[FFmpeg-devel] [PATCH 1/2] avutil/random_seed: use arc4random() when available

[Ganesh Ajjanagadde]

On Wed, Dec 9, 2015 at 8:08 AM, Ganesh Ajjanagadde
<gajjanagadde at gmail.com> wrote:
> On Sun, Dec 6, 2015 at 10:56 PM, Ganesh Ajjanagadde
> <gajjanagadde at gmail.com> wrote:
>> arc4random() was designed as a superior interface for system random
>> number generation, designed for OpenBSD. It is thus an improvement to
>> use it whenever available.
>>
>> As a side note, this may or may not get included in glibc, and there is
>> a proposal to create a posix_random family based on these ideas:
>> http://austingroupbugs.net/view.php?id=859.
>>
>> Untested as I lack OpenBSD.
>>
>> Signed-off-by: Ganesh Ajjanagadde <gajjanagadde at gmail.com>
>> ---
>>  configure               | 2 ++
>>  libavutil/random_seed.c | 4 ++++
>>  2 files changed, 6 insertions(+)
>>
>> diff --git a/configure b/configure
>> index 7530c88..e676269 100755
>> --- a/configure
>> +++ b/configure
>> @@ -1840,6 +1840,7 @@ MATH_FUNCS="
>>  SYSTEM_FUNCS="
>>      access
>>      aligned_malloc
>> +    arc4random
>>      clock_gettime
>>      closesocket
>>      CommandLineToArgvW
>> @@ -5218,6 +5219,7 @@ check_func  ${malloc_prefix}memalign            && enable memalign
>>  check_func  ${malloc_prefix}posix_memalign      && enable posix_memalign
>>
>>  check_func  access
>> +check_func  arc4random
>>  check_func_headers time.h clock_gettime || { check_func_headers time.h clock_gettime -lrt && add_extralibs -lrt && LIBRT="-lrt"; }
>>  check_func  fcntl
>>  check_func  fork
>> diff --git a/libavutil/random_seed.c b/libavutil/random_seed.c
>> index 8aa8c38..205a636 100644
>> --- a/libavutil/random_seed.c
>> +++ b/libavutil/random_seed.c
>> @@ -121,6 +121,10 @@ uint32_t av_get_random_seed(void)
>>      }
>>  #endif
>>
>> +#if HAVE_ARC4RANDOM
>> +    return arc4random();
>> +#endif
>> +
>>      if (read_random(&seed, "/dev/urandom") == sizeof(seed))
>>          return seed;
>>      if (read_random(&seed, "/dev/random")  == sizeof(seed))
>> --
>> 2.6.3
>>
>
> any objections to this: all objections so far have been for 2/2, i.e
> for the more controversial one that harnesses libbsd for non-native
> functionality?

To clarify, none of wm4's comments apply here. It will provide
benefits not just to OpenBSD, but also to other BSD's, Mac OS X, and
some alternative libc's (not glibc, which refused and continues to
refuse arc4random).

@Ronald, or anyone else with a Mac: can you please try at least a build test?

I have now shelved 2/2, as it has got drawn into a stalemate.