1 //===-- llvm/lib/CodeGen/AsmPrinter/DebugHandlerBase.cpp -------*- C++ -*--===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // Common functionality for different debug information format backends.
10 // LLVM currently supports DWARF and CodeView.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "llvm/CodeGen/DebugHandlerBase.h"
15 #include "llvm/ADT/Optional.h"
16 #include "llvm/ADT/Twine.h"
17 #include "llvm/CodeGen/AsmPrinter.h"
18 #include "llvm/CodeGen/MachineFunction.h"
19 #include "llvm/CodeGen/MachineInstr.h"
20 #include "llvm/CodeGen/MachineModuleInfo.h"
21 #include "llvm/CodeGen/TargetSubtargetInfo.h"
22 #include "llvm/IR/DebugInfo.h"
23 #include "llvm/MC/MCStreamer.h"
24 #include "llvm/Support/CommandLine.h"
25 
26 using namespace llvm;
27 
28 #define DEBUG_TYPE "dwarfdebug"
29 
30 /// If true, we drop variable location ranges which exist entirely outside the
31 /// variable's lexical scope instruction ranges.
32 static cl::opt<bool> TrimVarLocs("trim-var-locs", cl::Hidden, cl::init(true));
33 
34 Optional<DbgVariableLocation>
35 DbgVariableLocation::extractFromMachineInstruction(
36     const MachineInstr &Instruction) {
37   DbgVariableLocation Location;
38   // Variables calculated from multiple locations can't be represented here.
39   if (Instruction.getNumDebugOperands() != 1)
40     return None;
41   if (!Instruction.getDebugOperand(0).isReg())
42     return None;
43   Location.Register = Instruction.getDebugOperand(0).getReg();
44   Location.FragmentInfo.reset();
45   // We only handle expressions generated by DIExpression::appendOffset,
46   // which doesn't require a full stack machine.
47   int64_t Offset = 0;
48   const DIExpression *DIExpr = Instruction.getDebugExpression();
49   auto Op = DIExpr->expr_op_begin();
50   // We can handle a DBG_VALUE_LIST iff it has exactly one location operand that
51   // appears exactly once at the start of the expression.
52   if (Instruction.isDebugValueList()) {
53     if (Instruction.getNumDebugOperands() == 1 &&
54         Op->getOp() == dwarf::DW_OP_LLVM_arg)
55       ++Op;
56     else
57       return None;
58   }
59   while (Op != DIExpr->expr_op_end()) {
60     switch (Op->getOp()) {
61     case dwarf::DW_OP_constu: {
62       int Value = Op->getArg(0);
63       ++Op;
64       if (Op != DIExpr->expr_op_end()) {
65         switch (Op->getOp()) {
66         case dwarf::DW_OP_minus:
67           Offset -= Value;
68           break;
69         case dwarf::DW_OP_plus:
70           Offset += Value;
71           break;
72         default:
73           continue;
74         }
75       }
76     } break;
77     case dwarf::DW_OP_plus_uconst:
78       Offset += Op->getArg(0);
79       break;
80     case dwarf::DW_OP_LLVM_fragment:
81       Location.FragmentInfo = {Op->getArg(1), Op->getArg(0)};
82       break;
83     case dwarf::DW_OP_deref:
84       Location.LoadChain.push_back(Offset);
85       Offset = 0;
86       break;
87     default:
88       return None;
89     }
90     ++Op;
91   }
92 
93   // Do one final implicit DW_OP_deref if this was an indirect DBG_VALUE
94   // instruction.
95   // FIXME: Replace these with DIExpression.
96   if (Instruction.isIndirectDebugValue())
97     Location.LoadChain.push_back(Offset);
98 
99   return Location;
100 }
101 
102 DebugHandlerBase::DebugHandlerBase(AsmPrinter *A) : Asm(A), MMI(Asm->MMI) {}
103 
104 void DebugHandlerBase::beginModule(Module *M) {
105   if (M->debug_compile_units().empty())
106     Asm = nullptr;
107 }
108 
109 // Each LexicalScope has first instruction and last instruction to mark
110 // beginning and end of a scope respectively. Create an inverse map that list
111 // scopes starts (and ends) with an instruction. One instruction may start (or
112 // end) multiple scopes. Ignore scopes that are not reachable.
113 void DebugHandlerBase::identifyScopeMarkers() {
114   SmallVector<LexicalScope *, 4> WorkList;
115   WorkList.push_back(LScopes.getCurrentFunctionScope());
116   while (!WorkList.empty()) {
117     LexicalScope *S = WorkList.pop_back_val();
118 
119     const SmallVectorImpl<LexicalScope *> &Children = S->getChildren();
120     if (!Children.empty())
121       WorkList.append(Children.begin(), Children.end());
122 
123     if (S->isAbstractScope())
124       continue;
125 
126     for (const InsnRange &R : S->getRanges()) {
127       assert(R.first && "InsnRange does not have first instruction!");
128       assert(R.second && "InsnRange does not have second instruction!");
129       requestLabelBeforeInsn(R.first);
130       requestLabelAfterInsn(R.second);
131     }
132   }
133 }
134 
135 // Return Label preceding the instruction.
136 MCSymbol *DebugHandlerBase::getLabelBeforeInsn(const MachineInstr *MI) {
137   MCSymbol *Label = LabelsBeforeInsn.lookup(MI);
138   assert(Label && "Didn't insert label before instruction");
139   return Label;
140 }
141 
142 // Return Label immediately following the instruction.
143 MCSymbol *DebugHandlerBase::getLabelAfterInsn(const MachineInstr *MI) {
144   return LabelsAfterInsn.lookup(MI);
145 }
146 
147 /// If this type is derived from a base type then return base type size.
148 uint64_t DebugHandlerBase::getBaseTypeSize(const DIType *Ty) {
149   assert(Ty);
150   const DIDerivedType *DDTy = dyn_cast<DIDerivedType>(Ty);
151   if (!DDTy)
152     return Ty->getSizeInBits();
153 
154   unsigned Tag = DDTy->getTag();
155 
156   if (Tag != dwarf::DW_TAG_member && Tag != dwarf::DW_TAG_typedef &&
157       Tag != dwarf::DW_TAG_const_type && Tag != dwarf::DW_TAG_volatile_type &&
158       Tag != dwarf::DW_TAG_restrict_type && Tag != dwarf::DW_TAG_atomic_type)
159     return DDTy->getSizeInBits();
160 
161   DIType *BaseType = DDTy->getBaseType();
162 
163   if (!BaseType)
164     return 0;
165 
166   // If this is a derived type, go ahead and get the base type, unless it's a
167   // reference then it's just the size of the field. Pointer types have no need
168   // of this since they're a different type of qualification on the type.
169   if (BaseType->getTag() == dwarf::DW_TAG_reference_type ||
170       BaseType->getTag() == dwarf::DW_TAG_rvalue_reference_type)
171     return Ty->getSizeInBits();
172 
173   return getBaseTypeSize(BaseType);
174 }
175 
176 bool DebugHandlerBase::isUnsignedDIType(const DIType *Ty) {
177   if (auto *CTy = dyn_cast<DICompositeType>(Ty)) {
178     // FIXME: Enums without a fixed underlying type have unknown signedness
179     // here, leading to incorrectly emitted constants.
180     if (CTy->getTag() == dwarf::DW_TAG_enumeration_type)
181       return false;
182 
183     // (Pieces of) aggregate types that get hacked apart by SROA may be
184     // represented by a constant. Encode them as unsigned bytes.
185     return true;
186   }
187 
188   if (auto *DTy = dyn_cast<DIDerivedType>(Ty)) {
189     dwarf::Tag T = (dwarf::Tag)Ty->getTag();
190     // Encode pointer constants as unsigned bytes. This is used at least for
191     // null pointer constant emission.
192     // FIXME: reference and rvalue_reference /probably/ shouldn't be allowed
193     // here, but accept them for now due to a bug in SROA producing bogus
194     // dbg.values.
195     if (T == dwarf::DW_TAG_pointer_type ||
196         T == dwarf::DW_TAG_ptr_to_member_type ||
197         T == dwarf::DW_TAG_reference_type ||
198         T == dwarf::DW_TAG_rvalue_reference_type)
199       return true;
200     assert(T == dwarf::DW_TAG_typedef || T == dwarf::DW_TAG_const_type ||
201            T == dwarf::DW_TAG_volatile_type ||
202            T == dwarf::DW_TAG_restrict_type || T == dwarf::DW_TAG_atomic_type);
203     assert(DTy->getBaseType() && "Expected valid base type");
204     return isUnsignedDIType(DTy->getBaseType());
205   }
206 
207   auto *BTy = cast<DIBasicType>(Ty);
208   unsigned Encoding = BTy->getEncoding();
209   assert((Encoding == dwarf::DW_ATE_unsigned ||
210           Encoding == dwarf::DW_ATE_unsigned_char ||
211           Encoding == dwarf::DW_ATE_signed ||
212           Encoding == dwarf::DW_ATE_signed_char ||
213           Encoding == dwarf::DW_ATE_float || Encoding == dwarf::DW_ATE_UTF ||
214           Encoding == dwarf::DW_ATE_boolean ||
215           (Ty->getTag() == dwarf::DW_TAG_unspecified_type &&
216            Ty->getName() == "decltype(nullptr)")) &&
217          "Unsupported encoding");
218   return Encoding == dwarf::DW_ATE_unsigned ||
219          Encoding == dwarf::DW_ATE_unsigned_char ||
220          Encoding == dwarf::DW_ATE_UTF || Encoding == dwarf::DW_ATE_boolean ||
221          Ty->getTag() == dwarf::DW_TAG_unspecified_type;
222 }
223 
224 static bool hasDebugInfo(const MachineModuleInfo *MMI,
225                          const MachineFunction *MF) {
226   if (!MMI->hasDebugInfo())
227     return false;
228   auto *SP = MF->getFunction().getSubprogram();
229   if (!SP)
230     return false;
231   assert(SP->getUnit());
232   auto EK = SP->getUnit()->getEmissionKind();
233   if (EK == DICompileUnit::NoDebug)
234     return false;
235   return true;
236 }
237 
238 void DebugHandlerBase::beginFunction(const MachineFunction *MF) {
239   PrevInstBB = nullptr;
240 
241   if (!Asm || !hasDebugInfo(MMI, MF)) {
242     skippedNonDebugFunction();
243     return;
244   }
245 
246   // Grab the lexical scopes for the function, if we don't have any of those
247   // then we're not going to be able to do anything.
248   LScopes.initialize(*MF);
249   if (LScopes.empty()) {
250     beginFunctionImpl(MF);
251     return;
252   }
253 
254   // Make sure that each lexical scope will have a begin/end label.
255   identifyScopeMarkers();
256 
257   // Calculate history for local variables.
258   assert(DbgValues.empty() && "DbgValues map wasn't cleaned!");
259   assert(DbgLabels.empty() && "DbgLabels map wasn't cleaned!");
260   calculateDbgEntityHistory(MF, Asm->MF->getSubtarget().getRegisterInfo(),
261                             DbgValues, DbgLabels);
262   InstOrdering.initialize(*MF);
263   if (TrimVarLocs)
264     DbgValues.trimLocationRanges(*MF, LScopes, InstOrdering);
265   LLVM_DEBUG(DbgValues.dump());
266 
267   // Request labels for the full history.
268   for (const auto &I : DbgValues) {
269     const auto &Entries = I.second;
270     if (Entries.empty())
271       continue;
272 
273     auto IsDescribedByReg = [](const MachineInstr *MI) {
274       return any_of(MI->debug_operands(),
275                     [](auto &MO) { return MO.isReg() && MO.getReg(); });
276     };
277 
278     // The first mention of a function argument gets the CurrentFnBegin label,
279     // so arguments are visible when breaking at function entry.
280     //
281     // We do not change the label for values that are described by registers,
282     // as that could place them above their defining instructions. We should
283     // ideally not change the labels for constant debug values either, since
284     // doing that violates the ranges that are calculated in the history map.
285     // However, we currently do not emit debug values for constant arguments
286     // directly at the start of the function, so this code is still useful.
287     // FIXME: If the first mention of an argument is in a unique section basic
288     // block, we cannot always assign the CurrentFnBeginLabel as it lies in a
289     // different section.  Temporarily, we disable generating loc list
290     // information or DW_AT_const_value when the block is in a different
291     // section.
292     const DILocalVariable *DIVar =
293         Entries.front().getInstr()->getDebugVariable();
294     if (DIVar->isParameter() &&
295         getDISubprogram(DIVar->getScope())->describes(&MF->getFunction()) &&
296         Entries.front().getInstr()->getParent()->sameSection(&MF->front())) {
297       if (!IsDescribedByReg(Entries.front().getInstr()))
298         LabelsBeforeInsn[Entries.front().getInstr()] = Asm->getFunctionBegin();
299       if (Entries.front().getInstr()->getDebugExpression()->isFragment()) {
300         // Mark all non-overlapping initial fragments.
301         for (auto I = Entries.begin(); I != Entries.end(); ++I) {
302           if (!I->isDbgValue())
303             continue;
304           const DIExpression *Fragment = I->getInstr()->getDebugExpression();
305           if (std::any_of(Entries.begin(), I,
306                           [&](DbgValueHistoryMap::Entry Pred) {
307                             return Pred.isDbgValue() &&
308                                    Fragment->fragmentsOverlap(
309                                        Pred.getInstr()->getDebugExpression());
310                           }))
311             break;
312           // The code that generates location lists for DWARF assumes that the
313           // entries' start labels are monotonically increasing, and since we
314           // don't change the label for fragments that are described by
315           // registers, we must bail out when encountering such a fragment.
316           if (IsDescribedByReg(I->getInstr()))
317             break;
318           LabelsBeforeInsn[I->getInstr()] = Asm->getFunctionBegin();
319         }
320       }
321     }
322 
323     for (const auto &Entry : Entries) {
324       if (Entry.isDbgValue())
325         requestLabelBeforeInsn(Entry.getInstr());
326       else
327         requestLabelAfterInsn(Entry.getInstr());
328     }
329   }
330 
331   // Ensure there is a symbol before DBG_LABEL.
332   for (const auto &I : DbgLabels) {
333     const MachineInstr *MI = I.second;
334     requestLabelBeforeInsn(MI);
335   }
336 
337   PrevInstLoc = DebugLoc();
338   PrevLabel = Asm->getFunctionBegin();
339   beginFunctionImpl(MF);
340 }
341 
342 void DebugHandlerBase::beginInstruction(const MachineInstr *MI) {
343   if (!Asm || !MMI->hasDebugInfo())
344     return;
345 
346   assert(CurMI == nullptr);
347   CurMI = MI;
348 
349   // Insert labels where requested.
350   DenseMap<const MachineInstr *, MCSymbol *>::iterator I =
351       LabelsBeforeInsn.find(MI);
352 
353   // No label needed.
354   if (I == LabelsBeforeInsn.end())
355     return;
356 
357   // Label already assigned.
358   if (I->second)
359     return;
360 
361   if (!PrevLabel) {
362     PrevLabel = MMI->getContext().createTempSymbol();
363     Asm->OutStreamer->emitLabel(PrevLabel);
364   }
365   I->second = PrevLabel;
366 }
367 
368 void DebugHandlerBase::endInstruction() {
369   if (!Asm || !MMI->hasDebugInfo())
370     return;
371 
372   assert(CurMI != nullptr);
373   // Don't create a new label after DBG_VALUE and other instructions that don't
374   // generate code.
375   if (!CurMI->isMetaInstruction()) {
376     PrevLabel = nullptr;
377     PrevInstBB = CurMI->getParent();
378   }
379 
380   DenseMap<const MachineInstr *, MCSymbol *>::iterator I =
381       LabelsAfterInsn.find(CurMI);
382   CurMI = nullptr;
383 
384   // No label needed.
385   if (I == LabelsAfterInsn.end())
386     return;
387 
388   // Label already assigned.
389   if (I->second)
390     return;
391 
392   // We need a label after this instruction.
393   if (!PrevLabel) {
394     PrevLabel = MMI->getContext().createTempSymbol();
395     Asm->OutStreamer->emitLabel(PrevLabel);
396   }
397   I->second = PrevLabel;
398 }
399 
400 void DebugHandlerBase::endFunction(const MachineFunction *MF) {
401   if (Asm && hasDebugInfo(MMI, MF))
402     endFunctionImpl(MF);
403   DbgValues.clear();
404   DbgLabels.clear();
405   LabelsBeforeInsn.clear();
406   LabelsAfterInsn.clear();
407   InstOrdering.clear();
408 }
409 
410 void DebugHandlerBase::beginBasicBlock(const MachineBasicBlock &MBB) {
411   if (!MBB.isBeginSection())
412     return;
413 
414   PrevLabel = MBB.getSymbol();
415 }
416 
417 void DebugHandlerBase::endBasicBlock(const MachineBasicBlock &MBB) {
418   if (!MBB.isEndSection())
419     return;
420 
421   PrevLabel = nullptr;
422 }
423