phc-discussions - RE: [FORGED] [PHC] Argon2 using memset still

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <9A043F3CF02CD34C8E74AC1594475C73F4AEC36C@uxcn10-5.UoA.auckland.ac.nz>
Date: Mon, 31 Aug 2015 11:53:48 +0000
From: Peter Gutmann <pgut001@...auckland.ac.nz>
To: "discussions@...sword-hashing.net" <discussions@...sword-hashing.net>
Subject: RE: [FORGED] [PHC] Argon2 using memset still

Bill Cox <waywardgeek@...il.com> writes:

>g++ has a habit of optimizing calls to memset away.  I doubt that Argon2
>succeed in it's attempts to clear memory.  I copied this function from Blake2:

You could also use environment-specific functions to do this:

#if defined( _MSC_VER ) && VC_GE_2005( _MSC_VER )
  /* This is just a mapping to RtlSecureZeroMemory() (via WinBase.h) which 
     is implemented as inline code implementing a loop on a pointer declared 
	 volatile, but unlike the corresponding RtlZeroMemory() there's a 
	 contract that this will always zeroise memory even in the face of 
	 compiler changes that would otherwise optimise away the access */
  #define zeroise( memory, size )	SecureZeroMemory( memory, size )
#elif defined( __STDC_LIB_EXT1__ )
  /* C11 defines a function memset_s() that guarantees that it won't be
	 optimised away, although this is quite well obfuscated in the spec,
	 "the memory indicated by [the memset parameters] may be accessible in 
	 the future and therefore must contain the values indicated by [the
	 value to set]", hopefully the implementers will know that this equates
	 to "the memset_s() call can't be optimised away" */
  #define zeroise( memory, size )	memset_s( memory, size, 0, size )
#elif defined( __OpenBSD__ )
  /* The OpenBSD folks defined their own won't-be-optimised-away bzero()
	 function */
  #define zeroise( memory, size )	explicit_bzero( memory, size )
#else
  #define zeroise( memory, size )	memset( memory, 0, size )
#endif /* Systems with distinct zeroise functions */

Peter.