[sheepdog] [PATCH] test: Implentment fast random on AES

Joseph Glanville joseph at cloudscaling.com
Tue May 7 13:56:40 CEST 2013


I like this version better than the one I come up with. You should update
your patch. :)


On Mon, May 6, 2013 at 9:00 AM, MORITA Kazutaka
<morita.kazutaka at gmail.com>wrote:

> At Mon, 06 May 2013 16:24:39 +0800,
> Liu Yuan wrote:
> >
> > On 05/06/2013 04:15 PM, Joseph Glanville wrote:
> > > +    _random | dd bs=1M | $COLLIE vdi write test$i -p 7000 &
> >
> > 'dd bs=1M' is redundant? I can't still pass the tests with your patch
> > yet. Still sees 'error writing output file' error
>
> I think it is because openssl fails to write data to stdout because collie
> vdi
> write has exited.  If so, we can safely ignore the error message.
>
> I've tried the following simpler version.  It looks working correctly
> without dd
> and shows much better performance than /dev/urandom on my environment.
>
> ==
> _random()
> {
>     openssl enc -aes-256-ctr -pass pass:"$(date)" -nosalt < /dev/zero
> 2>/dev/null
> }
> ==
>
> Thanks,
>
> Kazutaka
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog/attachments/20130507/27e66657/attachment-0004.html>


More information about the sheepdog mailing list