[llvm] bb50693 - [SystemZ] Improvement of emitMemMemWrapper()

[Jonas Paulsson via llvm-commits]

Author: Jonas Paulsson
Date: 2021-10-26T17:03:01+02:00
New Revision: bb506938be820df8c7992b7c567fa93afaab2bc4

URL: https://github.com/llvm/llvm-project/commit/bb506938be820df8c7992b7c567fa93afaab2bc4
DIFF: https://github.com/llvm/llvm-project/commit/bb506938be820df8c7992b7c567fa93afaab2bc4.diff

LOG: [SystemZ] Improvement of emitMemMemWrapper()

It was discovered that an extra register COPY remained when expanding a
(variable length) memory operation with a loop and there was another use of
the involved address register(s) afterwards.

A simple fix for this is to COPY the address registers before the loop and
use that new vreg instead.

Review: Ulrich Weigand

Differential Revision: https://reviews.llvm.org/D112065

Added: 
    

Modified: 
    llvm/lib/Target/SystemZ/SystemZISelLowering.cpp
    llvm/test/CodeGen/SystemZ/memset-05.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/SystemZ/SystemZISelLowering.cpp b/llvm/lib/Target/SystemZ/SystemZISelLowering.cpp
index a9bc4f30fff6..90d2794feb1e 100644
--- a/llvm/lib/Target/SystemZ/SystemZISelLowering.cpp
+++ b/llvm/lib/Target/SystemZ/SystemZISelLowering.cpp
@@ -7145,13 +7145,19 @@ SystemZTargetLowering::getStackProbeSize(MachineFunction &MF) const {
 // Force base value Base into a register before MI.  Return the register.
 static Register forceReg(MachineInstr &MI, MachineOperand &Base,
                          const SystemZInstrInfo *TII) {
-  if (Base.isReg())
-    return Base.getReg();
-
   MachineBasicBlock *MBB = MI.getParent();
   MachineFunction &MF = *MBB->getParent();
   MachineRegisterInfo &MRI = MF.getRegInfo();
 
+  if (Base.isReg()) {
+    // Copy Base into a new virtual register to help register coalescing in
+    // cases with multiple uses.
+    Register Reg = MRI.createVirtualRegister(&SystemZ::ADDR64BitRegClass);
+    BuildMI(*MBB, MI, MI.getDebugLoc(), TII->get(SystemZ::COPY), Reg)
+      .add(Base);
+    return Reg;
+  }
+
   Register Reg = MRI.createVirtualRegister(&SystemZ::ADDR64BitRegClass);
   BuildMI(*MBB, MI, MI.getDebugLoc(), TII->get(SystemZ::LA), Reg)
       .add(Base)

diff  --git a/llvm/test/CodeGen/SystemZ/memset-05.ll b/llvm/test/CodeGen/SystemZ/memset-05.ll
index 6d3aa315f8bd..077a1f6da566 100644
--- a/llvm/test/CodeGen/SystemZ/memset-05.ll
+++ b/llvm/test/CodeGen/SystemZ/memset-05.ll
@@ -48,43 +48,39 @@ define void @fun2(i8* %Addr, i32 %Len) {
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    llgfr %r1, %r3
 ; CHECK-NEXT:    aghi %r1, -1
-; CHECK-NEXT:    cgije %r1, -1, .LBB2_5
+; CHECK-NEXT:    cgije %r1, -1, .LBB2_4
 ; CHECK-NEXT:  # %bb.1:
 ; CHECK-NEXT:    srlg %r0, %r1, 8
 ; CHECK-NEXT:    lgr %r3, %r2
-; CHECK-NEXT:    cgije %r0, 0, .LBB2_4
-; CHECK-NEXT:  # %bb.2:
-; CHECK-NEXT:    lgr %r3, %r2
-; CHECK-NEXT:  .LBB2_3: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    cgije %r0, 0, .LBB2_3
+; CHECK-NEXT:  .LBB2_2: # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    xc 0(256,%r3), 0(%r3)
 ; CHECK-NEXT:    la %r3, 256(%r3)
-; CHECK-NEXT:    brctg %r0, .LBB2_3
-; CHECK-NEXT:  .LBB2_4:
+; CHECK-NEXT:    brctg %r0, .LBB2_2
+; CHECK-NEXT:  .LBB2_3:
 ; CHECK-NEXT:    exrl %r1, .Ltmp1
-; CHECK-NEXT:  .LBB2_5:
-; CHECK-NEXT:    cgije %r1, -1, .LBB2_10
-; CHECK-NEXT:  # %bb.6:
+; CHECK-NEXT:  .LBB2_4:
+; CHECK-NEXT:    cgije %r1, -1, .LBB2_8
+; CHECK-NEXT:  # %bb.5:
 ; CHECK-NEXT:    srlg %r0, %r1, 8
 ; CHECK-NEXT:    lgr %r3, %r2
-; CHECK-NEXT:    cgije %r0, 0, .LBB2_9
-; CHECK-NEXT:  # %bb.7:
-; CHECK-NEXT:    lgr %r3, %r2
-; CHECK-NEXT:  .LBB2_8: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    cgije %r0, 0, .LBB2_7
+; CHECK-NEXT:  .LBB2_6: # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    xc 0(256,%r3), 0(%r3)
 ; CHECK-NEXT:    la %r3, 256(%r3)
-; CHECK-NEXT:    brctg %r0, .LBB2_8
-; CHECK-NEXT:  .LBB2_9:
+; CHECK-NEXT:    brctg %r0, .LBB2_6
+; CHECK-NEXT:  .LBB2_7:
 ; CHECK-NEXT:    exrl %r1, .Ltmp1
-; CHECK-NEXT:  .LBB2_10:
+; CHECK-NEXT:  .LBB2_8:
 ; CHECK-NEXT:    cgibe %r1, -1, 0(%r14)
-; CHECK-NEXT:  .LBB2_11:
+; CHECK-NEXT:  .LBB2_9:
 ; CHECK-NEXT:    srlg %r0, %r1, 8
-; CHECK-NEXT:    cgije %r0, 0, .LBB2_13
-; CHECK-NEXT:  .LBB2_12: # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    cgije %r0, 0, .LBB2_11
+; CHECK-NEXT:  .LBB2_10: # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    xc 0(256,%r2), 0(%r2)
 ; CHECK-NEXT:    la %r2, 256(%r2)
-; CHECK-NEXT:    brctg %r0, .LBB2_12
-; CHECK-NEXT:  .LBB2_13:
+; CHECK-NEXT:    brctg %r0, .LBB2_10
+; CHECK-NEXT:  .LBB2_11:
 ; CHECK-NEXT:    exrl %r1, .Ltmp0
 ; CHECK-NEXT:    br %r14
   tail call void @llvm.memset.p0i8.i32(i8* %Addr, i8 0, i32 %Len, i1 false)