MIPS32: Improve method invocation

Improvements include:
- CodeGeneratorMIPS::GenerateStaticOrDirectCall() supports:
  - MethodLoadKind::kDirectAddressWithFixup (via literals)
  - CodePtrLocation::kCallDirectWithFixup (via literals)
  - MethodLoadKind::kDexCachePcRelative
- 32-bit literals to support the above (not ready for general-
  purpose applications yet because RA is not saved in leaf
  methods, but is clobbered on MIPS32R2 when simulating
  PC-relative addressing (MIPS32R6 is OK because it has
  PC-relative addressing with the lwpc instruction))
- shorter instruction sequences for recursive static/direct
  calls

Tested:
- test-art-host-gtest
- test-art-target-gtest and test-art-target-run-test-optimizing on:
  - MIPS32R2 QEMU
  - CI20 board
  - MIPS32R6 (2nd arch) QEMU

Change-Id: Id5b137ad32d5590487fd154c9a01d3b3e7e044ff
diff --git a/compiler/optimizing/pc_relative_fixups_mips.cc b/compiler/optimizing/pc_relative_fixups_mips.cc
new file mode 100644
index 0000000..ba405cd
--- /dev/null
+++ b/compiler/optimizing/pc_relative_fixups_mips.cc
@@ -0,0 +1,133 @@
+/*
+ * Copyright (C) 2016 The Android Open Source Project
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include "pc_relative_fixups_mips.h"
+#include "code_generator_mips.h"
+#include "intrinsics_mips.h"
+
+namespace art {
+namespace mips {
+
+/**
+ * Finds instructions that need the constant area base as an input.
+ */
+class PCRelativeHandlerVisitor : public HGraphVisitor {
+ public:
+  PCRelativeHandlerVisitor(HGraph* graph, CodeGenerator* codegen)
+      : HGraphVisitor(graph),
+        codegen_(down_cast<CodeGeneratorMIPS*>(codegen)),
+        base_(nullptr) {}
+
+  void MoveBaseIfNeeded() {
+    if (base_ != nullptr) {
+      // Bring the base closer to the first use (previously, it was in the
+      // entry block) and relieve some pressure on the register allocator
+      // while avoiding recalculation of the base in a loop.
+      base_->MoveBeforeFirstUserAndOutOfLoops();
+    }
+  }
+
+ private:
+  void VisitInvokeStaticOrDirect(HInvokeStaticOrDirect* invoke) OVERRIDE {
+    HandleInvoke(invoke);
+  }
+
+  void InitializePCRelativeBasePointer() {
+    // Ensure we only initialize the pointer once.
+    if (base_ != nullptr) {
+      return;
+    }
+    // Insert the base at the start of the entry block, move it to a better
+    // position later in MoveBaseIfNeeded().
+    base_ = new (GetGraph()->GetArena()) HMipsComputeBaseMethodAddress();
+    HBasicBlock* entry_block = GetGraph()->GetEntryBlock();
+    entry_block->InsertInstructionBefore(base_, entry_block->GetFirstInstruction());
+    DCHECK(base_ != nullptr);
+  }
+
+  void HandleInvoke(HInvoke* invoke) {
+    // If this is an invoke-static/-direct with PC-relative dex cache array
+    // addressing, we need the PC-relative address base.
+    HInvokeStaticOrDirect* invoke_static_or_direct = invoke->AsInvokeStaticOrDirect();
+    if (invoke_static_or_direct != nullptr) {
+      HInvokeStaticOrDirect::MethodLoadKind method_load_kind =
+          invoke_static_or_direct->GetMethodLoadKind();
+      HInvokeStaticOrDirect::CodePtrLocation code_ptr_location =
+          invoke_static_or_direct->GetCodePtrLocation();
+
+      bool has_extra_input =
+          (method_load_kind == HInvokeStaticOrDirect::MethodLoadKind::kDirectAddressWithFixup) ||
+          (code_ptr_location == HInvokeStaticOrDirect::CodePtrLocation::kCallDirectWithFixup);
+
+      // We can't add a pointer to the constant area if we already have a current
+      // method pointer. This may arise when sharpening doesn't remove the current
+      // method pointer from the invoke.
+      if (invoke_static_or_direct->HasCurrentMethodInput()) {
+        DCHECK(!invoke_static_or_direct->HasPcRelativeDexCache());
+        CHECK(!has_extra_input);  // TODO: review this.
+        return;
+      }
+
+      if (has_extra_input && !WillHaveCallFreeIntrinsicsCodeGen(invoke)) {
+        InitializePCRelativeBasePointer();
+        // Add the extra parameter base_.
+        invoke_static_or_direct->AddSpecialInput(base_);
+      }
+    }
+  }
+
+  bool WillHaveCallFreeIntrinsicsCodeGen(HInvoke* invoke) {
+    if (invoke->GetIntrinsic() != Intrinsics::kNone) {
+      // This invoke may have intrinsic code generation defined. However, we must
+      // now also determine if this code generation is truly there and call-free
+      // (not unimplemented, no bail on instruction features, or call on slow path).
+      // This is done by actually calling the locations builder on the instruction
+      // and clearing out the locations once result is known. We assume this
+      // call only has creating locations as side effects!
+      IntrinsicLocationsBuilderMIPS builder(codegen_);
+      bool success = builder.TryDispatch(invoke) && !invoke->GetLocations()->CanCall();
+      invoke->SetLocations(nullptr);
+      return success;
+    }
+    return false;
+  }
+
+  CodeGeneratorMIPS* codegen_;
+
+  // The generated HMipsComputeBaseMethodAddress in the entry block needed as an
+  // input to the HMipsLoadFromConstantTable instructions.
+  HMipsComputeBaseMethodAddress* base_;
+};
+
+void PcRelativeFixups::Run() {
+  CodeGeneratorMIPS* mips_codegen = down_cast<CodeGeneratorMIPS*>(codegen_);
+  if (mips_codegen->GetInstructionSetFeatures().IsR6()) {
+    // Do nothing for R6 because it has PC-relative addressing.
+    // TODO: review. Move this check into RunArchOptimizations()?
+    return;
+  }
+  if (graph_->HasIrreducibleLoops()) {
+    // Do not run this optimization, as irreducible loops do not work with an instruction
+    // that can be live-in at the irreducible loop header.
+    return;
+  }
+  PCRelativeHandlerVisitor visitor(graph_, codegen_);
+  visitor.VisitInsertionOrder();
+  visitor.MoveBaseIfNeeded();
+}
+
+}  // namespace mips
+}  // namespace art