Re: [Iscsitarget-devel] Re: [ANNOUNCE] High-Performance Unix Initiator 450MB/sec

[Dmitry Y. <dmi...@ya...>]

Thanks Karen,

Not exactly the main reason why we are here. Our Initiator is not only
designed for Linux. Overtime it hopefully will also work on FreeBSD,
OpenSolaris and so on. But yes we know about PyX announce.

We also pushing our code to the "key" kernel guys. Two weeks back we've
sent our code for review. Based on that input we've implemented bunch of
interesting suggestions. For example, our WRITE path is totally lock-
less due to using of kfifo.(see include/linux/kfifo.h). Christoph
Hellwig contributed nice Open-iSCSI Interface module split. His patch
makes possible to register other transports like SCTP, IB/iSER with
Open-iSCSI Interface.

Regards,
Dima

On Thu, 2005-02-24 at 22:23 -0800, Karen Shaeffer wrote:
> On Thu, Feb 24, 2005 at 10:05:29PM -0800, Dmitry Yusupov wrote:
> > 
> > it is our third iSCSI Initiator implementation since 2001. So, we kind
> > of knew what we were doing. But only just few weeks ago we decided to
> > open the code under GPL. As far as Linux development is concerned, our
> > main goal is mainline kernel acceptance. :) Since we truly believe that
> > our implementation will out-perform any other existing Linux
> > implementations and our design is the cleanest ever. :)
> > Though we do have some bugs...
> 
> 
> Hi,
> 
> FYI, the Pyx Technology group just submitted a new design for an initiator
> to the Linux kernel folks. Here is the mail header for that email:
> 
> Subject: [ANNOUNCE] iSCSI Initiator Core Stack v1.6.1.20
> From:   "Nicholas A. Bellinger" <ni...@py...>
> To:     James Bottomley <James.Bottomley@SteelEye.com>,
>         Christoph Hellwig <hc...@ls...>,
>         Mike Christie <mi...@us...>,
>         Andre Hedrick <an...@py...>
> Cc:     linux-scsi <lin...@vg...>
> Date:   Wed, 23 Feb 2005 11:45:35 -0800
> 
> 
> If yours is so good, then you might want to get it to the kernel folks
> now for a comparative review. I'm sure two original submissions for
> initiators at the same point in time would be interesting to the kernel
> folks.
> 
> If you already know about this, then well, I don't mean to waste your time.
> I am just thinking this would be a time to act.
> 
> Thanks,
> Karen