[PATCH] D39840: [MC][X86] Code padding for performance stability - Branch instructions and targets alignment

[Craig Topper via Phabricator via llvm-commits]

craig.topper added inline comments.


================
Comment at: lib/Target/X86/MCTargetDesc/X86AsmBackend.cpp:79
   X86AsmBackend(const Target &T, StringRef CPU)
-      : MCAsmBackend(), CPU(CPU),
+      : MCAsmBackend(std::make_unique<X86::X86MCCodePadder>(T, CPU)), CPU(CPU),
         MaxNopLength((CPU == "slm" || CPU == "silvermont") ? 7 : 15) {
----------------
Use llvm::make_unique not std::make_unique. std::make_unique is a C++14 feature. LLVM code is required to build in C++11.


================
Comment at: lib/Target/X86/MCTargetDesc/X86MCCodePadder.cpp:33
+      CPU != "core-avx-i" && CPU != "haswell" && CPU != "core-avx2" &&
+      CPU != "broadwell" && CPU != "skylake")
+    return;
----------------
I didn't see an answer about skylake-avx512.


https://reviews.llvm.org/D39840