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/CodeGen/AsmPrinter.h"
16 #include "llvm/CodeGen/MachineFunction.h"
17 #include "llvm/CodeGen/MachineInstr.h"
18 #include "llvm/CodeGen/MachineModuleInfo.h"
19 #include "llvm/CodeGen/TargetSubtargetInfo.h"
20 #include "llvm/IR/DebugInfo.h"
21 #include "llvm/MC/MCStreamer.h"
22 #include "llvm/Support/CommandLine.h"
23 
24 using namespace llvm;
25 
26 #define DEBUG_TYPE "dwarfdebug"
27 
28 /// If true, we drop variable location ranges which exist entirely outside the
29 /// variable's lexical scope instruction ranges.
30 static cl::opt<bool> TrimVarLocs("trim-var-locs", cl::Hidden, cl::init(true));
31 
32 std::optional<DbgVariableLocation>
33 DbgVariableLocation::extractFromMachineInstruction(
34     const MachineInstr &Instruction) {
35   DbgVariableLocation Location;
36   // Variables calculated from multiple locations can't be represented here.
37   if (Instruction.getNumDebugOperands() != 1)
38     return std::nullopt;
39   if (!Instruction.getDebugOperand(0).isReg())
40     return std::nullopt;
41   Location.Register = Instruction.getDebugOperand(0).getReg();
42   Location.FragmentInfo.reset();
43   // We only handle expressions generated by DIExpression::appendOffset,
44   // which doesn't require a full stack machine.
45   int64_t Offset = 0;
46   const DIExpression *DIExpr = Instruction.getDebugExpression();
47   auto Op = DIExpr->expr_op_begin();
48   // We can handle a DBG_VALUE_LIST iff it has exactly one location operand that
49   // appears exactly once at the start of the expression.
50   if (Instruction.isDebugValueList()) {
51     if (Instruction.getNumDebugOperands() == 1 &&
52         Op->getOp() == dwarf::DW_OP_LLVM_arg)
53       ++Op;
54     else
55       return std::nullopt;
56   }
57   while (Op != DIExpr->expr_op_end()) {
58     switch (Op->getOp()) {
59     case dwarf::DW_OP_constu: {
60       int Value = Op->getArg(0);
61       ++Op;
62       if (Op != DIExpr->expr_op_end()) {
63         switch (Op->getOp()) {
64         case dwarf::DW_OP_minus:
65           Offset -= Value;
66           break;
67         case dwarf::DW_OP_plus:
68           Offset += Value;
69           break;
70         default:
71           continue;
72         }
73       }
74     } break;
75     case dwarf::DW_OP_plus_uconst:
76       Offset += Op->getArg(0);
77       break;
78     case dwarf::DW_OP_LLVM_fragment:
79       Location.FragmentInfo = {Op->getArg(1), Op->getArg(0)};
80       break;
81     case dwarf::DW_OP_deref:
82       Location.LoadChain.push_back(Offset);
83       Offset = 0;
84       break;
85     default:
86       return std::nullopt;
87     }
88     ++Op;
89   }
90 
91   // Do one final implicit DW_OP_deref if this was an indirect DBG_VALUE
92   // instruction.
93   // FIXME: Replace these with DIExpression.
94   if (Instruction.isIndirectDebugValue())
95     Location.LoadChain.push_back(Offset);
96 
97   return Location;
98 }
99 
100 DebugHandlerBase::DebugHandlerBase(AsmPrinter *A) : Asm(A), MMI(Asm->MMI) {}
101 
102 void DebugHandlerBase::beginModule(Module *M) {
103   if (M->debug_compile_units().empty())
104     Asm = nullptr;
105 }
106 
107 // Each LexicalScope has first instruction and last instruction to mark
108 // beginning and end of a scope respectively. Create an inverse map that list
109 // scopes starts (and ends) with an instruction. One instruction may start (or
110 // end) multiple scopes. Ignore scopes that are not reachable.
111 void DebugHandlerBase::identifyScopeMarkers() {
112   SmallVector<LexicalScope *, 4> WorkList;
113   WorkList.push_back(LScopes.getCurrentFunctionScope());
114   while (!WorkList.empty()) {
115     LexicalScope *S = WorkList.pop_back_val();
116 
117     const SmallVectorImpl<LexicalScope *> &Children = S->getChildren();
118     if (!Children.empty())
119       WorkList.append(Children.begin(), Children.end());
120 
121     if (S->isAbstractScope())
122       continue;
123 
124     for (const InsnRange &R : S->getRanges()) {
125       assert(R.first && "InsnRange does not have first instruction!");
126       assert(R.second && "InsnRange does not have second instruction!");
127       requestLabelBeforeInsn(R.first);
128       requestLabelAfterInsn(R.second);
129     }
130   }
131 }
132 
133 // Return Label preceding the instruction.
134 MCSymbol *DebugHandlerBase::getLabelBeforeInsn(const MachineInstr *MI) {
135   MCSymbol *Label = LabelsBeforeInsn.lookup(MI);
136   assert(Label && "Didn't insert label before instruction");
137   return Label;
138 }
139 
140 // Return Label immediately following the instruction.
141 MCSymbol *DebugHandlerBase::getLabelAfterInsn(const MachineInstr *MI) {
142   return LabelsAfterInsn.lookup(MI);
143 }
144 
145 /// If this type is derived from a base type then return base type size.
146 uint64_t DebugHandlerBase::getBaseTypeSize(const DIType *Ty) {
147   assert(Ty);
148   const DIDerivedType *DDTy = dyn_cast<DIDerivedType>(Ty);
149   if (!DDTy)
150     return Ty->getSizeInBits();
151 
152   unsigned Tag = DDTy->getTag();
153 
154   if (Tag != dwarf::DW_TAG_member && Tag != dwarf::DW_TAG_typedef &&
155       Tag != dwarf::DW_TAG_const_type && Tag != dwarf::DW_TAG_volatile_type &&
156       Tag != dwarf::DW_TAG_restrict_type && Tag != dwarf::DW_TAG_atomic_type &&
157       Tag != dwarf::DW_TAG_immutable_type)
158     return DDTy->getSizeInBits();
159 
160   DIType *BaseType = DDTy->getBaseType();
161 
162   if (!BaseType)
163     return 0;
164 
165   // If this is a derived type, go ahead and get the base type, unless it's a
166   // reference then it's just the size of the field. Pointer types have no need
167   // of this since they're a different type of qualification on the type.
168   if (BaseType->getTag() == dwarf::DW_TAG_reference_type ||
169       BaseType->getTag() == dwarf::DW_TAG_rvalue_reference_type)
170     return Ty->getSizeInBits();
171 
172   return getBaseTypeSize(BaseType);
173 }
174 
175 bool DebugHandlerBase::isUnsignedDIType(const DIType *Ty) {
176   if (isa<DIStringType>(Ty)) {
177     // Some transformations (e.g. instcombine) may decide to turn a Fortran
178     // character object into an integer, and later ones (e.g. SROA) may
179     // further inject a constant integer in a llvm.dbg.value call to track
180     // the object's value. Here we trust the transformations are doing the
181     // right thing, and treat the constant as unsigned to preserve that value
182     // (i.e. avoid sign extension).
183     return true;
184   }
185 
186   if (auto *CTy = dyn_cast<DICompositeType>(Ty)) {
187     if (CTy->getTag() == dwarf::DW_TAG_enumeration_type) {
188       if (!(Ty = CTy->getBaseType()))
189         // FIXME: Enums without a fixed underlying type have unknown signedness
190         // here, leading to incorrectly emitted constants.
191         return false;
192     } else
193       // (Pieces of) aggregate types that get hacked apart by SROA may be
194       // represented by a constant. Encode them as unsigned bytes.
195       return true;
196   }
197 
198   if (auto *DTy = dyn_cast<DIDerivedType>(Ty)) {
199     dwarf::Tag T = (dwarf::Tag)Ty->getTag();
200     // Encode pointer constants as unsigned bytes. This is used at least for
201     // null pointer constant emission.
202     // FIXME: reference and rvalue_reference /probably/ shouldn't be allowed
203     // here, but accept them for now due to a bug in SROA producing bogus
204     // dbg.values.
205     if (T == dwarf::DW_TAG_pointer_type ||
206         T == dwarf::DW_TAG_ptr_to_member_type ||
207         T == dwarf::DW_TAG_reference_type ||
208         T == dwarf::DW_TAG_rvalue_reference_type)
209       return true;
210     assert(T == dwarf::DW_TAG_typedef || T == dwarf::DW_TAG_const_type ||
211            T == dwarf::DW_TAG_volatile_type ||
212            T == dwarf::DW_TAG_restrict_type || T == dwarf::DW_TAG_atomic_type ||
213            T == dwarf::DW_TAG_immutable_type);
214     assert(DTy->getBaseType() && "Expected valid base type");
215     return isUnsignedDIType(DTy->getBaseType());
216   }
217 
218   auto *BTy = cast<DIBasicType>(Ty);
219   unsigned Encoding = BTy->getEncoding();
220   assert((Encoding == dwarf::DW_ATE_unsigned ||
221           Encoding == dwarf::DW_ATE_unsigned_char ||
222           Encoding == dwarf::DW_ATE_signed ||
223           Encoding == dwarf::DW_ATE_signed_char ||
224           Encoding == dwarf::DW_ATE_float || Encoding == dwarf::DW_ATE_UTF ||
225           Encoding == dwarf::DW_ATE_boolean ||
226           Encoding == dwarf::DW_ATE_complex_float ||
227           (Ty->getTag() == dwarf::DW_TAG_unspecified_type &&
228            Ty->getName() == "decltype(nullptr)")) &&
229          "Unsupported encoding");
230   return Encoding == dwarf::DW_ATE_unsigned ||
231          Encoding == dwarf::DW_ATE_unsigned_char ||
232          Encoding == dwarf::DW_ATE_UTF || Encoding == dwarf::DW_ATE_boolean ||
233          Ty->getTag() == dwarf::DW_TAG_unspecified_type;
234 }
235 
236 static bool hasDebugInfo(const MachineModuleInfo *MMI,
237                          const MachineFunction *MF) {
238   if (!MMI->hasDebugInfo())
239     return false;
240   auto *SP = MF->getFunction().getSubprogram();
241   if (!SP)
242     return false;
243   assert(SP->getUnit());
244   auto EK = SP->getUnit()->getEmissionKind();
245   if (EK == DICompileUnit::NoDebug)
246     return false;
247   return true;
248 }
249 
250 void DebugHandlerBase::beginFunction(const MachineFunction *MF) {
251   PrevInstBB = nullptr;
252 
253   if (!Asm || !hasDebugInfo(MMI, MF)) {
254     skippedNonDebugFunction();
255     return;
256   }
257 
258   // Grab the lexical scopes for the function, if we don't have any of those
259   // then we're not going to be able to do anything.
260   LScopes.initialize(*MF);
261   if (LScopes.empty()) {
262     beginFunctionImpl(MF);
263     return;
264   }
265 
266   // Make sure that each lexical scope will have a begin/end label.
267   identifyScopeMarkers();
268 
269   // Calculate history for local variables.
270   assert(DbgValues.empty() && "DbgValues map wasn't cleaned!");
271   assert(DbgLabels.empty() && "DbgLabels map wasn't cleaned!");
272   calculateDbgEntityHistory(MF, Asm->MF->getSubtarget().getRegisterInfo(),
273                             DbgValues, DbgLabels);
274   InstOrdering.initialize(*MF);
275   if (TrimVarLocs)
276     DbgValues.trimLocationRanges(*MF, LScopes, InstOrdering);
277   LLVM_DEBUG(DbgValues.dump(MF->getName()));
278 
279   // Request labels for the full history.
280   for (const auto &I : DbgValues) {
281     const auto &Entries = I.second;
282     if (Entries.empty())
283       continue;
284 
285     auto IsDescribedByReg = [](const MachineInstr *MI) {
286       return any_of(MI->debug_operands(),
287                     [](auto &MO) { return MO.isReg() && MO.getReg(); });
288     };
289 
290     // The first mention of a function argument gets the CurrentFnBegin label,
291     // so arguments are visible when breaking at function entry.
292     //
293     // We do not change the label for values that are described by registers,
294     // as that could place them above their defining instructions. We should
295     // ideally not change the labels for constant debug values either, since
296     // doing that violates the ranges that are calculated in the history map.
297     // However, we currently do not emit debug values for constant arguments
298     // directly at the start of the function, so this code is still useful.
299     const DILocalVariable *DIVar =
300         Entries.front().getInstr()->getDebugVariable();
301     if (DIVar->isParameter() &&
302         getDISubprogram(DIVar->getScope())->describes(&MF->getFunction())) {
303       if (!IsDescribedByReg(Entries.front().getInstr()))
304         LabelsBeforeInsn[Entries.front().getInstr()] = Asm->getFunctionBegin();
305       if (Entries.front().getInstr()->getDebugExpression()->isFragment()) {
306         // Mark all non-overlapping initial fragments.
307         for (const auto *I = Entries.begin(); I != Entries.end(); ++I) {
308           if (!I->isDbgValue())
309             continue;
310           const DIExpression *Fragment = I->getInstr()->getDebugExpression();
311           if (std::any_of(Entries.begin(), I,
312                           [&](DbgValueHistoryMap::Entry Pred) {
313                             return Pred.isDbgValue() &&
314                                    Fragment->fragmentsOverlap(
315                                        Pred.getInstr()->getDebugExpression());
316                           }))
317             break;
318           // The code that generates location lists for DWARF assumes that the
319           // entries' start labels are monotonically increasing, and since we
320           // don't change the label for fragments that are described by
321           // registers, we must bail out when encountering such a fragment.
322           if (IsDescribedByReg(I->getInstr()))
323             break;
324           LabelsBeforeInsn[I->getInstr()] = Asm->getFunctionBegin();
325         }
326       }
327     }
328 
329     for (const auto &Entry : Entries) {
330       if (Entry.isDbgValue())
331         requestLabelBeforeInsn(Entry.getInstr());
332       else
333         requestLabelAfterInsn(Entry.getInstr());
334     }
335   }
336 
337   // Ensure there is a symbol before DBG_LABEL.
338   for (const auto &I : DbgLabels) {
339     const MachineInstr *MI = I.second;
340     requestLabelBeforeInsn(MI);
341   }
342 
343   PrevInstLoc = DebugLoc();
344   PrevLabel = Asm->getFunctionBegin();
345   beginFunctionImpl(MF);
346 }
347 
348 void DebugHandlerBase::beginInstruction(const MachineInstr *MI) {
349   if (!Asm || !MMI->hasDebugInfo())
350     return;
351 
352   assert(CurMI == nullptr);
353   CurMI = MI;
354 
355   // Insert labels where requested.
356   DenseMap<const MachineInstr *, MCSymbol *>::iterator I =
357       LabelsBeforeInsn.find(MI);
358 
359   // No label needed.
360   if (I == LabelsBeforeInsn.end())
361     return;
362 
363   // Label already assigned.
364   if (I->second)
365     return;
366 
367   if (!PrevLabel) {
368     PrevLabel = MMI->getContext().createTempSymbol();
369     Asm->OutStreamer->emitLabel(PrevLabel);
370   }
371   I->second = PrevLabel;
372 }
373 
374 void DebugHandlerBase::endInstruction() {
375   if (!Asm || !MMI->hasDebugInfo())
376     return;
377 
378   assert(CurMI != nullptr);
379   // Don't create a new label after DBG_VALUE and other instructions that don't
380   // generate code.
381   if (!CurMI->isMetaInstruction()) {
382     PrevLabel = nullptr;
383     PrevInstBB = CurMI->getParent();
384   }
385 
386   DenseMap<const MachineInstr *, MCSymbol *>::iterator I =
387       LabelsAfterInsn.find(CurMI);
388 
389   // No label needed or label already assigned.
390   if (I == LabelsAfterInsn.end() || I->second) {
391     CurMI = nullptr;
392     return;
393   }
394 
395   // We need a label after this instruction.  With basic block sections, just
396   // use the end symbol of the section if this is the last instruction of the
397   // section.  This reduces the need for an additional label and also helps
398   // merging ranges.
399   if (CurMI->getParent()->isEndSection() && CurMI->getNextNode() == nullptr) {
400     PrevLabel = CurMI->getParent()->getEndSymbol();
401   } else if (!PrevLabel) {
402     PrevLabel = MMI->getContext().createTempSymbol();
403     Asm->OutStreamer->emitLabel(PrevLabel);
404   }
405   I->second = PrevLabel;
406   CurMI = nullptr;
407 }
408 
409 void DebugHandlerBase::endFunction(const MachineFunction *MF) {
410   if (Asm && hasDebugInfo(MMI, MF))
411     endFunctionImpl(MF);
412   DbgValues.clear();
413   DbgLabels.clear();
414   LabelsBeforeInsn.clear();
415   LabelsAfterInsn.clear();
416   InstOrdering.clear();
417 }
418 
419 void DebugHandlerBase::beginBasicBlockSection(const MachineBasicBlock &MBB) {
420   EpilogBeginBlock = nullptr;
421   if (!MBB.isEntryBlock())
422     PrevLabel = MBB.getSymbol();
423 }
424 
425 void DebugHandlerBase::endBasicBlockSection(const MachineBasicBlock &MBB) {
426   PrevLabel = nullptr;
427 }
428