[ARM] Fix for PR36577

[samparker]

Summary

Don't PerformSHLSimplify if the given node is used by a node that also uses a constant because we may get stuck in an infinite combine loop.

bugzilla: https://bugs.llvm.org/show_bug.cgi?id=36577

[thopre]

lib/Target/ARM/ARMISelLowering.cpp

10452

its -> it's while you are at it?

[samparker]

Fixed typo.

[john.brawn]

LGTM.

lib/Target/ARM/ARMISelLowering.cpp

       return SDValue();
     case ISD::SUB:
     case ISD::ADD:
     case ISD::AND:
     case ISD::OR:
     case ISD::XOR:
     case ISD::SETCC:
     case ARMISD::CMP:
-      // Check that its not already using a shl.
+      // Check that the user isn't already using a constant because there
+      // aren't any instructions that support an immediate operand and a
+      // shifted operand.
+      if (isa<ConstantSDNode>(U->getOperand(0)) ||
+          isa<ConstantSDNode>(U->getOperand(1)))
+        return SDValue();
+
+      // Check that it's not already using a shift.
       if (U->getOperand(0).getOpcode() == ISD::SHL ||
           U->getOperand(1).getOpcode() == ISD::SHL)
         return SDValue();
       break;
     }
   }
 
   if (N->getOpcode() != ISD::ADD && N->getOpcode() != ISD::OR &&
       N->getOpcode() != ISD::XOR && N->getOpcode() != ISD::AND)
     return SDValue();
 
   if (N->getOperand(0).getOpcode() != ISD::SHL)
     return SDValue();
 
   SDValue SHL = N->getOperand(0);
 
   auto *C1ShlC2 = dyn_cast<ConstantSDNode>(N->getOperand(1));
   auto *C2 = dyn_cast<ConstantSDNode>(SHL.getOperand(1));
   if (!C1ShlC2 || !C2)
     return SDValue();
 
-  DEBUG(dbgs() << "Trying to simplify shl: "; N->dump());
-
   APInt C2Int = C2->getAPIntValue();
   APInt C1Int = C1ShlC2->getAPIntValue();
 
   // Check that performing a lshr will not lose any information.
   APInt Mask = APInt::getHighBitsSet(C2Int.getBitWidth(),
                                      C2Int.getBitWidth() - C2->getZExtValue());
   if ((C1Int & Mask) != C1Int)
     return SDValue();
 
   // Shift the first constant.
   C1Int.lshrInPlace(C2Int);
 
   // The immediates are encoded as an 8-bit value that can be rotated.
-  unsigned Zeros = C1Int.countLeadingZeros() + C1Int.countTrailingZeros();
-  if (C1Int.getBitWidth() - Zeros > 8)
-    return SDValue();
+  auto LargeImm = [](const APInt &Imm) {
+    unsigned Zeros = Imm.countLeadingZeros() + Imm.countTrailingZeros();
+    return Imm.getBitWidth() - Zeros > 8;
+  };
 
-  Zeros = C2Int.countLeadingZeros() + C2Int.countTrailingZeros();
-  if (C2Int.getBitWidth() - Zeros > 8)
+  if (LargeImm(C1Int) || LargeImm(C2Int))
     return SDValue();
 
   SelectionDAG &DAG = DCI.DAG;
   SDLoc dl(N);
   SDValue X = SHL.getOperand(0);
   SDValue BinOp = DAG.getNode(N->getOpcode(), dl, MVT::i32, X,
                               DAG.getConstant(C1Int, dl, MVT::i32));
   // Shift left to compensate for the lshr of C1Int.
   SDValue Res = DAG.getNode(ISD::SHL, dl, MVT::i32, BinOp, SHL.getOperand(1));
 
+  DEBUG(dbgs() << "Simplify shl use:\n"; SHL.getOperand(0).dump(); SHL.dump();
+        N->dump());
+  DEBUG(dbgs() << "Into:\n"; X.dump(); BinOp.dump(); Res.dump());
+
   DAG.ReplaceAllUsesWith(SDValue(N, 0), Res);
   return SDValue(N, 0);
 }
 
 
 /// PerformADDCombine - Target-specific dag combine xforms for ISD::ADD.
 ///
 static SDValue PerformADDCombine(SDNode *N,

test/CodeGen/ARM/pr36577.ll

+; RUN: llc -mtriple armv6t2 %s -o - | FileCheck %s
+; RUN: llc -mtriple thumbv6t2 %s -o - | FileCheck %s --check-prefix=CHECK-T2
+; RUN: llc -mtriple armv7 %s -o - | FileCheck %s
+; RUN: llc -mtriple thumbv7 %s -o - | FileCheck %s --check-prefix=CHECK-T2
+; RUN: llc -mtriple thumbv7m %s -o - | FileCheck %s --check-prefix=CHECK-T2
+; RUN: llc -mtriple thumbv8m.main %s -o - | FileCheck %s --check-prefix=CHECK-T2
+
+@a = common dso_local local_unnamed_addr global i16 0, align 2
+
+; CHECK-LABEL: pr36577
+; CHECK: ldrh r0, [r0]
+; CHECK: bic r0, r1, r0, lsr #5
+; CHECK: mvn r1, #7
+; CHECK: orr r0, r0, r1
+; CHECK-T2: ldrh r0, [r0]
+; CHECK-T2: bic.w r0, r1, r0, lsr #5
+; CHECK-T2: orn r0, r0, #7
+define dso_local arm_aapcscc i32** @pr36577() {
+entry:
+  %0 = load i16, i16* @a, align 2
+  %1 = lshr i16 %0, 7
+  %2 = and i16 %1, 1
+  %3 = zext i16 %2 to i32
+  %4 = xor i32 %3, -1
+  %add.ptr = getelementptr inbounds i32*, i32** null, i32 %4
+  ret i32** %add.ptr
+}
+