Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANWtx02u3us8epAX3aE80pQ2gNyyh11-3pJyENK36_g-osoVuw@mail.gmail.com>
Date: Mon, 6 May 2013 19:15:02 -0400
From: Rich Rumble <richrumble@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: --fork under Cygwin

On Mon, May 6, 2013 at 6:46 PM, <jfoug@....net> wrote:

> I would REALLY like someone else that has worked with this look at it,
> before I poopoo it.  I have never used fork before, and simply tried to run
> with a 1 hash file, and -fork -inc:all.   It wrote some fork messages out
> (4 of them, I think).  There were numerous johns running, some listed as
> <defunct> some said they were running. Also numerous bashes.  It 'appeared'
> like what I would expect for a forking process.  However, things were not
> running, and were NOT killable.  The processes would not even go away with
> a kill from another shell, or any amount of ^C's or ^D's   I had to use a
> win32 process manager to clean up the mess.
>
I've had that happen, john forks into it's processes, loads the hashes, and
6 out of 8 (for example) never use any CPU. The very first or second
contest version I had the least trouble with, but it was also used the
least. The main issue initially as I saw it was the forks were not saving
cracks to the pot, but would print to the screen, so if I diff'd the screen
output and the POT file, I'd have very different results. magnum(i think?)
introduced a switch in the config to write cracks to the stats file which
works well. I need to compile a recent version of Fork/Node on cygwin and
report back. I'm out for a bit, but I should have more in 24hrs or so to
report on my own testing. Might want to search out some of my old Fork/Node
emails...
-rich

Content of type "text/html" skipped

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.