netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Roland Dreier <roland@topspin.com>
To: "David S. Miller" <davem@davemloft.net>
Cc: Dmitry Yusupov <dmitry_yus@yahoo.com>,
	mpm@selenic.com, andrea@suse.de, michaelc@cs.wisc.edu,
	open-iscsi@googlegroups.com,
	James.Bottomley@HansenPartnership.com,
	ksummit-2005-discuss@thunk.org, netdev@oss.sgi.com
Subject: Re: [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics
Date: Mon, 28 Mar 2005 11:45:19 -0800	[thread overview]
Message-ID: <52vf7bwo4w.fsf@topspin.com> (raw)
In-Reply-To: <20050326224621.61f6d917.davem@davemloft.net> (David S. Miller's message of "Sat, 26 Mar 2005 22:46:21 -0800")

Let me slightly hijack this thread to throw out another topic that I
think is worth talking about at the kernel summit: handling remote DMA
(RDMA) network technologies.

As some of you might know, I'm one of the main authors of the
InfiniBand support in the kernel, and I think we have things fairly
well in hand there, although handling direct userspace access to RDMA
capabilities may raise some issues worth talking about.

However, there is also RDMA-over-TCP hardware beginning to be used,
based on the specs from the IETF rddp working group and the RDMA
Consortium.  I would hope that we can abstract out the common pieces
for InfiniBand and RDMA NIC (RNIC) support and morph
drivers/infiniband into a more general drivers/rdma.

This is not _that_ offtopic, since RDMA NICs provide another way of
handling OOM for iSCSI.  By having the NIC handle the network
transport through something like iSER, you avoid a lot of the issues
in this thread.  Having to reconnect to a target while OOM is still a
problem, but it seems no worse in principal than the issues with a
dump FC card that needs the host driver to handling fabric login.

I know that in the InfiniBand world, people have been able to run
stress tests of storage over SCSI RDMA Protocol (SRP) with very heavy
swapping going on and no deadlocks.  SRP is in effect network storage
with the transport handled by the IB hardware.

However there are some sticky points that I would be interested in
discussing.  For example, the IETF rddp drafts envisage what they call
a "dual stack" model: TCP connections are set up by the usual network
stack and run for a while in "streaming" mode until the application is
ready to start using RDMA.  At that point there is an "MPA"
negotiation and then the socket is handed over to the RNIC.  Clearly
moving the state from the kernel's stack to the RNIC is not trivial.

Other developers who have more direct experience with RNIC hardware or
perhaps just strong opinions may have other things in this area that
they'd like to talk about.

Thanks,
  Roland

  parent reply	other threads:[~2005-03-28 19:45 UTC|newest]

Thread overview: 88+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <4241D106.8050302@cs.wisc.edu>
     [not found] ` <20050324101622S.fujita.tomonori@lab.ntt.co.jp>
     [not found]   ` <1111628393.1548.307.camel@beastie>
     [not found]     ` <20050324113312W.fujita.tomonori@lab.ntt.co.jp>
     [not found]       ` <1111633846.1548.318.camel@beastie>
     [not found]         ` <20050324215922.GT14202@opteron.random>
     [not found]           ` <424346FE.20704@cs.wisc.edu>
     [not found]             ` <20050324233921.GZ14202@opteron.random>
     [not found]               ` <20050325034341.GV32638@waste.org>
     [not found]                 ` <20050327035149.GD4053@g5.random>
2005-03-27  5:48                   ` [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics Matt Mackall
2005-03-27  6:04                     ` Andrea Arcangeli
2005-03-27  6:38                       ` Matt Mackall
2005-03-27 14:50                         ` Andrea Arcangeli
2005-03-27  6:33                     ` Dmitry Yusupov
2005-03-27  6:46                       ` David S. Miller
2005-03-27  7:05                         ` Dmitry Yusupov
2005-03-27  7:57                           ` David S. Miller
2005-03-27  8:18                             ` Dmitry Yusupov
2005-03-27 18:26                               ` Mike Christie
2005-03-27 18:31                                 ` David S. Miller
2005-03-27 19:58                                   ` Matt Mackall
2005-03-27 21:49                                   ` Dmitry Yusupov
2005-03-27 18:47                                 ` Dmitry Yusupov
2005-03-27 21:14                         ` Alex Aizman
     [not found]                         ` <20050327211506.85EDA16022F6@mx1.suse.de>
2005-03-28  0:15                           ` Andrea Arcangeli
2005-03-28  3:54                         ` Rik van Riel
2005-03-28  4:34                           ` David S. Miller
2005-03-28  4:50                             ` Rik van Riel
2005-03-28  6:58                           ` Alex Aizman
2005-03-28 16:12                           ` Andi Kleen
2005-03-28 16:22                             ` Andrea Arcangeli
2005-03-28 16:24                             ` Rik van Riel
2005-03-29 15:11                               ` Andi Kleen
2005-03-29 15:29                                 ` Rik van Riel
2005-03-29 17:03                                 ` Matt Mackall
2005-03-28 16:28                             ` James Bottomley
2005-03-29 15:20                               ` Andi Kleen
2005-03-29 15:56                                 ` James Bottomley
2005-03-29 17:19                                 ` Dmitry Yusupov
2005-03-29 21:08                                   ` jamal
2005-03-29 22:00                                     ` Rik van Riel
2005-03-29 22:17                                       ` Matt Mackall
2005-03-29 23:30                                         ` jamal
2005-03-29 23:00                                       ` jamal
2005-03-29 23:25                                         ` Matt Mackall
2005-03-30  0:30                                           ` H. Peter Anvin
2005-03-30 15:24                                         ` Andi Kleen
2005-03-29 22:03                                     ` Rick Jones
2005-03-29 23:13                                       ` jamal
2005-03-30  2:28                                         ` Alex Aizman
     [not found]                                         ` <E1DGSwp-0004ZE-00@thunker.thunk.org>
2005-03-30 17:16                                           ` Grant Grundler
2005-03-30 18:46                                         ` Dmitry Yusupov
2005-03-30 15:22                                     ` Andi Kleen
2005-03-30 15:33                                       ` Andrea Arcangeli
2005-03-30 15:38                                         ` Rik van Riel
2005-03-30 15:39                                         ` Andi Kleen
2005-03-30 15:44                                           ` Andrea Arcangeli
2005-03-30 15:50                                             ` Rik van Riel
2005-03-30 16:04                                               ` James Bottomley
2005-03-30 17:48                                                 ` H. Peter Anvin
2005-03-30 16:02                                             ` Andi Kleen
2005-03-30 16:15                                               ` Andrea Arcangeli
2005-03-30 16:55                                                 ` jamal
2005-03-30 18:42                                                   ` Rik van Riel
2005-03-30 19:28                                                 ` Alex Aizman
2005-03-31 11:41                                                 ` Andi Kleen
2005-03-31 12:12                                                   ` Rik van Riel
2005-03-31 18:59                                                     ` Andi Kleen
2005-03-31 19:04                                                       ` Rik van Riel
2005-03-31 15:35                                                   ` Grant Grundler
2005-03-31 19:15                                                   ` Alex Aizman
2005-03-31 19:34                                                   ` Andi Kleen
2005-03-31 19:39                                                     ` Rik van Riel
2005-03-31 11:45                                                 ` Andi Kleen
2005-03-31 11:50                                                 ` Andi Kleen
2005-03-31 17:09                                                   ` Andrea Arcangeli
2005-03-31 22:05                                                     ` Dmitry Yusupov
2005-03-30 17:24                                       ` Matt Mackall
2005-03-30 17:39                                         ` Dmitry Yusupov
2005-03-30 20:10                                           ` Mike Christie
2005-03-30 17:07                                     ` Grant Grundler
2005-03-30  5:12                                   ` H. Peter Anvin
2005-03-28 16:37                             ` Dmitry Yusupov
2005-03-28 19:45                         ` Roland Dreier [this message]
2005-03-28 20:32                           ` Topic: Remote DMA network technologies Gerrit Huizenga
2005-03-28 20:36                             ` Roland Dreier
     [not found]                           ` <1112042936.5088.22.camel@beastie>
2005-03-28 22:32                             ` [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics Benjamin LaHaise
2005-03-29  3:19                               ` Linux support for RDMA (was: [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics) Roland Dreier
2005-03-30 16:00                                 ` Benjamin LaHaise
2005-03-31  1:08                                   ` Linux support for RDMA H. Peter Anvin
2005-04-02 18:08                               ` [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics Dmitry Yusupov
2005-04-02 19:13                                 ` Ming Zhang
2005-04-04  6:31                                 ` Grant Grundler
2005-04-04 18:57                                 ` Rick Jones
2005-03-29  3:14                             ` Linux support for RDMA (was: [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics) Roland Dreier
     [not found] <42472259.2866086e.3169.318fSMTPIN_ADDED@mx.googlegroups.com>
2005-03-27 21:18 ` [Ksummit-2005-discuss] Summary of 2005 Kernel Summit Proposed Topics Alex Aizman
2005-03-27 21:53 Alex Aizman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=52vf7bwo4w.fsf@topspin.com \
    --to=roland@topspin.com \
    --cc=James.Bottomley@HansenPartnership.com \
    --cc=andrea@suse.de \
    --cc=davem@davemloft.net \
    --cc=dmitry_yus@yahoo.com \
    --cc=ksummit-2005-discuss@thunk.org \
    --cc=michaelc@cs.wisc.edu \
    --cc=mpm@selenic.com \
    --cc=netdev@oss.sgi.com \
    --cc=open-iscsi@googlegroups.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).