[PATCH] D17625: Do not select EhPad BB in MachineBlockPlacement when there is regular BB to schedule

David Li via llvm-commits llvm-commits at lists.llvm.org
Fri Mar 11 14:22:06 PST 2016


davidxl added inline comments.

================
Comment at: lib/CodeGen/MachineBlockPlacement.cpp:621
@@ -597,3 +620,3 @@
 
   if (Chain.UnscheduledPredecessors != 0)
     return;
----------------
suggest this change:

if (Chain.UnschedulePredecesors == 0) {
   MBB = *Chain.begin();
   auto &WorkList = (MBB->isEHPad() ? EHPadWorkList: BlockWorkList);
   WorkList.push_back(MBB);
}

================
Comment at: test/CodeGen/X86/block-placement.ll:1145
@@ +1144,3 @@
+; CHECK: %exit
+; CHECK: %innerlp
+; CHECK: %outerlp
----------------
Where is innnercleanup?

================
Comment at: test/CodeGen/X86/block-placement.ll:1159
@@ +1158,3 @@
+
+innerlp:
+  %2 = landingpad { i8*, i32 }
----------------
Can you shuffle the lp labels into different orders in the original source?


http://reviews.llvm.org/D17625





More information about the llvm-commits mailing list