phc-discussions - RE: [PHC] "clear_in_asap" & "Parallel Threads" -- was "Submission deadline extension (March 31)"

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>] [day] [month] [year] [list]

Date: Mon, 6 Jan 2014 19:22:12 -0500
From: "Gary W. Hvizdak" <gary.hvizdak@....rr.com>
To: <discussions@...sword-hashing.net>
Subject: RE: [PHC] "clear_in_asap" & "Parallel Threads" -- was "Submission deadline extension (March 31)"

Hey Bill,

    I'm another newbie - pleased to meet you - and haven't been reading many
posts since subscribing just a couple weeks ago.  Thank you for commenting
about a "clear_in_asap" flag.  That's something that I hadn't considered,
although the way my algorithm works, it probably won't buy me very much
since most of the execution time will have elapsed before it can be cleared.

    While I'm writing . I don't know the first thing about using parallel
threads.  Further, my API includes an optional flag which explicitly
requires that the calculation be done single-threaded.  Still, I'd be
curious how you do "that" and to know whether it's "portable".  Can you
provide a link and or source code fragment?

Thanks,
Gary

P.S.  In addition to being a crypto newbie, I'm also a newcomer to PHC,
having only learned of it a few months back, and having only installed a
suitable IDE (MS Visual Studio Express) in the past three weeks.

From: Bill Cox [mailto:waywardgeek@...il.com] 
Sent: Monday 06 Jan 2014 1906
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Submission deadline extension (March 31)

Nice.  Here are parameter's I've added to NoelKDF:

num_threads     - the number of threads to run in parallel
parallelism     - number of inner loops allowed to run in parallel - should
match
                  user's machine for best protection
num_hash_rounds - number of SHA256 rounds to compute the intermediate key
clear_in        - when true, overwrite the in buffer with 0's early on

Also, I "borrowed" the excellent Catena idea for "garlic", and use t_cost in
the outer loop, where I double memory each time through the loop.  Users
wont need it, but server admins will like it.  I also have a flag like yours
for clearing and freeing memory.

I have to say that people on this forum have been extremely generous with
their ideas and feedback.  As a crypto noob, I've benefitted tremendously.

Now... as for your chart.  That looks like your telling me the password
length based on the runtime of the hash.  Is this right?  A snooper for
run-time is trivial and runs in user mode.

Bill

Content of type "text/html" skipped