linux, dump_syms: Filter module entries outside specified ranges

Partitioned libraries generated with lld and llvm-objcopy currently
contain a superset of debug information, beyond what applies to the
library itself. This is because objcopy cannot split up debug
information by partition - instead, it places a copy of all debug
information into each partition.

In lieu of potential future support for lld or objcopy becoming able to
split up debug information, let dump_syms do the next best thing:

- Find the address ranges of all PT_LOAD segments in the lib.
- Supply these to the Module being generated.
- Filter additions to the Module based on these ranges.

Bug: 990190
Change-Id: Ib5f279f42e3f6ea79eed9665efbcc23c3c5d25dc
Reviewed-on: https://chromium-review.googlesource.com/c/breakpad/breakpad/+/1884699
Reviewed-by: Joshua Peraza <jperaza@chromium.org>

1 files changed, 25 insertions, 0 deletions

diff --git a/src/common/linux/dump_symbols.cc b/src/common/linux/dump_symbols.cc
index 660f133e..e561ad94 100644
--- a/src/common/linux/dump_symbols.cc
+++ b/src/common/linux/dump_symbols.cc
@@ -182,6 +182,23 @@ typename ElfClass::Addr GetLoadingAddress(
   return 0;
 }
 
+// Find the set of address ranges for all PT_LOAD segments.
+template <typename ElfClass>
+vector<Module::Range> GetPtLoadSegmentRanges(
+    const typename ElfClass::Phdr* program_headers,
+    int nheader) {
+  typedef typename ElfClass::Phdr Phdr;
+  vector<Module::Range> ranges;
+
+  for (int i = 0; i < nheader; ++i) {
+    const Phdr& header = program_headers[i];
+    if (header.p_type == PT_LOAD) {
+      ranges.push_back(Module::Range(header.p_vaddr, header.p_memsz));
+    }
+  }
+  return ranges;
+}
+
 #ifndef NO_STABS_SUPPORT
 template<typename ElfClass>
 bool LoadStabs(const typename ElfClass::Ehdr* elf_header,
@@ -649,6 +666,14 @@ bool LoadSymbols(const string& obj_file,
   module->SetLoadAddress(loading_addr);
   info->set_loading_addr(loading_addr, obj_file);
 
+  // Allow filtering of extraneous debug information in partitioned libraries.
+  // Such libraries contain debug information for all libraries extracted from
+  // the same combined library, implying extensive duplication.
+  vector<Module::Range> address_ranges = GetPtLoadSegmentRanges<ElfClass>(
+      GetOffset<ElfClass, Phdr>(elf_header, elf_header->e_phoff),
+      elf_header->e_phnum);
+  module->SetAddressRanges(address_ranges);
+
   const Shdr* sections =
       GetOffset<ElfClass, Shdr>(elf_header, elf_header->e_shoff);
   const Shdr* section_names = sections + elf_header->e_shstrndx;