmp LBB16_1 ; ENABLE-NEXT: LBB16_2: ## %split ; ENABLE-NEXT: movq _irreducibleCFGb@GOTPCREL(%rip), %rax ; ENABLE-NEXT: cmpl $0, (%rax) ; ENABLE-NEXT: je LBB16_3 ; ENABLE-NEXT: ## %bb.4: ## %for.body4.i ; ENABLE-NEXT: movq _irreducibleCFGa@GOTPCREL(%rip), %rax ; ENABLE-NEXT: movl (%rax), %edi ; ENABLE-NEXT: xorl %ebx, %ebx ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: callq _something ; ENABLE-NEXT: jmp LBB16_5 ; ENABLE-NEXT: LBB16_3: ; ENABLE-NEXT: xorl %ebx, %ebx ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB16_5: ## %for.inc ; ENABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: incl %ebx ; ENABLE-NEXT: cmpl $7, %ebx ; ENABLE-NEXT: jl LBB16_5 ; ENABLE-NEXT: ## %bb.6: ## %fn1.exit ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: addq $8, %rsp ; ENABLE-NEXT: popq %rbx ; ENABLE-NEXT: popq %rbp ; ENABLE-NEXT: retq ; ; DISABLE-LABEL: irreducibleCFG: ; DISABLE: ## %bb.0: ## %entry ; DISABLE-NEXT: pushq %rbp ; DISABLE-NEXT: .cfi_def_cfa_offset 16 ; DISABLE-NEXT: .cfi_offset %rbp, -16 ; DISABLE-NEXT: movq %rsp, %rbp ; DISABLE-NEXT: .cfi_def_cfa_register %rbp ; DISABLE-NEXT: pushq %rbx ; DISABLE-NEXT: pushq %rax ; DISABLE-NEXT: .cfi_offset %rbx, -24 ; DISABLE-NEXT: movq _irreducibleCFGf@GOTPCREL(%rip), %rax ; DISABLE-NEXT: cmpb $0, (%rax) ; DISABLE-NEXT: je LBB16_2 ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB16_1: ## %preheader ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: jmp LBB16_1 ; DISABLE-NEXT: LBB16_2: ## %split ; DISABLE-NEXT: movq _irreducibleCFGb@GOTPCREL(%rip), %rax ; DISABLE-NEXT: cmpl $0, (%rax) ; DISABLE-NEXT: je LBB16_3 ; DISABLE-NEXT: ## %bb.4: ## %for.body4.i ; DISABLE-NEXT: movq _irreducibleCFGa@GOTPCREL(%rip), %rax ; DISABLE-NEXT: movl (%rax), %edi ; DISABLE-NEXT: xorl %ebx, %ebx ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: callq _something ; DISABLE-NEXT: jmp LBB16_5 ; DISABLE-NEXT: LBB16_3: ; DISABLE-NEXT: xorl %ebx, %ebx ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB16_5: ## %for.inc ; DISABLE-NEXT: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: incl %ebx ; DISABLE-NEXT: cmpl $7, %ebx ; DISABLE-NEXT: jl LBB16_5 ; DISABLE-NEXT: ## %bb.6: ## %fn1.exit ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: addq $8, %rsp ; DISABLE-NEXT: popq %rbx ; DISABLE-NEXT: popq %rbp ; DISABLE-NEXT: retq entry: %i0 = load i32, ptr @irreducibleCFGa, align 4 %.pr = load i8, ptr @irreducibleCFGf, align 1 %bool = icmp eq i8 %.pr, 0 br i1 %bool, label %split, label %preheader preheader: br label %preheader split: %i1 = load i32, ptr @irreducibleCFGb, align 4 %tobool1.i = icmp ne i32 %i1, 0 br i1 %tobool1.i, label %for.body4.i, label %for.cond8.i.preheader for.body4.i: %call.i = tail call i32 (...) @something(i32 %i0) br label %for.cond8 for.cond8: %p1 = phi i32 [ %inc18.i, %for.inc ], [ 0, %for.body4.i ] %.pr1.pr = load i32, ptr @irreducibleCFGb, align 4 br label %for.cond8.i.preheader for.cond8.i.preheader: %.pr1 = phi i32 [ %.pr1.pr, %for.cond8 ], [ %i1, %split ] %p13 = phi i32 [ %p1, %for.cond8 ], [ 0, %split ] br label %for.inc fn1.exit: ret i32 0 for.inc: %inc18.i = add nuw nsw i32 %p13, 1 %cmp = icmp slt i32 %inc18.i, 7 br i1 %cmp, label %for.cond8, label %fn1.exit } attributes #4 = { "frame-pointer"="all" } @x = external global i32, align 4 @y = external global i32, align 4 ; The post-dominator tree does not include the branch containing the infinite ; loop, which can occur into a misplacement of the restore block, if we're ; looking for the nearest common post-dominator of an "unreachable" block. define void @infiniteLoopNoSuccessor() #5 { ; ENABLE-LABEL: infiniteLoopNoSuccessor: ; ENABLE: ## %bb.0: ; ENABLE-NEXT: pushq %rbp ; ENABLE-NEXT: movq %rsp, %rbp ; ENABLE-NEXT: movq _x@GOTPCREL(%rip), %rax ; ENABLE-NEXT: cmpl $0, (%rax) ; ENABLE-NEXT: je LBB17_2 ; ENABLE-NEXT: ## %bb.1: ; ENABLE-NEXT: movl $0, (%rax) ; ENABLE-NEXT: LBB17_2: ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: callq _somethingElse ; ENABLE-NEXT: movq _y@GOTPCREL(%rip), %rax ; ENABLE-NEXT: cmpl $0, (%rax) ; ENABLE-NEXT: je LBB17_3 ; ENABLE-NEXT: ## %bb.5: ; ENABLE-NEXT: popq %rbp ; ENABLE-NEXT: retq ; ENABLE-NEXT: LBB17_3: ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: callq _something ; ENABLE-NEXT: .p2align 4, 0x90 ; ENABLE-NEXT: LBB17_4: ## =>This Inner Loop Header: Depth=1 ; ENABLE-NEXT: xorl %eax, %eax ; ENABLE-NEXT: callq _somethingElse ; ENABLE-NEXT: jmp LBB17_4 ; ; DISABLE-LABEL: infiniteLoopNoSuccessor: ; DISABLE: ## %bb.0: ; DISABLE-NEXT: pushq %rbp ; DISABLE-NEXT: movq %rsp, %rbp ; DISABLE-NEXT: movq _x@GOTPCREL(%rip), %rax ; DISABLE-NEXT: cmpl $0, (%rax) ; DISABLE-NEXT: je LBB17_2 ; DISABLE-NEXT: ## %bb.1: ; DISABLE-NEXT: movl $0, (%rax) ; DISABLE-NEXT: LBB17_2: ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: callq _somethingElse ; DISABLE-NEXT: movq _y@GOTPCREL(%rip), %rax ; DISABLE-NEXT: cmpl $0, (%rax) ; DISABLE-NEXT: je LBB17_3 ; DISABLE-NEXT: ## %bb.5: ; DISABLE-NEXT: popq %rbp ; DISABLE-NEXT: retq ; DISABLE-NEXT: LBB17_3: ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: callq _something ; DISABLE-NEXT: .p2align 4, 0x90 ; DISABLE-NEXT: LBB17_4: ## =>This Inner Loop Header: Depth=1 ; DISABLE-NEXT: xorl %eax, %eax ; DISABLE-NEXT: callq _somethingElse ; DISABLE-NEXT: jmp LBB17_4 %1 = load i32, ptr @x, align 4 %2 = icmp ne i32 %1, 0 br i1 %2, label %3, label %4 ; <label>:3: store i32 0, ptr @x, align 4 br label %4 ; <label>:4: call void (...) @somethingElse() %5 = load i32, ptr @y, align 4 %6 = icmp ne i32 %5, 0 br i1 %6, label %10, label %7 ; <label>:7: %8 = call i32 (...) @something() br label %9 ; <label>:9: call void (...) @somethingElse() br label %9 ; <label>:10: ret void } declare void @somethingElse(...) attributes #5 = { nounwind "frame-pointer"="non-leaf" }