[Lldb-commits] [lldb] Improve type and namespace lookup using parent chain (PR #108907)

[via lldb-commits]

================
@@ -374,25 +377,40 @@ void DebugNamesDWARFIndex::GetFullyQualifiedType(
   m_fallback.GetFullyQualifiedType(context, callback);
 }
 
+bool DebugNamesDWARFIndex::SameAsEntryATName(
+    llvm::StringRef query_name, const DebugNames::Entry &entry) const {
+  auto maybe_dieoffset = entry.getDIEUnitOffset();
+  if (!maybe_dieoffset)
+    return false;
+
+  // [Optimization] instead of parsing the entry from dwo file, we simply
+  // check if the query_name can point to an entry of the same DIE offset.
+  // This greatly reduced number of dwo file parsed and thus improved the
+  // performance.
+  for (const DebugNames::Entry &query_entry :
+       entry.getNameIndex()->equal_range(query_name)) {
+    auto query_dieoffset = query_entry.getDIEUnitOffset();
+    if (!query_dieoffset)
+      continue;
+
+    if (*query_dieoffset == *maybe_dieoffset) {
+      return true;
+    } else if (*query_dieoffset > *maybe_dieoffset) {
+      // The pool entries of the same name are sequentially cluttered together
+      // so if the query name from `query_name` is after the target entry, this
+      // is definitely not the correct one, we can stop searching.
+      return false;
+    }
+  }
+  return false;
+}
+
 bool DebugNamesDWARFIndex::SameParentChain(
     llvm::ArrayRef<llvm::StringRef> parent_names,
     llvm::ArrayRef<DebugNames::Entry> parent_entries) const {
-
   if (parent_entries.size() != parent_names.size())
     return false;
 
-  auto SameAsEntryATName = [this](llvm::StringRef name,
-                                  const DebugNames::Entry &entry) {
-    // Peek at the AT_name of `entry` and test equality to `name`.
-    auto maybe_dieoffset = entry.getDIEUnitOffset();
-    if (!maybe_dieoffset)
-      return false;
-    DWARFUnit *unit = GetNonSkeletonUnit(entry);
-    if (!unit)
-      return false;
-    return name == unit->PeekDIEName(unit->GetOffset() + *maybe_dieoffset);
----------------
jeffreytan81 wrote:

@felipepiovezan , I am having trouble finding a context without split dwarf to measure the performance. Our target is way too big to build without split dwarf (linker error). I am trying to debug clang++ per your benchmark, but the evaluation is very fast `293ms` which I do not think it is slow enough to measure perf. How do you get 5~6 seconds from this? 

```
devbig623:~/llvm-sand/build/Debug/fbcode-x86_64/toolchain$ lldb -o "b CodeGenFunction::GenerateCode" -o r -o "timeit expr Fn" --  ~/llvm-sand/build/Debug/fbcode-x86_64/toolchain/bin/clang++ -std=c++17 -pthread -glldb -c ~/personal/main.cpp -o ~/personal/a
(lldb) target create "/home/jeffreytan/llvm-sand/build/Debug/fbcode-x86_64/toolchain/bin/clang++"
Current executable set to '/home/jeffreytan/llvm-sand/build/Debug/fbcode-x86_64/toolchain/bin/clang++' (x86_64).
(lldb) settings set -- target.run-args  "-std=c++17" "-pthread" "-glldb" "-c" "/home/jeffreytan/personal/main.cpp" "-o" "/home/jeffreytan/personal/a"
(lldb) b CodeGenFunction::GenerateCode
Breakpoint 1: where = clang++`clang::CodeGen::CodeGenFunction::GenerateCode(clang::GlobalDecl, llvm::Function*, clang::CodeGen::CGFunctionInfo const&) + 59 at CodeGenFunction.cpp:1440:3, address = 0x00000000083c507b
(lldb) r
Process 233320 launched: '/home/jeffreytan/llvm-sand/build/Debug/fbcode-x86_64/toolchain/bin/clang++' (x86_64)
Process 233320 stopped
* thread #1, name = 'clang++', stop reason = breakpoint 1.1
    frame #0: 0x000055555d91907b clang++`clang::CodeGen::CodeGenFunction::GenerateCode(this=0x00007fffffff5bb0, GD=GlobalDecl @ 0x00007fffffff5b48, Fn=0x000002921787f228, FnInfo=0x0000029218308200) at CodeGenFunction.cpp:1440:3
   1437
   1438	void CodeGenFunction::GenerateCode(GlobalDecl GD, llvm::Function *Fn,
   1439	                                   const CGFunctionInfo &FnInfo) {
-> 1440	  assert(Fn && "generating code for null Function");
   1441	  const FunctionDecl *FD = cast<FunctionDecl>(GD.getDecl());
   1442	  CurGD = GD;
   1443
(lldb) timeit expr Fn
(llvm::Function *) $0 = 0x000002921787f228

time: 0m0.293s
```

https://github.com/llvm/llvm-project/pull/108907