[RFC] [PATCH 0/4] librdmacm: Rsockets API and implementation

[RFC] [PATCH 0/4] librdmacm: Rsockets API and implementation

[Hefty, Sean]

The following patch set contains an initial implementation of rsockets as
presented at the 2012 OpenFabrics Workshop.  A copy of that presentation
is available at:

	https://www.openfabrics.org/downloads/rdmacm/rsockets-ofa12.pptx

and a video of the presentation can be found here:

	http://www.youtube.com/watch?v=2IRrXdIsJuU

For those of you who don't like clicking on links in an email, rsockets
is a socket-like API layered over verbs.  It is conceptually similar to a
user space implementation of SDP, except that it uses custom algorithms and
protocols.

The source code has been pushed into an rsocket branch in my git tree at:

	git://git.openfabrics.org/~shefty/librdmacm.git rsocket

All comments are welcome, but mostly verbal praise.  Although comments
on the wire protocol before an official release would be the most useful.

Signed-off-by: Sean Hefty <sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC] [PATCH 0/4] librdmacm: Rsockets API and implementation

[Bart Van Assche]

On 04/11/12 18:29, Hefty, Sean wrote:

> The following patch set contains an initial implementation of rsockets as
> presented at the 2012 OpenFabrics Workshop.  A copy of that presentation
> is available at:
> 
> 	https://www.openfabrics.org/downloads/rdmacm/rsockets-ofa12.pptx
> 
> and a video of the presentation can be found here:
> 
> 	http://www.youtube.com/watch?v=2IRrXdIsJuU
> 
> For those of you who don't like clicking on links in an email, rsockets
> is a socket-like API layered over verbs.  It is conceptually similar to a
> user space implementation of SDP, except that it uses custom algorithms and
> protocols.
> 
> The source code has been pushed into an rsocket branch in my git tree at:
> 
> 	git://git.openfabrics.org/~shefty/librdmacm.git rsocket
> 
> All comments are welcome, but mostly verbal praise.  Although comments
> on the wire protocol before an official release would be the most useful.


Interesting work. Regarding direct data placement: have the io_submit()
and io_getevents() system calls been considered ? Those are the
foundation of libaio.

Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: [RFC] [PATCH 0/4] librdmacm: Rsockets API and implementation

[Hefty, Sean]

> Interesting work. Regarding direct data placement: have the io_submit()
> and io_getevents() system calls been considered ? Those are the
> foundation of libaio.

I did look at aio and other interfaces, but I didn't try to move beyond the idea phase.  With aio, I didn't see how the calls could easily be used by a socket app without recoding, or even redesigning the app around the aio interface.

Are you suggesting calls similar to io_submit() and io_getevents() that work with a socket rather than an aio_context_t?

My intent is to start a more general discussion on whether direct data placement is needed (I'm assuming yes) and the best way to add it.  I wasn't comfortable enough with just my own musing on things to dedicate any time implementing something.

- Sean
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html