panthema / 2013 / parallel-string-sorting / parallel-string-sorting-0.5 / minitbb / tbb / machine / macos_common.h (Download File)
/*
    Copyright 2005-2012 Intel Corporation.  All Rights Reserved.

    This file is part of Threading Building Blocks.

    Threading Building Blocks is free software; you can redistribute it
    and/or modify it under the terms of the GNU General Public License
    version 2 as published by the Free Software Foundation.

    Threading Building Blocks 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 for more details.

    You should have received a copy of the GNU General Public License
    along with Threading Building Blocks; if not, write to the Free Software
    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

    As a special exception, you may use this file as part of a free software
    library without restriction.  Specifically, if other files instantiate
    templates or use macros or inline functions from this file, or you compile
    this file and link it with other files to produce an executable, this
    file does not by itself cause the resulting executable to be covered by
    the GNU General Public License.  This exception does not however
    invalidate any other reasons why the executable file might be covered by
    the GNU General Public License.
*/

#if !defined(__TBB_machine_H) || defined(__TBB_machine_macos_common_H)
#error Do not #include this internal file directly; use public TBB headers instead.
#endif

#define __TBB_machine_macos_common_H

#include <sched.h>
#define __TBB_Yield()  sched_yield()

// __TBB_HardwareConcurrency

#include <sys/types.h>
#include <sys/sysctl.h>

static inline int __TBB_macos_available_cpu() {
    int name[2] = {CTL_HW, HW_AVAILCPU};
    int ncpu;
    size_t size = sizeof(ncpu);
    sysctl( name, 2, &ncpu, &size, NULL, 0 );
    return ncpu;
}

#define __TBB_HardwareConcurrency() __TBB_macos_available_cpu()

#ifndef __TBB_full_memory_fence
    // TBB has not recognized the architecture (none of the architecture abstraction
    // headers was included).
    #define __TBB_UnknownArchitecture 1
#endif

#if __TBB_UnknownArchitecture || (__TBB_WORDSIZE==4 && !TBB_USE_ICC_BUILTINS)
// In case of IA32 this is a workaround for compiler bugs triggered by inline
// assembly implementation of __TBB_machine_cmpswp8 in linux_ia32.h, which may
// lead to incorrect codegen (gcc) or compilation failures (any icc including 12.0.4).

// Implementation of atomic operations based on OS provided primitives
#include <libkern/OSAtomic.h>

static inline int64_t __TBB_machine_cmpswp8_OsX(volatile void *ptr, int64_t value, int64_t comparand)
{
    __TBB_ASSERT( !((uintptr_t)ptr&0x7), "address not properly aligned for Mac OS atomics");
    int64_t* address = (int64_t*)ptr;
    while( !OSAtomicCompareAndSwap64Barrier(comparand, value, address) ){
#if __TBB_WORDSIZE==8
        int64_t snapshot = *address;
#else
        int64_t snapshot = OSAtomicAdd64( 0, address );
#endif
        if( snapshot!=comparand ) return snapshot;
    }
    return comparand;
}

#define __TBB_machine_cmpswp8 __TBB_machine_cmpswp8_OsX

#endif /* __TBB_UnknownArchitecture || (__TBB_WORDSIZE==4 && !TBB_USE_ICC_BUILTINS) */

#if __TBB_UnknownArchitecture

#ifndef __TBB_WORDSIZE
#define __TBB_WORDSIZE 4
#endif

#ifdef __TBB_BIG_ENDIAN
    // Already determined based on hardware architecture.
#elif __BIG_ENDIAN__
    #define __TBB_BIG_ENDIAN 1
#elif __LITTLE_ENDIAN__
    #define __TBB_BIG_ENDIAN 0
#else
    #define __TBB_BIG_ENDIAN -1 // not currently supported
#endif

/** As this generic implementation has absolutely no information about underlying
    hardware, its performance most likely will be sub-optimal because of full memory
    fence usages where a more lightweight synchronization means (or none at all)
    could suffice. Thus if you use this header to enable TBB on a new platform,
    consider forking it and relaxing below helpers as appropriate. **/
#define __TBB_control_consistency_helper() OSMemoryBarrier()
#define __TBB_acquire_consistency_helper() OSMemoryBarrier()
#define __TBB_release_consistency_helper() OSMemoryBarrier()
#define __TBB_full_memory_fence()          OSMemoryBarrier()

static inline int32_t __TBB_machine_cmpswp4(volatile void *ptr, int32_t value, int32_t comparand)
{
    __TBB_ASSERT( !((uintptr_t)ptr&0x3), "address not properly aligned for Mac OS atomics");
    int32_t* address = (int32_t*)ptr;
    while( !OSAtomicCompareAndSwap32Barrier(comparand, value, address) ){
        int32_t snapshot = *address;
        if( snapshot!=comparand ) return snapshot;
    }
    return comparand;
}

static inline int32_t __TBB_machine_fetchadd4(volatile void *ptr, int32_t addend)
{
    __TBB_ASSERT( !((uintptr_t)ptr&0x3), "address not properly aligned for Mac OS atomics");
    return OSAtomicAdd32Barrier(addend, (int32_t*)ptr) - addend;
}

static inline int64_t __TBB_machine_fetchadd8(volatile void *ptr, int64_t addend)
{
    __TBB_ASSERT( !((uintptr_t)ptr&0x7), "address not properly aligned for Mac OS atomics");
    return OSAtomicAdd64Barrier(addend, (int64_t*)ptr) - addend;
}

#define __TBB_USE_GENERIC_PART_WORD_CAS                     1
#define __TBB_USE_GENERIC_PART_WORD_FETCH_ADD               1
#define __TBB_USE_GENERIC_FETCH_STORE                       1
#define __TBB_USE_GENERIC_HALF_FENCED_LOAD_STORE            1
#define __TBB_USE_GENERIC_RELAXED_LOAD_STORE                1
#if __TBB_WORDSIZE == 4
    #define __TBB_USE_GENERIC_DWORD_LOAD_STORE              1
#endif
#define __TBB_USE_GENERIC_SEQUENTIAL_CONSISTENCY_LOAD_STORE 1

#endif /* __TBB_UnknownArchitecture */