/*

 * Copyright (c) 2003, 2013, Oracle and/or its affiliates. All rights reserved.

 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.

 *

 * This code is free software; you can redistribute it and/or modify it

 * under the terms of the GNU General Public License version 2 only, as

 * published by the Free Software Foundation.

 *

 * This code is distributed in the hope that it will be useful, but WITHOUT

 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or

 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License

 * version 2 for more details (a copy is included in the LICENSE file that

 * accompanied this code).

 *

 * You should have received a copy of the GNU General Public License version

 * 2 along with this work; if not, write to the Free Software Foundation,

 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.

 *

 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA

 * or visit www.oracle.com if you need additional information or have any

 * questions.

 *

 */

#include "precompiled.hpp"

#include "interp_masm_x86.hpp"

#include "interpreter/interpreter.hpp"

#include "interpreter/interpreterRuntime.hpp"

#include "oops/arrayOop.hpp"

#include "oops/markOop.hpp"

#include "oops/methodData.hpp"

#include "oops/method.hpp"

#include "prims/jvmtiExport.hpp"

#include "prims/jvmtiRedefineClassesTrace.hpp"

#include "prims/jvmtiThreadState.hpp"

#include "runtime/basicLock.hpp"

#include "runtime/biasedLocking.hpp"

#include "runtime/sharedRuntime.hpp"

#include "runtime/thread.inline.hpp"

// Implementation of InterpreterMacroAssembler

#ifdef CC_INTERP

void InterpreterMacroAssembler::get_method(Register reg) {

  movptr(reg, Address(rbp, -((int)sizeof(BytecodeInterpreter) + 2 * wordSize)));

  movptr(reg, Address(reg, byte_offset_of(BytecodeInterpreter, _method)));

}

#endif // CC_INTERP

#ifndef CC_INTERP

void InterpreterMacroAssembler::call_VM_leaf_base(address entry_point,

                                                  int number_of_arguments) {

  // interpreter specific

  //

  // Note: No need to save/restore bcp & locals (r13 & r14) pointer

  //       since these are callee saved registers and no blocking/

  //       GC can happen in leaf calls.

  // Further Note: DO NOT save/restore bcp/locals. If a caller has

  // already saved them so that it can use esi/edi as temporaries

  // then a save/restore here will DESTROY the copy the caller

  // saved! There used to be a save_bcp() that only happened in

  // the ASSERT path (no restore_bcp). Which caused bizarre failures

  // when jvm built with ASSERTs.

#ifdef ASSERT

  {

    Label L;

    cmpptr(Address(rbp, frame::interpreter_frame_last_sp_offset * wordSize), (int32_t)NULL_WORD);

    jcc(Assembler::equal, L);

    stop("InterpreterMacroAssembler::call_VM_leaf_base:"

         " last_sp != NULL");

    bind(L);

  }

#endif

  // super call

  MacroAssembler::call_VM_leaf_base(entry_point, number_of_arguments);

  // interpreter specific

  // Used to ASSERT that r13/r14 were equal to frame's bcp/locals

  // but since they may not have been saved (and we don't want to

  // save thme here (see note above) the assert is invalid.

}

void InterpreterMacroAssembler::call_VM_base(Register oop_result,

                                             Register java_thread,

                                             Register last_java_sp,

                                             address  entry_point,

                                             int      number_of_arguments,

                                             bool     check_exceptions) {

  // interpreter specific

  //

  // Note: Could avoid restoring locals ptr (callee saved) - however doesn't

  //       really make a difference for these runtime calls, since they are

  //       slow anyway. Btw., bcp must be saved/restored since it may change

  //       due to GC.

  // assert(java_thread == noreg , "not expecting a precomputed java thread");

  save_bcp();

#ifdef ASSERT

  {

    Label L;

    cmpptr(Address(rbp, frame::interpreter_frame_last_sp_offset * wordSize), (int32_t)NULL_WORD);

    jcc(Assembler::equal, L);

    stop("InterpreterMacroAssembler::call_VM_leaf_base:"

         " last_sp != NULL");

    bind(L);

  }

#endif /* ASSERT */

  // super call

  MacroAssembler::call_VM_base(oop_result, noreg, last_java_sp,

                               entry_point, number_of_arguments,

                               check_exceptions);

  // interpreter specific

  restore_bcp();

  restore_locals();

}

void InterpreterMacroAssembler::check_and_handle_popframe(Register java_thread) {

  if (JvmtiExport::can_pop_frame()) {

    Label L;

    // Initiate popframe handling only if it is not already being

    // processed.  If the flag has the popframe_processing bit set, it

    // means that this code is called *during* popframe handling - we

    // don't want to reenter.

    // This method is only called just after the call into the vm in

    // call_VM_base, so the arg registers are available.

    movl(c_rarg0, Address(r15_thread, JavaThread::popframe_condition_offset()));

    testl(c_rarg0, JavaThread::popframe_pending_bit);

    jcc(Assembler::zero, L);

    testl(c_rarg0, JavaThread::popframe_processing_bit);

    jcc(Assembler::notZero, L);

    // Call Interpreter::remove_activation_preserving_args_entry() to get the

    // address of the same-named entrypoint in the generated interpreter code.

    call_VM_leaf(CAST_FROM_FN_PTR(address, Interpreter::remove_activation_preserving_args_entry));

    jmp(rax);

    bind(L);

  }

}

void InterpreterMacroAssembler::load_earlyret_value(TosState state) {

  movptr(rcx, Address(r15_thread, JavaThread::jvmti_thread_state_offset()));

  const Address tos_addr(rcx, JvmtiThreadState::earlyret_tos_offset());

  const Address oop_addr(rcx, JvmtiThreadState::earlyret_oop_offset());

  const Address val_addr(rcx, JvmtiThreadState::earlyret_value_offset());

  switch (state) {

    case atos: movptr(rax, oop_addr);

               movptr(oop_addr, (int32_t)NULL_WORD);

               verify_oop(rax, state);              break;

    case ltos: movptr(rax, val_addr);                 break;

    case btos:                                   // fall through

    case ctos:                                   // fall through

    case stos:                                   // fall through

    case itos: movl(rax, val_addr);                 break;

    case ftos: movflt(xmm0, val_addr);              break;

    case dtos: movdbl(xmm0, val_addr);              break;

    case vtos: /* nothing to do */                  break;

    default  : ShouldNotReachHere();

  }

  // Clean up tos value in the thread object

  movl(tos_addr,  (int) ilgl);

  movl(val_addr,  (int32_t) NULL_WORD);

}

void InterpreterMacroAssembler::check_and_handle_earlyret(Register java_thread) {

  if (JvmtiExport::can_force_early_return()) {

    Label L;

    movptr(c_rarg0, Address(r15_thread, JavaThread::jvmti_thread_state_offset()));

    testptr(c_rarg0, c_rarg0);

    jcc(Assembler::zero, L); // if (thread->jvmti_thread_state() == NULL) exit;

    // Initiate earlyret handling only if it is not already being processed.

    // If the flag has the earlyret_processing bit set, it means that this code

    // is called *during* earlyret handling - we don't want to reenter.

    movl(c_rarg0, Address(c_rarg0, JvmtiThreadState::earlyret_state_offset()));

    cmpl(c_rarg0, JvmtiThreadState::earlyret_pending);

    jcc(Assembler::notEqual, L);

    // Call Interpreter::remove_activation_early_entry() to get the address of the

    // same-named entrypoint in the generated interpreter code.

    movptr(c_rarg0, Address(r15_thread, JavaThread::jvmti_thread_state_offset()));

    movl(c_rarg0, Address(c_rarg0, JvmtiThreadState::earlyret_tos_offset()));

    call_VM_leaf(CAST_FROM_FN_PTR(address, Interpreter::remove_activation_early_entry), c_rarg0);

    jmp(rax);

    bind(L);

  }

}

void InterpreterMacroAssembler::get_unsigned_2_byte_index_at_bcp(

  Register reg,

  int bcp_offset) {

  assert(bcp_offset >= 0, "bcp is still pointing to start of bytecode");

  load_unsigned_short(reg, Address(r13, bcp_offset));

  bswapl(reg);

  shrl(reg, 16);

}

void InterpreterMacroAssembler::get_cache_index_at_bcp(Register index,

                                                       int bcp_offset,

                                                       size_t index_size) {

  assert(bcp_offset > 0, "bcp is still pointing to start of bytecode");

  if (index_size == sizeof(u2)) {

    load_unsigned_short(index, Address(r13, bcp_offset));

  } else if (index_size == sizeof(u4)) {

    assert(EnableInvokeDynamic, "giant index used only for JSR 292");

    movl(index, Address(r13, bcp_offset));

    // Check if the secondary index definition is still ~x, otherwise

    // we have to change the following assembler code to calculate the

    // plain index.

    assert(ConstantPool::decode_invokedynamic_index(~123) == 123, "else change next line");

    notl(index);  // convert to plain index

  } else if (index_size == sizeof(u1)) {

    load_unsigned_byte(index, Address(r13, bcp_offset));

  } else {

    ShouldNotReachHere();

  }

}

void InterpreterMacroAssembler::get_cache_and_index_at_bcp(Register cache,

                                                           Register index,

                                                           int bcp_offset,

                                                           size_t index_size) {

  assert_different_registers(cache, index);

  get_cache_index_at_bcp(index, bcp_offset, index_size);

  movptr(cache, Address(rbp, frame::interpreter_frame_cache_offset * wordSize));

  assert(sizeof(ConstantPoolCacheEntry) == 4 * wordSize, "adjust code below");

  // convert from field index to ConstantPoolCacheEntry index

  assert(exact_log2(in_words(ConstantPoolCacheEntry::size())) == 2, "else change next line");

  shll(index, 2);

}

void InterpreterMacroAssembler::get_cache_and_index_and_bytecode_at_bcp(Register cache,

                                                                        Register index,

                                                                        Register bytecode,

                                                                        int byte_no,

                                                                        int bcp_offset,

                                                                        size_t index_size) {

  get_cache_and_index_at_bcp(cache, index, bcp_offset, index_size);

  // We use a 32-bit load here since the layout of 64-bit words on

  // little-endian machines allow us that.

  movl(bytecode, Address(cache, index, Address::times_ptr, ConstantPoolCache::base_offset() + ConstantPoolCacheEntry::indices_offset()));

  const int shift_count = (1 + byte_no) * BitsPerByte;

  assert((byte_no == TemplateTable::f1_byte && shift_count == ConstantPoolCacheEntry::bytecode_1_shift) ||

         (byte_no == TemplateTable::f2_byte && shift_count == ConstantPoolCacheEntry::bytecode_2_shift),

         "correct shift count");

  shrl(bytecode, shift_count);

  assert(ConstantPoolCacheEntry::bytecode_1_mask == ConstantPoolCacheEntry::bytecode_2_mask, "common mask");

  andl(bytecode, ConstantPoolCacheEntry::bytecode_1_mask);

}

void InterpreterMacroAssembler::get_cache_entry_pointer_at_bcp(Register cache,

                                                               Register tmp,

                                                               int bcp_offset,

                                                               size_t index_size) {

  assert(cache != tmp, "must use different register");

  get_cache_index_at_bcp(tmp, bcp_offset, index_size);

  assert(sizeof(ConstantPoolCacheEntry) == 4 * wordSize, "adjust code below");

  // convert from field index to ConstantPoolCacheEntry index

  // and from word offset to byte offset

  assert(exact_log2(in_bytes(ConstantPoolCacheEntry::size_in_bytes())) == 2 + LogBytesPerWord, "else change next line");

  shll(tmp, 2 + LogBytesPerWord);

  movptr(cache, Address(rbp, frame::interpreter_frame_cache_offset * wordSize));

  // skip past the header

  addptr(cache, in_bytes(ConstantPoolCache::base_offset()));

  addptr(cache, tmp);  // construct pointer to cache entry

}

// Load object from cpool->resolved_references(index)

void InterpreterMacroAssembler::load_resolved_reference_at_index(

                                           Register result, Register index) {

  assert_different_registers(result, index);

  // convert from field index to resolved_references() index and from

  // word index to byte offset. Since this is a java object, it can be compressed

  Register tmp = index;  // reuse

  shll(tmp, LogBytesPerHeapOop);

  get_constant_pool(result);

  // load pointer for resolved_references[] objArray

  movptr(result, Address(result, ConstantPool::resolved_references_offset_in_bytes()));

  // JNIHandles::resolve(obj);

  movptr(result, Address(result, 0));

  // Add in the index

  addptr(result, tmp);

  load_heap_oop(result, Address(result, arrayOopDesc::base_offset_in_bytes(T_OBJECT)));

}

// Generate a subtype check: branch to ok_is_subtype if sub_klass is a

// subtype of super_klass.

//

// Args:

//      rax: superklass

//      Rsub_klass: subklass

//

// Kills:

//      rcx, rdi

void InterpreterMacroAssembler::gen_subtype_check(Register Rsub_klass,

                                                  Label& ok_is_subtype) {

  assert(Rsub_klass != rax, "rax holds superklass");

  assert(Rsub_klass != r14, "r14 holds locals");

  assert(Rsub_klass != r13, "r13 holds bcp");

  assert(Rsub_klass != rcx, "rcx holds 2ndary super array length");

  assert(Rsub_klass != rdi, "rdi holds 2ndary super array scan ptr");

  // Profile the not-null value's klass.

  profile_typecheck(rcx, Rsub_klass, rdi); // blows rcx, reloads rdi

  // Do the check.

  check_klass_subtype(Rsub_klass, rax, rcx, ok_is_subtype); // blows rcx

  // Profile the failure of the check.

  profile_typecheck_failed(rcx); // blows rcx

}

// Java Expression Stack

void InterpreterMacroAssembler::pop_ptr(Register r) {

  pop(r);

}

void InterpreterMacroAssembler::pop_i(Register r) {

  // XXX can't use pop currently, upper half non clean

  movl(r, Address(rsp, 0));

  addptr(rsp, wordSize);

}

void InterpreterMacroAssembler::pop_l(Register r) {

  movq(r, Address(rsp, 0));

  addptr(rsp, 2 * Interpreter::stackElementSize);

}

void InterpreterMacroAssembler::pop_f(XMMRegister r) {

  movflt(r, Address(rsp, 0));

  addptr(rsp, wordSize);

}

void InterpreterMacroAssembler::pop_d(XMMRegister r) {

  movdbl(r, Address(rsp, 0));

  addptr(rsp, 2 * Interpreter::stackElementSize);

}

void InterpreterMacroAssembler::push_ptr(Register r) {

  push(r);

}

void InterpreterMacroAssembler::push_i(Register r) {

  push(r);

}

void InterpreterMacroAssembler::push_l(Register r) {

  subptr(rsp, 2 * wordSize);

  movq(Address(rsp, 0), r);

}

void InterpreterMacroAssembler::push_f(XMMRegister r) {

  subptr(rsp, wordSize);

  movflt(Address(rsp, 0), r);

}

void InterpreterMacroAssembler::push_d(XMMRegister r) {

  subptr(rsp, 2 * wordSize);

  movdbl(Address(rsp, 0), r);

}

void InterpreterMacroAssembler::pop(TosState state) {

  switch (state) {

  case atos: pop_ptr();                 break;

  case btos:

  case ctos:

  case stos:

  case itos: pop_i();                   break;

  case ltos: pop_l();                   break;

  case ftos: pop_f();                   break;

  case dtos: pop_d();                   break;

  case vtos: /* nothing to do */        break;

  default:   ShouldNotReachHere();

  }

  verify_oop(rax, state);

}

void InterpreterMacroAssembler::push(TosState state) {

  verify_oop(rax, state);

  switch (state) {

  case atos: push_ptr();                break;

  case btos:

  case ctos:

  case stos:

  case itos: push_i();                  break;

  case ltos: push_l();                  break;

  case ftos: push_f();                  break;

  case dtos: push_d();                  break;

  case vtos: /* nothing to do */        break;

  default  : ShouldNotReachHere();

  }

}

// Helpers for swap and dup

void InterpreterMacroAssembler::load_ptr(int n, Register val) {

  movptr(val, Address(rsp, Interpreter::expr_offset_in_bytes(n)));

}

void InterpreterMacroAssembler::store_ptr(int n, Register val) {

  movptr(Address(rsp, Interpreter::expr_offset_in_bytes(n)), val);

}

void InterpreterMacroAssembler::prepare_to_jump_from_interpreted() {

  // set sender sp

  lea(r13, Address(rsp, wordSize));

  // record last_sp

  movptr(Address(rbp, frame::interpreter_frame_last_sp_offset * wordSize), r13);

}

// Jump to from_interpreted entry of a call unless single stepping is possible

// in this thread in which case we must call the i2i entry

void InterpreterMacroAssembler::jump_from_interpreted(Register method, Register temp) {

  prepare_to_jump_from_interpreted();

  if (JvmtiExport::can_post_interpreter_events()) {

    Label run_compiled_code;

    // JVMTI events, such as single-stepping, are implemented partly by avoiding running

    // compiled code in threads for which the event is enabled.  Check here for

    // interp_only_mode if these events CAN be enabled.

    // interp_only is an int, on little endian it is sufficient to test the byte only

    // Is a cmpl faster?

    cmpb(Address(r15_thread, JavaThread::interp_only_mode_offset()), 0);

    jccb(Assembler::zero, run_compiled_code);

    jmp(Address(method, Method::interpreter_entry_offset()));

    bind(run_compiled_code);

  }

  jmp(Address(method, Method::from_interpreted_offset()));

}

// The following two routines provide a hook so that an implementation

// can schedule the dispatch in two parts.  amd64 does not do this.

void InterpreterMacroAssembler::dispatch_prolog(TosState state, int step) {

  // Nothing amd64 specific to be done here

}

void InterpreterMacroAssembler::dispatch_epilog(TosState state, int step) {

  dispatch_next(state, step);

}

void InterpreterMacroAssembler::dispatch_base(TosState state,

                                              address* table,

                                              bool verifyoop) {

  verify_FPU(1, state);

  if (VerifyActivationFrameSize) {

    Label L;

    mov(rcx, rbp);

    subptr(rcx, rsp);

    int32_t min_frame_size =

      (frame::link_offset - frame::interpreter_frame_initial_sp_offset) *

      wordSize;

    cmpptr(rcx, (int32_t)min_frame_size);

    jcc(Assembler::greaterEqual, L);

    stop("broken stack frame");

    bind(L);

  }

  if (verifyoop) {

    verify_oop(rax, state);

  }

  lea(rscratch1, ExternalAddress((address)table));

  jmp(Address(rscratch1, rbx, Address::times_8));

}

void InterpreterMacroAssembler::dispatch_only(TosState state) {

  dispatch_base(state, Interpreter::dispatch_table(state));

}

void InterpreterMacroAssembler::dispatch_only_normal(TosState state) {

  dispatch_base(state, Interpreter::normal_table(state));

}

void InterpreterMacroAssembler::dispatch_only_noverify(TosState state) {

  dispatch_base(state, Interpreter::normal_table(state), false);

}

void InterpreterMacroAssembler::dispatch_next(TosState state, int step) {

  // load next bytecode (load before advancing r13 to prevent AGI)

  load_unsigned_byte(rbx, Address(r13, step));

  // advance r13

  increment(r13, step);

  dispatch_base(state, Interpreter::dispatch_table(state));

}

void InterpreterMacroAssembler::dispatch_via(TosState state, address* table) {

  // load current bytecode

  load_unsigned_byte(rbx, Address(r13, 0));

  dispatch_base(state, table);

}

// remove activation

//

// Unlock the receiver if this is a synchronized method.

// Unlock any Java monitors from syncronized blocks.