Process BUNDLE in tail duplication

When tail duplication estimates a size of tail it uses instruction
count. Account for a number of instrictions in a bundle too.

Differential Revision: https://reviews.llvm.org/D72783

llvm/lib/CodeGen/TailDuplicator.cpp

@@ -625,7 +625,9 @@ bool TailDuplicator::shouldTailDuplicate(bool IsSimple,
     if (PreRegAlloc && MI.isCall())
       return false;
 
-    if (!MI.isPHI() && !MI.isMetaInstruction())
+    if (MI.isBundle())
+      InstrCount += MI.getBundleSize();
+    else if (!MI.isPHI() && !MI.isMetaInstruction())
       InstrCount += 1;
 
     if (InstrCount > MaxDuplicateCount)

llvm/test/CodeGen/AMDGPU/tail-dup-bundle.mir

@@ -0,0 +1,38 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -march=amdgcn -run-pass=block-placement -tail-dup-placement-threshold=2 -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
+
+# Check that tail duplication correctly counts instructions in a bundle.
+# The bundle below shall not be duplicated.
+
+---
+name:            tail_dup_bundle
+body:             |
+  ; GCN-LABEL: name: tail_dup_bundle
+  ; GCN: bb.0:
+  ; GCN:   successors: %bb.2(0x80000000)
+  ; GCN: bb.2:
+  ; GCN:   BUNDLE {
+  ; GCN:     S_NOP 0
+  ; GCN:     S_NOP 0
+  ; GCN:   }
+  ; GCN:   S_ENDPGM 0
+  ; GCN: bb.1:
+  ; GCN:   successors: %bb.2(0x80000000)
+  ; GCN:   S_BRANCH %bb.2
+  bb.0:
+    successors: %bb.2(0x80000000)
+
+    S_BRANCH %bb.2
+
+  bb.1:
+    successors: %bb.2(0x80000000)
+
+  bb.2:
+
+    BUNDLE {
+      S_NOP 0
+      S_NOP 0
+    }
+    S_ENDPGM 0
+
+...