summaryrefslogtreecommitdiff
path: root/lib/Target/ARM/ARMTargetMachine.cpp
blob: 718311ba1decc483d6f106c7af620db934d2629b (plain)
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
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
//===-- ARMTargetMachine.cpp - Define TargetMachine for ARM ---------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
//
//===----------------------------------------------------------------------===//

#include "ARMTargetMachine.h"
#include "ARM.h"
#include "ARMFrameLowering.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/MC/MCAsmInfo.h"
#include "llvm/PassManager.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/FormattedStream.h"
#include "llvm/Support/TargetRegistry.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/Transforms/Scalar.h"
using namespace llvm;

static cl::opt<bool>
EnableGlobalMerge("global-merge", cl::Hidden,
                  cl::desc("Enable global merge pass"),
                  cl::init(true));

static cl::opt<bool>
DisableA15SDOptimization("disable-a15-sd-optimization", cl::Hidden,
                   cl::desc("Inhibit optimization of S->D register accesses on A15"),
                   cl::init(false));

extern "C" void LLVMInitializeARMTarget() {
  // Register the target.
  RegisterTargetMachine<ARMTargetMachine> X(TheARMTarget);
  RegisterTargetMachine<ThumbTargetMachine> Y(TheThumbTarget);
}


/// TargetMachine ctor - Create an ARM architecture model.
///
ARMBaseTargetMachine::ARMBaseTargetMachine(const Target &T, StringRef TT,
                                           StringRef CPU, StringRef FS,
                                           const TargetOptions &Options,
                                           Reloc::Model RM, CodeModel::Model CM,
                                           CodeGenOpt::Level OL)
  : LLVMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
    Subtarget(TT, CPU, FS, Options),
    JITInfo(),
    InstrItins(Subtarget.getInstrItineraryData()) {

  // Default to triple-appropriate float ABI
  if (Options.FloatABIType == FloatABI::Default)
    this->Options.FloatABIType =
        Subtarget.isTargetHardFloat() ? FloatABI::Hard : FloatABI::Soft;
}

void ARMBaseTargetMachine::addAnalysisPasses(PassManagerBase &PM) {
  // Add first the target-independent BasicTTI pass, then our ARM pass. This
  // allows the ARM pass to delegate to the target independent layer when
  // appropriate.
  PM.add(createBasicTargetTransformInfoPass(this));
  PM.add(createARMTargetTransformInfoPass(this));
}


void ARMTargetMachine::anchor() { }

static std::string computeDataLayout(ARMSubtarget &ST) {
  // Little endian.
  std::string Ret = "e";

  Ret += DataLayout::getManglingComponent(ST.getTargetTriple());

  // Pointers are 32 bits and aligned to 32 bits.
  Ret += "-p:32:32";

  // On thumb, i16,i18 and i1 have natural aligment requirements, but we try to
  // align to 32.
  if (ST.isThumb())
    Ret += "-i1:8:32-i8:8:32-i16:16:32";

  // ABIs other than APC have 64 bit integers with natural alignment.
  if (!ST.isAPCS_ABI())
    Ret += "-i64:64";

  // We have 64 bits floats. The APCS ABI requires them to be aligned to 32
  // bits, others to 64 bits. We always try to align to 64 bits.
  if (ST.isAPCS_ABI())
    Ret += "-f64:32:64";

  // We have 128 and 64 bit vectors. The APCS ABI aligns them to 32 bits, others
  // to 64. We always ty to give them natural alignment.
  if (ST.isAPCS_ABI())
    Ret += "-v64:32:64-v128:32:128";
  else
    Ret += "-v128:64:128";

  // On thumb and APCS, only try to align aggregates to 32 bits (the default is
  // 64 bits).
  if (ST.isThumb() || ST.isAPCS_ABI())
    Ret += "-a:0:32";

  // Integer registers are 32 bits.
  Ret += "-n32";

  // The stack is 128 bit aligned on NaCl, 64 bit aligned on AAPCS and 32 bit
  // aligned everywhere else.
  if (ST.isTargetNaCl())
    Ret += "-S128";
  else if (ST.isAAPCS_ABI())
    Ret += "-S64";
  else
    Ret += "-S32";

  return Ret;
}

ARMTargetMachine::ARMTargetMachine(const Target &T, StringRef TT,
                                   StringRef CPU, StringRef FS,
                                   const TargetOptions &Options,
                                   Reloc::Model RM, CodeModel::Model CM,
                                   CodeGenOpt::Level OL)
  : ARMBaseTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
    InstrInfo(Subtarget),
    DL(computeDataLayout(Subtarget)),
    TLInfo(*this),
    TSInfo(*this),
    FrameLowering(Subtarget) {
  initAsmInfo();
  if (!Subtarget.hasARMOps())
    report_fatal_error("CPU: '" + Subtarget.getCPUString() + "' does not "
                       "support ARM mode execution!");
}

void ThumbTargetMachine::anchor() { }

ThumbTargetMachine::ThumbTargetMachine(const Target &T, StringRef TT,
                                       StringRef CPU, StringRef FS,
                                       const TargetOptions &Options,
                                       Reloc::Model RM, CodeModel::Model CM,
                                       CodeGenOpt::Level OL)
  : ARMBaseTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
    InstrInfo(Subtarget.hasThumb2()
              ? ((ARMBaseInstrInfo*)new Thumb2InstrInfo(Subtarget))
              : ((ARMBaseInstrInfo*)new Thumb1InstrInfo(Subtarget))),
    DL(computeDataLayout(Subtarget)),
    TLInfo(*this),
    TSInfo(*this),
    FrameLowering(Subtarget.hasThumb2()
              ? new ARMFrameLowering(Subtarget)
              : (ARMFrameLowering*)new Thumb1FrameLowering(Subtarget)) {
  initAsmInfo();
}

namespace {
/// ARM Code Generator Pass Configuration Options.
class ARMPassConfig : public TargetPassConfig {
public:
  ARMPassConfig(ARMBaseTargetMachine *TM, PassManagerBase &PM)
    : TargetPassConfig(TM, PM) {}

  ARMBaseTargetMachine &getARMTargetMachine() const {
    return getTM<ARMBaseTargetMachine>();
  }

  const ARMSubtarget &getARMSubtarget() const {
    return *getARMTargetMachine().getSubtargetImpl();
  }

  virtual bool addPreISel();
  virtual bool addInstSelector();
  virtual bool addPreRegAlloc();
  virtual bool addPreSched2();
  virtual bool addPreEmitPass();
};
} // namespace

TargetPassConfig *ARMBaseTargetMachine::createPassConfig(PassManagerBase &PM) {
  return new ARMPassConfig(this, PM);
}

bool ARMPassConfig::addPreISel() {
  if (TM->getOptLevel() != CodeGenOpt::None && EnableGlobalMerge)
    addPass(createGlobalMergePass(TM));

  return false;
}

bool ARMPassConfig::addInstSelector() {
  addPass(createARMISelDag(getARMTargetMachine(), getOptLevel()));

  const ARMSubtarget *Subtarget = &getARMSubtarget();
  if (Subtarget->isTargetELF() && !Subtarget->isThumb1Only() &&
      TM->Options.EnableFastISel)
    addPass(createARMGlobalBaseRegPass());
  return false;
}

bool ARMPassConfig::addPreRegAlloc() {
  // FIXME: temporarily disabling load / store optimization pass for Thumb1.
  if (getOptLevel() != CodeGenOpt::None && !getARMSubtarget().isThumb1Only())
    addPass(createARMLoadStoreOptimizationPass(true));
  if (getOptLevel() != CodeGenOpt::None && getARMSubtarget().isCortexA9())
    addPass(createMLxExpansionPass());
  // Since the A15SDOptimizer pass can insert VDUP instructions, it can only be
  // enabled when NEON is available.
  if (getOptLevel() != CodeGenOpt::None && getARMSubtarget().isCortexA15() &&
    getARMSubtarget().hasNEON() && !DisableA15SDOptimization) {
    addPass(createA15SDOptimizerPass());
  }
  return true;
}

bool ARMPassConfig::addPreSched2() {
  // FIXME: temporarily disabling load / store optimization pass for Thumb1.
  if (getOptLevel() != CodeGenOpt::None) {
    if (!getARMSubtarget().isThumb1Only()) {
      addPass(createARMLoadStoreOptimizationPass());
      printAndVerify("After ARM load / store optimizer");
    }
    if (getARMSubtarget().hasNEON())
      addPass(createExecutionDependencyFixPass(&ARM::DPRRegClass));
  }

  // Expand some pseudo instructions into multiple instructions to allow
  // proper scheduling.
  addPass(createARMExpandPseudoPass());

  if (getOptLevel() != CodeGenOpt::None) {
    if (!getARMSubtarget().isThumb1Only()) {
      // in v8, IfConversion depends on Thumb instruction widths
      if (getARMSubtarget().restrictIT() &&
          !getARMSubtarget().prefers32BitThumb())
        addPass(createThumb2SizeReductionPass());
      addPass(&IfConverterID);
    }
  }
  if (getARMSubtarget().isThumb2())
    addPass(createThumb2ITBlockPass());

  return true;
}

bool ARMPassConfig::addPreEmitPass() {
  if (getARMSubtarget().isThumb2()) {
    if (!getARMSubtarget().prefers32BitThumb())
      addPass(createThumb2SizeReductionPass());

    // Constant island pass work on unbundled instructions.
    addPass(&UnpackMachineBundlesID);
  }

  addPass(createARMConstantIslandPass());

  return true;
}

bool ARMBaseTargetMachine::addCodeEmitter(PassManagerBase &PM,
                                          JITCodeEmitter &JCE) {
  // Machine code emitter pass for ARM.
  PM.add(createARMJITCodeEmitterPass(*this, JCE));
  return false;
}