/*
 * Copyright (c) 1999, 2017, 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.
 *
 */

#ifndef OS_CPU_WINDOWS_X86_VM_ATOMIC_WINDOWS_X86_HPP
#define OS_CPU_WINDOWS_X86_VM_ATOMIC_WINDOWS_X86_HPP

#include "runtime/os.hpp"

// The following alternative implementations are needed because
// Windows 95 doesn't support (some of) the corresponding Windows NT
// calls. Furthermore, these versions allow inlining in the caller.
// (More precisely: The documentation for InterlockedExchange says
// it is supported for Windows 95. However, when single-stepping
// through the assembly code we cannot step into the routine and
// when looking at the routine address we see only garbage code.
// Better safe then sorry!). Was bug 7/31/98 (gri).
//
// Performance note: On uniprocessors, the 'lock' prefixes are not
// necessary (and expensive). We should generate separate cases if
// this becomes a performance problem.

#pragma warning(disable: 4035) // Disables warnings reporting missing return statement

#ifdef AMD64

template <>
inline int32_t Atomic::specialized_add<int32_t>(int32_t add_value, volatile int32_t* dest) {
  return (int32_t)(*os::atomic_add_func)(add_value, dest);
}

template <>
inline int64_t Atomic::specialized_add<int64_t>(int64_t add_value, volatile int64_t* dest) {
  return (intptr_t)(*os::atomic_add_ptr_func)(add_value, dest);
}

template <>
inline int32_t Atomic::specialized_xchg<int32_t>(int32_t exchange_value, volatile int32_t* dest) {
  return (int32_t)(*os::atomic_xchg_func)(exchange_value, dest);
}

template <>
inline int64_t Atomic::specialized_xchg<int64_t>(int64_t exchange_value, volatile int64_t* dest) {
  return (intptr_t)(os::atomic_xchg_ptr_func)(exchange_value, dest);
}

template <>
inline int32_t Atomic::specialized_cmpxchg<int32_t>(int32_t exchange_value, volatile int32_t* dest, int32_t compare_value, cmpxchg_memory_order order) {
  return (*os::atomic_cmpxchg_func)(exchange_value, dest, compare_value);
}

#define VM_HAS_SPECIALIZED_CMPXCHG_BYTE
template <>
inline int8_t Atomic::specialized_cmpxchg<int8_t>(int8_t exchange_value, volatile int8_t* dest, int8_t compare_value, cmpxchg_memory_order order) {
    return (*os::atomic_cmpxchg_byte_func)(exchange_value, dest, compare_value);
}

template <>
inline int64_t Atomic::specialized_cmpxchg<int64_t>(int64_t exchange_value, volatile int64_t* dest, int64_t compare_value, cmpxchg_memory_order order) {
  return (*os::atomic_cmpxchg_long_func)(exchange_value, dest, compare_value);
}

#else // !AMD64

template <>
inline int32_t Atomic::specialized_add<int32_t>(int32_t add_value, volatile int32_t* dest) {
  __asm {
    mov edx, dest;
    mov eax, add_value;
    mov ecx, eax;
    lock xadd dword ptr [edx], eax;
    add eax, ecx;
  }
}

template <>
inline void Atomic::specialized_inc<int32_t>(volatile int32_t* dest) {
  // alternative for InterlockedIncrement
  __asm {
    mov edx, dest;
    lock add dword ptr [edx], 1;
  }
}

template <>
inline void Atomic::specialized_dec<int32_t>(volatile int32_t* dest) {
  // alternative for InterlockedDecrement
  __asm {
    mov edx, dest;
    lock sub dword ptr [edx], 1;
  }
}

template <>
inline int32_t Atomic::specialized_xchg<int32_t>(int32_t exchange_value, volatile int32_t* dest) {
  // alternative for InterlockedExchange
  __asm {
    mov eax, exchange_value;
    mov ecx, dest;
    xchg eax, dword ptr [ecx];
  }
}

#define VM_HAS_SPECIALIZED_CMPXCHG_BYTE
template <>
inline int8_t Atomic::specialized_cmpxchg<int8_t>(int8_t exchange_value, volatile int8_t* dest, int8_t compare_value, cmpxchg_memory_order order) {
  // alternative for InterlockedCompareExchange
  __asm {
    mov edx, dest
    mov cl, exchange_value
    mov al, compare_value
    lock cmpxchg byte ptr [edx], cl
  }
}

template <>
inline int32_t Atomic::specialized_cmpxchg<int32_t>(int32_t exchange_value, volatile int32_t* dest, int32_t compare_value, cmpxchg_memory_order order) {
  // alternative for InterlockedCompareExchange
  __asm {
    mov edx, dest
    mov ecx, exchange_value
    mov eax, compare_value
    lock cmpxchg dword ptr [edx], ecx
  }
}

template <>
inline int64_t Atomic::specialized_cmpxchg<int64_t>(int64_t exchange_value, volatile int64_t* dest, int64_t compare_value, cmpxchg_memory_order order) {
  int32_t ex_lo  = (int32_t)exchange_value;
  int32_t ex_hi  = *( ((int32_t*)&exchange_value) + 1 );
  int32_t cmp_lo = (int32_t)compare_value;
  int32_t cmp_hi = *( ((int32_t*)&compare_value) + 1 );
  __asm {
    push ebx
    push edi
    mov eax, cmp_lo
    mov edx, cmp_hi
    mov edi, dest
    mov ebx, ex_lo
    mov ecx, ex_hi
    lock cmpxchg8b qword ptr [edi]
    pop edi
    pop ebx
  }
}

template <>
inline int64_t Atomic::specialized_load<int64_t>(const volatile int64_t* src) {
  volatile int64_t dest;
  volatile int64_t* pdest = &dest;
  __asm {
    mov eax, src
    fild     qword ptr [eax]
    mov eax, pdest
    fistp    qword ptr [eax]
  }
  return dest;
}

template <>
inline void Atomic::specialized_store<int64_t>(int64_t store_value, volatile int64_t* dest) {
  volatile int64_t* src = &store_value;
  __asm {
    mov eax, src
    fild     qword ptr [eax]
    mov eax, dest
    fistp    qword ptr [eax]
  }
}

#endif // AMD64

#pragma warning(default: 4035) // Enables warnings reporting missing return statement

#endif // OS_CPU_WINDOWS_X86_VM_ATOMIC_WINDOWS_X86_HPP