MIPS: Use sltiu instead of LoadConst32() + sltu
Bltu is sltu + bnez so we can use sltiu + bnez without
loading constant (if it is 16-bit constant).
Additionally, in VisitInvokeInterface() LoadConst32() is moved
to before Jalr(T9) so the load can be taken into the delay slot.
Test: ./testrunner.py --target --optimizing in QEMU
Change-Id: Ic19f251aeba015be38b7d3690e78b2fe59e7c5ae
diff --git a/compiler/optimizing/code_generator_mips.cc b/compiler/optimizing/code_generator_mips.cc
index d01b895..87e6d68 100644
--- a/compiler/optimizing/code_generator_mips.cc
+++ b/compiler/optimizing/code_generator_mips.cc
@@ -1929,8 +1929,8 @@
enum_cast<uint32_t>(ClassStatus::kInitialized) << (status_lsb_position % kBitsPerByte);
__ LoadFromOffset(kLoadUnsignedByte, TMP, class_reg, status_byte_offset);
- __ LoadConst32(AT, shifted_initialized_value);
- __ Bltu(TMP, AT, slow_path->GetEntryLabel());
+ __ Sltiu(TMP, TMP, shifted_initialized_value);
+ __ Bnez(TMP, slow_path->GetEntryLabel());
// Even if the initialized flag is set, we need to ensure consistent memory ordering.
__ Sync(0);
__ Bind(slow_path->GetExitLabel());
@@ -7635,10 +7635,6 @@
uint32_t class_offset = mirror::Object::ClassOffset().Int32Value();
Offset entry_point = ArtMethod::EntryPointFromQuickCompiledCodeOffset(kMipsPointerSize);
- // Set the hidden argument.
- __ LoadConst32(invoke->GetLocations()->GetTemp(1).AsRegister<Register>(),
- invoke->GetDexMethodIndex());
-
// temp = object->GetClass();
if (receiver.IsStackSlot()) {
__ LoadFromOffset(kLoadWord, temp, SP, receiver.GetStackIndex());
@@ -7663,6 +7659,9 @@
__ LoadFromOffset(kLoadWord, temp, temp, method_offset);
// T9 = temp->GetEntryPoint();
__ LoadFromOffset(kLoadWord, T9, temp, entry_point.Int32Value());
+ // Set the hidden argument.
+ __ LoadConst32(invoke->GetLocations()->GetTemp(1).AsRegister<Register>(),
+ invoke->GetDexMethodIndex());
// T9();
__ Jalr(T9);
__ NopIfNoReordering();