From: Eric Blake <eblake@redhat.com>
To: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>,
Wouter Verhelst <w@uter.be>, Alex Bligh <alex@alex.org.uk>,
nbd list <nbd@other.debian.org>,
qemu-devel <qemu-devel@nongnu.org>
Subject: Re: [Qemu-devel] NBD BLOCK_STATUS
Date: Fri, 10 Nov 2017 10:06:55 -0600 [thread overview]
Message-ID: <bf31a83e-7cd6-4f26-595b-3b1a2def883f@redhat.com> (raw)
In-Reply-To: <37193dd7-a0fc-bfa6-3870-ecbcfefac4e5@virtuozzo.com>
[-- Attachment #1: Type: text/plain, Size: 2133 bytes --]
On 11/09/2017 06:42 AM, Vladimir Sementsov-Ogievskiy wrote:
> Hi!
>
> Interesting fact: list/set_meta_context options are per-export,
> so, in the server we should keep context selection per client per export.
>
> And it is possible for client to set contexts for one export and than
> proceed
> to transmission phase with another one.
However, we also documented in the spec that
+ A client MUST NOT send `NBD_CMD_BLOCK_STATUS` unless
+ within the negotiation phase it sent `NBD_OPT_SET_META_CONTEXT`
+ at least once, and the final time it was sent, it referred
+ to the export name that was ultimately selected, the server
+ responded without an error, and returned at least one metadata
+ context.
My take on this is if the client calls:
NBD_OPT_SET_META_CONTEXT (export name "foo")
NBD_OPT_GO (export name "bar")
then the server has no contexts to remember, and the client must not
call NBD_CMD_BLOCK_STATUS. That is, a server is free to track only a
SINGLE list of context ids (namely, the context ids it replied in the
last response to NBD_OPT_SET_META_CONTEXT), and then toss that list on
any further NBD_OPT_SET_META_CONTEXT or if NBD_OPT_EXPORT_NAME/GO does
not select the same export name, rather than having to track multiple
separate per-export lists while waiting for the client to pick which
export to finally go with.
>
> This is not really touched in spec, for example
> "Change the set of active metadata contexts. Issuing this command
> replaces all previously-set metadata contexts"
>
> looks like lacking "for specified export" suffix. May be there are other
> places.
No, it is intentional that it replaces ALL previous contexts (not just
the per-export context), because of the argument that the only context
that matters as a client is the one you use with NBD_OPT_GO, and if you
choose GO with a different export than your last SET_META_CONTEXT, then
you shouldn't be calling NBD_CMD_BLOCK_STATUS.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 619 bytes --]
next prev parent reply other threads:[~2017-11-10 16:07 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-09 12:42 [Qemu-devel] NBD BLOCK_STATUS Vladimir Sementsov-Ogievskiy
2017-11-10 16:06 ` Eric Blake [this message]
2017-11-10 16:51 ` Vladimir Sementsov-Ogievskiy
2017-11-13 14:13 ` Eric Blake
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=bf31a83e-7cd6-4f26-595b-3b1a2def883f@redhat.com \
--to=eblake@redhat.com \
--cc=alex@alex.org.uk \
--cc=nbd@other.debian.org \
--cc=qemu-devel@nongnu.org \
--cc=vsementsov@virtuozzo.com \
--cc=w@uter.be \
/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).