netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jarek Poplawski <jarkao2@gmail.com>
To: David Miller <davem@davemloft.net>
Cc: dada1@cosmosbay.com, netdev@vger.kernel.org
Subject: Re: [RFC,PATCH] loopback: calls netif_receive_skb() instead of netif_rx()
Date: Wed, 27 Feb 2008 07:20:41 +0000	[thread overview]
Message-ID: <20080227072041.GA4000@ff.dom.local> (raw)
In-Reply-To: <20080226.182120.183405235.davem@davemloft.net>

On 27-02-2008 03:21, David Miller wrote:
> From: Eric Dumazet <dada1@cosmosbay.com>
> Date: Thu, 21 Feb 2008 19:51:52 +0100
> 
>> Following patch directly calls netif_receive_skb() and avoids lot of 
>> atomic operations.
>> (atomic_inc(&dev->refcnt), set_and_set_bit(NAPI_STATE_SCHED, &n->state), ...
>>   atomic_dec(&dev->refcnt)...), cache line ping-pongs on device refcnt, 
>> but also softirq overhead.
>>
>> This gives a nice boost on tbench for example (5 % on my machine)
> 
> My only concern is stack usage.
...

I wonder why overloading with net processing is no concern here?
There would be no napi control around this  netif_receive_skb().

Another concern might be a code which depends on softirq context
here (unless it was checked already)?

Regards,
Jarek P.

  reply	other threads:[~2008-02-27  7:19 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-02-21 18:51 [RFC,PATCH] loopback: calls netif_receive_skb() instead of netif_rx() Eric Dumazet
2008-02-21 20:14 ` Daniel Lezcano
2008-02-21 23:19   ` Eric Dumazet
2008-02-22 10:19     ` Daniel Lezcano
2008-02-27  2:21 ` David Miller
2008-02-27  7:20   ` Jarek Poplawski [this message]
2008-02-27  7:23     ` David Miller
2008-02-27  7:34       ` Jarek Poplawski
2008-03-01 10:26   ` Eric Dumazet
2008-03-04  4:55     ` David Miller
2008-03-04  5:15       ` Stephen Hemminger
2008-03-04  6:27       ` Eric Dumazet
2008-03-23 10:29     ` David Miller
2008-03-23 18:48       ` Eric Dumazet
2008-03-23 19:15         ` Andi Kleen
2008-03-29  1:36         ` David Miller
2008-03-29  8:18           ` Eric Dumazet
2008-03-29 23:54             ` David Miller
2008-03-31  6:38               ` Eric Dumazet
2008-03-31  9:48         ` Ingo Molnar
2008-03-31 10:01           ` Eric Dumazet
2008-03-31 10:12             ` Ingo Molnar
2008-04-01  9:19               ` Eric Dumazet
2008-04-03 14:06                 ` Pavel Machek
2008-04-03 16:19                   ` Eric Dumazet
2008-03-31 10:08           ` David Miller
2008-03-31 10:44             ` Ingo Molnar
2008-03-31 11:02               ` David Miller
2008-03-31 11:36                 ` poor network loopback performance and scalability (was: Re: [RFC,PATCH] loopback: calls netif_receive_skb() instead of netif_rx()) Ingo Molnar
2008-04-21  3:24                   ` Herbert Xu
2008-04-21  3:38                     ` poor network loopback performance and scalability David Miller
2008-04-21  8:11                       ` Ingo Molnar
2008-04-21  8:16                         ` David Miller
2008-04-21 10:19                           ` Herbert Xu
2008-04-21 10:22                             ` David Miller

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=20080227072041.GA4000@ff.dom.local \
    --to=jarkao2@gmail.com \
    --cc=dada1@cosmosbay.com \
    --cc=davem@davemloft.net \
    --cc=netdev@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).