From: Chuck Lever <chucklever@gmail.com>
To: Jason Gunthorpe <jgg@ziepe.ca>,
Ka-Cheong Poon <ka-cheong.poon@oracle.com>
Cc: Leon Romanovsky <leon@kernel.org>,
linux-rdma <linux-rdma@vger.kernel.org>
Subject: Re: RDMA subsystem namespace related questions (was Re: Finding the namespace of a struct ib_device)
Date: Fri, 9 Oct 2020 11:52:37 -0400 [thread overview]
Message-ID: <06F8D44C-BB1E-4D72-AFCE-FFD52E20C9CF@gmail.com> (raw)
In-Reply-To: <20201009153406.GA5177@ziepe.ca>
> On Oct 9, 2020, at 11:34 AM, Jason Gunthorpe <jgg@ziepe.ca> wrote:
>
> On Fri, Oct 09, 2020 at 11:27:44AM -0400, Chuck Lever wrote:
>
>> Therefore I think the approach is going to be "one RDS listener per
>> net namespace". The problem Ka-Cheong is trying to address is how to
>> manage the destruction of a listener-namespace pair. The extra
>> reference count on the cm_id is pinning the namespace so it cannot
>> be destroyed.
>
> I really don't think this idea of just loading a kernel module and it
> immediately creates a network visibile listening socket in every
> namespace is very good.
>
>> Understood, but it doesn't seem like there is enough useful overlap
>> between the NFS and RDS usage scenarios. With NFS, I would expect
>> an explicit listener shutdown from userland prior to namespace
>> destruction.
>
> Yes, because namespaces are fundamentally supposed to be anchored in
> the processes inside the namespace.
Aye, the container model.
> Having the kernel jump in and start opening holes as soon as a
> namespace is created is just wrong.
>
> At a bare minimum the listener should not exist until something in the
> namespace is willing to work with RDS.
I was thinking that too, but I'm not sure if that change would have
ramifications to existing RDS applications. There's quite a bit of
legacy to deal with.
An alternative would be to add a user daemon to RDS to manage the
listener lifecycle, rather than having the endpoint created by
module load. That might help the listener-namespace destruction
issue, and should be entirely application-transparent.
--
Chuck Lever
chucklever@gmail.com
next prev parent reply other threads:[~2020-10-09 15:52 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-03 14:02 Finding the namespace of a struct ib_device Ka-Cheong Poon
2020-09-03 17:39 ` Jason Gunthorpe
2020-09-04 4:01 ` Ka-Cheong Poon
2020-09-04 11:32 ` Jason Gunthorpe
2020-09-04 14:02 ` Ka-Cheong Poon
2020-09-06 7:44 ` Leon Romanovsky
2020-09-07 3:33 ` Ka-Cheong Poon
2020-09-07 7:18 ` Leon Romanovsky
2020-09-07 8:24 ` Ka-Cheong Poon
2020-09-07 9:04 ` Leon Romanovsky
2020-09-07 9:28 ` Ka-Cheong Poon
2020-09-07 10:22 ` Leon Romanovsky
2020-09-07 13:48 ` Ka-Cheong Poon
2020-09-29 16:57 ` RDMA subsystem namespace related questions (was Re: Finding the namespace of a struct ib_device) Ka-Cheong Poon
2020-09-29 17:40 ` Jason Gunthorpe
2020-09-30 10:32 ` Ka-Cheong Poon
2020-10-02 14:04 ` Jason Gunthorpe
2020-10-05 10:27 ` Ka-Cheong Poon
2020-10-05 13:16 ` Jason Gunthorpe
2020-10-05 13:57 ` Ka-Cheong Poon
2020-10-05 14:25 ` Jason Gunthorpe
2020-10-05 15:02 ` Ka-Cheong Poon
2020-10-05 15:45 ` Jason Gunthorpe
2020-10-06 9:36 ` Ka-Cheong Poon
2020-10-06 12:46 ` Jason Gunthorpe
2020-10-07 8:38 ` Ka-Cheong Poon
2020-10-07 11:16 ` Leon Romanovsky
2020-10-08 10:22 ` Ka-Cheong Poon
2020-10-08 10:36 ` Leon Romanovsky
2020-10-08 11:08 ` Ka-Cheong Poon
2020-10-08 16:08 ` Jason Gunthorpe
2020-10-08 16:21 ` Chuck Lever
2020-10-08 16:46 ` Jason Gunthorpe
2020-10-09 4:49 ` Ka-Cheong Poon
2020-10-09 14:39 ` Jason Gunthorpe
2020-10-09 14:48 ` Chuck Lever
2020-10-09 14:57 ` Jason Gunthorpe
2020-10-09 15:00 ` Chuck Lever
2020-10-09 15:07 ` Jason Gunthorpe
2020-10-09 15:27 ` Chuck Lever
2020-10-09 15:34 ` Jason Gunthorpe
2020-10-09 15:52 ` Chuck Lever [this message]
2020-10-12 8:20 ` Ka-Cheong Poon
2020-10-16 18:54 ` Jason Gunthorpe
2020-10-16 20:49 ` Chuck Lever
2020-10-19 18:31 ` Jason Gunthorpe
2020-10-07 12:28 ` Jason Gunthorpe
2020-10-08 10:49 ` Ka-Cheong Poon
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=06F8D44C-BB1E-4D72-AFCE-FFD52E20C9CF@gmail.com \
--to=chucklever@gmail.com \
--cc=jgg@ziepe.ca \
--cc=ka-cheong.poon@oracle.com \
--cc=leon@kernel.org \
--cc=linux-rdma@vger.kernel.org \
/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).