Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 21 May 2013 15:08:40 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Parallella: bcrypt

Hi Katja,

On Tue, May 21, 2013 at 10:41:39AM +0200, Katja Malvoni wrote:
> I was testing bcrypt implementation but on only one Epiphany core.

Did it run?  Did it produce the correct results?  How fast was it?

> Than I
> tried to test it on all 16 cores but I allowed all cores to write to the
> same buffer in the shared ram (which is also read by the host). And I
> rashly ran the code. After a few moments the system stopped responding and
> now I'm not able to connect any more. I am sorry for this and for not going
> through my code before executing it.

Actually, this is good news (as long as it does not happen too often) -
it shows that you're trying new stuff.  We'll get the board back online
shortly.  In fact, there is a remote power-cycle feature in that setup,
but it turned out to be broken - chances are that it'll be fixed by next
time we need it. ;-)

Connecting to the USB UART port, which is configured as the serial
console in the Linux kernel, I got this one line of output:

7>[drm:output_poll_execute], <7>[drm:output_poll_execute], <7>[

This must be a previously buffered portion of the kernel's output.

No response to input from that serial port, indeed - the system is
locked up good.

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ