From 36c53fe147378109ddebed3d3f62fafcccca105a Mon Sep 17 00:00:00 2001
From: Jun Bum Lim <junbuml@codeaurora.org>
Date: Fri, 25 Mar 2016 21:15:35 +0000
Subject: [PATCH] [MachineCopyPropagation] Expose more dead copies across
 instructions with regmasks

When encountering instructions with regmasks, instead of cleaning up all the
elements in MaybeDeadCopies map, remove only the instructions erased. By keeping
more instruction in MaybeDeadCopies, this change will expose more dead copies
across instructions with regmasks.

llvm-svn: 264462
---
 llvm/lib/CodeGen/MachineCopyPropagation.cpp        | 17 +++++-
 .../test/CodeGen/MIR/AArch64/machine-dead-copy.mir | 67 ++++++++++++++++++++++
 2 files changed, 81 insertions(+), 3 deletions(-)
 create mode 100644 llvm/test/CodeGen/MIR/AArch64/machine-dead-copy.mir

diff --git a/llvm/lib/CodeGen/MachineCopyPropagation.cpp b/llvm/lib/CodeGen/MachineCopyPropagation.cpp
index 2bfd650..3a6f2b5 100644
--- a/llvm/lib/CodeGen/MachineCopyPropagation.cpp
+++ b/llvm/lib/CodeGen/MachineCopyPropagation.cpp
@@ -285,18 +285,28 @@ void MachineCopyPropagation::CopyPropagateBlock(MachineBasicBlock &MBB) {
     // defined registers.
     if (RegMask) {
       // Erase any MaybeDeadCopies whose destination register is clobbered.
-      for (MachineInstr *MaybeDead : MaybeDeadCopies) {
+      for (SmallSetVector<MachineInstr *, 8>::iterator DI =
+               MaybeDeadCopies.begin();
+           DI != MaybeDeadCopies.end();) {
+        MachineInstr *MaybeDead = *DI;
         unsigned Reg = MaybeDead->getOperand(0).getReg();
         assert(!MRI->isReserved(Reg));
-        if (!RegMask->clobbersPhysReg(Reg))
+
+        if (!RegMask->clobbersPhysReg(Reg)) {
+          ++DI;
           continue;
+        }
+
         DEBUG(dbgs() << "MCP: Removing copy due to regmask clobbering: ";
               MaybeDead->dump());
+
+        // erase() will return the next valid iterator pointing to the next
+        // element after the erased one.
+        DI = MaybeDeadCopies.erase(DI);
         MaybeDead->eraseFromParent();
         Changed = true;
         ++NumDeletes;
       }
-      MaybeDeadCopies.clear();
 
       removeClobberedRegsFromMap(AvailCopyMap, *RegMask);
       removeClobberedRegsFromMap(CopyMap, *RegMask);
@@ -348,3 +358,4 @@ bool MachineCopyPropagation::runOnMachineFunction(MachineFunction &MF) {
 
   return Changed;
 }
+
diff --git a/llvm/test/CodeGen/MIR/AArch64/machine-dead-copy.mir b/llvm/test/CodeGen/MIR/AArch64/machine-dead-copy.mir
new file mode 100644
index 0000000..1257948
--- /dev/null
+++ b/llvm/test/CodeGen/MIR/AArch64/machine-dead-copy.mir
@@ -0,0 +1,67 @@
+
+# RUN: llc -mtriple=aarch64-none-linux-gnu -run-pass machine-cp  -verify-machineinstrs  -o /dev/null %s 2>&1 | FileCheck %s
+
+--- |
+  define i32 @copyprop1(i32 %a, i32 %b) { ret i32 %a }
+  define i32 @copyprop2(i32 %a, i32 %b) { ret i32 %a }
+  define i32 @copyprop3(i32 %a, i32 %b) { ret i32 %a }
+  define i32 @copyprop4(i32 %a, i32 %b) { ret i32 %a }
+  declare i32 @foo(i32)
+...
+---
+# The first copy is dead copy which is not used.
+# CHECK-LABEL: name: copyprop1
+# CHECK: bb.0:
+# CHECK-NOT: %w20 = COPY
+name:            copyprop1
+body:             |
+  bb.0:
+    liveins: %w0,  %w1
+    %w20 = COPY %w1
+    BL @foo, csr_aarch64_aapcs, implicit %w0, implicit-def %w0
+    RET_ReallyLR implicit %w0
+...
+---
+# The first copy is not a dead copy which is used in the second copy after the
+# call.
+# CHECK-LABEL: name: copyprop2
+# CHECK: bb.0:
+# CHECK: %w20 = COPY
+name:            copyprop2
+body:             |
+  bb.0:
+    liveins: %w0,  %w1
+    %w20 = COPY %w1
+    BL @foo, csr_aarch64_aapcs, implicit %w0, implicit-def %w0
+    %w0 = COPY %w20
+    RET_ReallyLR implicit %w0
+...
+---
+# Both the first and second copy are dead copies which are not used.
+# CHECK-LABEL: name: copyprop3
+# CHECK: bb.0:
+# CHECK-NOT: COPY
+name:            copyprop3
+body:             |
+  bb.0:
+    liveins: %w0,  %w1
+    %w20 = COPY %w1
+    BL @foo, csr_aarch64_aapcs, implicit %w0, implicit-def %w0
+    %w20 = COPY %w0
+    RET_ReallyLR implicit %w0
+...
+# The second copy is removed as a NOP copy, after then the first copy become
+# dead which should be removed as well.
+# CHECK-LABEL: name: copyprop4
+# CHECK: bb.0:
+# CHECK-NOT: COPY
+name:            copyprop4
+body:             |
+  bb.0:
+    liveins: %w0,  %w1
+    %w20 = COPY %w0
+    %w0 = COPY %w20
+    BL @foo, csr_aarch64_aapcs, implicit %w0, implicit-def %w0
+    RET_ReallyLR implicit %w0
+...
+
-- 
2.7.4