1 //===- CodeEmitterGen.cpp - Code Emitter Generator ------------------------===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // CodeEmitterGen uses the descriptions of instructions and their fields to
11 // construct an automated code emitter: a function that, given a MachineInstr,
12 // returns the (currently, 32-bit unsigned) value of the instruction.
13 //
14 //===----------------------------------------------------------------------===//
15 
16 #include "CodeGenInstruction.h"
17 #include "CodeGenTarget.h"
18 #include "SubtargetFeatureInfo.h"
19 #include "llvm/ADT/ArrayRef.h"
20 #include "llvm/ADT/StringExtras.h"
21 #include "llvm/Support/Casting.h"
22 #include "llvm/Support/raw_ostream.h"
23 #include "llvm/TableGen/Record.h"
24 #include "llvm/TableGen/TableGenBackend.h"
25 #include <cassert>
26 #include <cstdint>
27 #include <map>
28 #include <set>
29 #include <string>
30 #include <utility>
31 #include <vector>
32 
33 using namespace llvm;
34 
35 namespace {
36 
37 class CodeEmitterGen {
38   RecordKeeper &Records;
39 
40 public:
41   CodeEmitterGen(RecordKeeper &R) : Records(R) {}
42 
43   void run(raw_ostream &o);
44 
45 private:
46   int getVariableBit(const std::string &VarName, BitsInit *BI, int bit);
47   std::string getInstructionCase(Record *R, CodeGenTarget &Target);
48   void AddCodeToMergeInOperand(Record *R, BitsInit *BI,
49                                const std::string &VarName,
50                                unsigned &NumberedOp,
51                                std::set<unsigned> &NamedOpIndices,
52                                std::string &Case, CodeGenTarget &Target);
53 
54 };
55 
56 // If the VarBitInit at position 'bit' matches the specified variable then
57 // return the variable bit position.  Otherwise return -1.
58 int CodeEmitterGen::getVariableBit(const std::string &VarName,
59                                    BitsInit *BI, int bit) {
60   if (VarBitInit *VBI = dyn_cast<VarBitInit>(BI->getBit(bit))) {
61     if (VarInit *VI = dyn_cast<VarInit>(VBI->getBitVar()))
62       if (VI->getName() == VarName)
63         return VBI->getBitNum();
64   } else if (VarInit *VI = dyn_cast<VarInit>(BI->getBit(bit))) {
65     if (VI->getName() == VarName)
66       return 0;
67   }
68 
69   return -1;
70 }
71 
72 void CodeEmitterGen::
73 AddCodeToMergeInOperand(Record *R, BitsInit *BI, const std::string &VarName,
74                         unsigned &NumberedOp,
75                         std::set<unsigned> &NamedOpIndices,
76                         std::string &Case, CodeGenTarget &Target) {
77   CodeGenInstruction &CGI = Target.getInstruction(R);
78 
79   // Determine if VarName actually contributes to the Inst encoding.
80   int bit = BI->getNumBits()-1;
81 
82   // Scan for a bit that this contributed to.
83   for (; bit >= 0; ) {
84     if (getVariableBit(VarName, BI, bit) != -1)
85       break;
86 
87     --bit;
88   }
89 
90   // If we found no bits, ignore this value, otherwise emit the call to get the
91   // operand encoding.
92   if (bit < 0) return;
93 
94   // If the operand matches by name, reference according to that
95   // operand number. Non-matching operands are assumed to be in
96   // order.
97   unsigned OpIdx;
98   if (CGI.Operands.hasOperandNamed(VarName, OpIdx)) {
99     // Get the machine operand number for the indicated operand.
100     OpIdx = CGI.Operands[OpIdx].MIOperandNo;
101     assert(!CGI.Operands.isFlatOperandNotEmitted(OpIdx) &&
102            "Explicitly used operand also marked as not emitted!");
103   } else {
104     unsigned NumberOps = CGI.Operands.size();
105     /// If this operand is not supposed to be emitted by the
106     /// generated emitter, skip it.
107     while (NumberedOp < NumberOps &&
108            (CGI.Operands.isFlatOperandNotEmitted(NumberedOp) ||
109               (!NamedOpIndices.empty() && NamedOpIndices.count(
110                 CGI.Operands.getSubOperandNumber(NumberedOp).first)))) {
111       ++NumberedOp;
112 
113       if (NumberedOp >= CGI.Operands.back().MIOperandNo +
114                         CGI.Operands.back().MINumOperands) {
115         errs() << "Too few operands in record " << R->getName() <<
116                   " (no match for variable " << VarName << "):\n";
117         errs() << *R;
118         errs() << '\n';
119 
120         return;
121       }
122     }
123 
124     OpIdx = NumberedOp++;
125   }
126 
127   std::pair<unsigned, unsigned> SO = CGI.Operands.getSubOperandNumber(OpIdx);
128   std::string &EncoderMethodName = CGI.Operands[SO.first].EncoderMethodName;
129 
130   // If the source operand has a custom encoder, use it. This will
131   // get the encoding for all of the suboperands.
132   if (!EncoderMethodName.empty()) {
133     // A custom encoder has all of the information for the
134     // sub-operands, if there are more than one, so only
135     // query the encoder once per source operand.
136     if (SO.second == 0) {
137       Case += "      // op: " + VarName + "\n" +
138               "      op = " + EncoderMethodName + "(MI, " + utostr(OpIdx);
139       Case += ", Fixups, STI";
140       Case += ");\n";
141     }
142   } else {
143     Case += "      // op: " + VarName + "\n" +
144       "      op = getMachineOpValue(MI, MI.getOperand(" + utostr(OpIdx) + ")";
145     Case += ", Fixups, STI";
146     Case += ");\n";
147   }
148 
149   for (; bit >= 0; ) {
150     int varBit = getVariableBit(VarName, BI, bit);
151 
152     // If this bit isn't from a variable, skip it.
153     if (varBit == -1) {
154       --bit;
155       continue;
156     }
157 
158     // Figure out the consecutive range of bits covered by this operand, in
159     // order to generate better encoding code.
160     int beginInstBit = bit;
161     int beginVarBit = varBit;
162     int N = 1;
163     for (--bit; bit >= 0;) {
164       varBit = getVariableBit(VarName, BI, bit);
165       if (varBit == -1 || varBit != (beginVarBit - N)) break;
166       ++N;
167       --bit;
168     }
169 
170     uint64_t opMask = ~(uint64_t)0 >> (64-N);
171     int opShift = beginVarBit - N + 1;
172     opMask <<= opShift;
173     opShift = beginInstBit - beginVarBit;
174 
175     if (opShift > 0) {
176       Case += "      Value |= (op & UINT64_C(" + utostr(opMask) + ")) << " +
177               itostr(opShift) + ";\n";
178     } else if (opShift < 0) {
179       Case += "      Value |= (op & UINT64_C(" + utostr(opMask) + ")) >> " +
180               itostr(-opShift) + ";\n";
181     } else {
182       Case += "      Value |= op & UINT64_C(" + utostr(opMask) + ");\n";
183     }
184   }
185 }
186 
187 std::string CodeEmitterGen::getInstructionCase(Record *R,
188                                                CodeGenTarget &Target) {
189   std::string Case;
190   BitsInit *BI = R->getValueAsBitsInit("Inst");
191   unsigned NumberedOp = 0;
192   std::set<unsigned> NamedOpIndices;
193 
194   // Collect the set of operand indices that might correspond to named
195   // operand, and skip these when assigning operands based on position.
196   if (Target.getInstructionSet()->
197        getValueAsBit("noNamedPositionallyEncodedOperands")) {
198     CodeGenInstruction &CGI = Target.getInstruction(R);
199     for (const RecordVal &RV : R->getValues()) {
200       unsigned OpIdx;
201       if (!CGI.Operands.hasOperandNamed(RV.getName(), OpIdx))
202         continue;
203 
204       NamedOpIndices.insert(OpIdx);
205     }
206   }
207 
208   // Loop over all of the fields in the instruction, determining which are the
209   // operands to the instruction.
210   for (const RecordVal &RV : R->getValues()) {
211     // Ignore fixed fields in the record, we're looking for values like:
212     //    bits<5> RST = { ?, ?, ?, ?, ? };
213     if (RV.getPrefix() || RV.getValue()->isComplete())
214       continue;
215 
216     AddCodeToMergeInOperand(R, BI, RV.getName(), NumberedOp,
217                             NamedOpIndices, Case, Target);
218   }
219 
220   StringRef PostEmitter = R->getValueAsString("PostEncoderMethod");
221   if (!PostEmitter.empty()) {
222     Case += "      Value = ";
223     Case += PostEmitter;
224     Case += "(MI, Value";
225     Case += ", STI";
226     Case += ");\n";
227   }
228 
229   return Case;
230 }
231 
232 void CodeEmitterGen::run(raw_ostream &o) {
233   CodeGenTarget Target(Records);
234   std::vector<Record*> Insts = Records.getAllDerivedDefinitions("Instruction");
235 
236   // For little-endian instruction bit encodings, reverse the bit order
237   Target.reverseBitsForLittleEndianEncoding();
238 
239   ArrayRef<const CodeGenInstruction*> NumberedInstructions =
240     Target.getInstructionsByEnumValue();
241 
242   // Emit function declaration
243   o << "uint64_t " << Target.getName();
244   o << "MCCodeEmitter::getBinaryCodeForInstr(const MCInst &MI,\n"
245     << "    SmallVectorImpl<MCFixup> &Fixups,\n"
246     << "    const MCSubtargetInfo &STI) const {\n";
247 
248   // Emit instruction base values
249   o << "  static const uint64_t InstBits[] = {\n";
250   for (const CodeGenInstruction *CGI : NumberedInstructions) {
251     Record *R = CGI->TheDef;
252 
253     if (R->getValueAsString("Namespace") == "TargetOpcode" ||
254         R->getValueAsBit("isPseudo")) {
255       o << "    UINT64_C(0),\n";
256       continue;
257     }
258 
259     BitsInit *BI = R->getValueAsBitsInit("Inst");
260 
261     // Start by filling in fixed values.
262     uint64_t Value = 0;
263     for (unsigned i = 0, e = BI->getNumBits(); i != e; ++i) {
264       if (BitInit *B = dyn_cast<BitInit>(BI->getBit(e-i-1)))
265         Value |= (uint64_t)B->getValue() << (e-i-1);
266     }
267     o << "    UINT64_C(" << Value << ")," << '\t' << "// " << R->getName() << "\n";
268   }
269   o << "    UINT64_C(0)\n  };\n";
270 
271   // Map to accumulate all the cases.
272   std::map<std::string, std::vector<std::string>> CaseMap;
273 
274   // Construct all cases statement for each opcode
275   for (std::vector<Record*>::iterator IC = Insts.begin(), EC = Insts.end();
276         IC != EC; ++IC) {
277     Record *R = *IC;
278     if (R->getValueAsString("Namespace") == "TargetOpcode" ||
279         R->getValueAsBit("isPseudo"))
280       continue;
281     std::string InstName =
282         (R->getValueAsString("Namespace") + "::" + R->getName()).str();
283     std::string Case = getInstructionCase(R, Target);
284 
285     CaseMap[Case].push_back(std::move(InstName));
286   }
287 
288   // Emit initial function code
289   o << "  const unsigned opcode = MI.getOpcode();\n"
290     << "  uint64_t Value = InstBits[opcode];\n"
291     << "  uint64_t op = 0;\n"
292     << "  (void)op;  // suppress warning\n"
293     << "  switch (opcode) {\n";
294 
295   // Emit each case statement
296   std::map<std::string, std::vector<std::string>>::iterator IE, EE;
297   for (IE = CaseMap.begin(), EE = CaseMap.end(); IE != EE; ++IE) {
298     const std::string &Case = IE->first;
299     std::vector<std::string> &InstList = IE->second;
300 
301     for (int i = 0, N = InstList.size(); i < N; i++) {
302       if (i) o << "\n";
303       o << "    case " << InstList[i]  << ":";
304     }
305     o << " {\n";
306     o << Case;
307     o << "      break;\n"
308       << "    }\n";
309   }
310 
311   // Default case: unhandled opcode
312   o << "  default:\n"
313     << "    std::string msg;\n"
314     << "    raw_string_ostream Msg(msg);\n"
315     << "    Msg << \"Not supported instr: \" << MI;\n"
316     << "    report_fatal_error(Msg.str());\n"
317     << "  }\n"
318     << "  return Value;\n"
319     << "}\n\n";
320 
321   const auto &All = SubtargetFeatureInfo::getAll(Records);
322   std::map<Record *, SubtargetFeatureInfo, LessRecordByID> SubtargetFeatures;
323   SubtargetFeatures.insert(All.begin(), All.end());
324 
325   o << "#ifdef ENABLE_INSTR_PREDICATE_VERIFIER\n"
326     << "#undef ENABLE_INSTR_PREDICATE_VERIFIER\n"
327     << "#include <sstream>\n\n";
328 
329   // Emit the subtarget feature enumeration.
330   SubtargetFeatureInfo::emitSubtargetFeatureFlagEnumeration(SubtargetFeatures,
331                                                             o);
332 
333   // Emit the name table for error messages.
334   o << "#ifndef NDEBUG\n";
335   SubtargetFeatureInfo::emitNameTable(SubtargetFeatures, o);
336   o << "#endif // NDEBUG\n";
337 
338   // Emit the available features compute function.
339   SubtargetFeatureInfo::emitComputeAssemblerAvailableFeatures(
340       Target.getName(), "MCCodeEmitter", "computeAvailableFeatures",
341       SubtargetFeatures, o);
342 
343   // Emit the predicate verifier.
344   o << "void " << Target.getName()
345     << "MCCodeEmitter::verifyInstructionPredicates(\n"
346     << "    const MCInst &Inst, uint64_t AvailableFeatures) const {\n"
347     << "#ifndef NDEBUG\n"
348     << "  static uint64_t RequiredFeatures[] = {\n";
349   unsigned InstIdx = 0;
350   for (const CodeGenInstruction *Inst : Target.getInstructionsByEnumValue()) {
351     o << "    ";
352     for (Record *Predicate : Inst->TheDef->getValueAsListOfDefs("Predicates")) {
353       const auto &I = SubtargetFeatures.find(Predicate);
354       if (I != SubtargetFeatures.end())
355         o << I->second.getEnumName() << " | ";
356     }
357     o << "0, // " << Inst->TheDef->getName() << " = " << InstIdx << "\n";
358     InstIdx++;
359   }
360   o << "  };\n\n";
361   o << "  assert(Inst.getOpcode() < " << InstIdx << ");\n";
362   o << "  uint64_t MissingFeatures =\n"
363     << "      (AvailableFeatures & RequiredFeatures[Inst.getOpcode()]) ^\n"
364     << "      RequiredFeatures[Inst.getOpcode()];\n"
365     << "  if (MissingFeatures) {\n"
366     << "    std::ostringstream Msg;\n"
367     << "    Msg << \"Attempting to emit \" << "
368        "MCII.getName(Inst.getOpcode()).str()\n"
369     << "        << \" instruction but the \";\n"
370     << "    for (unsigned i = 0; i < 8 * sizeof(MissingFeatures); ++i)\n"
371     << "      if (MissingFeatures & (1ULL << i))\n"
372     << "        Msg << SubtargetFeatureNames[i] << \" \";\n"
373     << "    Msg << \"predicate(s) are not met\";\n"
374     << "    report_fatal_error(Msg.str());\n"
375     << "  }\n"
376     << "#else\n"
377     << "// Silence unused variable warning on targets that don't use MCII for "
378        "other purposes (e.g. BPF).\n"
379     << "(void)MCII;\n"
380     << "#endif // NDEBUG\n";
381   o << "}\n";
382   o << "#endif\n";
383 }
384 
385 } // end anonymous namespace
386 
387 namespace llvm {
388 
389 void EmitCodeEmitter(RecordKeeper &RK, raw_ostream &OS) {
390   emitSourceFileHeader("Machine Code Emitter", OS);
391   CodeEmitterGen(RK).run(OS);
392 }
393 
394 } // end namespace llvm
395