[LLDB] Fix debuginfo ELF files overwriting Unified Section List (llvm#166635)

Recently I've been deep diving ELF cores in LLDB, aspiring to move LLDB
closer to GDB in capability. One issue I encountered was a system lib
losing it's unwind plan when loading the debuginfo. The reason for this
was the debuginfo has the eh_frame section stripped and the main
executable did not.

The root cause of this was this line in
[ObjectFileElf](https://github.com/llvm/llvm-project/blob/163933e9e7099f352ff8df1973f9a9c3d7def6c5/lldb/source/Plugins/ObjectFile/ELF/ObjectFileELF.cpp#L1972)
```
  // For eTypeDebugInfo files, the Symbol Vendor will take care of updating the
  // unified section list.
  if (GetType() != eTypeDebugInfo)
    unified_section_list = *m_sections_up;
```

This would always be executed because CalculateType can never return an
eTypeDebugInfo
```
ObjectFile::Type ObjectFileELF::CalculateType() {
  switch (m_header.e_type) {
  case llvm::ELF::ET_NONE:
    // 0 - No file type
    return eTypeUnknown;

  case llvm::ELF::ET_REL:
    // 1 - Relocatable file
    return eTypeObjectFile;

  case llvm::ELF::ET_EXEC:
    // 2 - Executable file
    return eTypeExecutable;

  case llvm::ELF::ET_DYN:
    // 3 - Shared object file
    return eTypeSharedLibrary;

  case ET_CORE:
    // 4 - Core file
    return eTypeCoreFile;

  default:
    break;
  }
  return eTypeUnknown;
}
```

This makes sense as there isn't a explicit sh_type to denote that this
file is a debuginfo. After some discussion with @clayborg and
@GeorgeHuyubo we settled on joining the exciting unified section list
with whatever new sections were being added. Adding each new unique
section, or taking the section with the maximum file size. We picked
this strategy to pick the section with the most information. In most
scenarios, LHS should be SHT_NOBITS and RHS would be SHT_PROGBITS.

Here is a diagram documenting the existing vs proposed new way.
<img width="1666" height="1093" alt="image"
src="https://github.com/user-attachments/assets/73ba9620-c737-439e-9934-ac350d88a3b5"
/>

Author: Jlalond

lldb/include/lldb/Core/Section.h

@@ -46,6 +46,8 @@ class SectionList {
   /// Create an empty list.
   SectionList() = default;
 
+  SectionList(const SectionList &lhs);
+
   SectionList &operator=(const SectionList &rhs);
 
   size_t AddSection(const lldb::SectionSP &section_sp);
@@ -96,6 +98,17 @@ class SectionList {
   /// information.
   uint64_t GetDebugInfoSize() const;
 
+  // Callback to decide which of two matching sections should be used in the
+  // merged output.
+  using MergeCallback =
+      std::function<lldb::SectionSP(lldb::SectionSP, lldb::SectionSP)>;
+
+  // Function that merges two different sections into a new output list. All
+  // unique sections will be checked for conflict and resolved using the
+  // supplied merging callback.
+  static SectionList Merge(SectionList &lhs, SectionList &rhs,
+                           MergeCallback filter);
+
 protected:
   collection m_sections;
 };

lldb/source/Core/Section.cpp

@@ -477,6 +477,8 @@ bool Section::IsGOTSection() const {
 
 #pragma mark SectionList
 
+SectionList::SectionList(const SectionList &rhs) : m_sections(rhs.m_sections) {}
+
 SectionList &SectionList::operator=(const SectionList &rhs) {
   if (this != &rhs)
     m_sections = rhs.m_sections;
@@ -687,6 +689,33 @@ uint64_t SectionList::GetDebugInfoSize() const {
   return debug_info_size;
 }
 
+SectionList SectionList::Merge(SectionList &lhs, SectionList &rhs,
+                               MergeCallback filter) {
+  SectionList output_list;
+
+  // Iterate through all the sections in lhs and see if we have matches in
+  // the rhs list.
+  for (const auto &lhs_section : lhs) {
+    auto rhs_section = rhs.FindSectionByName(lhs_section->GetName());
+    if (rhs_section)
+      output_list.AddSection(filter(lhs_section, rhs_section));
+    else
+      output_list.AddSection(lhs_section);
+  }
+
+  // Now that we've visited all possible duplicates, we can iterate over
+  // the rhs and take any values not in lhs.
+  for (const auto &rhs_section : rhs) {
+    auto lhs_section = lhs.FindSectionByName(rhs_section->GetName());
+    // Because we already visited everything overlapping between rhs
+    // and lhs, any section not in lhs is unique and can be output.
+    if (!lhs_section)
+      output_list.AddSection(rhs_section);
+  }
+
+  return output_list;
+}
+
 namespace llvm {
 namespace json {
 

lldb/source/Plugins/ObjectFile/ELF/ObjectFileELF.cpp

@@ -130,6 +130,29 @@ class ELFRelocation {
 
   RelocUnion reloc;
 };
+
+lldb::SectionSP MergeSections(lldb::SectionSP lhs, lldb::SectionSP rhs) {
+  assert(lhs && rhs);
+
+  lldb::ModuleSP lhs_module_parent = lhs->GetModule();
+  lldb::ModuleSP rhs_module_parent = rhs->GetModule();
+  assert(lhs_module_parent && rhs_module_parent);
+
+  // Do a sanity check, these should be the same.
+  if (lhs->GetFileAddress() != rhs->GetFileAddress())
+    lhs_module_parent->ReportWarning(
+        "Mismatch addresses for section {0} when "
+        "merging with {1}, expected: {2:x}, "
+        "actual: {3:x}",
+        lhs->GetTypeAsCString(),
+        rhs_module_parent->GetFileSpec().GetPathAsConstString().GetCString(),
+        lhs->GetByteSize(), rhs->GetByteSize());
+
+  // We want to take the greater of two sections. If LHS and RHS are both
+  // SHT_NOBITS, we should default to LHS. If RHS has a bigger section,
+  // indicating it has data that wasn't stripped, we should take that instead.
+  return rhs->GetFileSize() > lhs->GetFileSize() ? rhs : lhs;
+}
 } // end anonymous namespace
 
 ELFRelocation::ELFRelocation(unsigned type) {
@@ -1967,10 +1990,10 @@ void ObjectFileELF::CreateSections(SectionList &unified_section_list) {
     provider.AddSection(std::move(*InfoOr), std::move(section_sp));
   }
 
-  // For eTypeDebugInfo files, the Symbol Vendor will take care of updating the
-  // unified section list.
-  if (GetType() != eTypeDebugInfo)
-    unified_section_list = *m_sections_up;
+  // Merge the two adding any new sections, and overwriting any existing
+  // sections that are SHT_NOBITS
+  unified_section_list =
+      SectionList::Merge(unified_section_list, *m_sections_up, MergeSections);
 
   // If there's a .gnu_debugdata section, we'll try to read the .symtab that's
   // embedded in there and replace the one in the original object file (if any).

lldb/test/API/python_api/unified_section_list/Makefile

@@ -0,0 +1,5 @@
+CXX_SOURCES := main.cpp
+
+SPLIT_DEBUG_SYMBOLS := YES
+
+include Makefile.rules