1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
|
//===- CodeMetrics.cpp - Code cost measurements ---------------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements code cost measurement utilities.
//
//===----------------------------------------------------------------------===//
#include "llvm/Analysis/CodeMetrics.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/Support/CallSite.h"
using namespace llvm;
/// callIsSmall - If a call is likely to lower to a single target instruction,
/// or is otherwise deemed small return true.
/// TODO: Perhaps calls like memcpy, strcpy, etc?
bool llvm::callIsSmall(ImmutableCallSite CS) {
if (isa<IntrinsicInst>(CS.getInstruction()))
return true;
const Function *F = CS.getCalledFunction();
if (!F) return false;
if (F->hasLocalLinkage()) return false;
if (!F->hasName()) return false;
StringRef Name = F->getName();
// These will all likely lower to a single selection DAG node.
if (Name == "copysign" || Name == "copysignf" || Name == "copysignl" ||
Name == "fabs" || Name == "fabsf" || Name == "fabsl" ||
Name == "sin" || Name == "sinf" || Name == "sinl" ||
Name == "cos" || Name == "cosf" || Name == "cosl" ||
Name == "sqrt" || Name == "sqrtf" || Name == "sqrtl" )
return true;
// These are all likely to be optimized into something smaller.
if (Name == "pow" || Name == "powf" || Name == "powl" ||
Name == "exp2" || Name == "exp2l" || Name == "exp2f" ||
Name == "floor" || Name == "floorf" || Name == "ceil" ||
Name == "round" || Name == "ffs" || Name == "ffsl" ||
Name == "abs" || Name == "labs" || Name == "llabs")
return true;
return false;
}
bool llvm::isInstructionFree(const Instruction *I, const DataLayout *TD) {
if (isa<PHINode>(I))
return true;
// If a GEP has all constant indices, it will probably be folded with
// a load/store.
if (const GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(I))
return GEP->hasAllConstantIndices();
if (const IntrinsicInst *II = dyn_cast<IntrinsicInst>(I)) {
switch (II->getIntrinsicID()) {
default:
return false;
case Intrinsic::dbg_declare:
case Intrinsic::dbg_value:
case Intrinsic::invariant_start:
case Intrinsic::invariant_end:
case Intrinsic::lifetime_start:
case Intrinsic::lifetime_end:
case Intrinsic::objectsize:
case Intrinsic::ptr_annotation:
case Intrinsic::var_annotation:
// These intrinsics don't count as size.
return true;
}
}
if (const CastInst *CI = dyn_cast<CastInst>(I)) {
// Noop casts, including ptr <-> int, don't count.
if (CI->isLosslessCast())
return true;
Value *Op = CI->getOperand(0);
// An inttoptr cast is free so long as the input is a legal integer type
// which doesn't contain values outside the range of a pointer.
if (isa<IntToPtrInst>(CI) && TD &&
TD->isLegalInteger(Op->getType()->getScalarSizeInBits()) &&
Op->getType()->getScalarSizeInBits() <= TD->getPointerSizeInBits())
return true;
// A ptrtoint cast is free so long as the result is large enough to store
// the pointer, and a legal integer type.
if (isa<PtrToIntInst>(CI) && TD &&
TD->isLegalInteger(Op->getType()->getScalarSizeInBits()) &&
Op->getType()->getScalarSizeInBits() >= TD->getPointerSizeInBits())
return true;
// trunc to a native type is free (assuming the target has compare and
// shift-right of the same width).
if (TD && isa<TruncInst>(CI) &&
TD->isLegalInteger(TD->getTypeSizeInBits(CI->getType())))
return true;
// Result of a cmp instruction is often extended (to be used by other
// cmp instructions, logical or return instructions). These are usually
// nop on most sane targets.
if (isa<CmpInst>(CI->getOperand(0)))
return true;
}
return false;
}
/// analyzeBasicBlock - Fill in the current structure with information gleaned
/// from the specified block.
void CodeMetrics::analyzeBasicBlock(const BasicBlock *BB,
const DataLayout *TD) {
++NumBlocks;
unsigned NumInstsBeforeThisBB = NumInsts;
for (BasicBlock::const_iterator II = BB->begin(), E = BB->end();
II != E; ++II) {
if (isInstructionFree(II, TD))
continue;
// Special handling for calls.
if (isa<CallInst>(II) || isa<InvokeInst>(II)) {
ImmutableCallSite CS(cast<Instruction>(II));
if (const Function *F = CS.getCalledFunction()) {
// If a function is both internal and has a single use, then it is
// extremely likely to get inlined in the future (it was probably
// exposed by an interleaved devirtualization pass).
if (!CS.isNoInline() && F->hasInternalLinkage() && F->hasOneUse())
++NumInlineCandidates;
// If this call is to function itself, then the function is recursive.
// Inlining it into other functions is a bad idea, because this is
// basically just a form of loop peeling, and our metrics aren't useful
// for that case.
if (F == BB->getParent())
isRecursive = true;
}
if (!callIsSmall(CS)) {
// Each argument to a call takes on average one instruction to set up.
NumInsts += CS.arg_size();
// We don't want inline asm to count as a call - that would prevent loop
// unrolling. The argument setup cost is still real, though.
if (!isa<InlineAsm>(CS.getCalledValue()))
++NumCalls;
}
}
if (const AllocaInst *AI = dyn_cast<AllocaInst>(II)) {
if (!AI->isStaticAlloca())
this->usesDynamicAlloca = true;
}
if (isa<ExtractElementInst>(II) || II->getType()->isVectorTy())
++NumVectorInsts;
if (const CallInst *CI = dyn_cast<CallInst>(II))
if (CI->hasFnAttr(Attribute::NoDuplicate))
notDuplicatable = true;
if (const InvokeInst *InvI = dyn_cast<InvokeInst>(II))
if (InvI->hasFnAttr(Attribute::NoDuplicate))
notDuplicatable = true;
++NumInsts;
}
if (isa<ReturnInst>(BB->getTerminator()))
++NumRets;
// We never want to inline functions that contain an indirectbr. This is
// incorrect because all the blockaddress's (in static global initializers
// for example) would be referring to the original function, and this indirect
// jump would jump from the inlined copy of the function into the original
// function which is extremely undefined behavior.
// FIXME: This logic isn't really right; we can safely inline functions
// with indirectbr's as long as no other function or global references the
// blockaddress of a block within the current function. And as a QOI issue,
// if someone is using a blockaddress without an indirectbr, and that
// reference somehow ends up in another function or global, we probably
// don't want to inline this function.
notDuplicatable |= isa<IndirectBrInst>(BB->getTerminator());
// Remember NumInsts for this BB.
NumBBInsts[BB] = NumInsts - NumInstsBeforeThisBB;
}
void CodeMetrics::analyzeFunction(Function *F, const DataLayout *TD) {
// If this function contains a call that "returns twice" (e.g., setjmp or
// _setjmp) and it isn't marked with "returns twice" itself, never inline it.
// This is a hack because we depend on the user marking their local variables
// as volatile if they are live across a setjmp call, and they probably
// won't do this in callers.
exposesReturnsTwice = F->callsFunctionThatReturnsTwice() &&
!F->getAttributes().hasAttribute(AttributeSet::FunctionIndex,
Attribute::ReturnsTwice);
// Look at the size of the callee.
for (Function::const_iterator BB = F->begin(), E = F->end(); BB != E; ++BB)
analyzeBasicBlock(&*BB, TD);
}
|
