ART: Make InstructionSet an enum class and add kLast.

Adding InstructionSet::kLast shall make it easier to encode
the InstructionSet in fewer bits using BitField<>. However,
introducing `kLast` into the `art` namespace is not a good
idea, so we change the InstructionSet to an enum class.
This also uncovered a case of InstructionSet::kNone being
erroneously used instead of vixl32::Condition::None(), so
it's good to remove `kNone` from the `art` namespace.

Test: m test-art-host-gtest
Test: testrunner.py --host --optimizing
Change-Id: I6fa6168dfba4ed6da86d021a69c80224f09997a6
diff --git a/compiler/optimizing/instruction_simplifier_arm.cc b/compiler/optimizing/instruction_simplifier_arm.cc
index 9422f9f..d41e49a 100644
--- a/compiler/optimizing/instruction_simplifier_arm.cc
+++ b/compiler/optimizing/instruction_simplifier_arm.cc
@@ -84,7 +84,7 @@
 bool InstructionSimplifierArmVisitor::TryMergeIntoShifterOperand(HInstruction* use,
                                                                  HInstruction* bitfield_op,
                                                                  bool do_merge) {
-  DCHECK(HasShifterOperand(use, kArm));
+  DCHECK(HasShifterOperand(use, InstructionSet::kArm));
   DCHECK(use->IsBinaryOperation());
   DCHECK(CanFitInShifterOperand(bitfield_op));
   DCHECK(!bitfield_op->HasEnvironmentUses());
@@ -166,7 +166,7 @@
   // Check whether we can merge the instruction in all its users' shifter operand.
   for (const HUseListNode<HInstruction*>& use : uses) {
     HInstruction* user = use.GetUser();
-    if (!HasShifterOperand(user, kArm)) {
+    if (!HasShifterOperand(user, InstructionSet::kArm)) {
       return false;
     }
     if (!CanMergeIntoShifterOperand(user, bitfield_op)) {
@@ -242,7 +242,7 @@
 }
 
 void InstructionSimplifierArmVisitor::VisitMul(HMul* instruction) {
-  if (TryCombineMultiplyAccumulate(instruction, kArm)) {
+  if (TryCombineMultiplyAccumulate(instruction, InstructionSet::kArm)) {
     RecordSimplification();
   }
 }