From: Rick Jones <rick.jones2@hp.com>
To: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>,
David Miller <davem@davemloft.net>,
netdev <netdev@vger.kernel.org>, Jesse Gross <jesse@nicira.com>,
Tom Herbert <therbert@google.com>,
Yuchung Cheng <ycheng@google.com>
Subject: Re: [PATCH] net: gro: selective flush of packets
Date: Mon, 08 Oct 2012 09:42:27 -0700 [thread overview]
Message-ID: <50730273.4000408@hp.com> (raw)
In-Reply-To: <1349681967.21172.2866.camel@edumazet-glaptop>
On 10/08/2012 12:39 AM, Eric Dumazet wrote:
> On Sun, 2012-10-07 at 07:29 +0200, Eric Dumazet wrote:
>> On Sun, 2012-10-07 at 08:32 +0800, Herbert Xu wrote:
>
>>> Why don't we just always flush everything?
>>
>
>> This is what I tried first, but it lowered performance on several
>> typical workloads.
>>
>> Using this simple heuristic increases performance.
>>
>>
>
> By the way, one of the beauty of GRO is it helps under load to aggregate
> packets and reduce cpu load. People wanting very low latencies should
> probably not use GRO, and if they use it or not, receiving a full 64
> packets batch on a particular NIC makes latencies very unpredictable.
>
> So if we consumed all budget in a napi->poll() handler, its because we
> are under load and we dont really want to cancel GRO aggregation.
Is that actually absolute, or does it depend on GRO aggregation actually
aggregating? In your opening message you talked about how with though
flows GRO is defeated but its overhead remains.
rick jones
next prev parent reply other threads:[~2012-10-08 16:42 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-09-27 12:48 [PATCH net-next 3/3] ipv4: gre: add GRO capability Eric Dumazet
2012-09-27 17:52 ` Jesse Gross
2012-09-27 18:08 ` Eric Dumazet
2012-09-27 18:19 ` Eric Dumazet
2012-09-27 22:03 ` Jesse Gross
2012-09-28 14:04 ` Eric Dumazet
2012-10-01 20:56 ` Jesse Gross
2012-10-05 14:52 ` [RFC] GRO scalability Eric Dumazet
2012-10-05 18:16 ` Rick Jones
2012-10-05 19:00 ` Eric Dumazet
2012-10-05 19:35 ` Rick Jones
2012-10-05 20:06 ` Eric Dumazet
2012-10-08 16:40 ` Rick Jones
2012-10-08 16:59 ` Eric Dumazet
2012-10-08 17:49 ` Rick Jones
2012-10-08 17:55 ` Eric Dumazet
2012-10-08 17:56 ` Eric Dumazet
2012-10-08 18:58 ` [RFC] napi: limit GRO latency Stephen Hemminger
2012-10-08 19:10 ` David Miller
2012-10-08 19:12 ` Stephen Hemminger
2012-10-08 19:30 ` Eric Dumazet
2012-10-08 19:40 ` Stephen Hemminger
2012-10-08 19:46 ` Eric Dumazet
2012-10-08 19:21 ` Eric Dumazet
2012-10-08 18:21 ` [RFC] GRO scalability Rick Jones
2012-10-08 18:28 ` Eric Dumazet
2012-10-06 4:11 ` Herbert Xu
2012-10-06 5:08 ` Eric Dumazet
2012-10-06 5:14 ` Herbert Xu
2012-10-06 6:22 ` Eric Dumazet
2012-10-06 7:00 ` Eric Dumazet
2012-10-06 10:56 ` Herbert Xu
2012-10-06 18:08 ` [PATCH] net: gro: selective flush of packets Eric Dumazet
2012-10-07 0:32 ` Herbert Xu
2012-10-07 5:29 ` Eric Dumazet
2012-10-08 7:39 ` Eric Dumazet
2012-10-08 16:42 ` Rick Jones [this message]
2012-10-08 17:10 ` Eric Dumazet
2012-10-08 18:52 ` David Miller
2012-09-27 22:03 ` [PATCH net-next 3/3] ipv4: gre: add GRO capability Jesse Gross
2012-10-01 21:04 ` 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=50730273.4000408@hp.com \
--to=rick.jones2@hp.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=herbert@gondor.apana.org.au \
--cc=jesse@nicira.com \
--cc=netdev@vger.kernel.org \
--cc=therbert@google.com \
--cc=ycheng@google.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).