[FFmpeg-devel] [PATCH v4] vvcdec: add thread executor

[James Almer]

On 8/22/2023 9:30 AM, Nuo Mi wrote:
> On Tue, Aug 22, 2023 at 8:23 PM James Almer <jamrial at gmail.com> wrote:
> 
>> On 8/22/2023 9:03 AM, Nuo Mi wrote:
>>> On Thu, Aug 17, 2023 at 12:42 AM Rémi Denis-Courmont <remi at remlab.net>
>>> wrote:
>>>
>>>> Le tiistaina 15. elokuuta 2023, 17.50.13 EEST Nuo Mi a écrit :
>>>>> The executor design pattern was inroduced by java
>>>>> <
>>>>
>> https://docs.oracle.com/en/java/javase/20/docs/api/java.base/java/util/conc
>>>>> urrent/Executor.html> it also adapted by python
>>>>> <https://docs.python.org/3/library/concurrent.futures.html>
>>>>> Compared to handcrafted thread pool management, it greatly simplifies
>> the
>>>>> thread code. ---
>>>>>    libavutil/Makefile   |   2 +
>>>>>    libavutil/executor.c | 201
>> +++++++++++++++++++++++++++++++++++++++++++
>>>>>    libavutil/executor.h |  67 +++++++++++++++
>>>>>    3 files changed, 270 insertions(+)
>>>>>    create mode 100644 libavutil/executor.c
>>>>>    create mode 100644 libavutil/executor.h
>>>>>
>>>>> diff --git a/libavutil/Makefile b/libavutil/Makefile
>>>>> index 7828c94dc5..4711f8cde8 100644
>>>>> --- a/libavutil/Makefile
>>>>> +++ b/libavutil/Makefile
>>>>> @@ -31,6 +31,7 @@ HEADERS = adler32.h
>>>>
>>>>>                \ encryption_info.h
>>>>
>>>>>    \ error.h                                                       \
>>>> eval.h
>>>>>                                                        \ +
>>>>> executor.h                                                    \ fifo.h
>>>>
>>>>>                                                     \ file.h
>>>>
>>>>>                                         \ frame.h
>>>>
>>>>>                             \ @@ -127,6 +128,7 @@ OBJS = adler32.o
>>>>
>>>>>                                               \ encryption_info.o
>>>>
>>>>>                                      \ error.o
>>>>
>>>>>                             \ eval.o
>>>>
>>>>>                    \ +       executor.o
>>>>
>>>>>                   \ fifo.o
>>>>
>>>>>          \ file.o
>>>>    \
>>>>> file_open.o                                                      \ diff
>>>>> --git a/libavutil/executor.c b/libavutil/executor.c
>>>>> new file mode 100644
>>>>> index 0000000000..6b47ba0bf2
>>>>> --- /dev/null
>>>>> +++ b/libavutil/executor.c
>>>>> @@ -0,0 +1,201 @@
>>>>> +/*
>>>>> + * Copyright (C) 2023 Nuo Mi
>>>>> + *
>>>>> + * This file is part of FFmpeg.
>>>>> + *
>>>>> + * FFmpeg is free software; you can redistribute it and/or
>>>>> + * modify it under the terms of the GNU Lesser General Public
>>>>> + * License as published by the Free Software Foundation; either
>>>>> + * version 2.1 of the License, or (at your option) any later version.
>>>>> + *
>>>>> + * FFmpeg 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
>>>>> + * Lesser General Public License for more details.
>>>>> + *
>>>>> + * You should have received a copy of the GNU Lesser General Public
>>>>> + * License along with FFmpeg; if not, write to the Free Software
>>>>> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
>>>> 02110-1301
>>>>> USA + */
>>>>> +#include "internal.h"
>>>>> +#include "mem.h"
>>>>> +#include "thread.h"
>>>>> +
>>>>> +#include "executor.h"
>>>>> +
>>>>> +#if !HAVE_THREADS
>>>>> +#define pthread_create(t, a, s, ar)     0
>>>>
>>>> Err, this should probably return an error value, e.g. ENOSYS or ENOTSUP.
>>>>
>>>>> +#define pthread_join(t, r)              do {} while(0)
>>>>
>>>> ...and this should be an unreachable(), assert(0), abort() or something
>>>> along
>>>> those lines.
>>>>
>>>>> +
>>>>> +#define pthread_cond_init(c, a)         0
>>>>> +#define pthread_cond_broadcast(c)       do {} while(0)
>>>>> +#define pthread_cond_signal(c)          do {} while(0)
>>>>> +#define pthread_cond_wait(c, m)         do {} while(0)
>>>>> +#define pthread_cond_destroy(c)         do {} while(0)
>>>>> +
>>>>> +#define pthread_mutex_init(m, a)        0
>>>>> +#define pthread_mutex_lock(l)           do {} while(0)
>>>>> +#define pthread_mutex_unlock(l)         do {} while(0)
>>>>> +#define pthread_mutex_destroy(l)        do {} while(0)
>>>>> +#endif
>>>>> +
>>>>> +typedef struct ThreadInfo {
>>>>> +    AVExecutor *e;
>>>>> +    pthread_t thread;
>>>>
>>>> Does this even work if !HAVE_THREAD ?
>>>
>>> Yes. The definition provided by the following include chain:
>>>
>>   "mem.h"->"avutil.h"->"common.h"->"<stdlib.h>"->"<sys/types.h>"->"<bits/pthreadtypes.h>"
>>
>> Why would bits/pthreadtypes.h be available if there's no pthreads?
>>
> I guess it is related to POSIX. If you check <sys/types.h>, you will find
> the following code:
> 
> /* Now add the thread types.  */
> #if defined __USE_POSIX199506 || defined __USE_UNIX98
> # include <bits/pthreadtypes.h>
> #endif

Well, not all targets are POSIX. If you look at the win32 and os/2 
wrappers, they need to define pthread_t. So it's not guaranteed to be 
defined when !HAVE_THREADS.