[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1427f6579e0fb2b6fe11acf5e4550289@smtp.hushmail.com>
Date: Wed, 04 Apr 2012 02:18:47 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: RAR OpenCL committed to git
On 04/04/2012 01:59 AM, Solar Designer wrote:
> On Wed, Apr 04, 2012 at 01:46:55AM +0200, magnum wrote:
>> BTW when converting to EVP I just implemented the thread-locking
>> callbacks that Dhiru first used in the ssh format. In RAR I haven't seen
>> any regression like you were discussing then.
>
> Are the callbacks ever called (in your case)? I'd expect the AES stuff
> to be too simple to need any locking.
They sure are, but I have no idea if they are really needed:
...
init_locks()
lock_callback(9, 2, ex_data.c, 203)
lock_callback(10, 2, ex_data.c, 206)
lock_callback(9, 2, ex_data.c, 270)
lock_callback(10, 2, ex_data.c, 274)
lock_callback(9, 2, ex_data.c, 304)
lock_callback(10, 2, ex_data.c, 325)
lock_callback(5, 2, ex_data.c, 413)
lock_callback(6, 2, ex_data.c, 424)
lock_callback(9, 30, eng_list.c, 284)
lock_callback(10, 30, eng_list.c, 291)
lock_callback(9, 30, eng_lib.c, 116)
lock_callback(10, 30, eng_lib.c, 116)
lock_callback(9, 1, err.c, 295)
lock_callback(10, 1, err.c, 298)
thread_id()
lock_callback(9, 1, err.c, 446)
lock_callback(10, 1, err.c, 458)
lock_callback(9, 1, err.c, 446)
lock_callback(10, 1, err.c, 458)
lock_callback(9, 1, err.c, 513)
...
Some 55,000 of these for a two-core OpenCL-CPU self-test.
magnum
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.
