module atomics

// add_i32 atomically adds delta to the value at dest and returns the new value.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn add_i32(dest &i32, delta i32) i32 {
    mut result := i32(0)
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, delta
        lock xadd [edx], eax
        add eax, delta
        mov result, eax
        2:
        ; =r (result)
        ; r (dest)
          r (delta)
        ; eax
          edx
          memory
    }
    return result
}

// swap_i32 atomically stores new value at dest and returns the old value.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn swap_i32(dest &i32, new i32) i32 {
    mut old := i32(0)
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, new
        xchg [edx], eax
        mov old, eax
        2:
        ; =m (old)
        ; r (dest)
          r (new)
        ; eax
          edx
          memory
    }
    return old
}

// store_i32 atomically stores value at dest.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn store_i32(dest &i32, value i32) {
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, value
        xchg eax, [edx]
        2:
        ; ; r (dest)
          r (value)
        ; eax
          edx
          memory
    }
}

// load_i32 atomically loads and returns the value at num.
// The operation is performed with sequential consistency.
// Panics if num is not 4-byte aligned.
pub fn load_i32(num &i32) i32 {
    mut out := i32(0)
    asm volatile i386 {
        mov edx, num
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [edx]
        mov out, eax
        2:
        ; =r (out)
        ; r (num)
        ; eax
          edx
          memory
    }
    return out
}

// cas_i32 performs a compare-and-swap operation.
// If the current value at addr equals old, it atomically stores new.
// Returns true if the swap was performed, false otherwise.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn cas_i32(addr &i32, old i32, new i32) bool {
    mut swapped := false
    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, old
        mov ecx, new
        lock cmpxchg [edx], ecx
        sete al
        mov swapped, al
        2:
        ; =m (swapped)
        ; r (addr)
          r (old)
          r (new)
        ; eax
          ecx
          edx
          memory
    }
    return swapped
}

// store_i64 atomically stores value at dest using MMX instructions.
// The operation is performed with sequential consistency.
// Requires MMX support. Panics if dest is not 8-byte aligned.
pub fn store_i64(dest &i64, value i64) {
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        movq mm0, value
        movq [esi], mm0
        emms
        xor eax, eax
        lock xaddl [esp], eax
        2:
        ; ; r (dest)
          m (value)
        ; esi
          eax
          mm0
          memory
    }
}

// load_i64 atomically loads and returns the value at num using MMX instructions.
// The operation is performed with sequential consistency.
// Requires MMX support. Panics if num is not 8-byte aligned.
pub fn load_i64(num &i64) i64 {
    mut out := i64(0)
    asm volatile i386 {
        mov esi, num
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        movq mm0, [esi]
        movq out, mm0
        emms
        2:
        ; =m (out)
        ; r (num)
        ; esi
          mm0
          memory
    }
    return out
}

// add_i64 atomically adds delta to the value at dest and returns the new value.
// Uses a compare-and-swap loop. The operation is performed with sequential consistency.
// Panics if dest is not 8-byte aligned.
pub fn add_i64(dest &i64, delta i64) i64 {
    mut delta_lo := u32(u64(delta) & 0xFFFF_FFFF)
    mut delta_hi := u32(u64(delta) >> 32)
    mut res_lo := u32(0)
    mut res_hi := u32(0)
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [esi]
        mov edx, [esi + 4]
        mov ebx, eax
        mov ecx, edx
        add ebx, delta_lo
        adc ecx, delta_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov res_lo, ebx
        mov res_hi, ecx
        2:
        ; ; r (dest)
          m (delta_lo)
          m (delta_hi)
          m (res_lo)
          m (res_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    return i64(u64(res_lo) | (u64(res_hi) << 32))
}

// swap_i64 atomically stores value at dest and returns the old value.
// Uses a compare-and-swap loop. The operation is performed with sequential consistency.
// Panics if dest is not 8-byte aligned.
pub fn swap_i64(dest &i64, value i64) i64 {
    mut value_lo := u32(u64(value) & 0xFFFF_FFFF)
    mut value_hi := u32(u64(value) >> 32)
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [esi]
        mov edx, [esi + 4]
        mov ebx, value_lo
        mov ecx, value_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov value_lo, eax
        mov value_hi, edx
        2:
        ; ; r (dest)
          m (value_lo)
          m (value_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    return i64(u64(value_lo) | (u64(value_hi) << 32))
}

// cas_i64 performs a compare-and-swap operation.
// If the current value at addr equals old, it atomically stores new.
// Returns true if the swap was performed, false otherwise.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn cas_i64(addr &i64, old i64, new i64) bool {
    mut swapped := false
    mut old_lo := u32(u64(old) & 0xFFFF_FFFF)
    mut old_hi := u32(u64(old) >> 32)
    mut new_lo := u32(u64(new) & 0xFFFF_FFFF)
    mut new_hi := u32(u64(new) >> 32)
    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, old_lo
        mov edx, old_hi
        mov ebx, new_lo
        mov ecx, new_hi
        lock cmpxchg8b [esi]
        sete al
        mov swapped, al
        2:
        ; =m (swapped)
        ; r (addr)
          m (old_lo)
          m (old_hi)
          m (new_lo)
          m (new_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    return swapped
}

// add_u32 atomically adds delta to the value at dest and returns the new value.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn add_u32(dest &u32, delta u32) u32 {
    mut result := u32(0)
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, delta
        lock xadd [edx], eax
        add eax, delta
        mov result, eax
        2:
        ; =r (result)
        ; r (dest)
          r (delta)
        ; eax
          edx
          memory
    }
    return result
}

// swap_u32 atomically stores new value at dest and returns the old value.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn swap_u32(dest &u32, new u32) u32 {
    mut old := u32(0)
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, new
        xchg [edx], eax
        mov old, eax
        2:
        ; =m (old)
        ; r (dest)
          r (new)
        ; eax
          edx
          memory
    }
    return old
}

// store_u32 atomically stores value at dest.
// The operation is performed with sequential consistency.
// Panics if dest is not 4-byte aligned.
pub fn store_u32(dest &u32, value u32) {
    asm volatile i386 {
        mov edx, dest
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, value
        xchg eax, [edx]
        2:
        ; ; r (dest)
          r (value)
        ; eax
          edx
          memory
    }
}

// load_u32 atomically loads and returns the value at num.
// The operation is performed with sequential consistency.
// Panics if num is not 4-byte aligned.
pub fn load_u32(num &u32) u32 {
    mut out := u32(0)
    asm volatile i386 {
        mov edx, num
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [edx]
        mov out, eax
        2:
        ; =r (out)
        ; r (num)
        ; eax
          edx
          memory
    }
    return out
}

// cas_u32 performs a compare-and-swap operation.
// If the current value at addr equals old, it atomically stores new.
// Returns true if the swap was performed, false otherwise.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn cas_u32(addr &u32, old u32, new u32) bool {
    mut swapped := false
    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, old
        mov ecx, new
        lock cmpxchg [edx], ecx
        sete al
        mov swapped, al
        2:
        ; =m (swapped)
        ; r (addr)
          r (old)
          r (new)
        ; eax
          ecx
          edx
          memory
    }
    return swapped
}

// load_u64 atomically loads and returns the value at num using MMX instructions.
// The operation is performed with sequential consistency.
// Requires MMX support. Panics if num is not 8-byte aligned.
pub fn load_u64(num &u64) u64 {
    mut out := u64(0)
    asm volatile i386 {
        mov esi, num
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        movq mm0, [esi]
        movq out, mm0
        emms
        2:
        ; =m (out)
        ; r (num)
        ; esi
          mm0
          memory
    }
    return out
}

// store_u64 atomically stores value at dest using MMX instructions.
// The operation is performed with sequential consistency.
// Requires MMX support. Panics if dest is not 8-byte aligned.
pub fn store_u64(dest &u64, value u64) {
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        movq mm0, value
        movq [esi], mm0
        emms
        xor eax, eax
        lock xaddl [esp], eax
        2:
        ; ; r (dest)
          m (value)
        ; eax
          mm0
          memory
    }
}

// add_u64 atomically adds delta to the value at dest and returns the new value.
// Uses a compare-and-swap loop. The operation is performed with sequential consistency.
// Panics if dest is not 8-byte aligned.
pub fn add_u64(dest &u64, delta u64) u64 {
    mut delta_lo := u32(delta & 0xFFFF_FFFF)
    mut delta_hi := u32(delta >> 32)
    mut res_lo := u32(0)
    mut res_hi := u32(0)
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [esi]
        mov edx, [esi + 4]
        mov ebx, eax
        mov ecx, edx
        add ebx, delta_lo
        adc ecx, delta_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov res_lo, ebx
        mov res_hi, ecx
        2:
        ; ; r (dest)
          m (delta_lo)
          m (delta_hi)
          m (res_lo)
          m (res_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    return u64(res_lo) | (u64(res_hi) << 32)
}

// swap_u64 atomically stores value at dest and returns the old value.
// Uses a compare-and-swap loop. The operation is performed with sequential consistency.
// Panics if dest is not 8-byte aligned.
pub fn swap_u64(dest &u64, value u64) u64 {
    mut old := u64(0)
    mut value_lo := u32(value & 0xFFFF_FFFF)
    mut value_hi := u32(value >> 32)
    asm volatile i386 {
        mov esi, dest
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [esi]
        mov edx, [esi + 4]
        mov ebx, value_lo
        mov ecx, value_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov value_lo, eax
        mov value_hi, edx
        2:
        ; ; r (dest)
          m (value_lo)
          m (value_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    old = u64(value_lo) | (u64(value_hi) << 32)
    return old
}

// cas_u64 performs a compare-and-swap operation.
// If the current value at addr equals old, it atomically stores new.
// Returns true if the swap was performed, false otherwise.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn cas_u64(addr &u64, old u64, new u64) bool {
    mut swapped := false
    mut old_lo := u32(old & 0xFFFF_FFFF)
    mut old_hi := u32(old >> 32)
    mut new_lo := u32(new & 0xFFFF_FFFF)
    mut new_hi := u32(new >> 32)
    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, old_lo
        mov edx, old_hi
        mov ebx, new_lo
        mov ecx, new_hi
        lock cmpxchg8b [esi]
        sete al
        mov swapped, al
        2:
        ; =m (swapped)
        ; r (addr)
          m (old_lo)
          m (old_hi)
          m (new_lo)
          m (new_hi)
        ; eax
          ebx
          ecx
          edx
          esi
          memory
    }
    return swapped
}

// and_i64 atomically performs a bitwise AND of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn and_i64(addr &i64, mask i64) i64 {
    mask_lo := u32(u64(mask) & 0xFFFF_FFFF)
    mask_hi := u32(u64(mask) >> 32)
    mut old_lo := u32(0)
    mut old_hi := u32(0)

    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [esi]
        mov edx, [esi + 4]
        3:
        mov ebx, eax
        and ebx, mask_lo
        mov ecx, edx
        and ecx, mask_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov old_lo, eax
        mov old_hi, edx
        2:
        ; =m (old_lo)
          =m (old_hi)
        ; m (mask_lo)
          m (mask_hi)
          r (addr)
        ; eax
          edx
          ecx
          ebx
          esi
    }
    return i64(u64(old_lo) | (u64(old_hi) << 32))
}

// and_u64 atomically performs a bitwise AND of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn and_u64(addr &u64, mask u64) u64 {
    mask_lo := u32(u64(mask) & 0xFFFF_FFFF)
    mask_hi := u32(u64(mask) >> 32)
    mut old_lo := u32(0)
    mut old_hi := u32(0)

    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [esi]
        mov edx, [esi + 4]
        3:
        mov ebx, eax
        and ebx, mask_lo
        mov ecx, edx
        and ecx, mask_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov old_lo, eax
        mov old_hi, edx
        2:
        ; =m (old_lo)
          =m (old_hi)
        ; m (mask_lo)
          m (mask_hi)
          r (addr)
        ; eax
          edx
          ecx
          ebx
          esi
    }
    return u64(old_lo) | (u64(old_hi) << 32)
}

// or_i64 atomically performs a bitwise OR of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn or_i64(addr &i64, mask i64) i64 {
    mask_lo := u32(u64(mask) & 0xFFFF_FFFF)
    mask_hi := u32(u64(mask) >> 32)
    mut old_lo := u32(0)
    mut old_hi := u32(0)

    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [esi]
        mov edx, [esi + 4]
        3:
        mov ebx, eax
        or ebx, mask_lo
        mov ecx, edx
        or ecx, mask_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov old_lo, eax
        mov old_hi, edx
        2:
        ; =m (old_lo)
          =m (old_hi)
        ; m (mask_lo)
          m (mask_hi)
          r (addr)
        ; eax
          edx
          ecx
          ebx
          esi
    }
    return i64(u64(old_lo) | (u64(old_hi) << 32))
}

// or_u64 atomically performs a bitwise OR of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 8-byte aligned.
pub fn or_u64(addr &u64, mask u64) u64 {
    mask_lo := u32(u64(mask) & 0xFFFF_FFFF)
    mask_hi := u32(u64(mask) >> 32)
    mut old_lo := u32(0)
    mut old_hi := u32(0)

    asm volatile i386 {
        mov esi, addr
        test esi, 7
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        mov eax, [esi]
        mov edx, [esi + 4]
        3:
        mov ebx, eax
        or ebx, mask_lo
        mov ecx, edx
        or ecx, mask_hi
        lock cmpxchg8b [esi]
        jnz '3b'
        mov old_lo, eax
        mov old_hi, edx
        2:
        ; =m (old_lo)
          =m (old_hi)
        ; m (mask_lo)
          m (mask_hi)
          r (addr)
        ; eax
          edx
          ecx
          ebx
          esi
    }
    return u64(old_lo) | (u64(old_hi) << 32)
}

// and_u32 atomically performs a bitwise AND of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn and_u32(addr &u32, mask u32) u32 {
    mut old := u32(0)

    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [edx]
        mov ecx, eax
        and ecx, mask
        lock cmpxchgl [edx], ecx
        jnz '3b'
        mov old, eax
        2:
        ; =r (old)
        ; r (addr)
          r (mask)
        ; edx
          eax
          ecx
          memory
    }
    return old
}

// and_i32 atomically performs a bitwise AND of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn and_i32(addr &i32, mask i32) i32 {
    mut old := i32(0)

    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [edx]
        mov ecx, eax
        and ecx, mask
        lock cmpxchgl [edx], ecx
        jnz '3b'
        mov old, eax
        2:
        ; =r (old)
        ; r (addr)
          r (mask)
        ; edx
          eax
          ecx
          memory
    }
    return old
}

// or_i32 atomically performs a bitwise OR of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn or_i32(addr &i32, mask i32) i32 {
    mut old := i32(0)

    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [edx]
        mov ecx, eax
        or ecx, mask
        lock cmpxchgl [edx], ecx
        jnz '3b'
        mov old, eax
        2:
        ; =r (old)
        ; r (addr)
          r (mask)
        ; edx
          eax
          ecx
          memory
    }
    return old
}

// or_u32 atomically performs a bitwise OR of the value at addr with mask and returns the old value.
// The operation is performed with sequential consistency.
// Panics if addr is not 4-byte aligned.
pub fn or_u32(addr &u32, mask u32) u32 {
    mut old := u32(0)

    asm volatile i386 {
        mov edx, addr
        test edx, 3
        jz '1f'
        call panicUnaligned
        jmp '2f'
        1:
        3:
        mov eax, [edx]
        mov ecx, eax
        or ecx, mask
        lock cmpxchgl [edx], ecx
        jnz '3b'
        mov old, eax
        2:
        ; =r (old)
        ; r (addr)
          r (mask)
        ; edx
          eax
          ecx
          memory
    }
    return old
}