[PATCH] D15025: [ThinLTO] Option to invoke ThinLTO backend passes and importing

[Mehdi AMINI via cfe-commits]

joker.eph added inline comments.

================
Comment at: lib/CodeGen/BackendUtil.cpp:308
@@ +307,3 @@
+    return;
+  }
+
----------------
It does not seem to be nicely integrated within the context of this function. What about all the options set just a few line below? It is not clear if `CodeGenOpts.DisableLLVMOpts` is well honored either.

================
Comment at: lib/CodeGen/CodeGenAction.cpp:30
@@ -29,2 +29,3 @@
 #include "llvm/IR/LLVMContext.h"
+#include "llvm/IR/FunctionInfo.h"
 #include "llvm/IR/Module.h"
----------------
Not well sorted :)

================
Comment at: lib/CodeGen/CodeGenAction.cpp:190
@@ -169,3 +189,3 @@
                                 [=](const DiagnosticInfo &DI) {
-                                  linkerDiagnosticHandler(DI, LinkModule);
+                                  linkerDiagnosticHandler(DI, LinkModule, Diags);
                                 },
----------------
Is this an unrelated change to the `-fthinlto-backend` one that could be committed separately?

================
Comment at: lib/CodeGen/CodeGenAction.cpp:821
@@ +820,3 @@
+                                linkerDiagnosticHandler(DI, TheModule.get(),
+                                                        CI.getDiagnostics());
+                              }, llvm::Linker::Flags::None, Index.get()))
----------------
This lambda is the same as the one just above for `getFunctionIndexForFile`, name it and define it once?

================
Comment at: lib/CodeGen/CodeGenAction.cpp:826
@@ +825,3 @@
+      TheModule = std::move(Combined);
+    }
+
----------------
So for the renaming we need to duplicate completely the module? It cannot be done by morphing the existing module in place? That seems quite inefficient :(


http://reviews.llvm.org/D15025