[llvm] [KernelInfo] Implement new LLVM IR pass for GPU code analysis (PR #102944)

[Matt Arsenault via llvm-commits]

================
@@ -26,6 +26,11 @@
 #include "llvm/Target/TargetLoweringObjectFile.h"
 using namespace llvm;
 
+cl::opt<bool> KernelInfoEndLTO(
----------------
arsenm wrote:

Can you just unconditionally run the pass, and early exit if !isAnalysisRemarkEnabled? 

https://github.com/llvm/llvm-project/pull/102944