[InstCombine] add multi-use demanded bits fold for add with low-bit mask

[spatel]

Summary

I noticed an add example like the one from D91343, so here's a similar patch.
The logic is based on existing code for the single-use demanded bits fold. But I only matched a constant instead of using compute known bits on the operands because that was the motivating pattern.

I think this will allow removing a special-case (but incomplete) dedicated fold within visitAnd(), but I need to untangle the existing code to be sure.

https://rise4fun.com/Alive/V6fP

Name: add with low mask 
Pre: (C1 & (-1 u>> countLeadingZeros(C2))) == 0
%a = add i8 %x, C1
%r = and i8 %a, C2
=>
%r = and i8 %x, C2

[RKSimon]

llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp

838

We could technically support non-uniform vectors here by matching m_Constant and using a similar ConstantExpr fold.

llvm/test/Transforms/InstCombine/and.ll

1052

Vector tests?

[spatel]

llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp

838

Ok - I didn't see any signs that we had tried non-uniform in this file yet. Do you think we should do that in this patch or as a TODO/follow-up?

llvm/test/Transforms/InstCombine/and.ll

1052

Basic splat is the last (positive) test.

[RKSimon]

LGTM with a few minor comments

llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp

838

I'm happy for a TODO for now - but ideally we'd start getting as many combines supporting non-uniform constants (inc undef elements) as soon as possible.

Unfortunately I don't think llvm::MaskedValueIsZero will work for undef elements.

llvm/test/Transforms/InstCombine/and.ll

1052

cheers - if you could add the non-uniform test coverage now, even if support isn't added yet that'd be great.

[vitalybuka]

There is some uninitialized memory read http://lab.llvm.org:8011/#/builders/74/builds/888/steps/9/logs/stdio

FAIL: Clang :: SemaCXX/builtin-is-constant-evaluated.cpp (16913 of 71556)
******************** TEST 'Clang :: SemaCXX/builtin-is-constant-evaluated.cpp' FAILED ********************
Script:
--
: 'RUN: at line 1';   /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm_build_msan/bin/clang -cc1 -internal-isystem /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm_build_msan/lib/clang/12.0.0/include -nostdsysteminc -std=c++2a -verify /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/test/SemaCXX/builtin-is-constant-evaluated.cpp -fcxx-exceptions -Wno-constant-evaluated -triple=x86_64-linux-gnu
--
Exit Code: 77
Command Output (stderr):
--
==14417==WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0x127ea7d4 in (anonymous namespace)::FloatExprEvaluator::VisitCastExpr(clang::CastExpr const*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:13686:3
    #1 0x127d7282 in clang::StmtVisitorBase<llvm::make_const_ptr, (anonymous namespace)::FloatExprEvaluator, bool>::Visit(clang::Stmt const*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/include/clang/AST/StmtVisitor.h
    #2 0x126f42c1 in EvaluateFloat /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:13509:43
    #3 0x126f42c1 in Evaluate(clang::APValue&, (anonymous namespace)::EvalInfo&, clang::Expr const*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:14440:10
    #4 0x126e239e in EvaluateInPlace(clang::APValue&, (anonymous namespace)::EvalInfo&, (anonymous namespace)::LValue const&, clang::Expr const*, bool) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:14525:10
    #5 0x126e4d4a in clang::Expr::EvaluateAsInitializer(clang::APValue&, clang::ASTContext const&, clang::VarDecl const*, llvm::SmallVectorImpl<std::__1::pair<clang::SourceLocation, clang::PartialDiagnostic> >&, bool) const /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:14820:10
    #6 0x124e0c6d in clang::VarDecl::evaluateValueImpl(llvm::SmallVectorImpl<std::__1::pair<clang::SourceLocation, clang::PartialDiagnostic> >&, bool) const /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/Decl.cpp:2414:23
    #7 0x124e12b2 in clang::VarDecl::checkForConstantInitialization(llvm::SmallVectorImpl<std::__1::pair<clang::SourceLocation, clang::PartialDiagnostic> >&) const /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/Decl.cpp:2485:7
    #8 0x10603eb6 in clang::Sema::CheckCompleteVariableDeclaration(clang::VarDecl*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Sema/SemaDecl.cpp:13021:27
    #9 0x105fe616 in clang::Sema::AddInitializerToDecl(clang::Decl*, clang::Expr*, bool) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Sema/SemaDecl.cpp:12406:3
    #10 0xfe892f6 in clang::Parser::ParseDeclarationAfterDeclaratorAndAttributes(clang::Declarator&, clang::Parser::ParsedTemplateInfo const&, clang::Parser::ForRangeInit*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseDecl.cpp:2311:17
    #11 0xfe819bd in clang::Parser::ParseDeclGroup(clang::ParsingDeclSpec&, clang::DeclaratorContext, clang::SourceLocation*, clang::Parser::ForRangeInit*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseDecl.cpp:2037:21
    #12 0xfdeed21 in clang::Parser::ParseDeclOrFunctionDefInternal(clang::Parser::ParsedAttributesWithRange&, clang::ParsingDeclSpec&, clang::AccessSpecifier) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/Parser.cpp:1134:10
    #13 0xfded513 in clang::Parser::ParseDeclarationOrFunctionDefinition(clang::Parser::ParsedAttributesWithRange&, clang::ParsingDeclSpec*, clang::AccessSpecifier) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/Parser.cpp:1150:12
    #14 0xfdea89b in clang::Parser::ParseExternalDeclaration(clang::Parser::ParsedAttributesWithRange&, clang::ParsingDeclSpec*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/Parser.cpp:970:12
    #15 0xfe0b864 in clang::Parser::ParseInnerNamespace(llvm::SmallVector<clang::Parser::InnerNamespaceInfo, 4u> const&, unsigned int, clang::SourceLocation&, clang::ParsedAttributes&, clang::BalancedDelimiterTracker&) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseDeclCXX.cpp:246:7
    #16 0xfe08823 in clang::Parser::ParseNamespace(clang::DeclaratorContext, clang::SourceLocation&, clang::SourceLocation) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseDeclCXX.cpp:224:3
    #17 0xfe70478 in clang::Parser::ParseDeclaration(clang::DeclaratorContext, clang::SourceLocation&, clang::Parser::ParsedAttributesWithRange&, clang::SourceLocation*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseDecl.cpp
    #18 0xfde94bb in clang::Parser::ParseExternalDeclaration(clang::Parser::ParsedAttributesWithRange&, clang::ParsingDeclSpec*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/Parser.cpp
    #19 0xfde5aec in clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<clang::DeclGroupRef>&, bool) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/Parser.cpp:716:12
    #20 0xfdd3c23 in clang::ParseAST(clang::Sema&, bool, bool) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Parse/ParseAST.cpp:158:20
    #21 0xbed5695 in clang::FrontendAction::Execute() /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Frontend/FrontendAction.cpp:949:8
    #22 0xbcf8803 in clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/Frontend/CompilerInstance.cpp:991:33
    #23 0xc15ec6a in clang::ExecuteCompilerInvocation(clang::CompilerInstance*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/FrontendTool/ExecuteCompilerInvocation.cpp:278:25
    #24 0x325cd45 in cc1_main(llvm::ArrayRef<char const*>, char const*, void*) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/tools/driver/cc1_main.cpp:240:15
    #25 0x3255076 in ExecuteCC1Tool(llvm::SmallVectorImpl<char const*>&) /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/tools/driver/driver.cpp:330:12
    #26 0x3253d3a in main /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/tools/driver/driver.cpp:407:12
    #27 0x7f2d7579f09a in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2409a)
    #28 0x31c8ef9 in _start (/b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm_build_msan/bin/clang-12+0x31c8ef9)
SUMMARY: MemorySanitizer: use-of-uninitialized-value /b/sanitizer-x86_64-linux-bootstrap-msan/build/llvm-project/clang/lib/AST/ExprConstant.cpp:13686:3 in (anonymous namespace)::FloatExprEvaluator::VisitCastExpr(clang::CastExpr const*)
Exiting

[spatel]

In D91415#2397086, @vitalybuka wrote:

There is some uninitialized memory read http://lab.llvm.org:8011/#/builders/74/builds/888/steps/9/logs/stdio

Is there anything I can re-use from that bot to reduce the repro burden? (I don't have a system set up to repro a 2-stage msan bug currently.)

[spatel]

I reverted to restore the bot. Getting nowhere fast trying to recreate that setup, so I'm not sure if it's worth the effort.
We already have a dedicated fold for the pattern shown in the tests, and it is artificially limited to single-use and scalar, so it's easier to just update that code.

llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp

Value *InstCombinerImpl::SimplifyMultipleUseDemandedBits(
   KnownBits LHSKnown(BitWidth);
   KnownBits RHSKnown(BitWidth);
 
   // Despite the fact that we can't simplify this instruction in all User's
   // context, we can at least compute the known bits, and we can
   // do simplifications that apply to *just* the one user if we know that
   // this instruction has a simpler value in that context.
   switch (I->getOpcode()) {
+  case Instruction::Add: {
+    const APInt *C;
+    if (match(I->getOperand(1), m_APInt(C))) {
+      // Right fill the demanded bits for this add to demand the most
+      // significant demanded bit and all those below it.
+      unsigned Ctlz = DemandedMask.countLeadingZeros();
+      APInt LowMask(APInt::getLowBitsSet(BitWidth, BitWidth - Ctlz));
+      // If we are adding zeros to every bit below the highest demanded bit,
+      // just return the add's variable operand.
+      if ((*C & LowMask).isNullValue())
+        return I->getOperand(0);
+    }
+    break;
+  }
   case Instruction::And: {
     // If either the LHS or the RHS are Zero, the result is zero.
     computeKnownBits(I->getOperand(1), RHSKnown, Depth + 1, CxtI);
     computeKnownBits(I->getOperand(0), LHSKnown, Depth + 1,
                      CxtI);
 
     Known = LHSKnown & RHSKnown;
 

llvm/test/Transforms/InstCombine/and.ll

 ;
   %l = shl <2 x i32> %x, <i32 20, i32 20>
   %r = ashr <2 x i32> %l, <i32 20, i32 20>
   store <2 x i32> %r, <2 x i32>* %p
   %and = and <2 x i32> %r, <i32 4095, i32 4095>
   ret <2 x i32> %and
 }
 
+; Multi-use demanded bits - 'add' doesn't change 'and'
+
 define i8 @lowmask_add(i8 %x) {
 ; CHECK-LABEL: @lowmask_add(
 ; CHECK-NEXT:    [[A:%.*]] = add i8 [[X:%.*]], -64
 ; CHECK-NEXT:    call void @use8(i8 [[A]])
-; CHECK-NEXT:    [[R:%.*]] = and i8 [[A]], 32
+; CHECK-NEXT:    [[R:%.*]] = and i8 [[X]], 32
 ; CHECK-NEXT:    ret i8 [[R]]
 ;
   %a = add i8 %x, -64 ; 0xc0
   call void @use8(i8 %a)
   %r = and i8 %a, 32 ; 0x20
   ret i8 %r
 }
 
+; Negative test - mask overlaps low bit of add
+
 define i8 @not_lowmask_add(i8 %x) {
 ; CHECK-LABEL: @not_lowmask_add(
 ; CHECK-NEXT:    [[A:%.*]] = add i8 [[X:%.*]], -64
 ; CHECK-NEXT:    call void @use8(i8 [[A]])
 ; CHECK-NEXT:    [[R:%.*]] = and i8 [[A]], 64
 ; CHECK-NEXT:    ret i8 [[R]]
 ;
   %a = add i8 %x, -64 ; 0xc0
   call void @use8(i8 %a)
   %r = and i8 %a, 64 ; 0x40
   ret i8 %r
 }
 
+; Negative test - mask overlaps low bit of add
+
 define i8 @not_lowmask_add2(i8 %x) {
 ; CHECK-LABEL: @not_lowmask_add2(
 ; CHECK-NEXT:    [[A:%.*]] = add i8 [[X:%.*]], -96
 ; CHECK-NEXT:    call void @use8(i8 [[A]])
 ; CHECK-NEXT:    [[R:%.*]] = and i8 [[A]], 63
 ; CHECK-NEXT:    ret i8 [[R]]
 ;
   %a = add i8 %x, -96 ; 0xe0
   call void @use8(i8 %a)
   %r = and i8 %a, 63 ; 0x3f
   ret i8 %r
 }
 
+; Multi-use demanded bits - 'add' doesn't change 'and'
+
 define <2 x i8> @lowmask_add_splat(<2 x i8> %x, <2 x i8>* %p) {
 ; CHECK-LABEL: @lowmask_add_splat(
 ; CHECK-NEXT:    [[A:%.*]] = add <2 x i8> [[X:%.*]], <i8 -64, i8 -64>
 ; CHECK-NEXT:    store <2 x i8> [[A]], <2 x i8>* [[P:%.*]], align 2
-; CHECK-NEXT:    [[R:%.*]] = and <2 x i8> [[A]], <i8 32, i8 32>
+; CHECK-NEXT:    [[R:%.*]] = and <2 x i8> [[X]], <i8 32, i8 32>
 ; CHECK-NEXT:    ret <2 x i8> [[R]]
 ;
   %a = add <2 x i8> %x, <i8 -64, i8 -64> ; 0xc0
   store <2 x i8> %a, <2 x i8>* %p
   %r = and <2 x i8> %a, <i8 32, i8 32> ; 0x20
   ret <2 x i8> %r
 }

llvm/test/Transforms/LoopVectorize/X86/float-induction-x86.ll

 ; AUTO_VEC-NEXT:    [[IND_END:%.*]] = fadd fast float [[TMP0]], 1.000000e+00
 ; AUTO_VEC-NEXT:    [[TMP1:%.*]] = add nsw i64 [[N_VEC]], -32
 ; AUTO_VEC-NEXT:    [[TMP2:%.*]] = lshr exact i64 [[TMP1]], 5
 ; AUTO_VEC-NEXT:    [[TMP3:%.*]] = add nuw nsw i64 [[TMP2]], 1
 ; AUTO_VEC-NEXT:    [[XTRAITER:%.*]] = and i64 [[TMP3]], 3
 ; AUTO_VEC-NEXT:    [[TMP4:%.*]] = icmp ult i64 [[TMP1]], 96
 ; AUTO_VEC-NEXT:    br i1 [[TMP4]], label [[MIDDLE_BLOCK_UNR_LCSSA:%.*]], label [[VECTOR_PH_NEW:%.*]]
 ; AUTO_VEC:       vector.ph.new:
-; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP3]], 1152921504606846972
+; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP3]], -4
 ; AUTO_VEC-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; AUTO_VEC:       vector.body:
 ; AUTO_VEC-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH_NEW]] ], [ [[INDEX_NEXT_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[VEC_IND:%.*]] = phi <8 x float> [ <float 1.000000e+00, float 1.500000e+00, float 2.000000e+00, float 2.500000e+00, float 3.000000e+00, float 3.500000e+00, float 4.000000e+00, float 4.500000e+00>, [[VECTOR_PH_NEW]] ], [ [[VEC_IND_NEXT_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[NITER:%.*]] = phi i64 [ [[UNROLL_ITER]], [[VECTOR_PH_NEW]] ], [ [[NITER_NSUB_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[STEP_ADD:%.*]] = fadd fast <8 x float> [[VEC_IND]], <float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00, float 4.000000e+00>
 ; AUTO_VEC-NEXT:    [[STEP_ADD2:%.*]] = fadd fast <8 x float> [[VEC_IND]], <float 8.000000e+00, float 8.000000e+00, float 8.000000e+00, float 8.000000e+00, float 8.000000e+00, float 8.000000e+00, float 8.000000e+00, float 8.000000e+00>
 ; AUTO_VEC-NEXT:    [[STEP_ADD3:%.*]] = fadd fast <8 x float> [[VEC_IND]], <float 1.200000e+01, float 1.200000e+01, float 1.200000e+01, float 1.200000e+01, float 1.200000e+01, float 1.200000e+01, float 1.200000e+01, float 1.200000e+01>

 ; AUTO_VEC-NEXT:    [[TMP1:%.*]] = fmul fast double [[CAST_CRD]], 3.000000e+00
 ; AUTO_VEC-NEXT:    [[TMP2:%.*]] = add nsw i64 [[N_VEC]], -16
 ; AUTO_VEC-NEXT:    [[TMP3:%.*]] = lshr exact i64 [[TMP2]], 4
 ; AUTO_VEC-NEXT:    [[TMP4:%.*]] = add nuw nsw i64 [[TMP3]], 1
 ; AUTO_VEC-NEXT:    [[XTRAITER:%.*]] = and i64 [[TMP4]], 3
 ; AUTO_VEC-NEXT:    [[TMP5:%.*]] = icmp ult i64 [[TMP2]], 48
 ; AUTO_VEC-NEXT:    br i1 [[TMP5]], label [[MIDDLE_BLOCK_UNR_LCSSA:%.*]], label [[VECTOR_PH_NEW:%.*]]
 ; AUTO_VEC:       vector.ph.new:
-; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP4]], 2305843009213693948
+; AUTO_VEC-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP4]], -4
 ; AUTO_VEC-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; AUTO_VEC:       vector.body:
 ; AUTO_VEC-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH_NEW]] ], [ [[INDEX_NEXT_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[VEC_IND:%.*]] = phi <4 x double> [ <double 0.000000e+00, double 3.000000e+00, double 6.000000e+00, double 9.000000e+00>, [[VECTOR_PH_NEW]] ], [ [[VEC_IND_NEXT_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[NITER:%.*]] = phi i64 [ [[UNROLL_ITER]], [[VECTOR_PH_NEW]] ], [ [[NITER_NSUB_3:%.*]], [[VECTOR_BODY]] ]
 ; AUTO_VEC-NEXT:    [[STEP_ADD:%.*]] = fadd fast <4 x double> [[VEC_IND]], <double 1.200000e+01, double 1.200000e+01, double 1.200000e+01, double 1.200000e+01>
 ; AUTO_VEC-NEXT:    [[STEP_ADD2:%.*]] = fadd fast <4 x double> [[VEC_IND]], <double 2.400000e+01, double 2.400000e+01, double 2.400000e+01, double 2.400000e+01>
 ; AUTO_VEC-NEXT:    [[STEP_ADD3:%.*]] = fadd fast <4 x double> [[VEC_IND]], <double 3.600000e+01, double 3.600000e+01, double 3.600000e+01, double 3.600000e+01>

llvm/test/Transforms/LoopVectorize/if-conversion-nest.ll

 ; CHECK-NEXT:    [[TMP5:%.*]] = add nuw nsw i64 [[TMP4]], 1
 ; CHECK-NEXT:    [[SCEVGEP:%.*]] = getelementptr i32, i32* [[A:%.*]], i64 [[TMP5]]
 ; CHECK-NEXT:    [[SCEVGEP4:%.*]] = getelementptr i32, i32* [[B:%.*]], i64 [[TMP5]]
 ; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ugt i32* [[SCEVGEP4]], [[A]]
 ; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ugt i32* [[SCEVGEP]], [[B]]
 ; CHECK-NEXT:    [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]]
 ; CHECK-NEXT:    br i1 [[FOUND_CONFLICT]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP2]], 8589934588
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP2]], -4
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[A]], i64 [[INDEX]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = bitcast i32* [[TMP6]] to <4 x i32>*
 ; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP7]], align 4, !alias.scope !0, !noalias !3
 ; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[B]], i64 [[INDEX]]
 ; CHECK-NEXT:    [[TMP9:%.*]] = bitcast i32* [[TMP8]] to <4 x i32>*

llvm/test/Transforms/LoopVectorize/runtime-check.ll

-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function foo
 ; RUN: opt < %s -loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -dce -instcombine -S | FileCheck %s
 ; RUN: opt < %s -loop-vectorize -disable-basic-aa -S -pass-remarks-analysis='loop-vectorize' 2>&1 | FileCheck %s -check-prefix=FORCED_OPTSIZE
 
 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-S128"
 
 ; Make sure we vectorize this loop:
 ; int foo(float *a, float *b, int n) {
 ;   for (int i=0; i<n; ++i)

 ; CHECK-NEXT:    [[TMP5:%.*]] = add nuw nsw i64 [[TMP4]], 1, [[DBG9]]
 ; CHECK-NEXT:    [[SCEVGEP:%.*]] = getelementptr float, float* [[A:%.*]], i64 [[TMP5]], [[DBG9]]
 ; CHECK-NEXT:    [[SCEVGEP4:%.*]] = getelementptr float, float* [[B:%.*]], i64 [[TMP5]], [[DBG9]]
 ; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ugt float* [[SCEVGEP4]], [[A]], [[DBG9]]
 ; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ugt float* [[SCEVGEP]], [[B]], [[DBG9]]
 ; CHECK-NEXT:    [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]], [[DBG9]]
 ; CHECK-NEXT:    br i1 [[FOUND_CONFLICT]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]], [[DBG9]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP2]], 8589934588, [[DBG9]]
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP2]], -4, [[DBG9]]
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]], [[DBG9]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ], [[DBG9]]
 ; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds float, float* [[B]], i64 [[INDEX]], [[DBG9]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = bitcast float* [[TMP6]] to <4 x float>*, [[DBG9]]
 ; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, <4 x float>* [[TMP7]], align 4, [[DBG9]], !alias.scope !10
 ; CHECK-NEXT:    [[TMP8:%.*]] = fmul <4 x float> [[WIDE_LOAD]], <float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00>, [[DBG9]]
 ; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds float, float* [[A]], i64 [[INDEX]], [[DBG9]]

llvm/test/Transforms/PhaseOrdering/X86/vdiv.ll

 ; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x double> [[BROADCAST_SPLATINSERT]], <4 x double> undef, <4 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP0:%.*]] = add nsw i64 [[N_VEC]], -4
 ; CHECK-NEXT:    [[TMP1:%.*]] = lshr exact i64 [[TMP0]], 2
 ; CHECK-NEXT:    [[TMP2:%.*]] = add nuw nsw i64 [[TMP1]], 1
 ; CHECK-NEXT:    [[XTRAITER:%.*]] = and i64 [[TMP2]], 3
 ; CHECK-NEXT:    [[TMP3:%.*]] = icmp ult i64 [[TMP0]], 12
 ; CHECK-NEXT:    br i1 [[TMP3]], label [[MIDDLE_BLOCK_UNR_LCSSA:%.*]], label [[VECTOR_PH_NEW:%.*]]
 ; CHECK:       vector.ph.new:
-; CHECK-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP2]], 9223372036854775804
+; CHECK-NEXT:    [[UNROLL_ITER:%.*]] = and i64 [[TMP2]], -4
 ; CHECK-NEXT:    [[TMP4:%.*]] = fdiv fast <4 x double> <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>, [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = fdiv fast <4 x double> <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>, [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP6:%.*]] = fdiv fast <4 x double> <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>, [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = fdiv fast <4 x double> <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>, [[BROADCAST_SPLAT]]
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH_NEW]] ], [ [[INDEX_NEXT_3:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[NITER:%.*]] = phi i64 [ [[UNROLL_ITER]], [[VECTOR_PH_NEW]] ], [ [[NITER_NSUB_3:%.*]], [[VECTOR_BODY]] ]