[PATCH] Fixing a heisenbug where the memory dependence analysis behaves differently with and without -g

David Blaikie dblaikie at gmail.com
Wed Nov 13 13:26:53 PST 2013

Comment at: test/Transforms/DeadStoreElimination/inst_limits.ll:338
@@ +337,3 @@
+  ; effect on the working of DSE in any way.
+  call void @llvm.dbg.value(metadata !12, i64 0, metadata !10)
+  call void @llvm.dbg.value(metadata !12, i64 0, metadata !10)
I assume you only need one dbg.value intrinsic to demonstrate the problem, right?

And you only really need it in the inside limit case - in the sense that the presence of that one intrinsic shouldn't tip you over the limit.


More information about the llvm-commits mailing list