[SimplifyCFG] Avoid threading for loop headers (#151142)

Updates SimplifyCFG to avoid jump threading through loop headers if
-keep-loops is requested. Canonical loop form requires a loop header
that dominates all blocks in the loop. If we thread through a header, we
risk breaking its domination of the loop. This change avoids this issue
by conservatively avoiding threading through headers entirely.

Fixes: #151144

Author: ArneStenkrona2

llvm/lib/Transforms/Utils/SimplifyCFG.cpp

@@ -291,6 +291,7 @@ class SimplifyCFGOpt {
   bool simplifyBranch(BranchInst *Branch, IRBuilder<> &Builder);
   bool simplifyUncondBranch(BranchInst *BI, IRBuilder<> &Builder);
   bool simplifyCondBranch(BranchInst *BI, IRBuilder<> &Builder);
+  bool foldCondBranchOnValueKnownInPredecessor(BranchInst *BI);
 
   bool tryToSimplifyUncondBranchWithICmpInIt(ICmpInst *ICI,
                                              IRBuilder<> &Builder);
@@ -3669,15 +3670,19 @@ foldCondBranchOnValueKnownInPredecessorImpl(BranchInst *BI, DomTreeUpdater *DTU,
   return false;
 }
 
-static bool foldCondBranchOnValueKnownInPredecessor(BranchInst *BI,
-                                                    DomTreeUpdater *DTU,
-                                                    const DataLayout &DL,
-                                                    AssumptionCache *AC) {
+bool SimplifyCFGOpt::foldCondBranchOnValueKnownInPredecessor(BranchInst *BI) {
+  // Note: If BB is a loop header then there is a risk that threading introduces
+  // a non-canonical loop by moving a back edge. So we avoid this optimization
+  // for loop headers if NeedCanonicalLoop is set.
+  if (Options.NeedCanonicalLoop && is_contained(LoopHeaders, BI->getParent()))
+    return false;
+
   std::optional<bool> Result;
   bool EverChanged = false;
   do {
     // Note that None means "we changed things, but recurse further."
-    Result = foldCondBranchOnValueKnownInPredecessorImpl(BI, DTU, DL, AC);
+    Result =
+        foldCondBranchOnValueKnownInPredecessorImpl(BI, DTU, DL, Options.AC);
     EverChanged |= Result == std::nullopt || *Result;
   } while (Result == std::nullopt);
   return EverChanged;
@@ -8099,7 +8104,7 @@ bool SimplifyCFGOpt::simplifyCondBranch(BranchInst *BI, IRBuilder<> &Builder) {
   // If this is a branch on something for which we know the constant value in
   // predecessors (e.g. a phi node in the current block), thread control
   // through this block.
-  if (foldCondBranchOnValueKnownInPredecessor(BI, DTU, DL, Options.AC))
+  if (foldCondBranchOnValueKnownInPredecessor(BI))
     return requestResimplify();
 
   // Scan predecessor blocks for conditional branches.

llvm/test/CodeGen/ARM/2013-05-05-IfConvertBug.ll

@@ -1,7 +1,7 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 | FileCheck %s
-; RUN: llc < %s -mtriple=thumbv8 | FileCheck -check-prefix=CHECK-V8 %s
-; RUN: llc < %s -mtriple=thumbv7 -arm-restrict-it | FileCheck -check-prefix=CHECK-RESTRICT-IT %s
+; RUN: llc -keep-loops="false" < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 | FileCheck %s
+; RUN: llc -keep-loops="false" < %s -mtriple=thumbv8 | FileCheck -check-prefix=CHECK-V8 %s
+; RUN: llc -keep-loops="false" < %s -mtriple=thumbv7 -arm-restrict-it | FileCheck -check-prefix=CHECK-RESTRICT-IT %s
 
 define i32 @t1(i32 %a, i32 %b, ptr %retaddr) {
 ; CHECK-LABEL: t1:

llvm/test/Transforms/SimplifyCFG/2008-07-13-InfLoopMiscompile.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 -S | FileCheck %s
+; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 -keep-loops="false" -S | FileCheck %s
 ; PR2540
 ; Outval should end up with a select from 0/2, not all constants.
 
@@ -52,4 +52,3 @@ func_1.exit:		; preds = %cowblock, %entry
 }
 
 declare i32 @printf(ptr, ...) nounwind
-

llvm/test/Transforms/SimplifyCFG/2025-07-29-non-canoncial-loop.ll

@@ -0,0 +1,62 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 --keep-loops="true" -S | FileCheck --check-prefix=NO-THREADING %s
+; Checks that we do not thread the control flow through the loop header loop_header as
+; that will introduce a non-canonical loop.
+
+; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 --keep-loops="false" -S | FileCheck --check-prefix=THREADING %s
+; Checks that we thread the control flow through the loop header loop_header since we
+; do not request --keep-loops.
+
+define void @__start(i1 %cond) {
+; NO-THREADING-LABEL: define void @__start(
+; NO-THREADING-SAME: i1 [[COND:%.*]]) {
+; NO-THREADING-NEXT:  [[ENTRY:.*:]]
+; NO-THREADING-NEXT:    br label %[[LOOP_HEADER:.*]]
+; NO-THREADING:       [[LOOP_HEADER]]:
+; NO-THREADING-NEXT:    br i1 [[COND]], label %[[LOOP_BODY_1:.*]], label %[[LOOP_BODY_0:.*]]
+; NO-THREADING:       [[LOOP_BODY_0]]:
+; NO-THREADING-NEXT:    [[_0_:%.*]] = add i16 0, 0
+; NO-THREADING-NEXT:    br label %[[LOOP_EXIT:.*]]
+; NO-THREADING:       [[LOOP_BODY_1]]:
+; NO-THREADING-NEXT:    [[_1_:%.*]] = add i32 0, 1
+; NO-THREADING-NEXT:    br label %[[LOOP_EXIT]]
+; NO-THREADING:       [[LOOP_EXIT]]:
+; NO-THREADING-NEXT:    br i1 [[COND]], label %[[LOOP_HEADER]], label %[[EXIT:.*]]
+; NO-THREADING:       [[EXIT]]:
+; NO-THREADING-NEXT:    ret void
+;
+; THREADING-LABEL: define void @__start(
+; THREADING-SAME: i1 [[COND:%.*]]) {
+; THREADING-NEXT:  [[ENTRY:.*:]]
+; THREADING-NEXT:    br i1 [[COND]], label %[[LOOP_BODY_1:.*]], label %[[LOOP_BODY_0:.*]]
+; THREADING:       [[LOOP_BODY_0]]:
+; THREADING-NEXT:    [[_0_:%.*]] = add i16 0, 0
+; THREADING-NEXT:    br label %[[LOOP_EXIT:.*]]
+; THREADING:       [[LOOP_BODY_1]]:
+; THREADING-NEXT:    [[_1_:%.*]] = add i32 0, 1
+; THREADING-NEXT:    br label %[[LOOP_EXIT]]
+; THREADING:       [[LOOP_EXIT]]:
+; THREADING-NEXT:    br i1 [[COND]], label %[[LOOP_BODY_1]], label %[[EXIT:.*]]
+; THREADING:       [[EXIT]]:
+; THREADING-NEXT:    ret void
+;
+entry:
+  br label %loop_header
+
+loop_header:                                            ; preds = %loop_exit, %entry
+  br i1 %cond, label %loop_body_1, label %loop_body_0
+
+loop_body_0:                                            ; preds = %loop_header
+  %_0_ = add i16 0, 0
+  br label %loop_exit
+
+loop_body_1:                                            ; preds = %loop_header
+  %_1_ = add i32 0, 1
+  br label %loop_exit
+
+loop_exit:                                            ; preds = %loop_body_1, %loop_body_0
+  br i1 %cond, label %loop_header, label %exit
+
+exit:                                            ; preds = %loop_exit
+  ret void
+}

llvm/test/Transforms/SimplifyCFG/branch-phi-thread.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt < %s -passes=simplifycfg,adce -simplifycfg-require-and-preserve-domtree=1 -S | FileCheck %s
+; RUN: opt < %s -passes=simplifycfg,adce -simplifycfg-require-and-preserve-domtree=1 -keep-loops="false" -S | FileCheck %s
 
 declare void @f1()
 

llvm/test/Transforms/SimplifyCFG/jump-threading.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -S -passes=simplifycfg < %s | FileCheck %s
+; RUN: opt -S -passes=simplifycfg -keep-loops="false" < %s | FileCheck %s
 
 declare void @foo()
 declare void @bar()

llvm/test/Transforms/SimplifyCFG/two-entry-phi-return.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 -S | FileCheck %s
+; RUN: opt < %s -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 -keep-loops="false" -S | FileCheck %s
 
 define i1 @qux(ptr %m, ptr %n, ptr %o, ptr %p) nounwind  {
 ; CHECK-LABEL: @qux(