From: Stephen Hemminger <shemminger@osdl.org>
To: Andi Kleen <ak@suse.de>
Cc: "David S. Miller" <davem@davemloft.net>,
mpm@selenic.com, sri@us.ibm.com, ak@suse.de,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
Subject: Re: [RFC][PATCH 0/3] TCP/IP Critical socket communication mechanism
Date: Wed, 14 Dec 2005 22:06:46 -0800 [thread overview]
Message-ID: <20051214220646.6952f936@localhost.localdomain> (raw)
In-Reply-To: <20051215054245.GD18862@brahms.suse.de>
On Thu, 15 Dec 2005 06:42:45 +0100
Andi Kleen <ak@suse.de> wrote:
> On Wed, Dec 14, 2005 at 08:30:23PM -0800, David S. Miller wrote:
> > From: Matt Mackall <mpm@selenic.com>
> > Date: Wed, 14 Dec 2005 19:39:37 -0800
> >
> > > I think we need a global receive pool and per-socket send pools.
> >
> > Mind telling everyone how you plan to make use of the global receive
> > pool when the allocation happens in the device driver and we have no
> > idea which socket the packet is destined for? What should be done for
>
> In theory one could use multiple receive queue on intelligent enough
> NIC with the NIC distingushing the sockets.
>
> But that would be still a nasty "you need advanced hardware FOO to avoid
> subtle problem Y" case. Also it would require lots of driver hacking.
>
> And most NICs seem to have limits on the size of the socket tables for this, which
> means you would end up in a "only N sockets supported safely" situation,
> with N likely being quite small on common hardware.
>
> I think the idea of the original poster was that just freeing non critical packets
> after a short time again would be good enough, but I'm a bit sceptical
> on that.
>
> > I truly dislike these patches being discussed because they are a
> > complete hack, and admittedly don't even solve the problem fully. I
>
> I agree.
>
> > I think GFP_ATOMIC memory pools are more powerful than they are given
> > credit for. There is nothing preventing the implementation of dynamic
>
> Their main problem is that they are used too widely and in a lot
> of situations that aren't really critical.
Most of the use of GFP_ATOMIC is by stuff that could fail but can't
sleep waiting for memory. How about adding a GFP_NORMAL for allocations
while holding a lock.
#define GFP_NORMAL (__GFP_NOMEMALLOC)
Then get people to change the unneeded GFP_ATOMIC's to GFP_NORMAL in
places where the error paths are reasonable.
next prev parent reply other threads:[~2005-12-15 6:04 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-12-14 9:12 [RFC][PATCH 0/3] TCP/IP Critical socket communication mechanism Sridhar Samudrala
2005-12-14 9:22 ` Andi Kleen
2005-12-14 17:55 ` Sridhar Samudrala
2005-12-14 18:41 ` Andi Kleen
2005-12-14 19:20 ` David Stevens
2005-12-15 3:39 ` Matt Mackall
2005-12-15 4:30 ` David S. Miller
2005-12-15 5:02 ` Matt Mackall
2005-12-15 5:23 ` David S. Miller
2005-12-15 5:48 ` Matt Mackall
2005-12-15 5:53 ` Nick Piggin
2005-12-15 5:56 ` Stephen Hemminger
2005-12-15 8:44 ` David Stevens
2005-12-15 8:58 ` David S. Miller
2005-12-15 9:27 ` David Stevens
2005-12-15 5:42 ` Andi Kleen
2005-12-15 6:06 ` Stephen Hemminger [this message]
2005-12-15 7:37 ` Sridhar Samudrala
2005-12-15 8:21 ` David S. Miller
2005-12-15 8:35 ` Arjan van de Ven
2005-12-15 8:55 ` [RFC] Fine-grained memory priorities and PI Kyle Moffett
2005-12-15 9:04 ` Andi Kleen
2005-12-15 12:51 ` Kyle Moffett
2005-12-15 13:31 ` Andi Kleen
2005-12-15 12:45 ` Con Kolivas
2005-12-15 12:58 ` Kyle Moffett
2005-12-15 13:02 ` Con Kolivas
2005-12-16 2:09 ` [RFC][PATCH 0/3] TCP/IP Critical socket communication mechanism Sridhar Samudrala
2005-12-16 17:48 ` Stephen Hemminger
2005-12-16 18:38 ` Sridhar Samudrala
2005-12-21 9:11 ` Pavel Machek
2005-12-21 9:39 ` David Stevens
2005-12-14 20:16 ` Jesper Juhl
2005-12-14 20:25 ` Ben Greear
2005-12-14 20:49 ` James Courtier-Dutton
2005-12-14 21:55 ` Sridhar Samudrala
2005-12-14 22:09 ` James Courtier-Dutton
2005-12-14 22:39 ` Ben Greear
2005-12-14 23:42 ` Sridhar Samudrala
2005-12-15 1:54 ` Mitchell Blank Jr
2005-12-15 11:38 ` James Courtier-Dutton
2005-12-15 11:47 ` Arjan van de Ven
2005-12-15 13:00 ` jamal
2005-12-15 13:07 ` Arjan van de Ven
2005-12-15 13:32 ` jamal
[not found] <5jUjW-8nu-7@gated-at.bofh.it>
[not found] ` <5jWYp-3K1-19@gated-at.bofh.it>
[not found] ` <5jXhZ-4kj-19@gated-at.bofh.it>
2005-12-16 8:35 ` Bodo Eggert
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=20051214220646.6952f936@localhost.localdomain \
--to=shemminger@osdl.org \
--cc=ak@suse.de \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=mpm@selenic.com \
--cc=netdev@vger.kernel.org \
--cc=sri@us.ibm.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).