[llvm] [BOLT][DWARF] Add support for .debug_names (PR #81062)

[Davide Italiano via llvm-commits]

================
@@ -0,0 +1,644 @@
+//===- bolt/Rewrite/DebugNames.cpp -------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "bolt/Core/DebugNames.h"
+#include "bolt/Core/BinaryContext.h"
+#include "llvm/DebugInfo/DWARF/DWARFExpression.h"
+#include "llvm/DebugInfo/DWARF/DWARFTypeUnit.h"
+#include "llvm/Support/EndianStream.h"
+#include "llvm/Support/LEB128.h"
+
+namespace llvm {
+namespace bolt {
+DWARF5AcceleratorTable::DWARF5AcceleratorTable(
+    const bool CreateDebugNames, BinaryContext &BC,
+    DebugStrWriter &MainBinaryStrWriter)
+    : MainBinaryStrWriter(MainBinaryStrWriter) {
+  NeedToCreate = CreateDebugNames || BC.getDebugNamesSection();
+  if (!NeedToCreate)
+    return;
+  FullTableBuffer = std::make_unique<DebugStrBufferVector>();
+  FullTableStream = std::make_unique<raw_svector_ostream>(*FullTableBuffer);
+  StrBuffer = std::make_unique<DebugStrBufferVector>();
+  StrStream = std::make_unique<raw_svector_ostream>(*StrBuffer);
+  EntriesBuffer = std::make_unique<DebugStrBufferVector>();
+  Entriestream = std::make_unique<raw_svector_ostream>(*EntriesBuffer);
+  AugStringBuffer = std::make_unique<DebugStrBufferVector>();
+  AugStringtream = std::make_unique<raw_svector_ostream>(*AugStringBuffer);
+
+  // Binary has split-dwarf CUs.
+  // Even thought for non-skeleton-cu all names are in .debug_str.dwo section,
+  // for the .debug_names contributions they are in .debug_str section.
+  if (BC.getNumDWOCUs()) {
+    DataExtractor StrData(BC.DwCtx->getDWARFObj().getStrSection(),
+                          BC.DwCtx->isLittleEndian(), 0);
+    uint64_t Offset = 0;
+    uint64_t StrOffset = 0;
+    while (StrData.isValidOffset(Offset)) {
+      Error Err = Error::success();
+      const char *CStr = StrData.getCStr(&Offset, &Err);
+      if (Err) {
+        NeedToCreate = false;
+        errs() << "BOLT-WARNING: [internal-dwarf-error]: Could not extract "
+                  "string from .debug_str section at offset: "
+               << Twine::utohexstr(StrOffset) << ".\n";
+        return;
+      }
+      auto R = StrCacheToOffsetMap.try_emplace(
+          llvm::hash_value(llvm::StringRef(CStr)), StrOffset);
+      if (!R.second)
+        errs() << "BOLT-WARNING: [internal-dwarf-error]: collision occured on "
+               << CStr << " at offset : 0x" << Twine::utohexstr(StrOffset)
+               << ". Previous string offset is: 0x"
+               << Twine::utohexstr(R.first->second) << ".\n";
+      StrOffset = Offset;
+    }
+  }
+}
+
+void DWARF5AcceleratorTable::setCurrentUnit(DWARFUnit &Unit,
+                                            const uint64_t UnitStartOffset) {
+  CurrentUnit = nullptr;
+  CurrentUnitOffset = UnitStartOffset;
+  std::optional<uint64_t> DWOID = Unit.getDWOId();
+  // We process Skelton CUs after DWO Units for it.
+  // Patching offset in CU list to correct one.
+  if (!Unit.isDWOUnit() && DWOID) {
+    auto Iter = CUOffsetsToPatch.find(*DWOID);
+    // Check in case no entries were added from non skeleton DWO section.
+    if (Iter != CUOffsetsToPatch.end())
+      CUList[Iter->second] = UnitStartOffset;
+  }
+}
+
+void DWARF5AcceleratorTable::addUnit(DWARFUnit &Unit,
+                                     const std::optional<uint64_t> &DWOID) {
+  StrOffsetsWriter.clear();
+  StrOffsetsWriter.initialize(Unit);
+  StrSection = Unit.getStringSection();
+  if (Unit.isTypeUnit()) {
+    if (DWOID) {
+      // We adding an entry for a DWO TU. The DWO CU might not have any entries,
+      // so need to add it to the list pre-emptively.
+      auto Iter = CUOffsetsToPatch.insert({*DWOID, CUList.size()});
+      if (Iter.second)
+        CUList.push_back(0xBADBAD);
+      ForeignTUList.push_back(
+          cast_or_null<DWARFTypeUnit>(&Unit)->getTypeHash());
+    } else {
+      LocalTUList.push_back(CurrentUnitOffset);
+    }
+  } else {
+    if (DWOID) {
+      // This is a path for split dwarf without type units.
+      // We process DWO Units before Skeleton CU. So at this point we don't know
+      // the offset of Skeleton CU. Adding CULit index to a map to patch later
+      // with the correct offset.
+      auto Iter = CUOffsetsToPatch.insert({*DWOID, CUList.size()});
+      if (Iter.second)
+        CUList.push_back(0xBADBAD);
+    } else {
+      CUList.push_back(CurrentUnitOffset);
+    }
+  }
+}
+
+// Returns true if DW_TAG_variable should be included in .debug-names based on
+// section 6.1.1.1 for DWARF5 spec.
+static bool shouldIncludeVariable(const DWARFUnit &Unit, const DIE &Die) {
+  if (Die.findAttribute(dwarf::Attribute::DW_AT_declaration))
+    return false;
+  const DIEValue LocAttrInfo =
+      Die.findAttribute(dwarf::Attribute::DW_AT_location);
+  if (!LocAttrInfo)
+    return false;
+  if (!(doesFormBelongToClass(LocAttrInfo.getForm(), DWARFFormValue::FC_Exprloc,
+                              Unit.getVersion()) ||
+        doesFormBelongToClass(LocAttrInfo.getForm(), DWARFFormValue::FC_Block,
+                              Unit.getVersion())))
+    return false;
+  std::vector<uint8_t> Sblock;
+  auto constructVect =
+      [&](const DIEValueList::const_value_range &Iter) -> void {
+    for (const DIEValue &Val : Iter)
+      Sblock.push_back(Val.getDIEInteger().getValue());
+  };
+  if (doesFormBelongToClass(LocAttrInfo.getForm(), DWARFFormValue::FC_Exprloc,
+                            Unit.getVersion()))
+    constructVect(LocAttrInfo.getDIELoc().values());
+  else
+    constructVect(LocAttrInfo.getDIEBlock().values());
+  ArrayRef<uint8_t> Expr = ArrayRef<uint8_t>(Sblock);
+  DataExtractor Data(StringRef((const char *)Expr.data(), Expr.size()),
+                     Unit.getContext().isLittleEndian(), 0);
+  DWARFExpression LocExpr(Data, Unit.getAddressByteSize(),
+                          Unit.getFormParams().Format);
+  for (const DWARFExpression::Operation &Expr : LocExpr)
----------------
dcci wrote:

trying to confirm -- is this correct? i.e. as long as you have at least (not all) the locexpr being `addrx` or `form_tls_address`, do they qualify to be included in `.debug_names` ?

https://github.com/llvm/llvm-project/pull/81062