[BranchFolding] Avoid moving blocks to fall through to an indirect target

Author: XChy

llvm/lib/CodeGen/BranchFolding.cpp

@@ -1776,8 +1776,8 @@ bool BranchFolder::OptimizeBlock(MachineBasicBlock *MBB) {
       // Okay, there is no really great place to put this block.  If, however,
       // the block before this one would be a fall-through if this block were
       // removed, move this block to the end of the function. There is no real
-      // advantage in "falling through" to an EH block, so we don't want to
-      // perform this transformation for that case.
+      // advantage in "falling through" to an EH block or an other indirect
+      // target, so we don't want to perform this transformation for that case.
       //
       // Also, Windows EH introduced the possibility of an arbitrary number of
       // successors to a given block.  The analyzeBranch call does not consider
@@ -1787,10 +1787,14 @@ bool BranchFolder::OptimizeBlock(MachineBasicBlock *MBB) {
       // below were performed for EH "FallThrough" blocks.  Therefore, even if
       // that appears not to be happening anymore, we should assume that it is
       // possible and not remove the "!FallThrough()->isEHPad" condition below.
+      //
+      // Similarly, indirect branches like callbr/indirectbr also introduced the
+      // possibility of infinite rotation, as there are an arbitrary number of
+      // successors to a given block.
       MachineBasicBlock *PrevTBB = nullptr, *PrevFBB = nullptr;
       SmallVector<MachineOperand, 4> PrevCond;
-      if (FallThrough != MF.end() &&
-          !FallThrough->isEHPad() &&
+      if (FallThrough != MF.end() && !FallThrough->isEHPad() &&
+          !FallThrough->hasAddressTaken() &&
           !TII->analyzeBranch(PrevBB, PrevTBB, PrevFBB, PrevCond, true) &&
           PrevBB.isSuccessor(&*FallThrough)) {
         MBB->moveAfter(&MF.back());

llvm/test/CodeGen/X86/callbr-asm-loop.ll

@@ -1,35 +1,29 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
 
-; RUN: llc -O0 -mtriple=i686-- < %s | FileCheck %s
+; RUN: llc -O1 -mtriple=i686-- < %s | FileCheck %s
 
 ; Test that causes multiple defs of %eax.
 ; FIXME: The testcase hangs with -O1/2/3 enabled.
 define i32 @loop1() nounwind {
 ; CHECK-LABEL: loop1:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    pushl %esi
-; CHECK-NEXT:    jmp .LBB0_1
+; CHECK-NEXT:    .p2align 4
 ; CHECK-NEXT:  .LBB0_1: # %tailrecurse
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    xorl %eax, %eax
 ; CHECK-NEXT:    movl $1, %edx
 ; CHECK-NEXT:    #APP
 ; CHECK-NEXT:    #NO_APP
-; CHECK-NEXT:    movl %eax, %ecx
-; CHECK-NEXT:    movl %edx, %esi
-; CHECK-NEXT:    jmp .LBB0_3
+; CHECK-NEXT:    jmp .LBB0_1
 ; CHECK-NEXT:  .LBB0_2: # Inline asm indirect target
-; CHECK-NEXT:    # %tailrecurse.tailrecurse.backedge_crit_edge
+; CHECK-NEXT:    # %tailrecurse.tailrecurse_crit_edge
 ; CHECK-NEXT:    # in Loop: Header=BB0_1 Depth=1
 ; CHECK-NEXT:    # Label of block must be emitted
-; CHECK-NEXT:  .LBB0_3: # %tailrecurse.backedge
-; CHECK-NEXT:    # in Loop: Header=BB0_1 Depth=1
 ; CHECK-NEXT:    jmp .LBB0_1
-; CHECK-NEXT:  .LBB0_4: # Inline asm indirect target
+; CHECK-NEXT:  .LBB0_3: # Inline asm indirect target
 ; CHECK-NEXT:    # %lab2.split
 ; CHECK-NEXT:    # Label of block must be emitted
 ; CHECK-NEXT:    movl %edx, %eax
-; CHECK-NEXT:    popl %esi
 ; CHECK-NEXT:    retl
 entry:
   br label %tailrecurse

llvm/test/CodeGen/X86/catchpad-reuse.ll

@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
 ; RUN: llc < %s | FileCheck %s
 
 ; IR generated by the following C++ source with modifications to reuse the 'v'
@@ -15,46 +16,8 @@
 ;   }
 ;   return 0;
 ; }
-
-; CHECK: $cppxdata$main:
-; CHECK-NEXT: .long   429065506               # MagicNumber
-; CHECK-NEXT: .long   4                       # MaxState
-; CHECK-NEXT: .long   $stateUnwindMap$main@IMGREL # UnwindMap
-; CHECK-NEXT: .long   2                       # NumTryBlocks
-; CHECK-NEXT: .long   $tryMap$main@IMGREL   # TryBlockMap
-; CHECK-NEXT: .long   5                       # IPMapEntries
-; CHECK-NEXT: .long   $ip2state$main@IMGREL # IPToStateXData
-; CHECK-NEXT: .long   32                      # UnwindHelp
-; CHECK-NEXT: .long   0                       # ESTypeList
-; CHECK-NEXT: .long   1                       # EHFlags
-
-; CHECK: $tryMap$main:
-; CHECK-NEXT: .long   1                       # TryLow
-; CHECK-NEXT: .long   1                       # TryHigh
-; CHECK-NEXT: .long   2                       # CatchHigh
-; CHECK-NEXT: .long   1                       # NumCatches
-; CHECK-NEXT: .long   $handlerMap$0$main@IMGREL # HandlerArray
-; CHECK-NEXT: .long   0                       # TryLow
-; CHECK-NEXT: .long   2                       # TryHigh
-; CHECK-NEXT: .long   3                       # CatchHigh
-; CHECK-NEXT: .long   1                       # NumCatches
-; CHECK-NEXT: .long   $handlerMap$1$main@IMGREL # HandlerArray
-
-; CHECK: $handlerMap$0$main:
-; CHECK-NEXT: .long   0                       # Adjectives
-; CHECK-NEXT: .long   "??_R0H@8"@IMGREL       # Type
-; CHECK-NEXT: .long   [[v_offset:[0-9]+]]     # CatchObjOffset
-; CHECK-NEXT: .long   "?catch$2@?0?main@4HA"@IMGREL # Handler
-; CHECK-NEXT: .long   {{.*}}                  # ParentFrameOffset
-
-; CHECK: $handlerMap$1$main:
-; CHECK-NEXT: .long   0                       # Adjectives
-; CHECK-NEXT: .long   "??_R0H@8"@IMGREL       # Type
-; CHECK-NEXT: .long   [[v_offset]]            # CatchObjOffset
-; CHECK-NEXT: .long   "?catch$4@?0?main@4HA"@IMGREL # Handler
-; CHECK-NEXT: .long   {{.*}}                      # ParentFrameOffset
-
 ; ModuleID = 't.cpp'
+
 source_filename = "t.cpp"
 target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-pc-windows-msvc19.0.24210"
@@ -68,6 +31,92 @@ $"\01??_R0H@8" = comdat any
 
 ; Function Attrs: norecurse uwtable
 define i32 @main() local_unnamed_addr personality ptr @__CxxFrameHandler3 {
+; CHECK-LABEL: main:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    pushq %rbp
+; CHECK-NEXT:    .seh_pushreg %rbp
+; CHECK-NEXT:    subq $48, %rsp
+; CHECK-NEXT:    .seh_stackalloc 48
+; CHECK-NEXT:    leaq {{[0-9]+}}(%rsp), %rbp
+; CHECK-NEXT:    .seh_setframe %rbp, 48
+; CHECK-NEXT:    .seh_endprologue
+; CHECK-NEXT:    movq $-2, -16(%rbp)
+; CHECK-NEXT:  .Ltmp0: # EH_LABEL
+; CHECK-NEXT:    callq maythrow
+; CHECK-NEXT:    nop
+; CHECK-NEXT:  .Ltmp1: # EH_LABEL
+; CHECK-NEXT:  .LBB0_1: # Block address taken
+; CHECK-NEXT:    # %try.cont6
+; CHECK-NEXT:  $ehgcr_0_1:
+; CHECK-NEXT:    xorl %eax, %eax
+; CHECK-NEXT:    .seh_startepilogue
+; CHECK-NEXT:    addq $48, %rsp
+; CHECK-NEXT:    popq %rbp
+; CHECK-NEXT:    .seh_endepilogue
+; CHECK-NEXT:    retq
+; CHECK-NEXT:    .seh_handlerdata
+; CHECK-NEXT:    .long $cppxdata$main@IMGREL
+; CHECK-NEXT:    .text
+; CHECK-NEXT:    .seh_endproc
+; CHECK-NEXT:    .def "?catch$2@?0?main@4HA";
+; CHECK-NEXT:    .scl 3;
+; CHECK-NEXT:    .type 32;
+; CHECK-NEXT:    .endef
+; CHECK-NEXT:    .p2align 4
+; CHECK-NEXT:  "?catch$2@?0?main@4HA":
+; CHECK-NEXT:  .seh_proc "?catch$2@?0?main@4HA"
+; CHECK-NEXT:    .seh_handler __CxxFrameHandler3, @unwind, @except
+; CHECK-NEXT:  .LBB0_2: # %catch3
+; CHECK-NEXT:    movq %rdx, {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    pushq %rbp
+; CHECK-NEXT:    .seh_pushreg %rbp
+; CHECK-NEXT:    subq $32, %rsp
+; CHECK-NEXT:    .seh_stackalloc 32
+; CHECK-NEXT:    leaq 48(%rdx), %rbp
+; CHECK-NEXT:    .seh_endprologue
+; CHECK-NEXT:    callq maythrow
+; CHECK-NEXT:    leaq .LBB0_1(%rip), %rax
+; CHECK-NEXT:    .seh_startepilogue
+; CHECK-NEXT:    addq $32, %rsp
+; CHECK-NEXT:    popq %rbp
+; CHECK-NEXT:    .seh_endepilogue
+; CHECK-NEXT:    retq # CATCHRET
+; CHECK-NEXT:    .seh_handlerdata
+; CHECK-NEXT:    .long $cppxdata$main@IMGREL
+; CHECK-NEXT:    .text
+; CHECK-NEXT:    .seh_endproc
+; CHECK-NEXT:    .def "?catch$3@?0?main@4HA";
+; CHECK-NEXT:    .scl 3;
+; CHECK-NEXT:    .type 32;
+; CHECK-NEXT:    .endef
+; CHECK-NEXT:    .p2align 4
+; CHECK-NEXT:  "?catch$3@?0?main@4HA":
+; CHECK-NEXT:  .seh_proc "?catch$3@?0?main@4HA"
+; CHECK-NEXT:    .seh_handler __CxxFrameHandler3, @unwind, @except
+; CHECK-NEXT:  .LBB0_3: # %catch
+; CHECK-NEXT:    movq %rdx, {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    pushq %rbp
+; CHECK-NEXT:    .seh_pushreg %rbp
+; CHECK-NEXT:    subq $32, %rsp
+; CHECK-NEXT:    .seh_stackalloc 32
+; CHECK-NEXT:    leaq 48(%rdx), %rbp
+; CHECK-NEXT:    .seh_endprologue
+; CHECK-NEXT:  .Ltmp2: # EH_LABEL
+; CHECK-NEXT:    callq maythrow
+; CHECK-NEXT:    nop
+; CHECK-NEXT:  .Ltmp3: # EH_LABEL
+; CHECK-NEXT:  # %bb.4: # %invoke.cont1
+; CHECK-NEXT:    leaq .LBB0_1(%rip), %rax
+; CHECK-NEXT:    .seh_startepilogue
+; CHECK-NEXT:    addq $32, %rsp
+; CHECK-NEXT:    popq %rbp
+; CHECK-NEXT:    .seh_endepilogue
+; CHECK-NEXT:    retq # CATCHRET
+; CHECK-NEXT:  .Lfunc_end0:
+; CHECK-NEXT:    .seh_handlerdata
+; CHECK-NEXT:    .long $cppxdata$main@IMGREL
+; CHECK-NEXT:    .text
+; CHECK-NEXT:    .seh_endproc
 entry:
   %v = alloca i32, align 4
   invoke void @maythrow()