[FFmpeg-devel] [PATCH v3 3/3] avutil/random_seed: add support for gcrypt and OpenSSL as source of randomness

[James Almer]

On 7/6/2023 2:56 PM, Marton Balint wrote:
> 
> 
> On Thu, 6 Jul 2023, James Almer wrote:
> 
>> On 7/6/2023 4:52 AM, Anton Khirnov wrote:
>>>  Quoting James Almer (2023-07-05 15:03:07)
>>>>  On 7/5/2023 9:56 AM, Anton Khirnov wrote:
>>>>>  Quoting James Almer (2023-07-05 01:26:14)
>>>>>>  Signed-off-by: James Almer <jamrial at gmail.com>
>>>>>>  ---
>>>>>>     configure               |  2 +-
>>>>>>     libavutil/random_seed.c | 16 ++++++++++++++++
>>>>>>     2 files changed, 17 insertions(+), 1 deletion(-)
>>>>>>
>>>>>>  diff --git a/configure b/configure
>>>>>>  index 107d533b3e..d6e78297fe 100755
>>>>>>  --- a/configure
>>>>>>  +++ b/configure
>>>>>>  @@ -3892,7 +3892,7 @@ avfilter_deps="avutil"
>>>>>>     avfilter_suggest="libm stdatomic"
>>>>>>     avformat_deps="avcodec avutil"
>>>>>>     avformat_suggest="libm network zlib stdatomic"
>>>>>>  -avutil_suggest="clock_gettime ffnvcodec libm libdrm libmfx opencl
>>>>>>  user32 vaapi vulkan videotoolbox corefoundation corevideo coremedia
>>>>>>  bcrypt stdatomic"
>>>>>>  +avutil_suggest="clock_gettime ffnvcodec gcrypt libm libdrm libmfx
>>>>>>  opencl openssl user32 vaapi vulkan videotoolbox corefoundation
>>>>>>  corevideo coremedia bcrypt stdatomic"
>>>>>
>>>>>  What will this do exactly?
>>>>
>>>>  It's a soft dependency. Only if OpenSSL or GCrypt are enabled they 
>>>> will
>>>>  be added to avutil's library dependencies.
>>>>  Notice the bcrypt entry, also used in random_seed, and all the 
>>>> hwcontext
>>>>  backends.
>>>
>>>  Ok then.
>>
>> Any opinion on the order of function calls? Should i leave 
>> /dev/urandom as last resort after all (if any) external lib 
>> implementations failed, considering what Remí mentioned?
> 
> I think it is fine as is. Buffering can be turned off with
> setvbuf(f, NULL, _IONBF, 0); so that should not relevant.

Ok, applied as is, then.