On 8/19/18 11:27, Oliver Pinter wrote:
>> Modified: head/crypto/heimdal/lib/roken/rand.c
>> ==============================================================================
>> --- head/crypto/heimdal/lib/roken/rand.c     Sun Aug 19 17:36:50 2018        
>> (r338058)
>> +++ head/crypto/heimdal/lib/roken/rand.c     Sun Aug 19 17:40:50 2018        
>> (r338059)
>> @@ -37,7 +37,6 @@ void ROKEN_LIB_FUNCTION
>>  rk_random_init(void)
>>  {
>>  #if defined(HAVE_ARC4RANDOM)
>> -    arc4random_stir();
>>  #elif defined(HAVE_SRANDOMDEV)
>>      srandomdev();
>>  #elif defined(HAVE_RANDOM)
> 
> This hunk of the patch looks weird.
> 
> Isn't it would better to undef HAVE_ARC4RANDOM without deleting the
> arc4random_stir()?
> Or delete the macro part of detection like

No that would be wrong.  The intention is to explicitly say "If
HAVE_ARC4RANDOM, do nothing for rk_random_init()".

undef HAVE_ARC4RANDOM is not an option because that means the code would
no longer use arc4random().

Cheers,

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to