[LTO] Avoid repeated hash lookups (NFC)

[kazutakahirata]

No description provided.

[llvmbot]

@llvm/pr-subscribers-lto

Author: Kazu Hirata (kazutakahirata)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/112470.diff

1 Files Affected:

• (modified) llvm/lib/LTO/LTO.cpp (+3-2)

diff --git a/llvm/lib/LTO/LTO.cpp b/llvm/lib/LTO/LTO.cpp
index 90c4e2c3cd131c..e6846ea9912ca8 100644
--- a/llvm/lib/LTO/LTO.cpp
+++ b/llvm/lib/LTO/LTO.cpp
@@ -635,11 +635,12 @@ void LTO::addModuleToGlobalRes(ArrayRef<InputFile::Symbol> Syms,
     SymbolResolution Res = *ResI++;
 
     StringRef SymbolName = Sym.getName();
+    auto [It, Inserted] = GlobalResolutions->try_emplace(SymbolName);
     // Keep copies of symbols if the client of LTO says so.
-    if (GlobalResolutionSymbolSaver && !GlobalResolutions->contains(SymbolName))
+    if (GlobalResolutionSymbolSaver && Inserted)
       SymbolName = GlobalResolutionSymbolSaver->save(SymbolName);
 
-    auto &GlobalRes = (*GlobalResolutions)[SymbolName];
+    auto &GlobalRes = It->second;
     GlobalRes.UnnamedAddr &= Sym.isUnnamedAddr();
     if (Res.Prevailing) {
       assert(!GlobalRes.Prevailing &&

[teresajohnson]

Is this functionally equivalent? Previously if the GlobalResolutions map didn't contain SymbolName then we got a new SymbolName from the symbol saver and created an entry in GlobalResolutions with that. Now we are only indexing GlobalResolutions with the old SymbolName.

[kazutakahirata]

Is this functionally equivalent? Previously if the GlobalResolutions map didn't contain SymbolName then we got a new SymbolName from the symbol saver and created an entry in GlobalResolutions with that. Now we are only indexing GlobalResolutions with the old SymbolName.

Thank you for catching this. You are right. I missed the update to SymbolName.