[FFmpeg-devel] [PATCH 1/6] lavu: add av_gettime_relative()

[LANGLOIS Olivier PIS -EXT]

> >  configure          |  2 ++
> >  libavdevice/v4l2.c | 14 ++------------
> >  libavutil/time.c   | 31 ++++++++++++++++++++++++++++++-
> >  libavutil/time.h   | 15 +++++++++++++++
> >  4 files changed, 49 insertions(+), 13 deletions(-)
>
> Did you want to keep the lavd change with the lavu change? I suppose they
> used to conflict, but they no longer do.

Yes. that was on purpose. While it is true that the function av_gettime_monotonic() from lavd does not conflict anymore since we renamed the function in lavu, they are still essentially identical so in order to avoid unnecessary code replication, I have made v4l2 use the new lavu function as in previous versions of the patch. One could argue that since the function av_gettime_relative() may not always be implemented as monotonic, in order to avoid a possible maintenance glitch in case that theoretically av_gettime_relative() on Linux would stop being identical to v4l2's av_gettime_monotonic(), it would be better to leave v4l2.c as is.

If that is what you have in mind, I can drop these changes. Just let me know if I read your comment correctly.
>
> >
> > diff --git a/configure b/configure
> > index e5ff72b..656ae3d 100755
> > --- a/configure
> > +++ b/configure
> > @@ -1659,6 +1659,7 @@ SYSTEM_FUNCS="
> >      access
> >      aligned_malloc
> >      clock_gettime
>
> > +    clock_nanosleep
>
> Seems not useful anymore.

Correct
>
> >      closesocket
> >      CommandLineToArgvW
> >      CoTaskMemFree
> > @@ -4489,6 +4490,7 @@ check_func  ${malloc_prefix}posix_memalign
> && enable posix_memalign
> >
> >  check_func  access
> >  check_func  clock_gettime || { check_func clock_gettime -lrt &&
> > add_extralibs -lrt; }
>
> > +check_func  clock_nanosleep || { check_func clock_nanosleep &&
> > +add_extralibs -lrt; }
>
> Ditto.

I agree again

> > --- a/libavutil/time.c
> > +++ b/libavutil/time.c
> > @@ -38,7 +38,15 @@
> >
> >  int64_t av_gettime(void)
> >  {
> > -#if HAVE_GETTIMEOFDAY
>
> > +#if HAVE_CLOCK_GETTIME
> > +    /*
> > +     * POSIX.1-2008 marks gettimeofday() as obsolete,
> > +     * recommending the use of clock_gettime(2) instead.
> > +     */
> > +    struct timespec ts;
> > +    clock_gettime(CLOCK_REALTIME, &ts);
> > +    return (int64_t)ts.tv_sec * 1000000 + ts.tv_nsec / 1000; #elif
> > +HAVE_GETTIMEOFDAY
>
> I really think this is unnecessary: gettimeofday() is not going away anytime
> soon, so this would just be more code in FFmpeg for no benefit.
>

No problem.

thank you for your comments Nicolas.

regards,
Olivier

Please ignore the confidentiality notice below.
It is automatically added without my consent.


________________________________
CONFIDENTIALITY : This e-mail and any attachments are confidential and may be privileged. If you are not a named recipient, please notify the sender immediately and do not disclose the contents to another person, use it for any purpose or store or copy the information in any medium.