Add -lowertypetests-bitsets-level to control bitsets generation.

[krasin]

Summary

Sometimes, bitsets could get really large (>300k entries) and
we might want to drop a check, as it would have a too much cost.

Adding a flag to control how much penalty are we willing to pay
for bitsets.

[krasin]

make the test a bit more strict.

[kcc]

LGTM with a nit.

Please also add a runable test to compiler-rt as a separate change.

lib/Transforms/IPO/LowerTypeTests.cpp

475

extend the comment:
// If the bit set is all ones (or we treat it as such), ...

[krasin]

Update a comment.

[pcc]

lib/Transforms/IPO/LowerTypeTests.cpp

52

It seems really bizarre for a command line flag to affect an intrinsic's semantics like this. If we want this at all it should probably be an intrinsic argument or something.

[pcc]

lib/Transforms/IPO/LowerTypeTests.cpp

52

Ping.

lib/Transforms/IPO/LowerTypeTests.cpp

 STATISTIC(NumTypeTestCallsLowered, "Number of type test calls lowered");
 STATISTIC(NumTypeIdDisjointSets, "Number of disjoint sets of type identifiers");
 
 static cl::opt<bool> AvoidReuse(
     "lowertypetests-avoid-reuse",
     cl::desc("Try to avoid reuse of byte array addresses using aliases"),
     cl::Hidden, cl::init(true));
 
+static cl::opt<unsigned> BitsetsLevel(
+    "lowertypetests-bitsets-level",
+    cl::desc("Whether to generate bitsets: 0 - never, 1 - only if no loads, 2 - always."),
+    cl::Hidden, cl::init(2));
+
 bool BitSetInfo::containsGlobalOffset(uint64_t Offset) const {
   if (Offset < ByteOffset)
     return false;
 
   if ((Offset - ByteOffset) % (uint64_t(1) << AlignLog2) != 0)
     return false;
 
   uint64_t BitOffset = (Offset - ByteOffset) >> AlignLog2;
Value *OffsetSHL = B.CreateShl(
         PtrOffset,
         ConstantInt::get(IntPtrTy, DL.getPointerSizeInBits(0) - BSI.AlignLog2));
     BitOffset = B.CreateOr(OffsetSHR, OffsetSHL);
   }
 
   Constant *BitSizeConst = ConstantInt::get(IntPtrTy, BSI.BitSize);
   Value *OffsetInRange = B.CreateICmpULT(BitOffset, BitSizeConst);
 
   // If the bit set is all ones, testing against it is unnecessary.
-  if (BSI.isAllOnes())
+  if (BSI.isAllOnes() || BitsetsLevel == 0 ||
+      (BitsetsLevel == 1 && BSI.BitSize > 64))
     return OffsetInRange;
 
   TerminatorInst *Term = SplitBlockAndInsertIfThen(OffsetInRange, CI, false);
   IRBuilder<> ThenB(Term);
 
   // Now that we know that the offset is in range and aligned, load the
   // appropriate bit from the bitset.
   Value *Bit = createBitSetTest(ThenB, BSI, BAI, BitOffset);

test/Transforms/LowerTypeTests/simple.ll

-; RUN: opt -S -lowertypetests < %s | FileCheck %s
+; RUN: opt -S -lowertypetests < %s | FileCheck --check-prefixes=CHECK,CHECK2 %s
+; RUN: opt -S -lowertypetests -lowertypetests-bitsets-level=0 < %s | FileCheck --check-prefixes=CHECK,CHECK0 %s
+; RUN: opt -S -lowertypetests -lowertypetests-bitsets-level=2 < %s | FileCheck --check-prefixes=CHECK,CHECK2 %s
 ; RUN: opt -S -lowertypetests -mtriple=x86_64-apple-macosx10.8.0 < %s | FileCheck -check-prefix=CHECK-DARWIN %s
 ; RUN: opt -S -O3 < %s | FileCheck -check-prefix=CHECK-NODISCARD %s
 
 target datalayout = "e-p:32:32"
 
 ; CHECK: [[G:@[^ ]*]] = private constant { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] } { i32 1, [0 x i8] zeroinitializer, [63 x i32] zeroinitializer, [4 x i8] zeroinitializer, i32 3, [0 x i8] zeroinitializer, [2 x i32] [i32 4, i32 5] }
 @a = constant i32 1, !type !0, !type !2
 @b = hidden constant [63 x i32] zeroinitializer, !type !0, !type !1
 @c = protected constant i32 3, !type !1, !type !2
 @d = constant [2 x i32] [i32 4, i32 5], !type !3
 
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 ; CHECK-NODISCARD: !type
 
-; CHECK: [[BA:@[^ ]*]] = private constant [68 x i8] c"\03\01\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\02\00\01"
+; CHECK2: [[BA:@[^ ]*]] = private constant [68 x i8] c"\03\01\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\02\00\01"
 
 ; Offset 0, 4 byte alignment
 !0 = !{i32 0, !"typeid1"}
 !3 = !{i32 4, !"typeid1"}
 
 ; Offset 4, 256 byte alignment
 !1 = !{i32 0, !"typeid2"}
 
 ; Offset 0, 4 byte alignment
 !2 = !{i32 0, !"typeid3"}
 
-; CHECK: @bits_use{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
-; CHECK: @bits_use.{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
-; CHECK: @bits_use.{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
+; CHECK2: @bits_use{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
+; CHECK2: @bits_use.{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
+; CHECK2: @bits_use.{{[0-9]*}} = private alias i8, i8* @bits{{[0-9]*}}
 
 ; CHECK: @a = alias i32, getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 0)
 ; CHECK: @b = hidden alias [63 x i32], getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 2)
 ; CHECK: @c = protected alias i32, getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 4)
 ; CHECK: @d = alias [2 x i32], getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 6)
 
 ; CHECK-DARWIN: @aptr = constant i32* getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G:@[^ ]*]], i32 0, i32 0)
 @aptr = constant i32* @a
 
 ; CHECK-DARWIN: @bptr = constant [63 x i32]* getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 2)
 @bptr = constant [63 x i32]* @b
 
 ; CHECK-DARWIN: @cptr = constant i32* getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 4)
 @cptr = constant i32* @c
 
 ; CHECK-DARWIN: @dptr = constant [2 x i32]* getelementptr inbounds ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }, { i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]], i32 0, i32 6)
 @dptr = constant [2 x i32]* @d
 
 ; CHECK-DARWIN: [[G]] = private constant
 
-; CHECK: @bits{{[0-9]*}} = private alias i8, getelementptr inbounds ([68 x i8], [68 x i8]* [[BA]], i32 0, i32 0)
-; CHECK: @bits.{{[0-9]*}} = private alias i8, getelementptr inbounds ([68 x i8], [68 x i8]* [[BA]], i32 0, i32 0)
+; CHECK2: @bits{{[0-9]*}} = private alias i8, getelementptr inbounds ([68 x i8], [68 x i8]* [[BA]], i32 0, i32 0)
+; CHECK2: @bits.{{[0-9]*}} = private alias i8, getelementptr inbounds ([68 x i8], [68 x i8]* [[BA]], i32 0, i32 0)
 
 declare i1 @llvm.type.test(i8* %ptr, metadata %bitset) nounwind readnone
 
 ; CHECK: @foo(i32* [[A0:%[^ ]*]])
 define i1 @foo(i32* %p) {
   ; CHECK-NOT: llvm.type.test
 
   ; CHECK: [[R0:%[^ ]*]] = bitcast i32* [[A0]] to i8*
   %pi8 = bitcast i32* %p to i8*
   ; CHECK: [[R1:%[^ ]*]] = ptrtoint i8* [[R0]] to i32
   ; CHECK: [[R2:%[^ ]*]] = sub i32 [[R1]], ptrtoint ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]] to i32)
   ; CHECK: [[R3:%[^ ]*]] = lshr i32 [[R2]], 2
   ; CHECK: [[R4:%[^ ]*]] = shl i32 [[R2]], 30
   ; CHECK: [[R5:%[^ ]*]] = or i32 [[R3]], [[R4]]
   ; CHECK: [[R6:%[^ ]*]] = icmp ult i32 [[R5]], 68
-  ; CHECK: br i1 [[R6]]
+  ; CHECK2: br i1 [[R6]]
+  ; CHECK0-NOT: br
 
-  ; CHECK: [[R8:%[^ ]*]] = getelementptr i8, i8* @bits_use.{{[0-9]*}}, i32 [[R5]]
-  ; CHECK: [[R9:%[^ ]*]] = load i8, i8* [[R8]]
-  ; CHECK: [[R10:%[^ ]*]] = and i8 [[R9]], 1
-  ; CHECK: [[R11:%[^ ]*]] = icmp ne i8 [[R10]], 0
+  ; CHECK2: [[R8:%[^ ]*]] = getelementptr i8, i8* @bits_use.{{[0-9]*}}, i32 [[R5]]
+  ; CHECK0-NOT: bits_use
+  ; CHECK2: [[R9:%[^ ]*]] = load i8, i8* [[R8]]
+  ; CHECK2: [[R10:%[^ ]*]] = and i8 [[R9]], 1
+  ; CHECK2: [[R11:%[^ ]*]] = icmp ne i8 [[R10]], 0
 
-  ; CHECK: [[R16:%[^ ]*]] = phi i1 [ false, {{%[^ ]*}} ], [ [[R11]], {{%[^ ]*}} ]
+  ; CHECK2: [[R16:%[^ ]*]] = phi i1 [ false, {{%[^ ]*}} ], [ [[R11]], {{%[^ ]*}} ]
   %x = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid1")
 
   ; CHECK-NOT: llvm.type.test
+  ; CHECK0-NOT: llvm.type.test
   %y = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid1")
 
-  ; CHECK: ret i1 [[R16]]
+  ; CHECK2: ret i1 [[R16]]
+  ; CHECK0: ret i1 [[R6]]
   ret i1 %x
 }
 
 ; CHECK: @bar(i32* [[B0:%[^ ]*]])
 define i1 @bar(i32* %p) {
   ; CHECK: [[S0:%[^ ]*]] = bitcast i32* [[B0]] to i8*
   %pi8 = bitcast i32* %p to i8*
   ; CHECK: [[S1:%[^ ]*]] = ptrtoint i8* [[S0]] to i32
define i1 @baz(i32* %p) {
   ; CHECK: [[T0:%[^ ]*]] = bitcast i32* [[C0]] to i8*
   %pi8 = bitcast i32* %p to i8*
   ; CHECK: [[T1:%[^ ]*]] = ptrtoint i8* [[T0]] to i32
   ; CHECK: [[T2:%[^ ]*]] = sub i32 [[T1]], ptrtoint ({ i32, [0 x i8], [63 x i32], [4 x i8], i32, [0 x i8], [2 x i32] }* [[G]] to i32)
   ; CHECK: [[T3:%[^ ]*]] = lshr i32 [[T2]], 2
   ; CHECK: [[T4:%[^ ]*]] = shl i32 [[T2]], 30
   ; CHECK: [[T5:%[^ ]*]] = or i32 [[T3]], [[T4]]
   ; CHECK: [[T6:%[^ ]*]] = icmp ult i32 [[T5]], 66
-  ; CHECK: br i1 [[T6]]
+  ; CHECK2: br i1 [[T6]]
+  ; CHECK0-NOT: br
 
-  ; CHECK: [[T8:%[^ ]*]] = getelementptr i8, i8* @bits_use{{(\.[0-9]*)?}}, i32 [[T5]]
-  ; CHECK: [[T9:%[^ ]*]] = load i8, i8* [[T8]]
-  ; CHECK: [[T10:%[^ ]*]] = and i8 [[T9]], 2
-  ; CHECK: [[T11:%[^ ]*]] = icmp ne i8 [[T10]], 0
+  ; CHECK2: [[T8:%[^ ]*]] = getelementptr i8, i8* @bits_use{{(\.[0-9]*)?}}, i32 [[T5]]
+  ; CHECK0-NOT: bits_use
+  ; CHECK2: [[T9:%[^ ]*]] = load i8, i8* [[T8]]
+  ; CHECK2: [[T10:%[^ ]*]] = and i8 [[T9]], 2
+  ; CHECK2: [[T11:%[^ ]*]] = icmp ne i8 [[T10]], 0
 
-  ; CHECK: [[T16:%[^ ]*]] = phi i1 [ false, {{%[^ ]*}} ], [ [[T11]], {{%[^ ]*}} ]
+  ; CHECK2: [[T16:%[^ ]*]] = phi i1 [ false, {{%[^ ]*}} ], [ [[T11]], {{%[^ ]*}} ]
   %x = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid3")
-  ; CHECK: ret i1 [[T16]]
+  ; CHECK2: ret i1 [[T16]]
+  ; CHECK0: ret i1 [[T6]]
   ret i1 %x
 }