Remove ArtMethod* parameter from dex cache entry points.

Load the ArtMethod* using an optimized stack walk instead.
This reduces the size of the generated code.

Three of the entry points are called only from a slow-path
and the fourth (InitializeTypeAndVerifyAccess) is rare and
already slow enough that the one or two extra loads
(depending on whether we already have the ArtMethod* in a
register) are insignificant. And as we're starting to use
PC-relative addressing of the dex cache arrays (already
done by Quick for the boot image), having the ArtMethod* in
a register becomes less likely anyway.

Change-Id: Ib19b9d204e355e13bf386662a8b158178bf8ad28
diff --git a/compiler/optimizing/code_generator_arm.cc b/compiler/optimizing/code_generator_arm.cc
index ae1fb53..8589f94 100644
--- a/compiler/optimizing/code_generator_arm.cc
+++ b/compiler/optimizing/code_generator_arm.cc
@@ -176,7 +176,6 @@
 
     InvokeRuntimeCallingConvention calling_convention;
     __ LoadImmediate(calling_convention.GetRegisterAt(0), cls_->GetTypeIndex());
-    arm_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1));
     int32_t entry_point_offset = do_clinit_
         ? QUICK_ENTRY_POINT(pInitializeStaticStorage)
         : QUICK_ENTRY_POINT(pInitializeType);
@@ -222,7 +221,6 @@
     SaveLiveRegisters(codegen, locations);
 
     InvokeRuntimeCallingConvention calling_convention;
-    arm_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1));
     __ LoadImmediate(calling_convention.GetRegisterAt(0), instruction_->GetStringIndex());
     arm_codegen->InvokeRuntime(
         QUICK_ENTRY_POINT(pResolveString), instruction_, instruction_->GetDexPc(), this);
diff --git a/compiler/optimizing/code_generator_arm64.cc b/compiler/optimizing/code_generator_arm64.cc
index 1c6debd..2a2f07f 100644
--- a/compiler/optimizing/code_generator_arm64.cc
+++ b/compiler/optimizing/code_generator_arm64.cc
@@ -173,14 +173,13 @@
 
     InvokeRuntimeCallingConvention calling_convention;
     __ Mov(calling_convention.GetRegisterAt(0).W(), cls_->GetTypeIndex());
-    arm64_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1).W());
     int32_t entry_point_offset = do_clinit_ ? QUICK_ENTRY_POINT(pInitializeStaticStorage)
                                             : QUICK_ENTRY_POINT(pInitializeType);
     arm64_codegen->InvokeRuntime(entry_point_offset, at_, dex_pc_, this);
     if (do_clinit_) {
-      CheckEntrypointTypes<kQuickInitializeStaticStorage, void*, uint32_t, mirror::ArtMethod*>();
+      CheckEntrypointTypes<kQuickInitializeStaticStorage, void*, uint32_t>();
     } else {
-      CheckEntrypointTypes<kQuickInitializeType, void*, uint32_t, mirror::ArtMethod*>();
+      CheckEntrypointTypes<kQuickInitializeType, void*, uint32_t>();
     }
 
     // Move the class to the desired location.
@@ -225,11 +224,10 @@
     SaveLiveRegisters(codegen, locations);
 
     InvokeRuntimeCallingConvention calling_convention;
-    arm64_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1).W());
     __ Mov(calling_convention.GetRegisterAt(0).W(), instruction_->GetStringIndex());
     arm64_codegen->InvokeRuntime(
         QUICK_ENTRY_POINT(pResolveString), instruction_, instruction_->GetDexPc(), this);
-    CheckEntrypointTypes<kQuickResolveString, void*, uint32_t, mirror::ArtMethod*>();
+    CheckEntrypointTypes<kQuickResolveString, void*, uint32_t>();
     Primitive::Type type = instruction_->GetType();
     arm64_codegen->MoveLocation(locations->Out(), calling_convention.GetReturnLocation(type), type);
 
diff --git a/compiler/optimizing/code_generator_x86.cc b/compiler/optimizing/code_generator_x86.cc
index c604842..021e2eb 100644
--- a/compiler/optimizing/code_generator_x86.cc
+++ b/compiler/optimizing/code_generator_x86.cc
@@ -174,7 +174,6 @@
     SaveLiveRegisters(codegen, locations);
 
     InvokeRuntimeCallingConvention calling_convention;
-    x86_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1));
     __ movl(calling_convention.GetRegisterAt(0), Immediate(instruction_->GetStringIndex()));
     __ fs()->call(Address::Absolute(QUICK_ENTRYPOINT_OFFSET(kX86WordSize, pResolveString)));
     RecordPcInfo(codegen, instruction_, instruction_->GetDexPc());
@@ -208,7 +207,6 @@
 
     InvokeRuntimeCallingConvention calling_convention;
     __ movl(calling_convention.GetRegisterAt(0), Immediate(cls_->GetTypeIndex()));
-    x86_codegen->LoadCurrentMethod(calling_convention.GetRegisterAt(1));
     __ fs()->call(Address::Absolute(do_clinit_
         ? QUICK_ENTRYPOINT_OFFSET(kX86WordSize, pInitializeStaticStorage)
         : QUICK_ENTRYPOINT_OFFSET(kX86WordSize, pInitializeType)));
diff --git a/compiler/optimizing/code_generator_x86_64.cc b/compiler/optimizing/code_generator_x86_64.cc
index 47425fb..0c97a6e 100644
--- a/compiler/optimizing/code_generator_x86_64.cc
+++ b/compiler/optimizing/code_generator_x86_64.cc
@@ -197,7 +197,6 @@
 
     InvokeRuntimeCallingConvention calling_convention;
     __ movl(CpuRegister(calling_convention.GetRegisterAt(0)), Immediate(cls_->GetTypeIndex()));
-    x64_codegen->LoadCurrentMethod(CpuRegister(calling_convention.GetRegisterAt(1)));
     __ gs()->call(Address::Absolute((do_clinit_
           ? QUICK_ENTRYPOINT_OFFSET(kX86_64WordSize, pInitializeStaticStorage)
           : QUICK_ENTRYPOINT_OFFSET(kX86_64WordSize, pInitializeType)) , true));
@@ -244,7 +243,6 @@
     SaveLiveRegisters(codegen, locations);
 
     InvokeRuntimeCallingConvention calling_convention;
-    x64_codegen->LoadCurrentMethod(CpuRegister(calling_convention.GetRegisterAt(1)));
     __ movl(CpuRegister(calling_convention.GetRegisterAt(0)),
             Immediate(instruction_->GetStringIndex()));
     __ gs()->call(Address::Absolute(