1//=== DWARFLinker.cpp -----------------------------------------------------===//
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#include "llvm/DWARFLinker/Classic/DWARFLinker.h"
10#include "llvm/ADT/ArrayRef.h"
11#include "llvm/ADT/BitVector.h"
12#include "llvm/ADT/STLExtras.h"
13#include "llvm/ADT/StringExtras.h"
14#include "llvm/CodeGen/NonRelocatableStringpool.h"
15#include "llvm/DWARFLinker/Classic/DWARFLinkerDeclContext.h"
16#include "llvm/DWARFLinker/Classic/DWARFStreamer.h"
17#include "llvm/DWARFLinker/Utils.h"
18#include "llvm/DebugInfo/DWARF/DWARFAbbreviationDeclaration.h"
19#include "llvm/DebugInfo/DWARF/DWARFAcceleratorTable.h"
20#include "llvm/DebugInfo/DWARF/DWARFContext.h"
21#include "llvm/DebugInfo/DWARF/DWARFDataExtractor.h"
22#include "llvm/DebugInfo/DWARF/DWARFDebugLine.h"
23#include "llvm/DebugInfo/DWARF/DWARFDebugMacro.h"
24#include "llvm/DebugInfo/DWARF/DWARFDebugRangeList.h"
25#include "llvm/DebugInfo/DWARF/DWARFDie.h"
26#include "llvm/DebugInfo/DWARF/DWARFFormValue.h"
27#include "llvm/DebugInfo/DWARF/DWARFSection.h"
28#include "llvm/DebugInfo/DWARF/DWARFUnit.h"
29#include "llvm/DebugInfo/DWARF/LowLevel/DWARFExpression.h"
30#include "llvm/MC/MCDwarf.h"
31#include "llvm/Support/DataExtractor.h"
32#include "llvm/Support/Error.h"
33#include "llvm/Support/ErrorHandling.h"
34#include "llvm/Support/ErrorOr.h"
35#include "llvm/Support/FormatVariadic.h"
36#include "llvm/Support/LEB128.h"
37#include "llvm/Support/Path.h"
38#include "llvm/Support/ThreadPool.h"
39#include <vector>
40
41namespace llvm {
42
43using namespace dwarf_linker;
44using namespace dwarf_linker::classic;
45
46/// Hold the input and output of the debug info size in bytes.
47struct DebugInfoSize {
48 uint64_t Input;
49 uint64_t Output;
50};
51
52/// Compute the total size of the debug info.
53static uint64_t getDebugInfoSize(DWARFContext &Dwarf) {
54 uint64_t Size = 0;
55 for (auto &Unit : Dwarf.compile_units()) {
56 Size += Unit->getLength();
57 }
58 return Size;
59}
60
61/// Similar to DWARFUnitSection::getUnitForOffset(), but returning our
62/// CompileUnit object instead.
63static CompileUnit *getUnitForOffset(const UnitListTy &Units, uint64_t Offset) {
64 auto CU = llvm::upper_bound(
65 Range: Units, Value&: Offset, C: [](uint64_t LHS, const std::unique_ptr<CompileUnit> &RHS) {
66 return LHS < RHS->getOrigUnit().getNextUnitOffset();
67 });
68 return CU != Units.end() ? CU->get() : nullptr;
69}
70
71/// Resolve the DIE attribute reference that has been extracted in \p RefValue.
72/// The resulting DIE might be in another CompileUnit which is stored into \p
73/// ReferencedCU. \returns null if resolving fails for any reason.
74DWARFDie DWARFLinker::resolveDIEReference(const DWARFFile &File,
75 const UnitListTy &Units,
76 const DWARFFormValue &RefValue,
77 const DWARFDie &DIE,
78 CompileUnit *&RefCU) {
79 assert(RefValue.isFormClass(DWARFFormValue::FC_Reference));
80 uint64_t RefOffset;
81 if (std::optional<uint64_t> Off = RefValue.getAsRelativeReference()) {
82 RefOffset = RefValue.getUnit()->getOffset() + *Off;
83 } else if (Off = RefValue.getAsDebugInfoReference(); Off) {
84 RefOffset = *Off;
85 } else {
86 reportWarning(Warning: "Unsupported reference type", File, DIE: &DIE);
87 return DWARFDie();
88 }
89 if ((RefCU = getUnitForOffset(Units, Offset: RefOffset)))
90 if (const auto RefDie = RefCU->getOrigUnit().getDIEForOffset(Offset: RefOffset)) {
91 // In a file with broken references, an attribute might point to a NULL
92 // DIE.
93 if (!RefDie.isNULL())
94 return RefDie;
95 }
96
97 reportWarning(Warning: "could not find referenced DIE", File, DIE: &DIE);
98 return DWARFDie();
99}
100
101/// \returns whether the passed \a Attr type might contain a DIE reference
102/// suitable for ODR uniquing.
103static bool isODRAttribute(uint16_t Attr) {
104 switch (Attr) {
105 default:
106 return false;
107 case dwarf::DW_AT_type:
108 case dwarf::DW_AT_containing_type:
109 case dwarf::DW_AT_specification:
110 case dwarf::DW_AT_abstract_origin:
111 case dwarf::DW_AT_import:
112 case dwarf::DW_AT_LLVM_alloc_type:
113 return true;
114 }
115 llvm_unreachable("Improper attribute.");
116}
117
118static bool isTypeTag(uint16_t Tag) {
119 switch (Tag) {
120 case dwarf::DW_TAG_array_type:
121 case dwarf::DW_TAG_class_type:
122 case dwarf::DW_TAG_enumeration_type:
123 case dwarf::DW_TAG_pointer_type:
124 case dwarf::DW_TAG_reference_type:
125 case dwarf::DW_TAG_string_type:
126 case dwarf::DW_TAG_structure_type:
127 case dwarf::DW_TAG_subroutine_type:
128 case dwarf::DW_TAG_template_alias:
129 case dwarf::DW_TAG_typedef:
130 case dwarf::DW_TAG_union_type:
131 case dwarf::DW_TAG_ptr_to_member_type:
132 case dwarf::DW_TAG_set_type:
133 case dwarf::DW_TAG_subrange_type:
134 case dwarf::DW_TAG_base_type:
135 case dwarf::DW_TAG_const_type:
136 case dwarf::DW_TAG_constant:
137 case dwarf::DW_TAG_file_type:
138 case dwarf::DW_TAG_namelist:
139 case dwarf::DW_TAG_packed_type:
140 case dwarf::DW_TAG_volatile_type:
141 case dwarf::DW_TAG_restrict_type:
142 case dwarf::DW_TAG_atomic_type:
143 case dwarf::DW_TAG_interface_type:
144 case dwarf::DW_TAG_unspecified_type:
145 case dwarf::DW_TAG_shared_type:
146 case dwarf::DW_TAG_immutable_type:
147 return true;
148 default:
149 break;
150 }
151 return false;
152}
153
154/// Recurse through the input DIE's canonical references until we find a
155/// DW_AT_name.
156llvm::StringRef
157DWARFLinker::DIECloner::getCanonicalDIEName(DWARFDie Die, const DWARFFile &File,
158 CompileUnit *Unit) {
159 if (!Die)
160 return {};
161
162 std::optional<DWARFFormValue> Ref;
163
164 auto GetDieName = [](const DWARFDie &D) -> llvm::StringRef {
165 auto NameForm = D.find(Attr: llvm::dwarf::DW_AT_name);
166 if (!NameForm)
167 return {};
168
169 auto NameOrErr = NameForm->getAsCString();
170 if (!NameOrErr) {
171 llvm::consumeError(Err: NameOrErr.takeError());
172 return {};
173 }
174
175 return *NameOrErr;
176 };
177
178 llvm::StringRef Name = GetDieName(Die);
179 if (!Name.empty())
180 return Name;
181
182 while (true) {
183 if (!(Ref = Die.find(Attr: llvm::dwarf::DW_AT_specification)) &&
184 !(Ref = Die.find(Attr: llvm::dwarf::DW_AT_abstract_origin)))
185 break;
186
187 Die = Linker.resolveDIEReference(File, Units: CompileUnits, RefValue: *Ref, DIE: Die, RefCU&: Unit);
188 if (!Die)
189 break;
190
191 assert(Unit);
192
193 unsigned SpecIdx = Unit->getOrigUnit().getDIEIndex(D: Die);
194 CompileUnit::DIEInfo &SpecInfo = Unit->getInfo(Idx: SpecIdx);
195 if (SpecInfo.Ctxt && SpecInfo.Ctxt->hasCanonicalDIE()) {
196 if (!SpecInfo.Ctxt->getCanonicalName().empty()) {
197 Name = SpecInfo.Ctxt->getCanonicalName();
198 break;
199 }
200 }
201
202 Name = GetDieName(Die);
203 if (!Name.empty())
204 break;
205 }
206
207 return Name;
208}
209
210bool DWARFLinker::DIECloner::getDIENames(
211 const DWARFDie &Die, AttributesInfo &Info, OffsetsStringPool &StringPool,
212 const DWARFFile &File, CompileUnit &Unit, bool StripTemplate) {
213 // This function will be called on DIEs having low_pcs and
214 // ranges. As getting the name might be more expansive, filter out
215 // blocks directly.
216 if (Die.getTag() == dwarf::DW_TAG_lexical_block)
217 return false;
218
219 // The mangled name of an specification DIE will by virtue of the
220 // uniquing algorithm be the same as the one it got uniqued into.
221 // So just use the input DIE's linkage name.
222 if (!Info.MangledName)
223 if (const char *MangledName = Die.getLinkageName())
224 Info.MangledName = StringPool.getEntry(S: MangledName);
225
226 // For subprograms with linkage names, we unique on the linkage name,
227 // so DW_AT_name's may differ between the input and canonical DIEs.
228 // Use the name of the canonical DIE.
229 if (!Info.Name)
230 if (llvm::StringRef Name = getCanonicalDIEName(Die, File, Unit: &Unit);
231 !Name.empty())
232 Info.Name = StringPool.getEntry(S: Name);
233
234 if (!Info.MangledName)
235 Info.MangledName = Info.Name;
236
237 if (StripTemplate && Info.Name && Info.MangledName != Info.Name) {
238 StringRef Name = Info.Name.getString();
239 if (std::optional<StringRef> StrippedName = StripTemplateParameters(Name))
240 Info.NameWithoutTemplate = StringPool.getEntry(S: *StrippedName);
241 }
242
243 return Info.Name || Info.MangledName;
244}
245
246/// Resolve the relative path to a build artifact referenced by DWARF by
247/// applying DW_AT_comp_dir.
248static void resolveRelativeObjectPath(SmallVectorImpl<char> &Buf, DWARFDie CU) {
249 sys::path::append(path&: Buf, a: dwarf::toString(V: CU.find(Attr: dwarf::DW_AT_comp_dir), Default: ""));
250}
251
252/// Collect references to parseable Swift interfaces in imported
253/// DW_TAG_module blocks.
254static void analyzeImportedModule(
255 const DWARFDie &DIE, CompileUnit &CU,
256 DWARFLinkerBase::SwiftInterfacesMapTy *ParseableSwiftInterfaces,
257 std::function<void(const Twine &, const DWARFDie &)> ReportWarning) {
258 if (CU.getLanguage() != dwarf::DW_LANG_Swift)
259 return;
260
261 if (!ParseableSwiftInterfaces)
262 return;
263
264 StringRef Path = dwarf::toStringRef(V: DIE.find(Attr: dwarf::DW_AT_LLVM_include_path));
265 if (!Path.ends_with(Suffix: ".swiftinterface"))
266 return;
267 // Don't track interfaces that are part of the SDK.
268 StringRef SysRoot = dwarf::toStringRef(V: DIE.find(Attr: dwarf::DW_AT_LLVM_sysroot));
269 if (SysRoot.empty())
270 SysRoot = CU.getSysRoot();
271 if (!SysRoot.empty() && Path.starts_with(Prefix: SysRoot))
272 return;
273 // Don't track interfaces that are part of the toolchain.
274 // For example: Swift, _Concurrency, ...
275 StringRef DeveloperDir = guessDeveloperDir(SysRoot);
276 if (!DeveloperDir.empty() && Path.starts_with(Prefix: DeveloperDir))
277 return;
278 if (isInToolchainDir(Path))
279 return;
280 std::optional<const char *> Name =
281 dwarf::toString(V: DIE.find(Attr: dwarf::DW_AT_name));
282 if (!Name)
283 return;
284 auto &Entry = (*ParseableSwiftInterfaces)[*Name];
285 // The prepend path is applied later when copying.
286 DWARFDie CUDie = CU.getOrigUnit().getUnitDIE();
287 SmallString<128> ResolvedPath;
288 if (sys::path::is_relative(path: Path))
289 resolveRelativeObjectPath(Buf&: ResolvedPath, CU: CUDie);
290 sys::path::append(path&: ResolvedPath, a: Path);
291 if (!Entry.empty() && Entry != ResolvedPath)
292 ReportWarning(Twine("Conflicting parseable interfaces for Swift Module ") +
293 *Name + ": " + Entry + " and " + Path,
294 DIE);
295 Entry = std::string(ResolvedPath);
296}
297
298/// The distinct types of work performed by the work loop in
299/// analyzeContextInfo.
300enum class ContextWorklistItemType : uint8_t {
301 AnalyzeContextInfo,
302 UpdateChildPruning,
303 UpdatePruning,
304};
305
306/// This class represents an item in the work list. The type defines what kind
307/// of work needs to be performed when processing the current item. Everything
308/// but the Type and Die fields are optional based on the type.
309struct ContextWorklistItem {
310 DWARFDie Die;
311 unsigned ParentIdx;
312 union {
313 CompileUnit::DIEInfo *OtherInfo;
314 DeclContext *Context;
315 };
316 ContextWorklistItemType Type;
317 bool InImportedModule;
318
319 ContextWorklistItem(DWARFDie Die, ContextWorklistItemType T,
320 CompileUnit::DIEInfo *OtherInfo = nullptr)
321 : Die(Die), ParentIdx(0), OtherInfo(OtherInfo), Type(T),
322 InImportedModule(false) {}
323
324 ContextWorklistItem(DWARFDie Die, DeclContext *Context, unsigned ParentIdx,
325 bool InImportedModule)
326 : Die(Die), ParentIdx(ParentIdx), Context(Context),
327 Type(ContextWorklistItemType::AnalyzeContextInfo),
328 InImportedModule(InImportedModule) {}
329};
330
331static bool updatePruning(const DWARFDie &Die, CompileUnit &CU,
332 uint64_t ModulesEndOffset) {
333 CompileUnit::DIEInfo &Info = CU.getInfo(Die);
334
335 // Prune this DIE if it is either a forward declaration inside a
336 // DW_TAG_module or a DW_TAG_module that contains nothing but
337 // forward declarations.
338 Info.Prune &= (Die.getTag() == dwarf::DW_TAG_module) ||
339 (isTypeTag(Tag: Die.getTag()) &&
340 dwarf::toUnsigned(V: Die.find(Attr: dwarf::DW_AT_declaration), Default: 0));
341
342 // Only prune forward declarations inside a DW_TAG_module for which a
343 // definition exists elsewhere.
344 if (ModulesEndOffset == 0)
345 Info.Prune &= Info.Ctxt && Info.Ctxt->getCanonicalDIEOffset();
346 else
347 Info.Prune &= Info.Ctxt && Info.Ctxt->getCanonicalDIEOffset() > 0 &&
348 Info.Ctxt->getCanonicalDIEOffset() <= ModulesEndOffset;
349
350 return Info.Prune;
351}
352
353static void updateChildPruning(const DWARFDie &Die, CompileUnit &CU,
354 CompileUnit::DIEInfo &ChildInfo) {
355 CompileUnit::DIEInfo &Info = CU.getInfo(Die);
356 Info.Prune &= ChildInfo.Prune;
357}
358
359/// Recursive helper to build the global DeclContext information and
360/// gather the child->parent relationships in the original compile unit.
361///
362/// This function uses the same work list approach as lookForDIEsToKeep.
363///
364/// \return true when this DIE and all of its children are only
365/// forward declarations to types defined in external clang modules
366/// (i.e., forward declarations that are children of a DW_TAG_module).
367static void analyzeContextInfo(
368 const DWARFDie &DIE, unsigned ParentIdx, CompileUnit &CU,
369 DeclContext *CurrentDeclContext, DeclContextTree &Contexts,
370 uint64_t ModulesEndOffset,
371 DWARFLinkerBase::SwiftInterfacesMapTy *ParseableSwiftInterfaces,
372 std::function<void(const Twine &, const DWARFDie &)> ReportWarning) {
373 // LIFO work list.
374 std::vector<ContextWorklistItem> Worklist;
375 Worklist.emplace_back(args: DIE, args&: CurrentDeclContext, args&: ParentIdx, args: false);
376
377 while (!Worklist.empty()) {
378 ContextWorklistItem Current = Worklist.back();
379 Worklist.pop_back();
380
381 switch (Current.Type) {
382 case ContextWorklistItemType::UpdatePruning:
383 updatePruning(Die: Current.Die, CU, ModulesEndOffset);
384 continue;
385 case ContextWorklistItemType::UpdateChildPruning:
386 updateChildPruning(Die: Current.Die, CU, ChildInfo&: *Current.OtherInfo);
387 continue;
388 case ContextWorklistItemType::AnalyzeContextInfo:
389 break;
390 }
391
392 unsigned Idx = CU.getOrigUnit().getDIEIndex(D: Current.Die);
393 CompileUnit::DIEInfo &Info = CU.getInfo(Idx);
394
395 // Clang imposes an ODR on modules(!) regardless of the language:
396 // "The module-id should consist of only a single identifier,
397 // which provides the name of the module being defined. Each
398 // module shall have a single definition."
399 //
400 // This does not extend to the types inside the modules:
401 // "[I]n C, this implies that if two structs are defined in
402 // different submodules with the same name, those two types are
403 // distinct types (but may be compatible types if their
404 // definitions match)."
405 //
406 // We treat non-C++ modules like namespaces for this reason.
407 if (Current.Die.getTag() == dwarf::DW_TAG_module &&
408 Current.ParentIdx == 0 &&
409 dwarf::toString(V: Current.Die.find(Attr: dwarf::DW_AT_name), Default: "") !=
410 CU.getClangModuleName()) {
411 Current.InImportedModule = true;
412 analyzeImportedModule(DIE: Current.Die, CU, ParseableSwiftInterfaces,
413 ReportWarning);
414 }
415
416 Info.ParentIdx = Current.ParentIdx;
417 Info.InModuleScope = CU.isClangModule() || Current.InImportedModule;
418 if (CU.hasODR() || Info.InModuleScope) {
419 if (Current.Context) {
420 auto PtrInvalidPair = Contexts.getChildDeclContext(
421 Context&: *Current.Context, DIE: Current.Die, Unit&: CU, InClangModule: Info.InModuleScope);
422 Current.Context = PtrInvalidPair.getPointer();
423 Info.Ctxt =
424 PtrInvalidPair.getInt() ? nullptr : PtrInvalidPair.getPointer();
425 if (Info.Ctxt)
426 Info.Ctxt->setDefinedInClangModule(Info.InModuleScope);
427 } else
428 Info.Ctxt = Current.Context = nullptr;
429 }
430
431 Info.Prune = Current.InImportedModule;
432 // Add children in reverse order to the worklist to effectively process
433 // them in order.
434 Worklist.emplace_back(args&: Current.Die, args: ContextWorklistItemType::UpdatePruning);
435 for (auto Child : reverse(C: Current.Die.children())) {
436 CompileUnit::DIEInfo &ChildInfo = CU.getInfo(Die: Child);
437 Worklist.emplace_back(
438 args&: Current.Die, args: ContextWorklistItemType::UpdateChildPruning, args: &ChildInfo);
439 Worklist.emplace_back(args&: Child, args&: Current.Context, args&: Idx,
440 args&: Current.InImportedModule);
441 }
442 }
443}
444
445static bool dieNeedsChildrenToBeMeaningful(uint32_t Tag) {
446 switch (Tag) {
447 default:
448 return false;
449 case dwarf::DW_TAG_class_type:
450 case dwarf::DW_TAG_common_block:
451 case dwarf::DW_TAG_enumeration_type:
452 case dwarf::DW_TAG_lexical_block:
453 case dwarf::DW_TAG_structure_type:
454 case dwarf::DW_TAG_subprogram:
455 case dwarf::DW_TAG_subroutine_type:
456 case dwarf::DW_TAG_union_type:
457 return true;
458 }
459 llvm_unreachable("Invalid Tag");
460}
461
462void DWARFLinker::cleanupAuxiliarryData(LinkContext &Context) {
463 Context.clear();
464
465 for (DIEBlock *I : DIEBlocks)
466 I->~DIEBlock();
467 for (DIELoc *I : DIELocs)
468 I->~DIELoc();
469
470 DIEBlocks.clear();
471 DIELocs.clear();
472 DIEAlloc.Reset();
473}
474
475static void constructSeqOffsettoOrigRowMapping(
476 CompileUnit &Unit, const DWARFDebugLine::LineTable &LT,
477 DenseMap<uint64_t, uint64_t> &SeqOffToOrigRow) {
478 // Collect this unit's DW_AT_LLVM_stmt_sequence attribute values
479 // (input offsets), sorted ascending and deduplicated, to drive the
480 // shared mapping builder.
481 auto StmtAttrs = Unit.getStmtSeqListAttributes();
482 SmallVector<uint64_t> SortedOffsets;
483 SortedOffsets.reserve(N: StmtAttrs.size());
484 for (const PatchLocation &P : StmtAttrs)
485 SortedOffsets.push_back(Elt: P.get());
486 llvm::sort(C&: SortedOffsets);
487 SortedOffsets.erase(CS: llvm::unique(R&: SortedOffsets), CE: SortedOffsets.end());
488
489 dwarf_linker::buildStmtSeqOffsetToFirstRowIndex(LT, SortedStmtSeqOffsets: SortedOffsets,
490 SeqOffToFirstRow&: SeqOffToOrigRow);
491}
492
493std::pair<bool, std::optional<int64_t>>
494DWARFLinker::getVariableRelocAdjustment(AddressesMap &RelocMgr,
495 const DWARFDie &DIE) {
496 assert((DIE.getTag() == dwarf::DW_TAG_variable ||
497 DIE.getTag() == dwarf::DW_TAG_constant) &&
498 "Wrong type of input die");
499
500 const auto *Abbrev = DIE.getAbbreviationDeclarationPtr();
501
502 // Check if DIE has DW_AT_location attribute.
503 DWARFUnit *U = DIE.getDwarfUnit();
504 std::optional<uint32_t> LocationIdx =
505 Abbrev->findAttributeIndex(attr: dwarf::DW_AT_location);
506 if (!LocationIdx)
507 return std::make_pair(x: false, y: std::nullopt);
508
509 // Get offset to the DW_AT_location attribute.
510 uint64_t AttrOffset =
511 Abbrev->getAttributeOffsetFromIndex(AttrIndex: *LocationIdx, DIEOffset: DIE.getOffset(), U: *U);
512
513 // Get value of the DW_AT_location attribute.
514 std::optional<DWARFFormValue> LocationValue =
515 Abbrev->getAttributeValueFromOffset(AttrIndex: *LocationIdx, Offset: AttrOffset, U: *U);
516 if (!LocationValue)
517 return std::make_pair(x: false, y: std::nullopt);
518
519 // Check that DW_AT_location attribute is of 'exprloc' class.
520 // Handling value of location expressions for attributes of 'loclist'
521 // class is not implemented yet.
522 std::optional<ArrayRef<uint8_t>> Expr = LocationValue->getAsBlock();
523 if (!Expr)
524 return std::make_pair(x: false, y: std::nullopt);
525
526 // Parse 'exprloc' expression.
527 DataExtractor Data(*Expr, U->getContext().isLittleEndian());
528 DWARFExpression Expression(Data, U->getAddressByteSize(),
529 U->getFormParams().Format);
530
531 bool HasLocationAddress = false;
532 uint64_t CurExprOffset = 0;
533 for (DWARFExpression::iterator It = Expression.begin();
534 It != Expression.end(); ++It) {
535 DWARFExpression::iterator NextIt = It;
536 ++NextIt;
537
538 const DWARFExpression::Operation &Op = *It;
539 switch (Op.getCode()) {
540 case dwarf::DW_OP_const2u:
541 case dwarf::DW_OP_const4u:
542 case dwarf::DW_OP_const8u:
543 case dwarf::DW_OP_const2s:
544 case dwarf::DW_OP_const4s:
545 case dwarf::DW_OP_const8s:
546 if (NextIt == Expression.end() ||
547 !dwarf::isTlsAddressOp(O: NextIt->getCode()))
548 break;
549 [[fallthrough]];
550 case dwarf::DW_OP_addr: {
551 HasLocationAddress = true;
552 // Check relocation for the address.
553 if (std::optional<int64_t> RelocAdjustment =
554 RelocMgr.getExprOpAddressRelocAdjustment(
555 U&: *U, Op, StartOffset: AttrOffset + CurExprOffset,
556 EndOffset: AttrOffset + Op.getEndOffset(), Verbose: Options.Verbose))
557 return std::make_pair(x&: HasLocationAddress, y&: *RelocAdjustment);
558 } break;
559 case dwarf::DW_OP_constx:
560 case dwarf::DW_OP_addrx: {
561 HasLocationAddress = true;
562 if (std::optional<uint64_t> AddressOffset =
563 DIE.getDwarfUnit()->getIndexedAddressOffset(
564 Index: Op.getRawOperand(Idx: 0))) {
565 // Check relocation for the address.
566 if (std::optional<int64_t> RelocAdjustment =
567 RelocMgr.getExprOpAddressRelocAdjustment(
568 U&: *U, Op, StartOffset: *AddressOffset,
569 EndOffset: *AddressOffset + DIE.getDwarfUnit()->getAddressByteSize(),
570 Verbose: Options.Verbose))
571 return std::make_pair(x&: HasLocationAddress, y&: *RelocAdjustment);
572 }
573 } break;
574 default: {
575 // Nothing to do.
576 } break;
577 }
578 CurExprOffset = Op.getEndOffset();
579 }
580
581 return std::make_pair(x&: HasLocationAddress, y: std::nullopt);
582}
583
584/// Check if a variable describing DIE should be kept.
585/// \returns updated TraversalFlags.
586unsigned DWARFLinker::shouldKeepVariableDIE(AddressesMap &RelocMgr,
587 const DWARFDie &DIE,
588 CompileUnit::DIEInfo &MyInfo,
589 unsigned Flags) {
590 const auto *Abbrev = DIE.getAbbreviationDeclarationPtr();
591
592 // Global variables with constant value can always be kept.
593 if (!(Flags & TF_InFunctionScope) &&
594 Abbrev->findAttributeIndex(attr: dwarf::DW_AT_const_value)) {
595 MyInfo.InDebugMap = true;
596 return Flags | TF_Keep;
597 }
598
599 // See if there is a relocation to a valid debug map entry inside this
600 // variable's location. The order is important here. We want to always check
601 // if the variable has a valid relocation, so that the DIEInfo is filled.
602 // However, we don't want a static variable in a function to force us to keep
603 // the enclosing function, unless requested explicitly.
604 std::pair<bool, std::optional<int64_t>> LocExprAddrAndRelocAdjustment =
605 getVariableRelocAdjustment(RelocMgr, DIE);
606
607 if (LocExprAddrAndRelocAdjustment.first)
608 MyInfo.HasLocationExpressionAddr = true;
609
610 if (!LocExprAddrAndRelocAdjustment.second)
611 return Flags;
612
613 MyInfo.AddrAdjust = *LocExprAddrAndRelocAdjustment.second;
614 MyInfo.InDebugMap = true;
615
616 if (((Flags & TF_InFunctionScope) &&
617 !LLVM_UNLIKELY(Options.KeepFunctionForStatic)))
618 return Flags;
619
620 if (Options.Verbose) {
621 outs() << "Keeping variable DIE:";
622 DIDumpOptions DumpOpts;
623 DumpOpts.ChildRecurseDepth = 0;
624 DumpOpts.Verbose = Options.Verbose;
625 DIE.dump(OS&: outs(), indent: 8 /* Indent */, DumpOpts);
626 }
627
628 return Flags | TF_Keep;
629}
630
631/// Check if a function describing DIE should be kept.
632/// \returns updated TraversalFlags.
633unsigned DWARFLinker::shouldKeepSubprogramDIE(
634 AddressesMap &RelocMgr, const DWARFDie &DIE, const DWARFFile &File,
635 CompileUnit &Unit, CompileUnit::DIEInfo &MyInfo, unsigned Flags) {
636 Flags |= TF_InFunctionScope;
637
638 auto LowPc = dwarf::toAddress(V: DIE.find(Attr: dwarf::DW_AT_low_pc));
639 if (!LowPc)
640 return Flags;
641
642 assert(LowPc && "low_pc attribute is not an address.");
643 std::optional<int64_t> RelocAdjustment =
644 RelocMgr.getSubprogramRelocAdjustment(DIE, Verbose: Options.Verbose);
645 if (!RelocAdjustment)
646 return Flags;
647
648 MyInfo.AddrAdjust = *RelocAdjustment;
649 MyInfo.InDebugMap = true;
650
651 if (Options.Verbose) {
652 outs() << "Keeping subprogram DIE:";
653 DIDumpOptions DumpOpts;
654 DumpOpts.ChildRecurseDepth = 0;
655 DumpOpts.Verbose = Options.Verbose;
656 DIE.dump(OS&: outs(), indent: 8 /* Indent */, DumpOpts);
657 }
658
659 if (DIE.getTag() == dwarf::DW_TAG_label) {
660 if (Unit.hasLabelAt(Addr: *LowPc))
661 return Flags;
662
663 DWARFUnit &OrigUnit = Unit.getOrigUnit();
664 // FIXME: dsymutil-classic compat. dsymutil-classic doesn't consider labels
665 // that don't fall into the CU's aranges. This is wrong IMO. Debug info
666 // generation bugs aside, this is really wrong in the case of labels, where
667 // a label marking the end of a function will have a PC == CU's high_pc.
668 if (dwarf::toAddress(V: OrigUnit.getUnitDIE().find(Attr: dwarf::DW_AT_high_pc))
669 .value_or(UINT64_MAX) <= LowPc)
670 return Flags;
671 // For assembly language files, try to preserve DWARF info by using
672 // function ranges when available, falling back to labels otherwise.
673 if (Unit.getLanguage() == dwarf::DW_LANG_Mips_Assembler ||
674 Unit.getLanguage() == dwarf::DW_LANG_Assembly) {
675 if (auto Range = RelocMgr.getSymbolRangeForAddress(Addr: *LowPc)) {
676 Unit.addFunctionRange(LowPC: Range->LowPC, HighPC: Range->HighPC, PCOffset: MyInfo.AddrAdjust);
677 } else {
678 Unit.addLabelLowPc(LabelLowPc: *LowPc, PcOffset: MyInfo.AddrAdjust);
679 }
680 } else {
681 Unit.addLabelLowPc(LabelLowPc: *LowPc, PcOffset: MyInfo.AddrAdjust);
682 }
683 return Flags | TF_Keep;
684 }
685
686 Flags |= TF_Keep;
687
688 std::optional<uint64_t> HighPc = DIE.getHighPC(LowPC: *LowPc);
689 if (!HighPc) {
690 reportWarning(Warning: "Function without high_pc. Range will be discarded.\n", File,
691 DIE: &DIE);
692 return Flags;
693 }
694 if (*LowPc > *HighPc) {
695 reportWarning(Warning: "low_pc greater than high_pc. Range will be discarded.\n",
696 File, DIE: &DIE);
697 return Flags;
698 }
699
700 // Replace the debug map range with a more accurate one.
701 Unit.addFunctionRange(
702 LowPC: *LowPc,
703 HighPC: RelocMgr.constrainCodeRangeHighPC(LowPC: *LowPc, HighPC: *HighPc, Adjustment: MyInfo.AddrAdjust),
704 PCOffset: MyInfo.AddrAdjust);
705 return Flags;
706}
707
708/// Check if a DIE should be kept.
709/// \returns updated TraversalFlags.
710unsigned DWARFLinker::shouldKeepDIE(AddressesMap &RelocMgr, const DWARFDie &DIE,
711 const DWARFFile &File, CompileUnit &Unit,
712 CompileUnit::DIEInfo &MyInfo,
713 unsigned Flags) {
714 switch (DIE.getTag()) {
715 case dwarf::DW_TAG_constant:
716 case dwarf::DW_TAG_variable:
717 return shouldKeepVariableDIE(RelocMgr, DIE, MyInfo, Flags);
718 case dwarf::DW_TAG_subprogram:
719 case dwarf::DW_TAG_label:
720 return shouldKeepSubprogramDIE(RelocMgr, DIE, File, Unit, MyInfo, Flags);
721 case dwarf::DW_TAG_base_type:
722 // DWARF Expressions may reference basic types, but scanning them
723 // is expensive. Basic types are tiny, so just keep all of them.
724 case dwarf::DW_TAG_imported_module:
725 case dwarf::DW_TAG_imported_declaration:
726 case dwarf::DW_TAG_imported_unit:
727 // We always want to keep these.
728 return Flags | TF_Keep;
729 default:
730 break;
731 }
732
733 return Flags;
734}
735
736/// Helper that updates the completeness of the current DIE based on the
737/// completeness of one of its children. It depends on the incompleteness of
738/// the children already being computed.
739static void updateChildIncompleteness(const DWARFDie &Die, CompileUnit &CU,
740 CompileUnit::DIEInfo &ChildInfo) {
741 switch (Die.getTag()) {
742 case dwarf::DW_TAG_structure_type:
743 case dwarf::DW_TAG_class_type:
744 case dwarf::DW_TAG_union_type:
745 break;
746 default:
747 return;
748 }
749
750 CompileUnit::DIEInfo &MyInfo = CU.getInfo(Die);
751
752 if (ChildInfo.Incomplete || ChildInfo.Prune)
753 MyInfo.Incomplete = true;
754}
755
756/// Helper that updates the completeness of the current DIE based on the
757/// completeness of the DIEs it references. It depends on the incompleteness of
758/// the referenced DIE already being computed.
759static void updateRefIncompleteness(const DWARFDie &Die, CompileUnit &CU,
760 CompileUnit::DIEInfo &RefInfo) {
761 switch (Die.getTag()) {
762 case dwarf::DW_TAG_typedef:
763 case dwarf::DW_TAG_member:
764 case dwarf::DW_TAG_reference_type:
765 case dwarf::DW_TAG_ptr_to_member_type:
766 case dwarf::DW_TAG_pointer_type:
767 break;
768 default:
769 return;
770 }
771
772 CompileUnit::DIEInfo &MyInfo = CU.getInfo(Die);
773
774 if (MyInfo.Incomplete)
775 return;
776
777 if (RefInfo.Incomplete)
778 MyInfo.Incomplete = true;
779}
780
781/// Look at the children of the given DIE and decide whether they should be
782/// kept.
783void DWARFLinker::lookForChildDIEsToKeep(
784 const DWARFDie &Die, CompileUnit &CU, unsigned Flags,
785 SmallVectorImpl<WorklistItem> &Worklist) {
786 // The TF_ParentWalk flag tells us that we are currently walking up the
787 // parent chain of a required DIE, and we don't want to mark all the children
788 // of the parents as kept (consider for example a DW_TAG_namespace node in
789 // the parent chain). There are however a set of DIE types for which we want
790 // to ignore that directive and still walk their children.
791 if (dieNeedsChildrenToBeMeaningful(Tag: Die.getTag()))
792 Flags &= ~DWARFLinker::TF_ParentWalk;
793
794 // We're finished if this DIE has no children or we're walking the parent
795 // chain.
796 if (!Die.hasChildren() || (Flags & DWARFLinker::TF_ParentWalk))
797 return;
798
799 // Add children in reverse order to the worklist to effectively process them
800 // in order.
801 for (auto Child : reverse(C: Die.children())) {
802 // Add a worklist item before every child to calculate incompleteness right
803 // after the current child is processed.
804 CompileUnit::DIEInfo &ChildInfo = CU.getInfo(Die: Child);
805 Worklist.emplace_back(Args: Die, Args&: CU, Args: WorklistItemType::UpdateChildIncompleteness,
806 Args: &ChildInfo);
807 Worklist.emplace_back(Args&: Child, Args&: CU, Args&: Flags);
808 }
809}
810
811static bool isODRCanonicalCandidate(const DWARFDie &Die, CompileUnit &CU) {
812 CompileUnit::DIEInfo &Info = CU.getInfo(Die);
813
814 if (!Info.Ctxt || (Die.getTag() == dwarf::DW_TAG_namespace))
815 return false;
816
817 if (!CU.hasODR() && !Info.InModuleScope)
818 return false;
819
820 return !Info.Incomplete && Info.Ctxt != CU.getInfo(Idx: Info.ParentIdx).Ctxt;
821}
822
823void DWARFLinker::markODRCanonicalDie(const DWARFDie &Die, CompileUnit &CU) {
824 CompileUnit::DIEInfo &Info = CU.getInfo(Die);
825
826 Info.ODRMarkingDone = true;
827 if (Info.Keep && isODRCanonicalCandidate(Die, CU) &&
828 !Info.Ctxt->hasCanonicalDIE())
829 Info.Ctxt->setHasCanonicalDIE();
830}
831
832/// Look at DIEs referenced by the given DIE and decide whether they should be
833/// kept. All DIEs referenced though attributes should be kept.
834void DWARFLinker::lookForRefDIEsToKeep(
835 const DWARFDie &Die, CompileUnit &CU, unsigned Flags,
836 const UnitListTy &Units, const DWARFFile &File,
837 SmallVectorImpl<WorklistItem> &Worklist) {
838 bool UseOdr = (Flags & DWARFLinker::TF_DependencyWalk)
839 ? (Flags & DWARFLinker::TF_ODR)
840 : CU.hasODR();
841 DWARFUnit &Unit = CU.getOrigUnit();
842 DWARFDataExtractor Data = Unit.getDebugInfoExtractor();
843 const auto *Abbrev = Die.getAbbreviationDeclarationPtr();
844 uint64_t Offset = Die.getOffset() + getULEB128Size(Value: Abbrev->getCode());
845
846 SmallVector<std::pair<DWARFDie, CompileUnit &>, 4> ReferencedDIEs;
847 for (const auto &AttrSpec : Abbrev->attributes()) {
848 DWARFFormValue Val(AttrSpec.Form);
849 if (!Val.isFormClass(FC: DWARFFormValue::FC_Reference) ||
850 AttrSpec.Attr == dwarf::DW_AT_sibling) {
851 DWARFFormValue::skipValue(Form: AttrSpec.Form, DebugInfoData: Data, OffsetPtr: &Offset,
852 FormParams: Unit.getFormParams());
853 continue;
854 }
855
856 Val.extractValue(Data, OffsetPtr: &Offset, FormParams: Unit.getFormParams(), U: &Unit);
857 CompileUnit *ReferencedCU;
858 if (auto RefDie =
859 resolveDIEReference(File, Units, RefValue: Val, DIE: Die, RefCU&: ReferencedCU)) {
860 CompileUnit::DIEInfo &Info = ReferencedCU->getInfo(Die: RefDie);
861 // If the referenced DIE has a DeclContext that has already been
862 // emitted, then do not keep the one in this CU. We'll link to
863 // the canonical DIE in cloneDieReferenceAttribute.
864 //
865 // FIXME: compatibility with dsymutil-classic. UseODR shouldn't
866 // be necessary and could be advantageously replaced by
867 // ReferencedCU->hasODR() && CU.hasODR().
868 //
869 // FIXME: compatibility with dsymutil-classic. There is no
870 // reason not to unique ref_addr references.
871 if (AttrSpec.Form != dwarf::DW_FORM_ref_addr &&
872 isODRAttribute(Attr: AttrSpec.Attr) && Info.Ctxt &&
873 Info.Ctxt->hasCanonicalDIE())
874 continue;
875
876 // Keep a module forward declaration if there is no definition.
877 if (!(isODRAttribute(Attr: AttrSpec.Attr) && Info.Ctxt &&
878 Info.Ctxt->hasCanonicalDIE()))
879 Info.Prune = false;
880 ReferencedDIEs.emplace_back(Args&: RefDie, Args&: *ReferencedCU);
881 }
882 }
883
884 unsigned ODRFlag = UseOdr ? DWARFLinker::TF_ODR : 0;
885
886 // Add referenced DIEs in reverse order to the worklist to effectively
887 // process them in order.
888 for (auto &P : reverse(C&: ReferencedDIEs)) {
889 // Add a worklist item before every child to calculate incompleteness right
890 // after the current child is processed.
891 CompileUnit::DIEInfo &Info = P.second.getInfo(Die: P.first);
892 Worklist.emplace_back(Args: Die, Args&: CU, Args: WorklistItemType::UpdateRefIncompleteness,
893 Args: &Info);
894 Worklist.emplace_back(Args&: P.first, Args&: P.second,
895 Args: DWARFLinker::TF_Keep |
896 DWARFLinker::TF_DependencyWalk | ODRFlag);
897 }
898}
899
900/// Look at the parent of the given DIE and decide whether they should be kept.
901void DWARFLinker::lookForParentDIEsToKeep(
902 unsigned AncestorIdx, CompileUnit &CU, unsigned Flags,
903 SmallVectorImpl<WorklistItem> &Worklist) {
904 // Stop if we encounter an ancestor that's already marked as kept.
905 if (CU.getInfo(Idx: AncestorIdx).Keep)
906 return;
907
908 DWARFUnit &Unit = CU.getOrigUnit();
909 DWARFDie ParentDIE = Unit.getDIEAtIndex(Index: AncestorIdx);
910 Worklist.emplace_back(Args&: CU.getInfo(Idx: AncestorIdx).ParentIdx, Args&: CU, Args&: Flags);
911 Worklist.emplace_back(Args&: ParentDIE, Args&: CU, Args&: Flags);
912}
913
914/// Recursively walk the \p DIE tree and look for DIEs to keep. Store that
915/// information in \p CU's DIEInfo.
916///
917/// This function is the entry point of the DIE selection algorithm. It is
918/// expected to walk the DIE tree in file order and (though the mediation of
919/// its helper) call hasValidRelocation() on each DIE that might be a 'root
920/// DIE' (See DwarfLinker class comment).
921///
922/// While walking the dependencies of root DIEs, this function is also called,
923/// but during these dependency walks the file order is not respected. The
924/// TF_DependencyWalk flag tells us which kind of traversal we are currently
925/// doing.
926///
927/// The recursive algorithm is implemented iteratively as a work list because
928/// very deep recursion could exhaust the stack for large projects. The work
929/// list acts as a scheduler for different types of work that need to be
930/// performed.
931///
932/// The recursive nature of the algorithm is simulated by running the "main"
933/// algorithm (LookForDIEsToKeep) followed by either looking at more DIEs
934/// (LookForChildDIEsToKeep, LookForRefDIEsToKeep, LookForParentDIEsToKeep) or
935/// fixing up a computed property (UpdateChildIncompleteness,
936/// UpdateRefIncompleteness).
937///
938/// The return value indicates whether the DIE is incomplete.
939void DWARFLinker::lookForDIEsToKeep(AddressesMap &AddressesMap,
940 const UnitListTy &Units,
941 const DWARFDie &Die, const DWARFFile &File,
942 CompileUnit &Cu, unsigned Flags) {
943 // LIFO work list.
944 SmallVector<WorklistItem, 4> Worklist;
945 Worklist.emplace_back(Args: Die, Args&: Cu, Args&: Flags);
946
947 while (!Worklist.empty()) {
948 WorklistItem Current = Worklist.pop_back_val();
949
950 // Look at the worklist type to decide what kind of work to perform.
951 switch (Current.Type) {
952 case WorklistItemType::UpdateChildIncompleteness:
953 updateChildIncompleteness(Die: Current.Die, CU&: Current.CU, ChildInfo&: *Current.OtherInfo);
954 continue;
955 case WorklistItemType::UpdateRefIncompleteness:
956 updateRefIncompleteness(Die: Current.Die, CU&: Current.CU, RefInfo&: *Current.OtherInfo);
957 continue;
958 case WorklistItemType::LookForChildDIEsToKeep:
959 lookForChildDIEsToKeep(Die: Current.Die, CU&: Current.CU, Flags: Current.Flags, Worklist);
960 continue;
961 case WorklistItemType::LookForRefDIEsToKeep:
962 lookForRefDIEsToKeep(Die: Current.Die, CU&: Current.CU, Flags: Current.Flags, Units, File,
963 Worklist);
964 continue;
965 case WorklistItemType::LookForParentDIEsToKeep:
966 lookForParentDIEsToKeep(AncestorIdx: Current.AncestorIdx, CU&: Current.CU, Flags: Current.Flags,
967 Worklist);
968 continue;
969 case WorklistItemType::MarkODRCanonicalDie:
970 markODRCanonicalDie(Die: Current.Die, CU&: Current.CU);
971 continue;
972 case WorklistItemType::LookForDIEsToKeep:
973 break;
974 }
975
976 unsigned Idx = Current.CU.getOrigUnit().getDIEIndex(D: Current.Die);
977 CompileUnit::DIEInfo &MyInfo = Current.CU.getInfo(Idx);
978
979 if (MyInfo.Prune) {
980 // We're walking the dependencies of a module forward declaration that was
981 // kept because there is no definition.
982 if (Current.Flags & TF_DependencyWalk)
983 MyInfo.Prune = false;
984 else
985 continue;
986 }
987
988 // If the Keep flag is set, we are marking a required DIE's dependencies.
989 // If our target is already marked as kept, we're all set.
990 bool AlreadyKept = MyInfo.Keep;
991 if ((Current.Flags & TF_DependencyWalk) && AlreadyKept)
992 continue;
993
994 if (!(Current.Flags & TF_DependencyWalk))
995 Current.Flags = shouldKeepDIE(RelocMgr&: AddressesMap, DIE: Current.Die, File, Unit&: Current.CU,
996 MyInfo, Flags: Current.Flags);
997
998 // We need to mark context for the canonical die in the end of normal
999 // traversing(not TF_DependencyWalk) or after normal traversing if die
1000 // was not marked as kept.
1001 if (!(Current.Flags & TF_DependencyWalk) ||
1002 (MyInfo.ODRMarkingDone && !MyInfo.Keep)) {
1003 if (Current.CU.hasODR() || MyInfo.InModuleScope)
1004 Worklist.emplace_back(Args&: Current.Die, Args&: Current.CU,
1005 Args: WorklistItemType::MarkODRCanonicalDie);
1006 }
1007
1008 // Finish by looking for child DIEs. Because of the LIFO worklist we need
1009 // to schedule that work before any subsequent items are added to the
1010 // worklist.
1011 Worklist.emplace_back(Args&: Current.Die, Args&: Current.CU, Args&: Current.Flags,
1012 Args: WorklistItemType::LookForChildDIEsToKeep);
1013
1014 if (AlreadyKept || !(Current.Flags & TF_Keep))
1015 continue;
1016
1017 // If it is a newly kept DIE mark it as well as all its dependencies as
1018 // kept.
1019 MyInfo.Keep = true;
1020
1021 // We're looking for incomplete types.
1022 MyInfo.Incomplete =
1023 Current.Die.getTag() != dwarf::DW_TAG_subprogram &&
1024 Current.Die.getTag() != dwarf::DW_TAG_member &&
1025 dwarf::toUnsigned(V: Current.Die.find(Attr: dwarf::DW_AT_declaration), Default: 0);
1026
1027 // After looking at the parent chain, look for referenced DIEs. Because of
1028 // the LIFO worklist we need to schedule that work before any subsequent
1029 // items are added to the worklist.
1030 Worklist.emplace_back(Args&: Current.Die, Args&: Current.CU, Args&: Current.Flags,
1031 Args: WorklistItemType::LookForRefDIEsToKeep);
1032
1033 bool UseOdr = (Current.Flags & TF_DependencyWalk) ? (Current.Flags & TF_ODR)
1034 : Current.CU.hasODR();
1035 unsigned ODRFlag = UseOdr ? TF_ODR : 0;
1036 unsigned ParFlags = TF_ParentWalk | TF_Keep | TF_DependencyWalk | ODRFlag;
1037
1038 // Now schedule the parent walk.
1039 Worklist.emplace_back(Args&: MyInfo.ParentIdx, Args&: Current.CU, Args&: ParFlags);
1040 }
1041}
1042
1043#ifndef NDEBUG
1044/// A broken link in the keep chain. By recording both the parent and the child
1045/// we can show only broken links for DIEs with multiple children.
1046struct BrokenLink {
1047 BrokenLink(DWARFDie Parent, DWARFDie Child) : Parent(Parent), Child(Child) {}
1048 DWARFDie Parent;
1049 DWARFDie Child;
1050};
1051
1052/// Verify the keep chain by looking for DIEs that are kept but who's parent
1053/// isn't.
1054static void verifyKeepChain(CompileUnit &CU) {
1055 std::vector<DWARFDie> Worklist;
1056 Worklist.push_back(CU.getOrigUnit().getUnitDIE());
1057
1058 // List of broken links.
1059 std::vector<BrokenLink> BrokenLinks;
1060
1061 while (!Worklist.empty()) {
1062 const DWARFDie Current = Worklist.back();
1063 Worklist.pop_back();
1064
1065 const bool CurrentDieIsKept = CU.getInfo(Current).Keep;
1066
1067 for (DWARFDie Child : reverse(Current.children())) {
1068 Worklist.push_back(Child);
1069
1070 const bool ChildDieIsKept = CU.getInfo(Child).Keep;
1071 if (!CurrentDieIsKept && ChildDieIsKept)
1072 BrokenLinks.emplace_back(Current, Child);
1073 }
1074 }
1075
1076 if (!BrokenLinks.empty()) {
1077 for (BrokenLink Link : BrokenLinks) {
1078 WithColor::error() << formatv(
1079 "Found invalid link in keep chain between {0:x} and {1:x}\n",
1080 Link.Parent.getOffset(), Link.Child.getOffset());
1081
1082 errs() << "Parent:";
1083 Link.Parent.dump(errs(), 0, {});
1084 CU.getInfo(Link.Parent).dump();
1085
1086 errs() << "Child:";
1087 Link.Child.dump(errs(), 2, {});
1088 CU.getInfo(Link.Child).dump();
1089 }
1090 report_fatal_error("invalid keep chain");
1091 }
1092}
1093#endif
1094
1095/// Assign an abbreviation number to \p Abbrev.
1096///
1097/// Our DIEs get freed after every DebugMapObject has been processed,
1098/// thus the FoldingSet we use to unique DIEAbbrevs cannot refer to
1099/// the instances hold by the DIEs. When we encounter an abbreviation
1100/// that we don't know, we create a permanent copy of it.
1101void DWARFLinker::assignAbbrev(DIEAbbrev &Abbrev) {
1102 // Check the set for priors.
1103 FoldingSetNodeID ID;
1104 Abbrev.Profile(ID);
1105 FoldingSetInsertToken Token;
1106 DIEAbbrev *InSet = AbbreviationsSet.lookup(ID, Token);
1107
1108 // If it's newly added.
1109 if (InSet) {
1110 // Assign existing abbreviation number.
1111 Abbrev.setNumber(InSet->getNumber());
1112 } else {
1113 // Add to abbreviation list.
1114 Abbreviations.push_back(
1115 x: std::make_unique<DIEAbbrev>(args: Abbrev.getTag(), args: Abbrev.hasChildren()));
1116 for (const auto &Attr : Abbrev.getData())
1117 Abbreviations.back()->AddAttribute(AbbrevData: Attr);
1118 AbbreviationsSet.insert(N: Abbreviations.back().get(), Token);
1119 // Assign the unique abbreviation number.
1120 Abbrev.setNumber(Abbreviations.size());
1121 Abbreviations.back()->setNumber(Abbreviations.size());
1122 }
1123}
1124
1125unsigned DWARFLinker::DIECloner::cloneStringAttribute(DIE &Die,
1126 AttributeSpec AttrSpec,
1127 const DWARFFormValue &Val,
1128 const DWARFUnit &U,
1129 AttributesInfo &Info) {
1130 std::optional<const char *> String = dwarf::toString(V: Val);
1131 if (!String)
1132 return 0;
1133 DwarfStringPoolEntryRef StringEntry;
1134 if (AttrSpec.Form == dwarf::DW_FORM_line_strp) {
1135 StringEntry = DebugLineStrPool.getEntry(S: *String);
1136 } else {
1137 StringEntry = DebugStrPool.getEntry(S: *String);
1138
1139 if (AttrSpec.Attr == dwarf::DW_AT_APPLE_origin) {
1140 Info.HasAppleOrigin = true;
1141 if (std::optional<StringRef> FileName =
1142 ObjFile.Addresses->getLibraryInstallName()) {
1143 StringEntry = DebugStrPool.getEntry(S: *FileName);
1144 }
1145 }
1146
1147 // Update attributes info.
1148 if (AttrSpec.Attr == dwarf::DW_AT_name)
1149 Info.Name = StringEntry;
1150 else if (AttrSpec.Attr == dwarf::DW_AT_MIPS_linkage_name ||
1151 AttrSpec.Attr == dwarf::DW_AT_linkage_name)
1152 Info.MangledName = StringEntry;
1153 if (U.getVersion() >= 5) {
1154 // Switch everything to DW_FORM_strx strings.
1155 auto StringOffsetIndex =
1156 StringOffsetPool.getValueIndex(Value: StringEntry.getOffset());
1157 return Die
1158 .addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1159 Form: dwarf::DW_FORM_strx, Value: DIEInteger(StringOffsetIndex))
1160 ->sizeOf(FormParams: U.getFormParams());
1161 }
1162 // Switch everything to out of line strings.
1163 AttrSpec.Form = dwarf::DW_FORM_strp;
1164 }
1165 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr), Form: AttrSpec.Form,
1166 Value: DIEInteger(StringEntry.getOffset()));
1167 return 4;
1168}
1169
1170unsigned DWARFLinker::DIECloner::cloneDieReferenceAttribute(
1171 DIE &Die, const DWARFDie &InputDIE, AttributeSpec AttrSpec,
1172 unsigned AttrSize, const DWARFFormValue &Val, const DWARFFile &File,
1173 CompileUnit &Unit) {
1174 const DWARFUnit &U = Unit.getOrigUnit();
1175 uint64_t Ref;
1176 if (std::optional<uint64_t> Off = Val.getAsRelativeReference())
1177 Ref = Val.getUnit()->getOffset() + *Off;
1178 else if (Off = Val.getAsDebugInfoReference(); Off)
1179 Ref = *Off;
1180 else
1181 return 0;
1182
1183 DIE *NewRefDie = nullptr;
1184 CompileUnit *RefUnit = nullptr;
1185
1186 DWARFDie RefDie =
1187 Linker.resolveDIEReference(File, Units: CompileUnits, RefValue: Val, DIE: InputDIE, RefCU&: RefUnit);
1188
1189 // If the referenced DIE is not found, drop the attribute.
1190 if (!RefDie || AttrSpec.Attr == dwarf::DW_AT_sibling)
1191 return 0;
1192
1193 CompileUnit::DIEInfo &RefInfo = RefUnit->getInfo(Die: RefDie);
1194
1195 // If we already have emitted an equivalent DeclContext, just point
1196 // at it.
1197 if (isODRAttribute(Attr: AttrSpec.Attr) && RefInfo.Ctxt &&
1198 RefInfo.Ctxt->getCanonicalDIEOffset()) {
1199 assert(RefInfo.Ctxt->hasCanonicalDIE() &&
1200 "Offset to canonical die is set, but context is not marked");
1201 DIEInteger Attr(RefInfo.Ctxt->getCanonicalDIEOffset());
1202 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1203 Form: dwarf::DW_FORM_ref_addr, Value&: Attr);
1204 return U.getRefAddrByteSize();
1205 }
1206
1207 if (!RefInfo.Clone) {
1208 // We haven't cloned this DIE yet. Just create an empty one and
1209 // store it. It'll get really cloned when we process it.
1210 RefInfo.UnclonedReference = true;
1211 RefInfo.Clone = DIE::get(Alloc&: DIEAlloc, Tag: dwarf::Tag(RefDie.getTag()));
1212 }
1213 NewRefDie = RefInfo.Clone;
1214
1215 if (AttrSpec.Form == dwarf::DW_FORM_ref_addr ||
1216 (Unit.hasODR() && isODRAttribute(Attr: AttrSpec.Attr))) {
1217 if (Ref < InputDIE.getOffset() && !RefInfo.UnclonedReference) {
1218 // Backward reference: the target DIE is already cloned and
1219 // parented in a unit tree, so DIEEntry can resolve the
1220 // absolute offset at emission time.
1221 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1222 Form: dwarf::DW_FORM_ref_addr, Value: DIEEntry(*NewRefDie));
1223 } else {
1224 // Forward reference: the target DIE may be a placeholder that
1225 // never gets adopted into a unit tree (e.g. due to ODR
1226 // pruning), so DIEEntry cannot safely resolve it. Use a
1227 // placeholder integer and fix it up after all units are cloned.
1228 Unit.noteForwardReference(
1229 Die: NewRefDie, RefUnit, Ctxt: RefInfo.Ctxt,
1230 Attr: Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1231 Form: dwarf::DW_FORM_ref_addr, Value: DIEInteger(UINT64_MAX)));
1232 }
1233 return U.getRefAddrByteSize();
1234 }
1235
1236 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1237 Form: dwarf::Form(AttrSpec.Form), Value: DIEEntry(*NewRefDie));
1238
1239 return AttrSize;
1240}
1241
1242void DWARFLinker::DIECloner::cloneExpression(
1243 DataExtractor &Data, DWARFExpression Expression, const DWARFFile &File,
1244 CompileUnit &Unit, SmallVectorImpl<uint8_t> &OutputBuffer,
1245 int64_t AddrRelocAdjustment, bool IsLittleEndian) {
1246 using Encoding = DWARFExpression::Operation::Encoding;
1247
1248 uint8_t OrigAddressByteSize = Unit.getOrigUnit().getAddressByteSize();
1249
1250 uint64_t OpOffset = 0;
1251 for (auto &Op : Expression) {
1252 if (Op.isError()) {
1253 // The operation could not be decoded, so neither it nor anything after
1254 // it can be located. Its end offset is the offset it started at, so the
1255 // slice copied below would be empty and the rest of the expression
1256 // would be silently dropped. Preserve the remaining bytes instead.
1257 Linker.reportWarning(
1258 Warning: "cannot decode a DW_OP, copying the rest of the expression "
1259 "unmodified.",
1260 File);
1261 StringRef Bytes = Data.getData().substr(Start: OpOffset);
1262 OutputBuffer.append(in_start: Bytes.begin(), in_end: Bytes.end());
1263 return;
1264 }
1265 auto Desc = Op.getDescription();
1266 // DW_OP_const_type is variable-length and has 3
1267 // operands. Thus far we only support 2.
1268 if ((Desc.Op.size() == 2 && Desc.Op[0] == Encoding::BaseTypeRef) ||
1269 (Desc.Op.size() == 2 && Desc.Op[1] == Encoding::BaseTypeRef &&
1270 Desc.Op[0] != Encoding::Size1))
1271 Linker.reportWarning(Warning: "Unsupported DW_OP encoding.", File);
1272
1273 if ((Desc.Op.size() == 1 && Desc.Op[0] == Encoding::BaseTypeRef) ||
1274 (Desc.Op.size() == 2 && Desc.Op[1] == Encoding::BaseTypeRef &&
1275 Desc.Op[0] == Encoding::Size1)) {
1276 // This code assumes that the other non-typeref operand fits into 1 byte.
1277 assert(OpOffset < Op.getEndOffset());
1278 uint32_t ULEBsize = Op.getEndOffset() - OpOffset - 1;
1279 assert(ULEBsize <= 16);
1280
1281 // Copy over the operation.
1282 assert(!Op.getSubCode() && "SubOps not yet supported");
1283 OutputBuffer.push_back(Elt: Op.getCode());
1284 uint64_t RefOffset;
1285 if (Desc.Op.size() == 1) {
1286 RefOffset = Op.getRawOperand(Idx: 0);
1287 } else {
1288 OutputBuffer.push_back(Elt: Op.getRawOperand(Idx: 0));
1289 RefOffset = Op.getRawOperand(Idx: 1);
1290 }
1291 uint32_t Offset = 0;
1292 // Look up the base type. For DW_OP_convert, the operand may be 0 to
1293 // instead indicate the generic type. The same holds for
1294 // DW_OP_reinterpret, which is currently not supported.
1295 if (RefOffset > 0 || Op.getCode() != dwarf::DW_OP_convert) {
1296 RefOffset += Unit.getOrigUnit().getOffset();
1297 auto RefDie = Unit.getOrigUnit().getDIEForOffset(Offset: RefOffset);
1298 CompileUnit::DIEInfo &Info = Unit.getInfo(Die: RefDie);
1299 if (DIE *Clone = Info.Clone)
1300 Offset = Clone->getOffset();
1301 else
1302 Linker.reportWarning(
1303 Warning: "base type ref doesn't point to DW_TAG_base_type.", File);
1304 }
1305 uint8_t ULEB[16];
1306 unsigned RealSize = encodeULEB128(Value: Offset, p: ULEB, PadTo: ULEBsize);
1307 if (RealSize > ULEBsize) {
1308 // Emit the generic type as a fallback.
1309 RealSize = encodeULEB128(Value: 0, p: ULEB, PadTo: ULEBsize);
1310 Linker.reportWarning(Warning: "base type ref doesn't fit.", File);
1311 }
1312 assert(RealSize == ULEBsize && "padding failed");
1313 ArrayRef<uint8_t> ULEBbytes(ULEB, ULEBsize);
1314 OutputBuffer.append(in_start: ULEBbytes.begin(), in_end: ULEBbytes.end());
1315 } else if (!Linker.Options.Update && Op.getCode() == dwarf::DW_OP_addrx) {
1316 if (std::optional<object::SectionedAddress> SA =
1317 Unit.getOrigUnit().getAddrOffsetSectionItem(
1318 Index: Op.getRawOperand(Idx: 0))) {
1319 // DWARFLinker does not use addrx forms since it generates relocated
1320 // addresses. Replace DW_OP_addrx with DW_OP_addr here.
1321 // Argument of DW_OP_addrx should be relocated here as it is not
1322 // processed by applyValidRelocs.
1323 OutputBuffer.push_back(Elt: dwarf::DW_OP_addr);
1324 uint64_t LinkedAddress =
1325 SA->Address +
1326 File.Addresses
1327 ->getAddrIndexRelocAdjustment(U&: Unit.getOrigUnit(), Op,
1328 Verbose: Linker.Options.Verbose)
1329 .value_or(u&: AddrRelocAdjustment);
1330 if (IsLittleEndian != sys::IsLittleEndianHost)
1331 sys::swapByteOrder(Value&: LinkedAddress);
1332 ArrayRef<uint8_t> AddressBytes(
1333 reinterpret_cast<const uint8_t *>(&LinkedAddress),
1334 OrigAddressByteSize);
1335 OutputBuffer.append(in_start: AddressBytes.begin(), in_end: AddressBytes.end());
1336 } else
1337 Linker.reportWarning(Warning: "cannot read DW_OP_addrx operand.", File);
1338 } else if (!Linker.Options.Update && Op.getCode() == dwarf::DW_OP_constx) {
1339 if (std::optional<object::SectionedAddress> SA =
1340 Unit.getOrigUnit().getAddrOffsetSectionItem(
1341 Index: Op.getRawOperand(Idx: 0))) {
1342 // DWARFLinker does not use constx forms since it generates relocated
1343 // addresses. Replace DW_OP_constx with DW_OP_const[*]u here.
1344 // Argument of DW_OP_constx should be relocated here as it is not
1345 // processed by applyValidRelocs.
1346 std::optional<uint8_t> OutOperandKind;
1347 switch (OrigAddressByteSize) {
1348 case 4:
1349 OutOperandKind = dwarf::DW_OP_const4u;
1350 break;
1351 case 8:
1352 OutOperandKind = dwarf::DW_OP_const8u;
1353 break;
1354 default:
1355 Linker.reportWarning(
1356 Warning: formatv(Fmt: ("unsupported address size: {0}."), Vals&: OrigAddressByteSize),
1357 File);
1358 break;
1359 }
1360
1361 if (OutOperandKind) {
1362 OutputBuffer.push_back(Elt: *OutOperandKind);
1363 uint64_t LinkedAddress =
1364 SA->Address +
1365 File.Addresses
1366 ->getAddrIndexRelocAdjustment(U&: Unit.getOrigUnit(), Op,
1367 Verbose: Linker.Options.Verbose)
1368 .value_or(u&: AddrRelocAdjustment);
1369 if (IsLittleEndian != sys::IsLittleEndianHost)
1370 sys::swapByteOrder(Value&: LinkedAddress);
1371 ArrayRef<uint8_t> AddressBytes(
1372 reinterpret_cast<const uint8_t *>(&LinkedAddress),
1373 OrigAddressByteSize);
1374 OutputBuffer.append(in_start: AddressBytes.begin(), in_end: AddressBytes.end());
1375 }
1376 } else
1377 Linker.reportWarning(Warning: "cannot read DW_OP_constx operand.", File);
1378 } else {
1379 // Copy over everything else unmodified.
1380 StringRef Bytes = Data.getData().slice(Start: OpOffset, End: Op.getEndOffset());
1381 OutputBuffer.append(in_start: Bytes.begin(), in_end: Bytes.end());
1382 }
1383 OpOffset = Op.getEndOffset();
1384 }
1385}
1386
1387unsigned DWARFLinker::DIECloner::cloneBlockAttribute(
1388 DIE &Die, const DWARFDie &InputDIE, const DWARFFile &File,
1389 CompileUnit &Unit, AttributeSpec AttrSpec, const DWARFFormValue &Val,
1390 bool IsLittleEndian) {
1391 DIEValueList *Attr;
1392 DIEValue Value;
1393 DIELoc *Loc = nullptr;
1394 DIEBlock *Block = nullptr;
1395 if (AttrSpec.Form == dwarf::DW_FORM_exprloc) {
1396 Loc = new (DIEAlloc) DIELoc;
1397 Linker.DIELocs.push_back(x: Loc);
1398 } else {
1399 Block = new (DIEAlloc) DIEBlock;
1400 Linker.DIEBlocks.push_back(x: Block);
1401 }
1402 Attr = Loc ? static_cast<DIEValueList *>(Loc)
1403 : static_cast<DIEValueList *>(Block);
1404
1405 DWARFUnit &OrigUnit = Unit.getOrigUnit();
1406 // If the block is a DWARF Expression, clone it into the temporary
1407 // buffer using cloneExpression(), otherwise copy the data directly.
1408 SmallVector<uint8_t, 32> Buffer;
1409 ArrayRef<uint8_t> Bytes = *Val.getAsBlock();
1410 if (DWARFAttribute::mayHaveLocationExpr(Attr: AttrSpec.Attr) &&
1411 (Val.isFormClass(FC: DWARFFormValue::FC_Block) ||
1412 Val.isFormClass(FC: DWARFFormValue::FC_Exprloc))) {
1413 DataExtractor Data(Bytes, IsLittleEndian);
1414 DWARFExpression Expr(Data, OrigUnit.getAddressByteSize(),
1415 OrigUnit.getFormParams().Format);
1416 cloneExpression(Data, Expression: Expr, File, Unit, OutputBuffer&: Buffer,
1417 AddrRelocAdjustment: Unit.getInfo(Die: InputDIE).AddrAdjust, IsLittleEndian);
1418 Bytes = Buffer;
1419 }
1420 for (auto Byte : Bytes)
1421 Attr->addValue(Alloc&: DIEAlloc, Attribute: static_cast<dwarf::Attribute>(0),
1422 Form: dwarf::DW_FORM_data1, Value: DIEInteger(Byte));
1423
1424 // FIXME: If DIEBlock and DIELoc just reuses the Size field of
1425 // the DIE class, this "if" could be replaced by
1426 // Attr->setSize(Bytes.size()).
1427 if (Loc)
1428 Loc->setSize(Bytes.size());
1429 else
1430 Block->setSize(Bytes.size());
1431
1432 if (Loc)
1433 Value = DIEValue(dwarf::Attribute(AttrSpec.Attr),
1434 dwarf::Form(AttrSpec.Form), Loc);
1435 else {
1436 // The expression location data might be updated and exceed the original
1437 // size. Check whether the new data fits into the original form.
1438 if ((AttrSpec.Form == dwarf::DW_FORM_block1 &&
1439 (Bytes.size() > UINT8_MAX)) ||
1440 (AttrSpec.Form == dwarf::DW_FORM_block2 &&
1441 (Bytes.size() > UINT16_MAX)) ||
1442 (AttrSpec.Form == dwarf::DW_FORM_block4 && (Bytes.size() > UINT32_MAX)))
1443 AttrSpec.Form = dwarf::DW_FORM_block;
1444
1445 Value = DIEValue(dwarf::Attribute(AttrSpec.Attr),
1446 dwarf::Form(AttrSpec.Form), Block);
1447 }
1448
1449 return Die.addValue(Alloc&: DIEAlloc, V: Value)->sizeOf(FormParams: OrigUnit.getFormParams());
1450}
1451
1452/// Returns \p InputDIE's DW_AT_high_pc value \p HighPC, constrained so the code
1453/// range it ends stays clear of the symbol the linker places next. \p IsLength
1454/// tells whether high_pc is encoded as a length rather than an address, and
1455/// \p PCOffset is the amount the range shifts by in the output.
1456///
1457/// A scope nested in a function inherits the overrun of the function, so it is
1458/// constrained as well.
1459static uint64_t constrainHighPC(const DWARFDie &InputDIE, uint64_t HighPC,
1460 bool IsLength, int64_t PCOffset,
1461 AddressesMap &Addresses) {
1462 std::optional<uint64_t> LowPC =
1463 dwarf::toAddress(V: InputDIE.find(Attr: dwarf::DW_AT_low_pc));
1464 if (!LowPC)
1465 return HighPC;
1466 uint64_t Constrained = Addresses.constrainCodeRangeHighPC(
1467 LowPC: *LowPC, HighPC: IsLength ? *LowPC + HighPC : HighPC, Adjustment: PCOffset);
1468 return IsLength ? Constrained - *LowPC : Constrained;
1469}
1470
1471unsigned DWARFLinker::DIECloner::cloneAddressAttribute(
1472 DIE &Die, const DWARFDie &InputDIE, AttributeSpec AttrSpec,
1473 unsigned AttrSize, const DWARFFormValue &Val, const CompileUnit &Unit,
1474 AttributesInfo &Info) {
1475 if (AttrSpec.Attr == dwarf::DW_AT_low_pc)
1476 Info.HasLowPc = true;
1477
1478 if (LLVM_UNLIKELY(Linker.Options.Update)) {
1479 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1480 Form: dwarf::Form(AttrSpec.Form), Value: DIEInteger(Val.getRawUValue()));
1481 return AttrSize;
1482 }
1483
1484 // Cloned Die may have address attributes relocated to a
1485 // totally unrelated value. This can happen:
1486 // - If high_pc is an address (Dwarf version == 2), then it might have been
1487 // relocated to a totally unrelated value (because the end address in the
1488 // object file might be start address of another function which got moved
1489 // independently by the linker).
1490 // - If address relocated in an inline_subprogram that happens at the
1491 // beginning of its inlining function.
1492 // To avoid above cases and to not apply relocation twice (in
1493 // applyValidRelocs and here), read address attribute from InputDIE and apply
1494 // Info.PCOffset here.
1495
1496 std::optional<DWARFFormValue> AddrAttribute = InputDIE.find(Attr: AttrSpec.Attr);
1497 if (!AddrAttribute)
1498 llvm_unreachable("Cann't find attribute.");
1499
1500 std::optional<uint64_t> Addr = AddrAttribute->getAsAddress();
1501 if (!Addr) {
1502 Linker.reportWarning(Warning: "Cann't read address attribute value.", File: ObjFile);
1503 return 0;
1504 }
1505
1506 if (InputDIE.getTag() == dwarf::DW_TAG_compile_unit &&
1507 AttrSpec.Attr == dwarf::DW_AT_low_pc) {
1508 if (std::optional<uint64_t> LowPC = Unit.getLowPc())
1509 Addr = *LowPC;
1510 else
1511 return 0;
1512 } else if (InputDIE.getTag() == dwarf::DW_TAG_compile_unit &&
1513 AttrSpec.Attr == dwarf::DW_AT_high_pc) {
1514 if (uint64_t HighPc = Unit.getHighPc())
1515 Addr = HighPc;
1516 else
1517 return 0;
1518 } else {
1519 if (AttrSpec.Attr == dwarf::DW_AT_high_pc)
1520 Addr = constrainHighPC(InputDIE, HighPC: *Addr, /*IsLength=*/false, PCOffset: Info.PCOffset,
1521 Addresses&: *ObjFile.Addresses);
1522 *Addr += Info.PCOffset;
1523 }
1524
1525 if (AttrSpec.Form == dwarf::DW_FORM_addr) {
1526 Die.addValue(Alloc&: DIEAlloc, Attribute: static_cast<dwarf::Attribute>(AttrSpec.Attr),
1527 Form: AttrSpec.Form, Value: DIEInteger(*Addr));
1528 return Unit.getOrigUnit().getAddressByteSize();
1529 }
1530
1531 auto AddrIndex = AddrPool.getValueIndex(Value: *Addr);
1532
1533 return Die
1534 .addValue(Alloc&: DIEAlloc, Attribute: static_cast<dwarf::Attribute>(AttrSpec.Attr),
1535 Form: dwarf::Form::DW_FORM_addrx, Value: DIEInteger(AddrIndex))
1536 ->sizeOf(FormParams: Unit.getOrigUnit().getFormParams());
1537}
1538
1539unsigned DWARFLinker::DIECloner::cloneScalarAttribute(
1540 DIE &Die, const DWARFDie &InputDIE, const DWARFFile &File,
1541 CompileUnit &Unit, AttributeSpec AttrSpec, const DWARFFormValue &Val,
1542 unsigned AttrSize, AttributesInfo &Info) {
1543 uint64_t Value;
1544
1545 // We don't emit any skeleton CUs with dsymutil. So avoid emitting
1546 // a redundant DW_AT_GNU_dwo_id on the non-skeleton CU.
1547 if (AttrSpec.Attr == dwarf::DW_AT_GNU_dwo_id ||
1548 AttrSpec.Attr == dwarf::DW_AT_dwo_id)
1549 return 0;
1550
1551 // Check for the offset to the macro table. If offset is incorrect then we
1552 // need to remove the attribute.
1553 if (AttrSpec.Attr == dwarf::DW_AT_macro_info) {
1554 if (std::optional<uint64_t> Offset = Val.getAsSectionOffset()) {
1555 const llvm::DWARFDebugMacro *Macro = File.Dwarf->getDebugMacinfo();
1556 if (Macro == nullptr || !Macro->hasEntryForOffset(Offset: *Offset))
1557 return 0;
1558 }
1559 }
1560
1561 if (AttrSpec.Attr == dwarf::DW_AT_macros) {
1562 if (std::optional<uint64_t> Offset = Val.getAsSectionOffset()) {
1563 const llvm::DWARFDebugMacro *Macro = File.Dwarf->getDebugMacro();
1564 if (Macro == nullptr || !Macro->hasEntryForOffset(Offset: *Offset))
1565 return 0;
1566 }
1567 }
1568
1569 if (AttrSpec.Attr == dwarf::DW_AT_str_offsets_base) {
1570 // DWARFLinker generates common .debug_str_offsets table used for all
1571 // compile units. The offset to the common .debug_str_offsets table is 8 on
1572 // DWARF32.
1573 Info.AttrStrOffsetBaseSeen = true;
1574 return Die
1575 .addValue(Alloc&: DIEAlloc, Attribute: dwarf::DW_AT_str_offsets_base,
1576 Form: dwarf::DW_FORM_sec_offset, Value: DIEInteger(8))
1577 ->sizeOf(FormParams: Unit.getOrigUnit().getFormParams());
1578 }
1579
1580 if (AttrSpec.Attr == dwarf::DW_AT_LLVM_stmt_sequence) {
1581 // If needed, we'll patch this sec_offset later with the correct offset.
1582 auto Patch = Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1583 Form: dwarf::DW_FORM_sec_offset,
1584 Value: DIEInteger(*Val.getAsSectionOffset()));
1585
1586 // Record this patch location so that it can be fixed up later.
1587 Unit.noteStmtSeqListAttribute(Attr: Patch);
1588
1589 return Unit.getOrigUnit().getFormParams().getDwarfOffsetByteSize();
1590 }
1591
1592 if (LLVM_UNLIKELY(Linker.Options.Update)) {
1593 if (auto OptionalValue = Val.getAsUnsignedConstant())
1594 Value = *OptionalValue;
1595 else if (auto OptionalValue = Val.getAsSignedConstant())
1596 Value = *OptionalValue;
1597 else if (auto OptionalValue = Val.getAsSectionOffset())
1598 Value = *OptionalValue;
1599 else {
1600 Linker.reportWarning(
1601 Warning: "Unsupported scalar attribute form. Dropping attribute.", File,
1602 DIE: &InputDIE);
1603 return 0;
1604 }
1605 if (AttrSpec.Attr == dwarf::DW_AT_declaration && Value)
1606 Info.IsDeclaration = true;
1607
1608 if (AttrSpec.Form == dwarf::DW_FORM_loclistx)
1609 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1610 Form: dwarf::Form(AttrSpec.Form), Value: DIELocList(Value));
1611 else
1612 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1613 Form: dwarf::Form(AttrSpec.Form), Value: DIEInteger(Value));
1614 return AttrSize;
1615 }
1616
1617 [[maybe_unused]] dwarf::Form OriginalForm = AttrSpec.Form;
1618 if (AttrSpec.Form == dwarf::DW_FORM_rnglistx) {
1619 // DWARFLinker does not generate .debug_addr table. Thus we need to change
1620 // all "addrx" related forms to "addr" version. Change DW_FORM_rnglistx
1621 // to DW_FORM_sec_offset here.
1622 std::optional<uint64_t> Index = Val.getAsSectionOffset();
1623 if (!Index) {
1624 Linker.reportWarning(Warning: "Cannot read the attribute. Dropping.", File,
1625 DIE: &InputDIE);
1626 return 0;
1627 }
1628 std::optional<uint64_t> Offset =
1629 Unit.getOrigUnit().getRnglistOffset(Index: *Index);
1630 if (!Offset) {
1631 Linker.reportWarning(Warning: "Cannot read the attribute. Dropping.", File,
1632 DIE: &InputDIE);
1633 return 0;
1634 }
1635
1636 Value = *Offset;
1637 AttrSpec.Form = dwarf::DW_FORM_sec_offset;
1638 AttrSize = Unit.getOrigUnit().getFormParams().getDwarfOffsetByteSize();
1639 } else if (AttrSpec.Form == dwarf::DW_FORM_loclistx) {
1640 // DWARFLinker does not generate .debug_addr table. Thus we need to change
1641 // all "addrx" related forms to "addr" version. Change DW_FORM_loclistx
1642 // to DW_FORM_sec_offset here.
1643 std::optional<uint64_t> Index = Val.getAsSectionOffset();
1644 if (!Index) {
1645 Linker.reportWarning(Warning: "Cannot read the attribute. Dropping.", File,
1646 DIE: &InputDIE);
1647 return 0;
1648 }
1649 std::optional<uint64_t> Offset =
1650 Unit.getOrigUnit().getLoclistOffset(Index: *Index);
1651 if (!Offset) {
1652 Linker.reportWarning(Warning: "Cannot read the attribute. Dropping.", File,
1653 DIE: &InputDIE);
1654 return 0;
1655 }
1656
1657 Value = *Offset;
1658 AttrSpec.Form = dwarf::DW_FORM_sec_offset;
1659 AttrSize = Unit.getOrigUnit().getFormParams().getDwarfOffsetByteSize();
1660 } else if (AttrSpec.Attr == dwarf::DW_AT_high_pc &&
1661 Die.getTag() == dwarf::DW_TAG_compile_unit) {
1662 std::optional<uint64_t> LowPC = Unit.getLowPc();
1663 if (!LowPC)
1664 return 0;
1665 // Dwarf >= 4 high_pc is an size, not an address.
1666 Value = Unit.getHighPc() - *LowPC;
1667 } else if (AttrSpec.Form == dwarf::DW_FORM_sec_offset)
1668 Value = *Val.getAsSectionOffset();
1669 else if (AttrSpec.Form == dwarf::DW_FORM_sdata)
1670 Value = *Val.getAsSignedConstant();
1671 else if (auto OptionalValue = Val.getAsUnsignedConstant())
1672 Value = *OptionalValue;
1673 else {
1674 Linker.reportWarning(
1675 Warning: "Unsupported scalar attribute form. Dropping attribute.", File,
1676 DIE: &InputDIE);
1677 return 0;
1678 }
1679
1680 // A compile unit's high_pc comes from the unit's own linked range and spans
1681 // every symbol in it.
1682 if (AttrSpec.Attr == dwarf::DW_AT_high_pc &&
1683 Die.getTag() != dwarf::DW_TAG_compile_unit)
1684 Value = constrainHighPC(InputDIE, HighPC: Value, /*IsLength=*/true, PCOffset: Info.PCOffset,
1685 Addresses&: *File.Addresses);
1686
1687 DIE::value_iterator Patch =
1688 Die.addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(AttrSpec.Attr),
1689 Form: dwarf::Form(AttrSpec.Form), Value: DIEInteger(Value));
1690 if (AttrSpec.Attr == dwarf::DW_AT_ranges ||
1691 AttrSpec.Attr == dwarf::DW_AT_start_scope) {
1692 Unit.noteRangeAttribute(Die, Attr: Patch);
1693 Info.HasRanges = true;
1694 } else if (DWARFAttribute::mayHaveLocationList(Attr: AttrSpec.Attr) &&
1695 dwarf::doesFormBelongToClass(Form: AttrSpec.Form,
1696 FC: DWARFFormValue::FC_SectionOffset,
1697 DwarfVersion: Unit.getOrigUnit().getVersion())) {
1698
1699 CompileUnit::DIEInfo &LocationDieInfo = Unit.getInfo(Die: InputDIE);
1700 Unit.noteLocationAttribute(Attr: {Patch, LocationDieInfo.InDebugMap
1701 ? LocationDieInfo.AddrAdjust
1702 : Info.PCOffset});
1703 } else if (AttrSpec.Attr == dwarf::DW_AT_declaration && Value)
1704 Info.IsDeclaration = true;
1705
1706 // check that all dwarf::DW_FORM_rnglistx are handled previously.
1707 assert((Info.HasRanges || (OriginalForm != dwarf::DW_FORM_rnglistx)) &&
1708 "Unhandled DW_FORM_rnglistx attribute");
1709
1710 return AttrSize;
1711}
1712
1713/// Clone \p InputDIE's attribute described by \p AttrSpec with
1714/// value \p Val, and add it to \p Die.
1715/// \returns the size of the cloned attribute.
1716unsigned DWARFLinker::DIECloner::cloneAttribute(
1717 DIE &Die, const DWARFDie &InputDIE, const DWARFFile &File,
1718 CompileUnit &Unit, const DWARFFormValue &Val, const AttributeSpec AttrSpec,
1719 unsigned AttrSize, AttributesInfo &Info, bool IsLittleEndian) {
1720 const DWARFUnit &U = Unit.getOrigUnit();
1721
1722 switch (AttrSpec.Form) {
1723 case dwarf::DW_FORM_strp:
1724 case dwarf::DW_FORM_line_strp:
1725 case dwarf::DW_FORM_string:
1726 case dwarf::DW_FORM_strx:
1727 case dwarf::DW_FORM_strx1:
1728 case dwarf::DW_FORM_strx2:
1729 case dwarf::DW_FORM_strx3:
1730 case dwarf::DW_FORM_strx4:
1731 return cloneStringAttribute(Die, AttrSpec, Val, U, Info);
1732 case dwarf::DW_FORM_ref_addr:
1733 case dwarf::DW_FORM_ref1:
1734 case dwarf::DW_FORM_ref2:
1735 case dwarf::DW_FORM_ref4:
1736 case dwarf::DW_FORM_ref8:
1737 return cloneDieReferenceAttribute(Die, InputDIE, AttrSpec, AttrSize, Val,
1738 File, Unit);
1739 case dwarf::DW_FORM_block:
1740 case dwarf::DW_FORM_block1:
1741 case dwarf::DW_FORM_block2:
1742 case dwarf::DW_FORM_block4:
1743 case dwarf::DW_FORM_exprloc:
1744 return cloneBlockAttribute(Die, InputDIE, File, Unit, AttrSpec, Val,
1745 IsLittleEndian);
1746 case dwarf::DW_FORM_addr:
1747 case dwarf::DW_FORM_addrx:
1748 case dwarf::DW_FORM_addrx1:
1749 case dwarf::DW_FORM_addrx2:
1750 case dwarf::DW_FORM_addrx3:
1751 case dwarf::DW_FORM_addrx4:
1752 return cloneAddressAttribute(Die, InputDIE, AttrSpec, AttrSize, Val, Unit,
1753 Info);
1754 case dwarf::DW_FORM_data1:
1755 case dwarf::DW_FORM_data2:
1756 case dwarf::DW_FORM_data4:
1757 case dwarf::DW_FORM_data8:
1758 case dwarf::DW_FORM_udata:
1759 case dwarf::DW_FORM_sdata:
1760 case dwarf::DW_FORM_sec_offset:
1761 case dwarf::DW_FORM_flag:
1762 case dwarf::DW_FORM_flag_present:
1763 case dwarf::DW_FORM_rnglistx:
1764 case dwarf::DW_FORM_loclistx:
1765 case dwarf::DW_FORM_implicit_const:
1766 return cloneScalarAttribute(Die, InputDIE, File, Unit, AttrSpec, Val,
1767 AttrSize, Info);
1768 default:
1769 Linker.reportWarning(Warning: "Unsupported attribute form " +
1770 dwarf::FormEncodingString(Encoding: AttrSpec.Form) +
1771 " in cloneAttribute. Dropping.",
1772 File, DIE: &InputDIE);
1773 }
1774
1775 return 0;
1776}
1777
1778void DWARFLinker::DIECloner::addObjCAccelerator(CompileUnit &Unit,
1779 const DIE *Die,
1780 DwarfStringPoolEntryRef Name,
1781 OffsetsStringPool &StringPool,
1782 bool SkipPubSection) {
1783 std::optional<ObjCSelectorNames> Names =
1784 getObjCNamesIfSelector(Name: Name.getString());
1785 if (!Names)
1786 return;
1787 Unit.addNameAccelerator(Die, Name: StringPool.getEntry(S: Names->Selector),
1788 SkipPubnamesSection: SkipPubSection);
1789 Unit.addObjCAccelerator(Die, Name: StringPool.getEntry(S: Names->ClassName),
1790 SkipPubnamesSection: SkipPubSection);
1791 if (Names->ClassNameNoCategory)
1792 Unit.addObjCAccelerator(
1793 Die, Name: StringPool.getEntry(S: *Names->ClassNameNoCategory), SkipPubnamesSection: SkipPubSection);
1794 if (Names->MethodNameNoCategory)
1795 Unit.addNameAccelerator(
1796 Die, Name: StringPool.getEntry(S: *Names->MethodNameNoCategory), SkipPubnamesSection: SkipPubSection);
1797}
1798
1799static bool
1800shouldSkipAttribute(bool Update,
1801 DWARFAbbreviationDeclaration::AttributeSpec AttrSpec,
1802 bool SkipPC) {
1803 switch (AttrSpec.Attr) {
1804 default:
1805 return false;
1806 case dwarf::DW_AT_low_pc:
1807 case dwarf::DW_AT_high_pc:
1808 case dwarf::DW_AT_ranges:
1809 return !Update && SkipPC;
1810 case dwarf::DW_AT_rnglists_base:
1811 // In case !Update the .debug_addr table is not generated/preserved.
1812 // Thus instead of DW_FORM_rnglistx the DW_FORM_sec_offset is used.
1813 // Since DW_AT_rnglists_base is used for only DW_FORM_rnglistx the
1814 // DW_AT_rnglists_base is removed.
1815 return !Update;
1816 case dwarf::DW_AT_loclists_base:
1817 // In case !Update the .debug_addr table is not generated/preserved.
1818 // Thus instead of DW_FORM_loclistx the DW_FORM_sec_offset is used.
1819 // Since DW_AT_loclists_base is used for only DW_FORM_loclistx the
1820 // DW_AT_loclists_base is removed.
1821 return !Update;
1822 case dwarf::DW_AT_location:
1823 case dwarf::DW_AT_frame_base:
1824 return !Update && SkipPC;
1825 }
1826}
1827
1828struct AttributeLinkedOffsetFixup {
1829 int64_t LinkedOffsetFixupVal;
1830 uint64_t InputAttrStartOffset;
1831 uint64_t InputAttrEndOffset;
1832};
1833
1834DIE *DWARFLinker::DIECloner::cloneDIE(const DWARFDie &InputDIE,
1835 const DWARFFile &File, CompileUnit &Unit,
1836 int64_t PCOffset, uint32_t OutOffset,
1837 unsigned Flags, bool IsLittleEndian,
1838 DIE *Die) {
1839 DWARFUnit &U = Unit.getOrigUnit();
1840 unsigned Idx = U.getDIEIndex(D: InputDIE);
1841 CompileUnit::DIEInfo &Info = Unit.getInfo(Idx);
1842
1843 // Should the DIE appear in the output?
1844 if (!Unit.getInfo(Idx).Keep)
1845 return nullptr;
1846
1847 uint64_t Offset = InputDIE.getOffset();
1848 assert(!(Die && Info.Clone) && "Can't supply a DIE and a cloned DIE");
1849 if (!Die) {
1850 // The DIE might have been already created by a forward reference
1851 // (see cloneDieReferenceAttribute()).
1852 if (!Info.Clone)
1853 Info.Clone = DIE::get(Alloc&: DIEAlloc, Tag: dwarf::Tag(InputDIE.getTag()));
1854 Die = Info.Clone;
1855 }
1856
1857 assert(Die->getTag() == InputDIE.getTag());
1858 Die->setOffset(OutOffset);
1859 if (isODRCanonicalCandidate(Die: InputDIE, CU&: Unit) && Info.Ctxt &&
1860 (Info.Ctxt->getCanonicalDIEOffset() == 0)) {
1861 if (!Info.Ctxt->hasCanonicalDIE())
1862 Info.Ctxt->setHasCanonicalDIE();
1863 // We are about to emit a DIE that is the root of its own valid
1864 // DeclContext tree. Make the current offset the canonical offset
1865 // for this context.
1866 Info.Ctxt->setCanonicalDIEOffset(OutOffset + Unit.getStartOffset());
1867 }
1868
1869 // Extract and clone every attribute.
1870 DWARFDataExtractor Data = U.getDebugInfoExtractor();
1871 // Point to the next DIE (generally there is always at least a NULL
1872 // entry after the current one). If this is a lone
1873 // DW_TAG_compile_unit without any children, point to the next unit.
1874 uint64_t NextOffset = (Idx + 1 < U.getNumDIEs())
1875 ? U.getDIEAtIndex(Index: Idx + 1).getOffset()
1876 : U.getNextUnitOffset();
1877 AttributesInfo AttrInfo;
1878
1879 // We could copy the data only if we need to apply a relocation to it. After
1880 // testing, it seems there is no performance downside to doing the copy
1881 // unconditionally, and it makes the code simpler.
1882 SmallString<40> DIECopy(Data.getData().substr(Start: Offset, N: NextOffset - Offset));
1883 Data =
1884 DWARFDataExtractor(DIECopy, Data.isLittleEndian(), Data.getAddressSize());
1885
1886 // Modify the copy with relocated addresses.
1887 ObjFile.Addresses->applyValidRelocs(Data: DIECopy, BaseOffset: Offset, IsLittleEndian: Data.isLittleEndian());
1888
1889 // Reset the Offset to 0 as we will be working on the local copy of
1890 // the data.
1891 Offset = 0;
1892
1893 const auto *Abbrev = InputDIE.getAbbreviationDeclarationPtr();
1894 Offset += getULEB128Size(Value: Abbrev->getCode());
1895
1896 // We are entering a subprogram. Get and propagate the PCOffset.
1897 if (Die->getTag() == dwarf::DW_TAG_subprogram)
1898 PCOffset = Info.AddrAdjust;
1899 AttrInfo.PCOffset = PCOffset;
1900
1901 if (Abbrev->getTag() == dwarf::DW_TAG_subprogram) {
1902 Flags |= TF_InFunctionScope;
1903 if (LLVM_LIKELY(!Update)) {
1904 if (Info.InDebugMap)
1905 Flags &= ~TF_SkipPC;
1906 else
1907 Flags |= TF_SkipPC;
1908 }
1909 } else if (Abbrev->getTag() == dwarf::DW_TAG_variable) {
1910 // Function-local globals could be in the debug map even when the function
1911 // is not, e.g., inlined functions.
1912 if ((Flags & TF_InFunctionScope) && Info.InDebugMap)
1913 Flags &= ~TF_SkipPC;
1914 // Location expressions referencing an address which is not in debug map
1915 // should be deleted.
1916 else if (!Info.InDebugMap && Info.HasLocationExpressionAddr &&
1917 LLVM_LIKELY(!Update))
1918 Flags |= TF_SkipPC;
1919 }
1920
1921 std::optional<StringRef> LibraryInstallName =
1922 ObjFile.Addresses->getLibraryInstallName();
1923 SmallVector<AttributeLinkedOffsetFixup> AttributesFixups;
1924 for (const auto &AttrSpec : Abbrev->attributes()) {
1925 if (shouldSkipAttribute(Update, AttrSpec, SkipPC: Flags & TF_SkipPC)) {
1926 DWARFFormValue::skipValue(Form: AttrSpec.Form, DebugInfoData: Data, OffsetPtr: &Offset,
1927 FormParams: U.getFormParams());
1928 continue;
1929 }
1930
1931 AttributeLinkedOffsetFixup CurAttrFixup;
1932 CurAttrFixup.InputAttrStartOffset = InputDIE.getOffset() + Offset;
1933 CurAttrFixup.LinkedOffsetFixupVal =
1934 Unit.getStartOffset() + OutOffset - CurAttrFixup.InputAttrStartOffset;
1935
1936 DWARFFormValue Val = AttrSpec.getFormValue();
1937 uint64_t AttrSize = Offset;
1938 Val.extractValue(Data, OffsetPtr: &Offset, FormParams: U.getFormParams(), U: &U);
1939 CurAttrFixup.InputAttrEndOffset = InputDIE.getOffset() + Offset;
1940 AttrSize = Offset - AttrSize;
1941
1942 uint64_t FinalAttrSize =
1943 cloneAttribute(Die&: *Die, InputDIE, File, Unit, Val, AttrSpec, AttrSize,
1944 Info&: AttrInfo, IsLittleEndian);
1945 if (FinalAttrSize != 0 && ObjFile.Addresses->needToSaveValidRelocs())
1946 AttributesFixups.push_back(Elt: CurAttrFixup);
1947
1948 OutOffset += FinalAttrSize;
1949 }
1950
1951 uint16_t Tag = InputDIE.getTag();
1952 // Add the DW_AT_APPLE_origin attribute to Compile Unit die if we have
1953 // an install name and the DWARF doesn't have the attribute yet.
1954 const bool NeedsAppleOrigin = (Tag == dwarf::DW_TAG_compile_unit) &&
1955 LibraryInstallName.has_value() &&
1956 !AttrInfo.HasAppleOrigin;
1957 if (NeedsAppleOrigin) {
1958 auto StringEntry = DebugStrPool.getEntry(S: LibraryInstallName.value());
1959 Die->addValue(Alloc&: DIEAlloc, Attribute: dwarf::Attribute(dwarf::DW_AT_APPLE_origin),
1960 Form: dwarf::DW_FORM_strp, Value: DIEInteger(StringEntry.getOffset()));
1961 AttrInfo.Name = StringEntry;
1962 OutOffset += 4;
1963 }
1964
1965 // Look for accelerator entries.
1966 // FIXME: This is slightly wrong. An inline_subroutine without a
1967 // low_pc, but with AT_ranges might be interesting to get into the
1968 // accelerator tables too. For now stick with dsymutil's behavior.
1969 if ((Info.InDebugMap || AttrInfo.HasLowPc || AttrInfo.HasRanges) &&
1970 Tag != dwarf::DW_TAG_compile_unit &&
1971 !(Tag == dwarf::DW_TAG_variable &&
1972 hasImplicitAddressLocation(Die: InputDIE)) &&
1973 getDIENames(Die: InputDIE, Info&: AttrInfo, StringPool&: DebugStrPool, File, Unit,
1974 StripTemplate: Tag != dwarf::DW_TAG_inlined_subroutine)) {
1975 if (AttrInfo.MangledName && AttrInfo.MangledName != AttrInfo.Name)
1976 Unit.addNameAccelerator(Die, Name: AttrInfo.MangledName,
1977 SkipPubnamesSection: Tag == dwarf::DW_TAG_inlined_subroutine);
1978 if (AttrInfo.Name) {
1979 if (AttrInfo.NameWithoutTemplate)
1980 Unit.addNameAccelerator(Die, Name: AttrInfo.NameWithoutTemplate,
1981 /* SkipPubSection */ SkipPubnamesSection: true);
1982 Unit.addNameAccelerator(Die, Name: AttrInfo.Name,
1983 SkipPubnamesSection: Tag == dwarf::DW_TAG_inlined_subroutine);
1984 }
1985 if (AttrInfo.Name)
1986 addObjCAccelerator(Unit, Die, Name: AttrInfo.Name, StringPool&: DebugStrPool,
1987 /* SkipPubSection =*/true);
1988
1989 } else if (Tag == dwarf::DW_TAG_namespace) {
1990 if (!AttrInfo.Name)
1991 AttrInfo.Name = DebugStrPool.getEntry(S: "(anonymous namespace)");
1992 Unit.addNamespaceAccelerator(Die, Name: AttrInfo.Name);
1993 } else if (Tag == dwarf::DW_TAG_imported_declaration && AttrInfo.Name) {
1994 Unit.addNamespaceAccelerator(Die, Name: AttrInfo.Name);
1995 } else if (isTypeTag(Tag) && !AttrInfo.IsDeclaration) {
1996 bool Success = getDIENames(Die: InputDIE, Info&: AttrInfo, StringPool&: DebugStrPool, File, Unit);
1997 uint64_t RuntimeLang =
1998 dwarf::toUnsigned(V: InputDIE.find(Attr: dwarf::DW_AT_APPLE_runtime_class))
1999 .value_or(u: 0);
2000 bool ObjCClassIsImplementation =
2001 (RuntimeLang == dwarf::DW_LANG_ObjC ||
2002 RuntimeLang == dwarf::DW_LANG_ObjC_plus_plus) &&
2003 dwarf::toUnsigned(V: InputDIE.find(Attr: dwarf::DW_AT_APPLE_objc_complete_type))
2004 .value_or(u: 0);
2005 if (Success && AttrInfo.Name && !AttrInfo.Name.getString().empty()) {
2006 uint32_t Hash = hashFullyQualifiedName(DIE: InputDIE, U&: Unit, File);
2007 Unit.addTypeAccelerator(Die, Name: AttrInfo.Name, ObjcClassImplementation: ObjCClassIsImplementation,
2008 QualifiedNameHash: Hash);
2009 }
2010
2011 // For Swift, mangled names are put into DW_AT_linkage_name.
2012 if (Success && AttrInfo.MangledName &&
2013 RuntimeLang == dwarf::DW_LANG_Swift &&
2014 !AttrInfo.MangledName.getString().empty() &&
2015 AttrInfo.MangledName != AttrInfo.Name) {
2016 auto Hash = djbHash(Buffer: AttrInfo.MangledName.getString().data());
2017 Unit.addTypeAccelerator(Die, Name: AttrInfo.MangledName,
2018 ObjcClassImplementation: ObjCClassIsImplementation, QualifiedNameHash: Hash);
2019 }
2020 }
2021
2022 // Determine whether there are any children that we want to keep.
2023 bool HasChildren = false;
2024 for (auto Child : InputDIE.children()) {
2025 unsigned Idx = U.getDIEIndex(D: Child);
2026 if (Unit.getInfo(Idx).Keep) {
2027 HasChildren = true;
2028 break;
2029 }
2030 }
2031
2032 if (Unit.getOrigUnit().getVersion() >= 5 && !AttrInfo.AttrStrOffsetBaseSeen &&
2033 Die->getTag() == dwarf::DW_TAG_compile_unit) {
2034 // No DW_AT_str_offsets_base seen, add it to the DIE.
2035 Die->addValue(Alloc&: DIEAlloc, Attribute: dwarf::DW_AT_str_offsets_base,
2036 Form: dwarf::DW_FORM_sec_offset, Value: DIEInteger(8));
2037 OutOffset += 4;
2038 }
2039
2040 DIEAbbrev NewAbbrev = Die->generateAbbrev();
2041 if (HasChildren)
2042 NewAbbrev.setChildrenFlag(dwarf::DW_CHILDREN_yes);
2043 // Assign a permanent abbrev number
2044 Linker.assignAbbrev(Abbrev&: NewAbbrev);
2045 Die->setAbbrevNumber(NewAbbrev.getNumber());
2046
2047 uint64_t AbbrevNumberSize = getULEB128Size(Value: Die->getAbbrevNumber());
2048
2049 // Add the size of the abbreviation number to the output offset.
2050 OutOffset += AbbrevNumberSize;
2051
2052 // Update fixups with the size of the abbreviation number
2053 for (AttributeLinkedOffsetFixup &F : AttributesFixups)
2054 F.LinkedOffsetFixupVal += AbbrevNumberSize;
2055
2056 for (AttributeLinkedOffsetFixup &F : AttributesFixups)
2057 ObjFile.Addresses->updateAndSaveValidRelocs(
2058 IsDWARF5: Unit.getOrigUnit().getVersion() >= 5, OriginalUnitOffset: Unit.getOrigUnit().getOffset(),
2059 LinkedOffset: F.LinkedOffsetFixupVal, StartOffset: F.InputAttrStartOffset, EndOffset: F.InputAttrEndOffset);
2060
2061 if (!HasChildren) {
2062 // Update our size.
2063 Die->setSize(OutOffset - Die->getOffset());
2064 return Die;
2065 }
2066
2067 // Recursively clone children.
2068 for (auto Child : InputDIE.children()) {
2069 if (DIE *Clone = cloneDIE(InputDIE: Child, File, Unit, PCOffset, OutOffset, Flags,
2070 IsLittleEndian)) {
2071 Die->addChild(Child: Clone);
2072 OutOffset = Clone->getOffset() + Clone->getSize();
2073 }
2074 }
2075
2076 // Account for the end of children marker.
2077 OutOffset += sizeof(int8_t);
2078 // Update our size.
2079 Die->setSize(OutOffset - Die->getOffset());
2080 return Die;
2081}
2082
2083/// Patch the input object file relevant debug_ranges or debug_rnglists
2084/// entries and emit them in the output file. Update the relevant attributes
2085/// to point at the new entries.
2086Error DWARFLinker::generateUnitRanges(CompileUnit &Unit, const DWARFFile &File,
2087 DebugDieValuePool &AddrPool) const {
2088 if (LLVM_UNLIKELY(Options.Update))
2089 return Error::success();
2090
2091 const auto &FunctionRanges = Unit.getFunctionRanges();
2092
2093 // Build set of linked address ranges for unit function ranges.
2094 AddressRanges LinkedFunctionRanges;
2095 for (const AddressRangeValuePair &Range : FunctionRanges)
2096 LinkedFunctionRanges.insert(
2097 Range: {Range.Range.start() + Range.Value, Range.Range.end() + Range.Value});
2098
2099 // Emit LinkedFunctionRanges into .debug_aranges
2100 if (!LinkedFunctionRanges.empty())
2101 TheDwarfEmitter->emitDwarfDebugArangesTable(Unit, LinkedRanges: LinkedFunctionRanges);
2102
2103 RngListAttributesTy AllRngListAttributes = Unit.getRangesAttributes();
2104 std::optional<PatchLocation> UnitRngListAttribute =
2105 Unit.getUnitRangesAttribute();
2106
2107 if (!AllRngListAttributes.empty() || UnitRngListAttribute) {
2108 std::optional<AddressRangeValuePair> CachedRange;
2109 MCSymbol *EndLabel = TheDwarfEmitter->emitDwarfDebugRangeListHeader(Unit);
2110
2111 // Read original address ranges, apply relocation value, emit linked address
2112 // ranges.
2113 for (PatchLocation &AttributePatch : AllRngListAttributes) {
2114 // Get ranges from the source DWARF corresponding to the current
2115 // attribute.
2116 AddressRanges LinkedRanges;
2117 if (Expected<DWARFAddressRangesVector> OriginalRanges =
2118 Unit.getOrigUnit().findRnglistFromOffset(Offset: AttributePatch.get())) {
2119 // Apply relocation adjustment.
2120 for (const auto &Range : *OriginalRanges) {
2121 if (!CachedRange || !CachedRange->Range.contains(Addr: Range.LowPC))
2122 CachedRange = FunctionRanges.getRangeThatContains(Addr: Range.LowPC);
2123
2124 // All range entries should lie in the function range.
2125 if (!CachedRange) {
2126 reportWarning(Warning: "inconsistent range data.", File);
2127 continue;
2128 }
2129
2130 // Store range for emiting.
2131 LinkedRanges.insert(Range: {Range.LowPC + CachedRange->Value,
2132 Range.HighPC + CachedRange->Value});
2133 }
2134 } else {
2135 llvm::consumeError(Err: OriginalRanges.takeError());
2136 reportWarning(Warning: "invalid range list ignored.", File);
2137 }
2138
2139 // Emit linked ranges.
2140 if (Error E = TheDwarfEmitter->emitDwarfDebugRangeListFragment(
2141 Unit, LinkedRanges, Patch: AttributePatch, AddrPool))
2142 return E;
2143 }
2144
2145 // Emit ranges for Unit AT_ranges attribute.
2146 if (UnitRngListAttribute.has_value())
2147 if (Error E = TheDwarfEmitter->emitDwarfDebugRangeListFragment(
2148 Unit, LinkedRanges: LinkedFunctionRanges, Patch: *UnitRngListAttribute, AddrPool))
2149 return E;
2150
2151 // Emit ranges footer.
2152 TheDwarfEmitter->emitDwarfDebugRangeListFooter(Unit, EndLabel);
2153 }
2154
2155 return Error::success();
2156}
2157
2158Error DWARFLinker::DIECloner::generateUnitLocations(
2159 CompileUnit &Unit, const DWARFFile &File,
2160 ExpressionHandlerRef ExprHandler) {
2161 if (LLVM_UNLIKELY(Linker.Options.Update))
2162 return Error::success();
2163
2164 const LocListAttributesTy &AllLocListAttributes =
2165 Unit.getLocationAttributes();
2166
2167 if (AllLocListAttributes.empty())
2168 return Error::success();
2169
2170 // Emit locations list table header.
2171 MCSymbol *EndLabel = Emitter->emitDwarfDebugLocListHeader(Unit);
2172
2173 for (auto &CurLocAttr : AllLocListAttributes) {
2174 // Get location expressions vector corresponding to the current attribute
2175 // from the source DWARF.
2176 Expected<DWARFLocationExpressionsVector> OriginalLocations =
2177 Unit.getOrigUnit().findLoclistFromOffset(Offset: CurLocAttr.get());
2178
2179 if (!OriginalLocations) {
2180 llvm::consumeError(Err: OriginalLocations.takeError());
2181 Linker.reportWarning(Warning: "Invalid location attribute ignored.", File);
2182 continue;
2183 }
2184
2185 DWARFLocationExpressionsVector LinkedLocationExpressions;
2186 for (DWARFLocationExpression &CurExpression : *OriginalLocations) {
2187 DWARFLocationExpression LinkedExpression;
2188
2189 if (CurExpression.Range) {
2190 // Relocate address range.
2191 LinkedExpression.Range = {
2192 CurExpression.Range->LowPC + CurLocAttr.RelocAdjustment,
2193 CurExpression.Range->HighPC + CurLocAttr.RelocAdjustment};
2194 }
2195
2196 // Clone expression.
2197 LinkedExpression.Expr.reserve(N: CurExpression.Expr.size());
2198 ExprHandler(CurExpression.Expr, LinkedExpression.Expr,
2199 CurLocAttr.RelocAdjustment);
2200
2201 LinkedLocationExpressions.push_back(x: LinkedExpression);
2202 }
2203
2204 // Emit locations list table fragment corresponding to the CurLocAttr.
2205 if (Error E = Emitter->emitDwarfDebugLocListFragment(
2206 Unit, LinkedLocationExpression: LinkedLocationExpressions, Patch: CurLocAttr, AddrPool))
2207 return E;
2208 }
2209
2210 // Emit locations list table footer.
2211 Emitter->emitDwarfDebugLocListFooter(Unit, EndLabel);
2212
2213 return Error::success();
2214}
2215
2216static void patchAddrBase(DIE &Die, DIEInteger Offset) {
2217 for (auto &V : Die.values())
2218 if (V.getAttribute() == dwarf::DW_AT_addr_base) {
2219 V = DIEValue(V.getAttribute(), V.getForm(), Offset);
2220 return;
2221 }
2222
2223 llvm_unreachable("Didn't find a DW_AT_addr_base in cloned DIE!");
2224}
2225
2226Error DWARFLinker::DIECloner::emitDebugAddrSection(
2227 CompileUnit &Unit, const uint16_t DwarfVersion) const {
2228
2229 if (LLVM_UNLIKELY(Linker.Options.Update))
2230 return Error::success();
2231
2232 if (DwarfVersion < 5)
2233 return Error::success();
2234
2235 if (AddrPool.getValues().empty())
2236 return Error::success();
2237
2238 MCSymbol *EndLabel = Emitter->emitDwarfDebugAddrsHeader(Unit);
2239 uint64_t AddrOffset = Emitter->getDebugAddrSectionSize();
2240 dwarf::FormParams FP = Unit.getOrigUnit().getFormParams();
2241 if (AddrOffset > FP.getDwarfMaxOffset())
2242 return createStringError(S: ".debug_addr section offset 0x" +
2243 Twine::utohexstr(Val: AddrOffset) + " exceeds the " +
2244 dwarf::FormatString(Format: FP.Format) + " limit");
2245 patchAddrBase(Die&: *Unit.getOutputUnitDIE(), Offset: DIEInteger(AddrOffset));
2246 Emitter->emitDwarfDebugAddrs(Addrs: AddrPool.getValues(),
2247 AddrSize: Unit.getOrigUnit().getAddressByteSize());
2248 Emitter->emitDwarfDebugAddrsFooter(Unit, EndLabel);
2249
2250 return Error::success();
2251}
2252
2253/// A helper struct to help keep track of the association between the input and
2254/// output rows during line table rewriting. This is used to patch
2255/// DW_AT_LLVM_stmt_sequence attributes, which reference a particular line table
2256/// row.
2257struct TrackedRow {
2258 DWARFDebugLine::Row Row;
2259 size_t OriginalRowIndex;
2260 bool isStartSeqInOutput;
2261};
2262
2263/// Insert the new line info sequence \p Seq into the current
2264/// set of already linked line info \p Rows.
2265static void insertLineSequence(std::vector<TrackedRow> &Seq,
2266 std::vector<TrackedRow> &Rows) {
2267 if (Seq.empty())
2268 return;
2269
2270 // Mark the first row in Seq to indicate it is the start of a sequence
2271 // in the output line table.
2272 Seq.front().isStartSeqInOutput = true;
2273
2274 if (!Rows.empty() && Rows.back().Row.Address < Seq.front().Row.Address) {
2275 llvm::append_range(C&: Rows, R&: Seq);
2276 Seq.clear();
2277 return;
2278 }
2279
2280 object::SectionedAddress Front = Seq.front().Row.Address;
2281 auto InsertPoint = partition_point(
2282 Range&: Rows, P: [=](const TrackedRow &O) { return O.Row.Address < Front; });
2283
2284 // FIXME: this only removes the unneeded end_sequence if the
2285 // sequences have been inserted in order. Using a global sort like
2286 // described in generateLineTableForUnit() and delaying the end_sequence
2287 // elimination to emitLineTableForUnit() we can get rid of all of them.
2288 if (InsertPoint != Rows.end() && InsertPoint->Row.Address == Front &&
2289 InsertPoint->Row.EndSequence) {
2290 *InsertPoint = Seq.front();
2291 Rows.insert(position: InsertPoint + 1, first: Seq.begin() + 1, last: Seq.end());
2292 } else {
2293 Rows.insert(position: InsertPoint, first: Seq.begin(), last: Seq.end());
2294 }
2295
2296 Seq.clear();
2297}
2298
2299static void patchStmtList(DIE &Die, DIEInteger Offset) {
2300 for (auto &V : Die.values())
2301 if (V.getAttribute() == dwarf::DW_AT_stmt_list) {
2302 V = DIEValue(V.getAttribute(), V.getForm(), Offset);
2303 return;
2304 }
2305
2306 llvm_unreachable("Didn't find DW_AT_stmt_list in cloned DIE!");
2307}
2308
2309void DWARFLinker::DIECloner::rememberUnitForMacroOffset(CompileUnit &Unit) {
2310 DWARFUnit &OrigUnit = Unit.getOrigUnit();
2311 DWARFDie OrigUnitDie = OrigUnit.getUnitDIE();
2312
2313 if (std::optional<uint64_t> MacroAttr =
2314 dwarf::toSectionOffset(V: OrigUnitDie.find(Attr: dwarf::DW_AT_macros))) {
2315 UnitMacroMap.insert(KV: std::make_pair(x&: *MacroAttr, y: &Unit));
2316 return;
2317 }
2318
2319 if (std::optional<uint64_t> MacroAttr =
2320 dwarf::toSectionOffset(V: OrigUnitDie.find(Attr: dwarf::DW_AT_macro_info))) {
2321 UnitMacroMap.insert(KV: std::make_pair(x&: *MacroAttr, y: &Unit));
2322 return;
2323 }
2324}
2325
2326Error DWARFLinker::DIECloner::generateLineTableForUnit(CompileUnit &Unit) {
2327 if (LLVM_UNLIKELY(Emitter == nullptr))
2328 return Error::success();
2329
2330 // Check whether DW_AT_stmt_list attribute is presented.
2331 DWARFDie CUDie = Unit.getOrigUnit().getUnitDIE();
2332 auto StmtList = dwarf::toSectionOffset(V: CUDie.find(Attr: dwarf::DW_AT_stmt_list));
2333 if (!StmtList)
2334 return Error::success();
2335
2336 // Update the cloned DW_AT_stmt_list with the correct debug_line offset.
2337 if (auto *OutputDIE = Unit.getOutputUnitDIE()) {
2338 uint64_t StmtOffset = Emitter->getLineSectionSize();
2339 dwarf::FormParams FP = Unit.getOrigUnit().getFormParams();
2340 if (StmtOffset > FP.getDwarfMaxOffset())
2341 return createStringError(S: ".debug_line section offset 0x" +
2342 Twine::utohexstr(Val: StmtOffset) + " exceeds the " +
2343 dwarf::FormatString(Format: FP.Format) + " limit");
2344 patchStmtList(Die&: *OutputDIE, Offset: DIEInteger(StmtOffset));
2345 }
2346
2347 if (const DWARFDebugLine::LineTable *LT =
2348 ObjFile.Dwarf->getLineTableForUnit(U: &Unit.getOrigUnit())) {
2349
2350 DWARFDebugLine::LineTable LineTable;
2351
2352 // Set Line Table header.
2353 LineTable.Prologue = LT->Prologue;
2354
2355 // Set Line Table Rows.
2356 if (Linker.Options.Update) {
2357 LineTable.Rows = LT->Rows;
2358 // If all the line table contains is a DW_LNE_end_sequence, clear the line
2359 // table rows, it will be inserted again in the DWARFStreamer.
2360 if (LineTable.Rows.size() == 1 && LineTable.Rows[0].EndSequence)
2361 LineTable.Rows.clear();
2362
2363 LineTable.Sequences = LT->Sequences;
2364
2365 Emitter->emitLineTableForUnit(LineTable, Unit, DebugStrPool,
2366 DebugLineStrPool);
2367 } else {
2368 // Create TrackedRow objects for all input rows.
2369 std::vector<TrackedRow> InputRows;
2370 InputRows.reserve(n: LT->Rows.size());
2371 for (size_t i = 0; i < LT->Rows.size(); i++)
2372 InputRows.emplace_back(args: TrackedRow{.Row: LT->Rows[i], .OriginalRowIndex: i, .isStartSeqInOutput: false});
2373
2374 // This vector is the output line table (still in TrackedRow form).
2375 std::vector<TrackedRow> OutputRows;
2376 OutputRows.reserve(n: InputRows.size());
2377
2378 // Current sequence of rows being extracted, before being inserted
2379 // in OutputRows.
2380 std::vector<TrackedRow> Seq;
2381 Seq.reserve(n: InputRows.size());
2382
2383 const auto &FunctionRanges = Unit.getFunctionRanges();
2384 std::optional<AddressRangeValuePair> CurrRange;
2385
2386 // FIXME: This logic is meant to generate exactly the same output as
2387 // Darwin's classic dsymutil. There is a nicer way to implement this
2388 // by simply putting all the relocated line info in OutputRows and simply
2389 // sorting OutputRows before passing it to emitLineTableForUnit. This
2390 // should be correct as sequences for a function should stay
2391 // together in the sorted output. There are a few corner cases that
2392 // look suspicious though, and that required to implement the logic
2393 // this way. Revisit that once initial validation is finished.
2394
2395 // Iterate over the object file line info and extract the sequences
2396 // that correspond to linked functions.
2397 for (size_t i = 0; i < InputRows.size(); i++) {
2398 TrackedRow TR = InputRows[i];
2399
2400 // Check whether we stepped out of the range. The range is
2401 // half-open, but consider accepting the end address of the range if
2402 // it is marked as end_sequence in the input (because in that
2403 // case, the relocation offset is accurate and that entry won't
2404 // serve as the start of another function).
2405 if (!CurrRange || !CurrRange->Range.contains(Addr: TR.Row.Address.Address)) {
2406 // We just stepped out of a known range. Insert an end_sequence
2407 // corresponding to the end of the range.
2408 uint64_t StopAddress =
2409 CurrRange ? CurrRange->Range.end() + CurrRange->Value : -1ULL;
2410 CurrRange =
2411 FunctionRanges.getRangeThatContains(Addr: TR.Row.Address.Address);
2412 if (StopAddress != -1ULL && !Seq.empty()) {
2413 // Insert end sequence row with the computed end address, but
2414 // the same line as the previous one.
2415 auto NextLine = Seq.back();
2416 NextLine.Row.Address.Address = StopAddress;
2417 NextLine.Row.EndSequence = 1;
2418 NextLine.Row.PrologueEnd = 0;
2419 NextLine.Row.BasicBlock = 0;
2420 NextLine.Row.EpilogueBegin = 0;
2421 Seq.push_back(x: NextLine);
2422 insertLineSequence(Seq, Rows&: OutputRows);
2423 }
2424
2425 if (!CurrRange)
2426 continue;
2427 }
2428
2429 // Ignore empty sequences.
2430 if (TR.Row.EndSequence && Seq.empty())
2431 continue;
2432
2433 // Relocate row address and add it to the current sequence.
2434 TR.Row.Address.Address += CurrRange->Value;
2435 Seq.push_back(x: TR);
2436
2437 if (TR.Row.EndSequence)
2438 insertLineSequence(Seq, Rows&: OutputRows);
2439 }
2440
2441 // Recompute isStartSeqInOutput based on the final row ordering.
2442 // A row is a sequence start (will have DW_LNE_set_address emitted) iff:
2443 // 1. It's the first row, OR
2444 // 2. The previous row has EndSequence = 1
2445 // This is necessary because insertLineSequence may merge sequences when
2446 // an EndSequence row is replaced by the start of a new sequence, which
2447 // removes the EndSequence marker and invalidates the original flag.
2448 if (!OutputRows.empty()) {
2449 OutputRows[0].isStartSeqInOutput = true;
2450 for (size_t i = 1; i < OutputRows.size(); ++i)
2451 OutputRows[i].isStartSeqInOutput = OutputRows[i - 1].Row.EndSequence;
2452 }
2453
2454 // Materialize the tracked rows into final DWARFDebugLine::Row objects.
2455 LineTable.Rows.clear();
2456 LineTable.Rows.reserve(n: OutputRows.size());
2457 for (auto &TR : OutputRows)
2458 LineTable.Rows.push_back(x: TR.Row);
2459
2460 // Use OutputRowOffsets to store the offsets of each line table row in the
2461 // output .debug_line section.
2462 std::vector<uint64_t> OutputRowOffsets;
2463
2464 // The unit might not have any DW_AT_LLVM_stmt_sequence attributes, so use
2465 // hasStmtSeq to skip the patching logic.
2466 bool hasStmtSeq = Unit.getStmtSeqListAttributes().size() > 0;
2467 Emitter->emitLineTableForUnit(LineTable, Unit, DebugStrPool,
2468 DebugLineStrPool,
2469 RowOffsets: hasStmtSeq ? &OutputRowOffsets : nullptr);
2470
2471 if (hasStmtSeq) {
2472 assert(OutputRowOffsets.size() == OutputRows.size() &&
2473 "must have an offset for each row");
2474
2475 // Create a map of stmt sequence offsets to original row indices.
2476 DenseMap<uint64_t, uint64_t> SeqOffToOrigRow;
2477 // The DWARF parser's discovery of sequences can be incomplete. To
2478 // ensure all DW_AT_LLVM_stmt_sequence attributes can be patched, we
2479 // build a map from both the parser's results and a manual
2480 // reconstruction.
2481 if (!LT->Rows.empty())
2482 constructSeqOffsettoOrigRowMapping(Unit, LT: *LT, SeqOffToOrigRow);
2483
2484 // Build two maps to handle stmt_sequence patching:
2485 // 1. OrigRowToOutputRow: maps original row indices to output row
2486 // indices (for all rows, not just sequence starts).
2487 // 2. OutputRowToSeqStart: maps each output row index to its sequence
2488 // start's output row index
2489 DenseMap<size_t, size_t> OrigRowToOutputRow;
2490 std::vector<size_t> OutputRowToSeqStart(OutputRows.size());
2491
2492 size_t CurrentSeqStart = 0;
2493 for (size_t i = 0; i < OutputRows.size(); ++i) {
2494 // Track the current sequence start.
2495 if (OutputRows[i].isStartSeqInOutput)
2496 CurrentSeqStart = i;
2497 OutputRowToSeqStart[i] = CurrentSeqStart;
2498
2499 // Map original row index to output row index.
2500 OrigRowToOutputRow[OutputRows[i].OriginalRowIndex] = i;
2501 }
2502
2503 // Patch DW_AT_LLVM_stmt_sequence attributes in the compile unit DIE
2504 // with the correct offset into the .debug_line section.
2505 for (const auto &StmtSeq : Unit.getStmtSeqListAttributes()) {
2506 uint64_t OrigStmtSeq = StmtSeq.get();
2507 // 1. Get the original row index from the stmt list offset.
2508 auto OrigRowIter = SeqOffToOrigRow.find(Val: OrigStmtSeq);
2509 const uint64_t InvalidOffset =
2510 Unit.getOrigUnit().getFormParams().getDwarfMaxOffset();
2511 // Check whether we have an output sequence for the StmtSeq offset.
2512 // Some sequences are discarded by the DWARFLinker if they are invalid
2513 // (empty).
2514 if (OrigRowIter == SeqOffToOrigRow.end()) {
2515 StmtSeq.set(InvalidOffset);
2516 continue;
2517 }
2518 size_t OrigRowIndex = OrigRowIter->second;
2519
2520 // 2. Find the output row for this original row.
2521 auto OutputRowIter = OrigRowToOutputRow.find(Val: OrigRowIndex);
2522 if (OutputRowIter == OrigRowToOutputRow.end()) {
2523 // Row was dropped during linking.
2524 StmtSeq.set(InvalidOffset);
2525 continue;
2526 }
2527 size_t OutputRowIdx = OutputRowIter->second;
2528
2529 // 3. Find the sequence start for this output row.
2530 // If the original row was a sequence start but got merged into
2531 // another sequence, this finds the correct sequence start.
2532 size_t SeqStartIdx = OutputRowToSeqStart[OutputRowIdx];
2533
2534 // 4. Get the offset of the sequence start in the output .debug_line
2535 // section. This offset points to the DW_LNE_set_address opcode.
2536 assert(SeqStartIdx < OutputRowOffsets.size() &&
2537 "Sequence start index out of bounds");
2538 uint64_t NewStmtSeqOffset = OutputRowOffsets[SeqStartIdx];
2539
2540 // 5. Patch the stmt_sequence attribute with the new offset.
2541 StmtSeq.set(NewStmtSeqOffset);
2542 }
2543 }
2544 }
2545
2546 } else
2547 Linker.reportWarning(Warning: "Cann't load line table.", File: ObjFile);
2548
2549 return Error::success();
2550}
2551
2552void DWARFLinker::emitAcceleratorEntriesForUnit(CompileUnit &Unit) {
2553 for (AccelTableKind AccelTableKind : Options.AccelTables) {
2554 switch (AccelTableKind) {
2555 case AccelTableKind::Apple: {
2556 // Add namespaces.
2557 for (const auto &Namespace : Unit.getNamespaces())
2558 AppleNamespaces.addName(Name: Namespace.Name, Args: Namespace.Die->getOffset() +
2559 Unit.getStartOffset());
2560 // Add names.
2561 for (const auto &Pubname : Unit.getPubnames())
2562 AppleNames.addName(Name: Pubname.Name,
2563 Args: Pubname.Die->getOffset() + Unit.getStartOffset());
2564 // Add types.
2565 for (const auto &Pubtype : Unit.getPubtypes())
2566 AppleTypes.addName(
2567 Name: Pubtype.Name, Args: Pubtype.Die->getOffset() + Unit.getStartOffset(),
2568 Args: Pubtype.Die->getTag(),
2569 Args: Pubtype.ObjcClassImplementation ? dwarf::DW_FLAG_type_implementation
2570 : 0,
2571 Args: Pubtype.QualifiedNameHash);
2572 // Add ObjC names.
2573 for (const auto &ObjC : Unit.getObjC())
2574 AppleObjc.addName(Name: ObjC.Name,
2575 Args: ObjC.Die->getOffset() + Unit.getStartOffset());
2576 } break;
2577 case AccelTableKind::Pub: {
2578 TheDwarfEmitter->emitPubNamesForUnit(Unit);
2579 TheDwarfEmitter->emitPubTypesForUnit(Unit);
2580 } break;
2581 case AccelTableKind::DebugNames: {
2582 for (const auto &Namespace : Unit.getNamespaces())
2583 DebugNames.addName(
2584 Name: Namespace.Name, Args: Namespace.Die->getOffset(),
2585 Args: DWARF5AccelTableData::getDefiningParentDieOffset(Die: *Namespace.Die),
2586 Args: Namespace.Die->getTag(), Args: Unit.getUniqueID(),
2587 Args: Unit.getTag() == dwarf::DW_TAG_type_unit);
2588 for (const auto &Pubname : Unit.getPubnames())
2589 DebugNames.addName(
2590 Name: Pubname.Name, Args: Pubname.Die->getOffset(),
2591 Args: DWARF5AccelTableData::getDefiningParentDieOffset(Die: *Pubname.Die),
2592 Args: Pubname.Die->getTag(), Args: Unit.getUniqueID(),
2593 Args: Unit.getTag() == dwarf::DW_TAG_type_unit);
2594 for (const auto &Pubtype : Unit.getPubtypes())
2595 DebugNames.addName(
2596 Name: Pubtype.Name, Args: Pubtype.Die->getOffset(),
2597 Args: DWARF5AccelTableData::getDefiningParentDieOffset(Die: *Pubtype.Die),
2598 Args: Pubtype.Die->getTag(), Args: Unit.getUniqueID(),
2599 Args: Unit.getTag() == dwarf::DW_TAG_type_unit);
2600 } break;
2601 }
2602 }
2603}
2604
2605/// Read the frame info stored in the object, and emit the
2606/// patched frame descriptions for the resulting file.
2607///
2608/// This is actually pretty easy as the data of the CIEs and FDEs can
2609/// be considered as black boxes and moved as is. The only thing to do
2610/// is to patch the addresses in the headers.
2611void DWARFLinker::patchFrameInfoForObject(LinkContext &Context) {
2612 DWARFContext &OrigDwarf = *Context.File.Dwarf;
2613 unsigned SrcAddrSize = OrigDwarf.getDWARFObj().getAddressSize();
2614
2615 StringRef FrameData = OrigDwarf.getDWARFObj().getFrameSection().Data;
2616 if (FrameData.empty())
2617 return;
2618
2619 RangesTy AllUnitsRanges;
2620 for (std::unique_ptr<CompileUnit> &Unit : Context.CompileUnits) {
2621 for (auto CurRange : Unit->getFunctionRanges())
2622 AllUnitsRanges.insert(Range: CurRange.Range, Value: CurRange.Value);
2623 }
2624
2625 DataExtractor Data(FrameData, OrigDwarf.isLittleEndian());
2626 uint64_t InputOffset = 0;
2627
2628 // Store the data of the CIEs defined in this object, keyed by their
2629 // offsets.
2630 DenseMap<uint64_t, StringRef> LocalCIES;
2631
2632 while (Data.isValidOffset(offset: InputOffset)) {
2633 uint64_t EntryOffset = InputOffset;
2634 uint32_t InitialLength = Data.getU32(offset_ptr: &InputOffset);
2635 if (InitialLength == 0xFFFFFFFF)
2636 return reportWarning(Warning: "Dwarf64 bits no supported", File: Context.File);
2637
2638 uint32_t CIEId = Data.getU32(offset_ptr: &InputOffset);
2639 if (CIEId == 0xFFFFFFFF) {
2640 // This is a CIE, store it.
2641 StringRef CIEData = FrameData.substr(Start: EntryOffset, N: InitialLength + 4);
2642 LocalCIES[EntryOffset] = CIEData;
2643 // The -4 is to account for the CIEId we just read.
2644 InputOffset += InitialLength - 4;
2645 continue;
2646 }
2647
2648 uint64_t Loc = Data.getUnsigned(offset_ptr: &InputOffset, byte_size: SrcAddrSize);
2649
2650 // Some compilers seem to emit frame info that doesn't start at
2651 // the function entry point, thus we can't just lookup the address
2652 // in the debug map. Use the AddressInfo's range map to see if the FDE
2653 // describes something that we can relocate.
2654 std::optional<AddressRangeValuePair> Range =
2655 AllUnitsRanges.getRangeThatContains(Addr: Loc);
2656 if (!Range) {
2657 // The +4 is to account for the size of the InitialLength field itself.
2658 InputOffset = EntryOffset + InitialLength + 4;
2659 continue;
2660 }
2661
2662 // This is an FDE, and we have a mapping.
2663 // Have we already emitted a corresponding CIE?
2664 StringRef CIEData = LocalCIES[CIEId];
2665 if (CIEData.empty())
2666 return reportWarning(Warning: "Inconsistent debug_frame content. Dropping.",
2667 File: Context.File);
2668
2669 // Look if we already emitted a CIE that corresponds to the
2670 // referenced one (the CIE data is the key of that lookup).
2671 auto IteratorInserted = EmittedCIEs.insert(
2672 KV: std::make_pair(x&: CIEData, y: TheDwarfEmitter->getFrameSectionSize()));
2673 // If there is no CIE yet for this ID, emit it.
2674 if (IteratorInserted.second) {
2675 LastCIEOffset = TheDwarfEmitter->getFrameSectionSize();
2676 IteratorInserted.first->getValue() = LastCIEOffset;
2677 TheDwarfEmitter->emitCIE(CIEBytes: CIEData);
2678 }
2679
2680 // Emit the FDE with updated address and CIE pointer.
2681 // (4 + AddrSize) is the size of the CIEId + initial_location
2682 // fields that will get reconstructed by emitFDE().
2683 unsigned FDERemainingBytes = InitialLength - (4 + SrcAddrSize);
2684 TheDwarfEmitter->emitFDE(CIEOffset: IteratorInserted.first->getValue(), AddreSize: SrcAddrSize,
2685 Address: Loc + Range->Value,
2686 Bytes: FrameData.substr(Start: InputOffset, N: FDERemainingBytes));
2687 InputOffset += FDERemainingBytes;
2688 }
2689}
2690
2691uint32_t DWARFLinker::DIECloner::hashFullyQualifiedName(DWARFDie DIE,
2692 CompileUnit &U,
2693 const DWARFFile &File,
2694 int ChildRecurseDepth) {
2695 const char *Name = nullptr;
2696 DWARFUnit *OrigUnit = &U.getOrigUnit();
2697 CompileUnit *CU = &U;
2698 std::optional<DWARFFormValue> Ref;
2699
2700 while (true) {
2701 if (const char *CurrentName = DIE.getName(Kind: DINameKind::ShortName))
2702 Name = CurrentName;
2703
2704 if (!(Ref = DIE.find(Attr: dwarf::DW_AT_specification)) &&
2705 !(Ref = DIE.find(Attr: dwarf::DW_AT_abstract_origin)))
2706 break;
2707
2708 if (!Ref->isFormClass(FC: DWARFFormValue::FC_Reference))
2709 break;
2710
2711 CompileUnit *RefCU;
2712 if (auto RefDIE =
2713 Linker.resolveDIEReference(File, Units: CompileUnits, RefValue: *Ref, DIE, RefCU)) {
2714 CU = RefCU;
2715 OrigUnit = &RefCU->getOrigUnit();
2716 DIE = RefDIE;
2717 }
2718 }
2719
2720 unsigned Idx = OrigUnit->getDIEIndex(D: DIE);
2721 if (!Name && DIE.getTag() == dwarf::DW_TAG_namespace)
2722 Name = "(anonymous namespace)";
2723
2724 if (CU->getInfo(Idx).ParentIdx == 0 ||
2725 // FIXME: dsymutil-classic compatibility. Ignore modules.
2726 CU->getOrigUnit().getDIEAtIndex(Index: CU->getInfo(Idx).ParentIdx).getTag() ==
2727 dwarf::DW_TAG_module)
2728 return djbHash(Buffer: Name ? Name : "", H: djbHash(Buffer: ChildRecurseDepth ? "" : "::"));
2729
2730 DWARFDie Die = OrigUnit->getDIEAtIndex(Index: CU->getInfo(Idx).ParentIdx);
2731 return djbHash(
2732 Buffer: (Name ? Name : ""),
2733 H: djbHash(Buffer: (Name ? "::" : ""),
2734 H: hashFullyQualifiedName(DIE: Die, U&: *CU, File, ChildRecurseDepth: ++ChildRecurseDepth)));
2735}
2736
2737static uint64_t getDwoId(const DWARFDie &CUDie) {
2738 auto DwoId = dwarf::toUnsigned(
2739 V: CUDie.find(Attrs: {dwarf::DW_AT_dwo_id, dwarf::DW_AT_GNU_dwo_id}));
2740 if (DwoId)
2741 return *DwoId;
2742 return 0;
2743}
2744
2745static std::string
2746remapPath(StringRef Path,
2747 const DWARFLinkerBase::ObjectPrefixMapTy &ObjectPrefixMap) {
2748 if (ObjectPrefixMap.empty())
2749 return Path.str();
2750
2751 SmallString<256> p = Path;
2752 for (const auto &Entry : ObjectPrefixMap)
2753 if (llvm::sys::path::replace_path_prefix(Path&: p, OldPrefix: Entry.first, NewPrefix: Entry.second))
2754 break;
2755 return p.str().str();
2756}
2757
2758static std::string
2759getPCMFile(const DWARFDie &CUDie,
2760 const DWARFLinkerBase::ObjectPrefixMapTy *ObjectPrefixMap) {
2761 std::string PCMFile = dwarf::toString(
2762 V: CUDie.find(Attrs: {dwarf::DW_AT_dwo_name, dwarf::DW_AT_GNU_dwo_name}), Default: "");
2763
2764 if (PCMFile.empty())
2765 return PCMFile;
2766
2767 if (ObjectPrefixMap)
2768 PCMFile = remapPath(Path: PCMFile, ObjectPrefixMap: *ObjectPrefixMap);
2769
2770 return PCMFile;
2771}
2772
2773std::pair<bool, bool> DWARFLinker::isClangModuleRef(const DWARFDie &CUDie,
2774 std::string &PCMFile,
2775 LinkContext &Context,
2776 unsigned Indent,
2777 bool Quiet) {
2778 if (PCMFile.empty())
2779 return std::make_pair(x: false, y: false);
2780
2781 // Clang module DWARF skeleton CUs abuse this for the path to the module.
2782 uint64_t DwoId = getDwoId(CUDie);
2783
2784 std::string Name = dwarf::toString(V: CUDie.find(Attr: dwarf::DW_AT_name), Default: "");
2785 if (Name.empty()) {
2786 if (!Quiet)
2787 reportWarning(Warning: "Anonymous module skeleton CU for " + PCMFile,
2788 File: Context.File);
2789 return std::make_pair(x: true, y: true);
2790 }
2791
2792 if (!Quiet && Options.Verbose) {
2793 outs().indent(NumSpaces: Indent);
2794 outs() << "Found clang module reference " << PCMFile;
2795 }
2796
2797 auto Cached = ClangModules.find(Key: PCMFile);
2798 if (Cached != ClangModules.end()) {
2799 // FIXME: Until PR27449 (https://llvm.org/bugs/show_bug.cgi?id=27449) is
2800 // fixed in clang, only warn about DWO_id mismatches in verbose mode.
2801 // ASTFileSignatures will change randomly when a module is rebuilt.
2802 if (!Quiet && Options.Verbose && (Cached->second != DwoId))
2803 reportWarning(Warning: Twine("hash mismatch: this object file was built against a "
2804 "different version of the module ") +
2805 PCMFile,
2806 File: Context.File);
2807 if (!Quiet && Options.Verbose)
2808 outs() << " [cached].\n";
2809 return std::make_pair(x: true, y: true);
2810 }
2811
2812 return std::make_pair(x: true, y: false);
2813}
2814
2815bool DWARFLinker::registerModuleReference(const DWARFDie &CUDie,
2816 LinkContext &Context,
2817 ObjFileLoaderTy Loader,
2818 CompileUnitHandlerTy OnCUDieLoaded,
2819 unsigned Indent) {
2820 std::string PCMFile = getPCMFile(CUDie, ObjectPrefixMap: Options.ObjectPrefixMap);
2821 std::pair<bool, bool> IsClangModuleRef =
2822 isClangModuleRef(CUDie, PCMFile, Context, Indent, Quiet: false);
2823
2824 if (!IsClangModuleRef.first)
2825 return false;
2826
2827 if (IsClangModuleRef.second)
2828 return true;
2829
2830 if (Options.Verbose)
2831 outs() << " ...\n";
2832
2833 // Cyclic dependencies are disallowed by Clang, but we still
2834 // shouldn't run into an infinite loop, so mark it as processed now.
2835 ClangModules.insert(KV: {PCMFile, getDwoId(CUDie)});
2836
2837 if (Error E = loadClangModule(Loader, CUDie, PCMFile, Context, OnCUDieLoaded,
2838 Indent: Indent + 2)) {
2839 consumeError(Err: std::move(E));
2840 return false;
2841 }
2842 return true;
2843}
2844
2845Error DWARFLinker::loadClangModule(
2846 ObjFileLoaderTy Loader, const DWARFDie &CUDie, const std::string &PCMFile,
2847 LinkContext &Context, CompileUnitHandlerTy OnCUDieLoaded, unsigned Indent) {
2848
2849 uint64_t DwoId = getDwoId(CUDie);
2850 std::string ModuleName = dwarf::toString(V: CUDie.find(Attr: dwarf::DW_AT_name), Default: "");
2851
2852 /// Using a SmallString<0> because loadClangModule() is recursive.
2853 SmallString<0> Path(Options.PrependPath);
2854 if (sys::path::is_relative(path: PCMFile))
2855 resolveRelativeObjectPath(Buf&: Path, CU: CUDie);
2856 sys::path::append(path&: Path, a: PCMFile);
2857 // Don't use the cached binary holder because we have no thread-safety
2858 // guarantee and the lifetime is limited.
2859
2860 if (Loader == nullptr) {
2861 reportError(Warning: "Could not load clang module: loader is not specified.\n",
2862 File: Context.File);
2863 return Error::success();
2864 }
2865
2866 auto ErrOrObj = Loader(Context.File.FileName, Path);
2867 if (!ErrOrObj)
2868 return Error::success();
2869
2870 std::unique_ptr<CompileUnit> Unit;
2871 for (const auto &CU : ErrOrObj->Dwarf->compile_units()) {
2872 OnCUDieLoaded(*CU);
2873 // Recursively get all modules imported by this one.
2874 auto ChildCUDie = CU->getUnitDIE();
2875 if (!ChildCUDie)
2876 continue;
2877 if (!registerModuleReference(CUDie: ChildCUDie, Context, Loader, OnCUDieLoaded,
2878 Indent)) {
2879 if (Unit) {
2880 std::string Err =
2881 (PCMFile +
2882 ": Clang modules are expected to have exactly 1 compile unit.\n");
2883 reportError(Warning: Err, File: Context.File);
2884 return make_error<StringError>(Args&: Err, Args: inconvertibleErrorCode());
2885 }
2886 // FIXME: Until PR27449 (https://llvm.org/bugs/show_bug.cgi?id=27449) is
2887 // fixed in clang, only warn about DWO_id mismatches in verbose mode.
2888 // ASTFileSignatures will change randomly when a module is rebuilt.
2889 uint64_t PCMDwoId = getDwoId(CUDie: ChildCUDie);
2890 if (PCMDwoId != DwoId) {
2891 if (Options.Verbose)
2892 reportWarning(
2893 Warning: Twine("hash mismatch: this object file was built against a "
2894 "different version of the module ") +
2895 PCMFile,
2896 File: Context.File);
2897 // Update the cache entry with the DwoId of the module loaded from disk.
2898 ClangModules[PCMFile] = PCMDwoId;
2899 }
2900
2901 // Add this module.
2902 Unit = std::make_unique<CompileUnit>(args&: *CU, args: UniqueUnitID++, args: !Options.NoODR,
2903 args&: ModuleName);
2904 }
2905 }
2906
2907 if (Unit)
2908 Context.ModuleUnits.emplace_back(args: RefModuleUnit{*ErrOrObj, std::move(Unit)});
2909
2910 return Error::success();
2911}
2912
2913Expected<uint64_t> DWARFLinker::DIECloner::cloneAllCompileUnits(
2914 DWARFContext &DwarfContext, const DWARFFile &File, bool IsLittleEndian) {
2915 uint64_t OutputDebugInfoSize =
2916 (Emitter == nullptr) ? 0 : Emitter->getDebugInfoSectionSize();
2917 const uint64_t StartOutputDebugInfoSize = OutputDebugInfoSize;
2918
2919 for (auto &CurrentUnit : CompileUnits) {
2920 const uint16_t DwarfVersion = CurrentUnit->getOrigUnit().getVersion();
2921 const uint32_t UnitHeaderSize = DwarfVersion >= 5 ? 12 : 11;
2922 auto InputDIE = CurrentUnit->getOrigUnit().getUnitDIE();
2923 CurrentUnit->setStartOffset(OutputDebugInfoSize);
2924 if (!InputDIE) {
2925 OutputDebugInfoSize = CurrentUnit->computeNextUnitOffset(DwarfVersion);
2926 continue;
2927 }
2928 if (CurrentUnit->getInfo(Idx: 0).Keep) {
2929 // Clone the InputDIE into your Unit DIE in our compile unit since it
2930 // already has a DIE inside of it.
2931 CurrentUnit->createOutputDIE();
2932 rememberUnitForMacroOffset(Unit&: *CurrentUnit);
2933 cloneDIE(InputDIE, File, Unit&: *CurrentUnit, PCOffset: 0 /* PC offset */, OutOffset: UnitHeaderSize,
2934 Flags: 0, IsLittleEndian, Die: CurrentUnit->getOutputUnitDIE());
2935 }
2936
2937 OutputDebugInfoSize = CurrentUnit->computeNextUnitOffset(DwarfVersion);
2938
2939 if (Emitter != nullptr) {
2940
2941 if (Error E = generateLineTableForUnit(Unit&: *CurrentUnit))
2942 return E;
2943
2944 Linker.emitAcceleratorEntriesForUnit(Unit&: *CurrentUnit);
2945
2946 if (LLVM_UNLIKELY(Linker.Options.Update))
2947 continue;
2948
2949 if (Error E = Linker.generateUnitRanges(Unit&: *CurrentUnit, File, AddrPool))
2950 return E;
2951
2952 auto ProcessExpr = [&](SmallVectorImpl<uint8_t> &SrcBytes,
2953 SmallVectorImpl<uint8_t> &OutBytes,
2954 int64_t RelocAdjustment) {
2955 DWARFUnit &OrigUnit = CurrentUnit->getOrigUnit();
2956 DataExtractor Data(SrcBytes, IsLittleEndian);
2957 cloneExpression(Data,
2958 Expression: DWARFExpression(Data, OrigUnit.getAddressByteSize(),
2959 OrigUnit.getFormParams().Format),
2960 File, Unit&: *CurrentUnit, OutputBuffer&: OutBytes, AddrRelocAdjustment: RelocAdjustment,
2961 IsLittleEndian);
2962 };
2963 if (Error E = generateUnitLocations(Unit&: *CurrentUnit, File, ExprHandler: ProcessExpr))
2964 return E;
2965 if (Error E = emitDebugAddrSection(Unit&: *CurrentUnit, DwarfVersion))
2966 return E;
2967 }
2968 AddrPool.clear();
2969 }
2970
2971 if (Emitter != nullptr) {
2972 assert(Emitter);
2973 // Emit macro tables.
2974 Emitter->emitMacroTables(Context: File.Dwarf.get(), UnitMacroMap, StringPool&: DebugStrPool);
2975
2976 // Emit all the compile unit's debug information.
2977 for (auto &CurrentUnit : CompileUnits) {
2978 CurrentUnit->fixupForwardReferences();
2979
2980 if (!CurrentUnit->getOutputUnitDIE())
2981 continue;
2982
2983 unsigned DwarfVersion = CurrentUnit->getOrigUnit().getVersion();
2984
2985 assert(Emitter->getDebugInfoSectionSize() ==
2986 CurrentUnit->getStartOffset());
2987 Emitter->emitCompileUnitHeader(Unit&: *CurrentUnit, DwarfVersion);
2988 Emitter->emitDIE(Die&: *CurrentUnit->getOutputUnitDIE());
2989 assert(Emitter->getDebugInfoSectionSize() ==
2990 CurrentUnit->computeNextUnitOffset(DwarfVersion));
2991 }
2992 }
2993
2994 return OutputDebugInfoSize - StartOutputDebugInfoSize;
2995}
2996
2997void DWARFLinker::copyInvariantDebugSection(DWARFContext &Dwarf) {
2998 TheDwarfEmitter->emitSectionContents(SecData: Dwarf.getDWARFObj().getLocSection().Data,
2999 SecKind: DebugSectionKind::DebugLoc);
3000 TheDwarfEmitter->emitSectionContents(
3001 SecData: Dwarf.getDWARFObj().getRangesSection().Data,
3002 SecKind: DebugSectionKind::DebugRange);
3003 TheDwarfEmitter->emitSectionContents(
3004 SecData: Dwarf.getDWARFObj().getFrameSection().Data, SecKind: DebugSectionKind::DebugFrame);
3005 TheDwarfEmitter->emitSectionContents(SecData: Dwarf.getDWARFObj().getArangesSection(),
3006 SecKind: DebugSectionKind::DebugARanges);
3007 TheDwarfEmitter->emitSectionContents(
3008 SecData: Dwarf.getDWARFObj().getAddrSection().Data, SecKind: DebugSectionKind::DebugAddr);
3009 TheDwarfEmitter->emitSectionContents(
3010 SecData: Dwarf.getDWARFObj().getRnglistsSection().Data,
3011 SecKind: DebugSectionKind::DebugRngLists);
3012 TheDwarfEmitter->emitSectionContents(
3013 SecData: Dwarf.getDWARFObj().getLoclistsSection().Data,
3014 SecKind: DebugSectionKind::DebugLocLists);
3015}
3016
3017void DWARFLinker::addObjectFile(DWARFFile &File, ObjFileLoaderTy Loader,
3018 CompileUnitHandlerTy OnCUDieLoaded) {
3019 ObjectContexts.emplace_back(args: LinkContext(File));
3020
3021 if (ObjectContexts.back().File.Dwarf) {
3022 for (const std::unique_ptr<DWARFUnit> &CU :
3023 ObjectContexts.back().File.Dwarf->compile_units()) {
3024 DWARFDie CUDie = CU->getUnitDIE();
3025
3026 if (!CUDie)
3027 continue;
3028
3029 OnCUDieLoaded(*CU);
3030
3031 if (!LLVM_UNLIKELY(Options.Update))
3032 registerModuleReference(CUDie, Context&: ObjectContexts.back(), Loader,
3033 OnCUDieLoaded);
3034 }
3035 }
3036}
3037
3038Error DWARFLinker::link() {
3039 assert((Options.TargetDWARFVersion != 0) &&
3040 "TargetDWARFVersion should be set");
3041
3042 // First populate the data structure we need for each iteration of the
3043 // parallel loop.
3044 unsigned NumObjects = ObjectContexts.size();
3045
3046 // This Dwarf string pool which is used for emission. It must be used
3047 // serially as the order of calling getStringOffset matters for
3048 // reproducibility.
3049 OffsetsStringPool DebugStrPool(true);
3050 OffsetsStringPool DebugLineStrPool(false);
3051 DebugDieValuePool StringOffsetPool;
3052
3053 // ODR Contexts for the optimize.
3054 DeclContextTree ODRContexts;
3055
3056 for (LinkContext &OptContext : ObjectContexts) {
3057 if (Options.Verbose)
3058 outs() << "DEBUG MAP OBJECT: " << OptContext.File.FileName << "\n";
3059
3060 if (!OptContext.File.Dwarf)
3061 continue;
3062
3063 if (Options.VerifyInputDWARF)
3064 verifyInput(File: OptContext.File);
3065
3066 // Look for relocations that correspond to address map entries.
3067
3068 // there was findvalidrelocations previously ... probably we need to gather
3069 // info here
3070 if (LLVM_LIKELY(!Options.Update) &&
3071 !OptContext.File.Addresses->hasValidRelocs()) {
3072 if (Options.Verbose)
3073 outs() << "No valid relocations found. Skipping.\n";
3074
3075 // Set "Skip" flag as a signal to other loops that we should not
3076 // process this iteration.
3077 OptContext.Skip = true;
3078 continue;
3079 }
3080
3081 // Setup access to the debug info.
3082 if (!OptContext.File.Dwarf)
3083 continue;
3084
3085 // Check whether type units are presented.
3086 if (!OptContext.File.Dwarf->types_section_units().empty()) {
3087 reportWarning(Warning: "type units are not currently supported: file will "
3088 "be skipped",
3089 File: OptContext.File);
3090 OptContext.Skip = true;
3091 continue;
3092 }
3093
3094 // Clone all the clang modules with requires extracting the DIE units. We
3095 // don't need the full debug info until the Analyze phase.
3096 OptContext.CompileUnits.reserve(
3097 n: OptContext.File.Dwarf->getNumCompileUnits());
3098 for (const auto &CU : OptContext.File.Dwarf->compile_units()) {
3099 auto CUDie = CU->getUnitDIE(/*ExtractUnitDIEOnly=*/true);
3100 if (Options.Verbose) {
3101 outs() << "Input compilation unit:";
3102 DIDumpOptions DumpOpts;
3103 DumpOpts.ChildRecurseDepth = 0;
3104 DumpOpts.Verbose = Options.Verbose;
3105 CUDie.dump(OS&: outs(), indent: 0, DumpOpts);
3106 }
3107 }
3108
3109 for (auto &CU : OptContext.ModuleUnits) {
3110 if (Error Err = cloneModuleUnit(Context&: OptContext, Unit&: CU, ODRContexts, DebugStrPool,
3111 DebugLineStrPool, StringOffsetPool))
3112 reportWarning(Warning: toString(E: std::move(Err)), File: CU.File);
3113 }
3114 }
3115
3116 // At this point we know how much data we have emitted. We use this value to
3117 // compare canonical DIE offsets in analyzeContextInfo to see if a definition
3118 // is already emitted, without being affected by canonical die offsets set
3119 // later. This prevents undeterminism when analyze and clone execute
3120 // concurrently, as clone set the canonical DIE offset and analyze reads it.
3121 const uint64_t ModulesEndOffset =
3122 (TheDwarfEmitter == nullptr) ? 0
3123 : TheDwarfEmitter->getDebugInfoSectionSize();
3124
3125 // These variables manage the list of processed object files.
3126 // The mutex and condition variable are to ensure that this is thread safe.
3127 std::mutex ProcessedFilesMutex;
3128 std::condition_variable ProcessedFilesConditionVariable;
3129 BitVector ProcessedFiles(NumObjects, false);
3130
3131 // Analyzing the context info is particularly expensive so it is executed in
3132 // parallel with emitting the previous compile unit.
3133 auto AnalyzeLambda = [&](size_t I) {
3134 auto &Context = ObjectContexts[I];
3135
3136 if (Context.Skip || !Context.File.Dwarf)
3137 return;
3138
3139 for (const auto &CU : Context.File.Dwarf->compile_units()) {
3140 // Previously we only extracted the unit DIEs. We need the full debug info
3141 // now.
3142 auto CUDie = CU->getUnitDIE(/*ExtractUnitDIEOnly=*/false);
3143 std::string PCMFile = getPCMFile(CUDie, ObjectPrefixMap: Options.ObjectPrefixMap);
3144
3145 if (!CUDie || LLVM_UNLIKELY(Options.Update) ||
3146 !isClangModuleRef(CUDie, PCMFile, Context, Indent: 0, Quiet: true).first) {
3147 Context.CompileUnits.push_back(x: std::make_unique<CompileUnit>(
3148 args&: *CU, args: UniqueUnitID++, args: !Options.NoODR && !Options.Update, args: ""));
3149 }
3150 }
3151
3152 // Now build the DIE parent links that we will use during the next phase.
3153 for (auto &CurrentUnit : Context.CompileUnits) {
3154 auto CUDie = CurrentUnit->getOrigUnit().getUnitDIE();
3155 if (!CUDie)
3156 continue;
3157 analyzeContextInfo(DIE: CurrentUnit->getOrigUnit().getUnitDIE(), ParentIdx: 0,
3158 CU&: *CurrentUnit, CurrentDeclContext: &ODRContexts.getRoot(), Contexts&: ODRContexts,
3159 ModulesEndOffset, ParseableSwiftInterfaces: Options.ParseableSwiftInterfaces,
3160 ReportWarning: [&](const Twine &Warning, const DWARFDie &DIE) {
3161 reportWarning(Warning, File: Context.File, DIE: &DIE);
3162 });
3163 }
3164 };
3165
3166 // For each object file map how many bytes were emitted.
3167 StringMap<DebugInfoSize> SizeByObject;
3168
3169 // And then the remaining work in serial again.
3170 // Note, although this loop runs in serial, it can run in parallel with
3171 // the analyzeContextInfo loop so long as we process files with indices >=
3172 // than those processed by analyzeContextInfo.
3173 auto CloneLambda = [&](size_t I, llvm::Error &CE) {
3174 auto &OptContext = ObjectContexts[I];
3175 if (OptContext.Skip || !OptContext.File.Dwarf)
3176 return;
3177
3178 // Then mark all the DIEs that need to be present in the generated output
3179 // and collect some information about them.
3180 // Note that this loop can not be merged with the previous one because
3181 // cross-cu references require the ParentIdx to be setup for every CU in
3182 // the object file before calling this.
3183 if (LLVM_UNLIKELY(Options.Update)) {
3184 for (auto &CurrentUnit : OptContext.CompileUnits)
3185 CurrentUnit->markEverythingAsKept();
3186 copyInvariantDebugSection(Dwarf&: *OptContext.File.Dwarf);
3187 } else {
3188 for (auto &CurrentUnit : OptContext.CompileUnits) {
3189 lookForDIEsToKeep(AddressesMap&: *OptContext.File.Addresses, Units: OptContext.CompileUnits,
3190 Die: CurrentUnit->getOrigUnit().getUnitDIE(),
3191 File: OptContext.File, Cu&: *CurrentUnit, Flags: 0);
3192#ifndef NDEBUG
3193 verifyKeepChain(*CurrentUnit);
3194#endif
3195 }
3196 }
3197
3198 // The calls to applyValidRelocs inside cloneDIE will walk the reloc
3199 // array again (in the same way findValidRelocsInDebugInfo() did). We
3200 // need to reset the NextValidReloc index to the beginning.
3201 if (OptContext.File.Addresses->hasValidRelocs() ||
3202 LLVM_UNLIKELY(Options.Update)) {
3203 SizeByObject[OptContext.File.FileName].Input =
3204 getDebugInfoSize(Dwarf&: *OptContext.File.Dwarf);
3205 Expected<uint64_t> SizeOrErr =
3206 DIECloner(*this, TheDwarfEmitter, OptContext.File, DIEAlloc,
3207 OptContext.CompileUnits, Options.Update, DebugStrPool,
3208 DebugLineStrPool, StringOffsetPool)
3209 .cloneAllCompileUnits(DwarfContext&: *OptContext.File.Dwarf, File: OptContext.File,
3210 IsLittleEndian: OptContext.File.Dwarf->isLittleEndian());
3211 if (!SizeOrErr) {
3212 CE = SizeOrErr.takeError();
3213 return;
3214 }
3215 SizeByObject[OptContext.File.FileName].Output = *SizeOrErr;
3216 }
3217 if ((TheDwarfEmitter != nullptr) && !OptContext.CompileUnits.empty() &&
3218 LLVM_LIKELY(!Options.Update))
3219 patchFrameInfoForObject(Context&: OptContext);
3220
3221 // Clean-up before starting working on the next object.
3222 cleanupAuxiliarryData(Context&: OptContext);
3223 };
3224
3225 auto EmitLambda = [&]() {
3226 // Emit everything that's global.
3227 if (TheDwarfEmitter != nullptr) {
3228 TheDwarfEmitter->emitAbbrevs(Abbrevs: Abbreviations, DwarfVersion: Options.TargetDWARFVersion);
3229 TheDwarfEmitter->emitStrings(Pool: DebugStrPool);
3230 TheDwarfEmitter->emitStringOffsets(StringOffsets: StringOffsetPool.getValues(),
3231 TargetDWARFVersion: Options.TargetDWARFVersion);
3232 TheDwarfEmitter->emitLineStrings(Pool: DebugLineStrPool);
3233 for (AccelTableKind TableKind : Options.AccelTables) {
3234 switch (TableKind) {
3235 case AccelTableKind::Apple:
3236 TheDwarfEmitter->emitAppleNamespaces(Table&: AppleNamespaces);
3237 TheDwarfEmitter->emitAppleNames(Table&: AppleNames);
3238 TheDwarfEmitter->emitAppleTypes(Table&: AppleTypes);
3239 TheDwarfEmitter->emitAppleObjc(Table&: AppleObjc);
3240 break;
3241 case AccelTableKind::Pub:
3242 // Already emitted by emitAcceleratorEntriesForUnit.
3243 // Already emitted by emitAcceleratorEntriesForUnit.
3244 break;
3245 case AccelTableKind::DebugNames:
3246 TheDwarfEmitter->emitDebugNames(Table&: DebugNames);
3247 break;
3248 }
3249 }
3250 }
3251 };
3252
3253 auto AnalyzeAll = [&]() {
3254 for (unsigned I = 0, E = NumObjects; I != E; ++I) {
3255 AnalyzeLambda(I);
3256
3257 std::unique_lock<std::mutex> LockGuard(ProcessedFilesMutex);
3258 ProcessedFiles.set(I);
3259 ProcessedFilesConditionVariable.notify_one();
3260 }
3261 };
3262
3263 auto CloneAll = [&](llvm::Error &CE) {
3264 for (unsigned I = 0, E = NumObjects; I != E; ++I) {
3265 {
3266 std::unique_lock<std::mutex> LockGuard(ProcessedFilesMutex);
3267 if (!ProcessedFiles[I]) {
3268 ProcessedFilesConditionVariable.wait(
3269 lock&: LockGuard, p: [&]() { return ProcessedFiles[I]; });
3270 }
3271 }
3272
3273 CloneLambda(I, CE);
3274 if (CE)
3275 return;
3276 }
3277 EmitLambda();
3278 };
3279
3280 Error CE = Error::success();
3281
3282 // To limit memory usage in the single threaded case, analyze and clone are
3283 // run sequentially so the OptContext is freed after processing each object
3284 // in endDebugObject.
3285 if (Options.Threads == 1) {
3286 for (unsigned I = 0, E = NumObjects; I != E; ++I) {
3287 AnalyzeLambda(I);
3288 CloneLambda(I, CE);
3289 if (CE)
3290 break;
3291 }
3292 if (!CE)
3293 EmitLambda();
3294 } else {
3295 DefaultThreadPool Pool(hardware_concurrency(ThreadCount: 2));
3296 Pool.async(F&: AnalyzeAll);
3297 Pool.async(F&: CloneAll, ArgList: std::reference_wrapper<Error>(CE));
3298 Pool.wait();
3299 }
3300
3301 if (CE)
3302 return CE;
3303
3304 if (Options.Statistics) {
3305 // Create a vector sorted in descending order by output size.
3306 std::vector<std::pair<StringRef, DebugInfoSize>> Sorted;
3307 for (auto &E : SizeByObject)
3308 Sorted.emplace_back(args: E.first(), args&: E.second);
3309 llvm::sort(C&: Sorted, Comp: [](auto &LHS, auto &RHS) {
3310 return LHS.second.Output > RHS.second.Output;
3311 });
3312
3313 auto ComputePercentange = [](int64_t Input, int64_t Output) -> float {
3314 const float Difference = Output - Input;
3315 const float Sum = Input + Output;
3316 if (Sum == 0)
3317 return 0;
3318 return (Difference / (Sum / 2));
3319 };
3320
3321 int64_t InputTotal = 0;
3322 int64_t OutputTotal = 0;
3323 const char *FormatStr = "{0,-45} {1,10}b {2,10}b {3,8:P}\n";
3324
3325 // Print header.
3326 outs() << ".debug_info section size (in bytes)\n";
3327 outs() << "----------------------------------------------------------------"
3328 "---------------\n";
3329 outs() << "Filename Object "
3330 " dSYM Change\n";
3331 outs() << "----------------------------------------------------------------"
3332 "---------------\n";
3333
3334 // Print body.
3335 for (auto &E : Sorted) {
3336 InputTotal += E.second.Input;
3337 OutputTotal += E.second.Output;
3338 llvm::outs() << formatv(
3339 Fmt: FormatStr, Vals: sys::path::filename(path: E.first).take_back(N: 45), Vals&: E.second.Input,
3340 Vals&: E.second.Output, Vals: ComputePercentange(E.second.Input, E.second.Output));
3341 }
3342 // Print total and footer.
3343 outs() << "----------------------------------------------------------------"
3344 "---------------\n";
3345 llvm::outs() << formatv(Fmt: FormatStr, Vals: "Total", Vals&: InputTotal, Vals&: OutputTotal,
3346 Vals: ComputePercentange(InputTotal, OutputTotal));
3347 outs() << "----------------------------------------------------------------"
3348 "---------------\n\n";
3349 }
3350
3351 return Error::success();
3352}
3353
3354Error DWARFLinker::cloneModuleUnit(LinkContext &Context, RefModuleUnit &Unit,
3355 DeclContextTree &ODRContexts,
3356 OffsetsStringPool &DebugStrPool,
3357 OffsetsStringPool &DebugLineStrPool,
3358 DebugDieValuePool &StringOffsetPool,
3359 unsigned Indent) {
3360 assert(Unit.Unit.get() != nullptr);
3361
3362 if (!Unit.Unit->getOrigUnit().getUnitDIE().hasChildren())
3363 return Error::success();
3364
3365 if (Options.Verbose) {
3366 outs().indent(NumSpaces: Indent);
3367 outs() << "cloning .debug_info from " << Unit.File.FileName << "\n";
3368 }
3369
3370 // Analyze context for the module.
3371 analyzeContextInfo(DIE: Unit.Unit->getOrigUnit().getUnitDIE(), ParentIdx: 0, CU&: *(Unit.Unit),
3372 CurrentDeclContext: &ODRContexts.getRoot(), Contexts&: ODRContexts, ModulesEndOffset: 0,
3373 ParseableSwiftInterfaces: Options.ParseableSwiftInterfaces,
3374 ReportWarning: [&](const Twine &Warning, const DWARFDie &DIE) {
3375 reportWarning(Warning, File: Context.File, DIE: &DIE);
3376 });
3377 // Keep everything.
3378 Unit.Unit->markEverythingAsKept();
3379
3380 // Clone unit.
3381 UnitListTy CompileUnits;
3382 CompileUnits.emplace_back(args: std::move(Unit.Unit));
3383 assert(TheDwarfEmitter);
3384 Expected<uint64_t> SizeOrErr =
3385 DIECloner(*this, TheDwarfEmitter, Unit.File, DIEAlloc, CompileUnits,
3386 Options.Update, DebugStrPool, DebugLineStrPool,
3387 StringOffsetPool)
3388 .cloneAllCompileUnits(DwarfContext&: *Unit.File.Dwarf, File: Unit.File,
3389 IsLittleEndian: Unit.File.Dwarf->isLittleEndian());
3390 if (!SizeOrErr)
3391 return SizeOrErr.takeError();
3392 return Error::success();
3393}
3394
3395void DWARFLinker::verifyInput(const DWARFFile &File) {
3396 assert(File.Dwarf);
3397
3398 std::string Buffer;
3399 raw_string_ostream OS(Buffer);
3400 DIDumpOptions DumpOpts;
3401 if (!File.Dwarf->verify(OS, DumpOpts: DumpOpts.noImplicitRecursion())) {
3402 if (Options.InputVerificationHandler)
3403 Options.InputVerificationHandler(File, OS.str());
3404 }
3405}
3406
3407} // namespace llvm
3408