Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 11 Jan 2013 19:12:59 -0200
From: Claudio André <claudioandre.br@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: SIGALRM and new 'done' method

Solve it is simple, but i have no idea what really has to be done inside 
sig_handle_timer.

Patch below.
BTW: 2 out of 228 tests have FAILED
- Benchmarking: ZIP-AES PBKDF2-HMAC-SHA-1 [OpenCL]... FAILED (cmp_all(1))
- Benchmarking: STRIP Password Manager PBKDF2-SHA1 [OpenCL]... FAILED 
(cmp_all(1))

-----

+static void sig_handle_timer(int signum)
+{
+
+}

void opencl_process_event(void)
{
     if (!bench_running) {
#if !OS_TIMER
+        signal(SIGALRM, sig_handle_timer);
         sig_timer_emu_tick();
#endif
         if (event_pending) {



Em 11-01-2013 16:03, Claudio André escreveu:
> Hi, just in case anyone has an answer.
>
> While as was trying my 'done' tree, i noticed that when i run './john
> -t', at some point, (no matter which [sequence of] formats) john exits
> with an "Alarm clock" message. Seems SIGALRM is not handled as
> expected in this multi-formats session.
>
> The message can be seen below. It happens on AMD CPU and GPU and
> bull's NVIDIA.
>
> Anyone has an idea or suggestion?
>
> Commenting 'opencl_process_event()' helps.
>
> Claudio
>
> Example:
> Local worksize (LWS) 1, Global worksize (GWS) 1024
> Benchmarking: WPA-PSK PBKDF2-HMAC-SHA-1 [OpenCL]... Alarm clock
> [exit]
>

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.