From: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Daniel Phillips <phillips@phunq.net>,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-fsdevel@vger.kernel.org,
Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Subject: Re: Distributed storage.
Date: Fri, 3 Aug 2007 17:49:44 +0400 [thread overview]
Message-ID: <20070803134943.GA21221@2ka.mipt.ru> (raw)
In-Reply-To: <1186144072.11797.55.camel@lappy>
On Fri, Aug 03, 2007 at 02:27:52PM +0200, Peter Zijlstra (peterz@infradead.org) wrote:
> On Fri, 2007-08-03 at 14:57 +0400, Evgeniy Polyakov wrote:
>
> > For receiving situation is worse, since system does not know in advance
> > to which socket given packet will belong to, so it must allocate from
> > global pool (and thus there must be independent global reserve), and
> > then exchange part of the socket's reserve to the global one (or just
> > copy packet to the new one, allocated from socket's reseve is it was
> > setup, or drop it otherwise). Global independent reserve is what I
> > proposed when stopped to advertise network allocator, but it seems that
> > it was not taken into account, and reserve was always allocated only
> > when system has serious memory pressure in Peter's patches without any
> > meaning for per-socket reservation.
>
> This is not true. I have a global reserve which is set-up a priori. You
> cannot allocate a reserve when under pressure, that does not make sense.
I probably did not cut enough details - my main position is to allocate
per socket reserve from socket's queue, and copy data there from main
reserve, all of which are allocated either in advance (global one) or
per sockoption, so that there would be no fairness issues what to mark
as special and what to not.
Say we have a page per socket, each socket can assign a reserve for
itself from own memory, this accounts both tx and rx side. Tx is not
interesting, it is simple, rx has global reserve (always allocated on
startup or sometime way before reclaim/oom)where data is originally
received (including skb, shared info and whatever is needed, page is
just an exmaple), then it is copied into per-socket reserve and reused
for the next packet. Having per-socket reserve allows to have progress
in any situation not only in cases where single action must be
received/processed, and allows to be completely fair for all users, but
not only special sockets, thus admin for example would be allowed to
login, ipsec would work and so on...
--
Evgeniy Polyakov
next prev parent reply other threads:[~2007-08-03 13:50 UTC|newest]
Thread overview: 86+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-07-31 17:13 Distributed storage Evgeniy Polyakov
2007-08-02 21:08 ` Daniel Phillips
2007-08-03 10:26 ` Evgeniy Polyakov
2007-08-03 10:57 ` Evgeniy Polyakov
2007-08-03 12:27 ` Peter Zijlstra
2007-08-03 13:49 ` Evgeniy Polyakov [this message]
2007-08-03 14:53 ` Peter Zijlstra
2007-08-03 19:48 ` Daniel Phillips
2007-08-03 19:41 ` Daniel Phillips
2007-08-04 1:19 ` Daniel Phillips
2007-08-04 16:37 ` Evgeniy Polyakov
2007-08-05 8:04 ` Daniel Phillips
2007-08-05 15:08 ` Evgeniy Polyakov
2007-08-05 21:23 ` Daniel Phillips
2007-08-06 8:25 ` Evgeniy Polyakov
2007-08-07 12:05 ` Jens Axboe
2007-08-07 18:24 ` Daniel Phillips
2007-08-07 20:55 ` Jens Axboe
2007-08-08 9:54 ` Block device throttling [Re: Distributed storage.] Evgeniy Polyakov
2007-08-08 10:17 ` [1/1] " Evgeniy Polyakov
2007-08-08 13:28 ` Evgeniy Polyakov
2007-08-12 23:16 ` Daniel Phillips
2007-08-13 8:18 ` Evgeniy Polyakov
2007-08-27 21:57 ` Daniel Phillips
2007-08-28 9:35 ` Evgeniy Polyakov
2007-08-28 17:27 ` Daniel Phillips
2007-08-28 17:54 ` Evgeniy Polyakov
2007-08-28 21:08 ` Daniel Phillips
2007-08-29 8:53 ` Evgeniy Polyakov
2007-08-30 23:20 ` Daniel Phillips
2007-08-31 17:33 ` Evgeniy Polyakov
2007-08-31 21:41 ` Alasdair G Kergon
2007-09-02 4:42 ` Daniel Phillips
2007-08-13 5:22 ` Daniel Phillips
2007-08-13 5:36 ` Daniel Phillips
2007-08-13 6:44 ` Daniel Phillips
2007-08-13 8:14 ` Evgeniy Polyakov
2007-08-13 11:04 ` Daniel Phillips
2007-08-13 12:04 ` Evgeniy Polyakov
2007-08-13 12:18 ` Daniel Phillips
2007-08-13 12:24 ` Evgeniy Polyakov
2007-08-13 8:23 ` Evgeniy Polyakov
2007-08-13 11:18 ` Daniel Phillips
2007-08-13 12:18 ` Evgeniy Polyakov
2007-08-13 13:04 ` Daniel Phillips
2007-08-14 8:46 ` Evgeniy Polyakov
2007-08-14 11:13 ` Daniel Phillips
2007-08-14 11:30 ` Evgeniy Polyakov
2007-08-14 11:35 ` Daniel Phillips
2007-08-14 11:50 ` Evgeniy Polyakov
2007-08-14 12:32 ` Daniel Phillips
2007-08-14 12:46 ` Evgeniy Polyakov
2007-08-14 12:54 ` Daniel Phillips
2007-08-12 23:36 ` Distributed storage Daniel Phillips
2007-08-13 7:28 ` Jens Axboe
2007-08-13 7:45 ` Jens Axboe
2007-08-13 9:08 ` Daniel Phillips
2007-08-13 9:13 ` Jens Axboe
2007-08-13 9:55 ` Daniel Phillips
2007-08-13 10:06 ` Jens Axboe
2007-08-13 10:15 ` Daniel Phillips
2007-08-13 10:22 ` Jens Axboe
2007-08-13 10:32 ` Daniel Phillips
2007-08-13 9:18 ` Evgeniy Polyakov
2007-08-13 10:12 ` Daniel Phillips
2007-08-13 11:03 ` Evgeniy Polyakov
2007-08-13 11:45 ` Daniel Phillips
2007-08-13 8:59 ` Daniel Phillips
2007-08-13 9:12 ` Jens Axboe
2007-08-13 23:27 ` Daniel Phillips
2007-08-03 4:09 ` Mike Snitzer
2007-08-03 10:42 ` Evgeniy Polyakov
2007-08-04 0:49 ` Daniel Phillips
2007-08-03 5:04 ` Manu Abraham
2007-08-03 10:44 ` Evgeniy Polyakov
2007-08-04 2:51 ` Dave Dillow
2007-08-04 3:44 ` Manu Abraham
2007-08-04 17:03 ` Evgeniy Polyakov
2007-08-28 17:19 ` Evgeniy Polyakov
2007-08-04 0:41 ` Daniel Phillips
2007-08-04 16:44 ` Evgeniy Polyakov
2007-08-05 8:06 ` Daniel Phillips
2007-08-05 15:01 ` Evgeniy Polyakov
2007-08-05 21:35 ` Daniel Phillips
2007-08-06 8:28 ` Evgeniy Polyakov
-- strict thread matches above, loose matches on Subject: below --
2008-08-27 16:07 Distributed STorage Evgeniy Polyakov
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=20070803134943.GA21221@2ka.mipt.ru \
--to=johnpol@2ka.mipt.ru \
--cc=acme@ghostprotocols.net \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=phillips@phunq.net \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.