From: David Miller <davem@davemloft.net>
To: andi@firstfloor.org
Cc: hadi@cyberus.ca, shemminger@linux-foundation.org,
jeff@garzik.org, johnpol@2ka.mipt.ru,
herbert@gondor.apana.org.au, gaagaan@gmail.com,
Robert.Olsson@data.slu.se, netdev@vger.kernel.org,
rdreier@cisco.com, peter.p.waskiewicz.jr@intel.com,
mcarlson@broadcom.com, jagana@us.ibm.com,
general@lists.openfabrics.org, mchan@broadcom.com, tgraf@suug.ch,
randy.dunlap@oracle.com, sri@us.ibm.com, kaber@trash.net
Subject: Re: [ofa-general] Re: [PATCH 2/3][NET_BATCH] net core use batching
Date: Wed, 10 Oct 2007 02:25:50 -0700 (PDT) [thread overview]
Message-ID: <20071010.022550.21928751.davem@davemloft.net> (raw)
In-Reply-To: <20071010091644.GA9807@one.firstfloor.org>
From: Andi Kleen <andi@firstfloor.org>
Date: Wed, 10 Oct 2007 11:16:44 +0200
> > A 256 entry TX hw queue fills up trivially on 1GB and 10GB, but if you
>
> With TSO really?
Yes.
> > increase the size much more performance starts to go down due to L2
> > cache thrashing.
>
> Another possibility would be to consider using cache avoidance
> instructions while updating the TX ring (e.g. write combining
> on x86)
The chip I was working with at the time (UltraSPARC-IIi) compressed
all the linear stores into 64-byte full cacheline transactions via
the store buffer.
It's true that it would allocate in the L2 cache on a miss, which
is different from your suggestion.
In fact, such a thing might not pan out well, because most of the time
you write a single descriptor or two, and that isn't a full cacheline,
which means a read/modify/write is the only coherent way to make such
a write to RAM.
Sure you could batch, but I'd rather give the chip work to do unless
I unequivocably knew I'd have enough pending to fill a cacheline's
worth of descriptors. And since you suggest we shouldn't queue in
software... :-)
next prev parent reply other threads:[~2007-10-10 9:25 UTC|newest]
Thread overview: 86+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-10-08 18:26 [PATCH 2/3][NET_BATCH] net core use batching jamal
2007-10-08 19:46 ` Waskiewicz Jr, Peter P
2007-10-08 20:48 ` jamal
2007-10-08 21:26 ` [ofa-general] " David Miller
2007-10-08 22:34 ` jamal
2007-10-08 22:36 ` [ofa-general] " Waskiewicz Jr, Peter P
2007-10-08 22:33 ` Waskiewicz Jr, Peter P
2007-10-08 23:40 ` jamal
2007-10-09 1:13 ` Jeff Garzik
2007-10-09 1:41 ` [ofa-general] " David Miller
2007-10-09 2:01 ` Herbert Xu
2007-10-09 2:03 ` Herbert Xu
2007-10-09 2:04 ` Herbert Xu
2007-10-09 2:15 ` jamal
2007-10-09 2:16 ` Herbert Xu
2007-10-09 2:19 ` [ofa-general] " jamal
2007-10-09 2:20 ` Herbert Xu
2007-10-09 2:45 ` [ofa-general] " David Miller
2007-10-09 2:43 ` David Miller
2007-10-09 2:46 ` Herbert Xu
2007-10-09 2:12 ` [ofa-general] " Jeff Garzik
2007-10-09 2:46 ` David Miller
2007-10-09 18:48 ` [ofa-general] " Waskiewicz Jr, Peter P
2007-10-09 19:04 ` Jeff Garzik
2007-10-09 19:07 ` Waskiewicz Jr, Peter P
2007-10-09 2:14 ` [ofa-general] " jamal
2007-10-09 2:16 ` Herbert Xu
2007-10-09 2:47 ` [ofa-general] " David Miller
2007-10-09 16:51 ` Andi Kleen
2007-10-09 18:22 ` Stephen Hemminger
2007-10-09 18:30 ` Andi Kleen
2007-10-09 20:43 ` David Miller
2007-10-09 20:53 ` Stephen Hemminger
2007-10-09 21:22 ` David Miller
2007-10-09 21:56 ` jamal
2007-10-10 0:04 ` David Miller
2007-10-10 0:37 ` Andi Kleen
2007-10-10 0:50 ` David Miller
2007-10-10 9:16 ` Andi Kleen
2007-10-10 9:25 ` David Miller [this message]
2007-10-10 10:23 ` Andi Kleen
2007-10-10 10:44 ` David Miller
2007-10-10 13:08 ` jamal
2007-10-10 22:37 ` David Miller
2007-10-10 15:35 ` Waskiewicz Jr, Peter P
2007-10-10 16:02 ` Andi Kleen
2007-10-10 16:42 ` Waskiewicz Jr, Peter P
2007-10-10 9:53 ` Herbert Xu
2007-10-12 16:08 ` Brandeburg, Jesse
2007-10-12 17:05 ` Stephen Hemminger
2007-10-12 18:29 ` Andi Kleen
2007-10-12 18:27 ` Andi Kleen
2007-10-10 16:02 ` Bill Fink
2007-10-10 22:53 ` David Miller
2007-10-11 6:52 ` Krishna Kumar2
2007-10-09 1:31 ` Jeff Garzik
2007-10-09 10:58 ` [ofa-general] " Krishna Kumar2
2007-10-09 11:02 ` David Miller
2007-10-09 11:20 ` [ofa-general] " Krishna Kumar2
2007-10-09 11:21 ` Krishna Kumar2
2007-10-09 11:24 ` David Miller
2007-10-09 12:44 ` [ofa-general] " Jeff Garzik
2007-10-09 12:55 ` Herbert Xu
2007-10-09 13:00 ` Jeff Garzik
2007-10-09 20:14 ` David Miller
2007-10-09 20:20 ` [ofa-general] " Jeff Garzik
2007-10-09 21:25 ` David Miller
2007-10-09 20:22 ` [ofa-general] " Roland Dreier
2007-10-09 20:51 ` David Miller
2007-10-09 21:40 ` Roland Dreier
2007-10-09 22:44 ` [ofa-general] " Roland Dreier
2007-10-09 22:46 ` [ofa-general] [PATCH 1/4] IPoIB: Fix unused variable warning Roland Dreier
2007-10-09 22:47 ` [ofa-general] [PATCH 2/4] ibm_emac: Convert to use napi_struct independent of struct net_device Roland Dreier
2007-10-09 22:47 ` [PATCH 3/4] ibm_new_emac: Nuke SET_MODULE_OWNER() use Roland Dreier
2007-10-09 22:48 ` [PATCH 4/4] ibm_emac: Convert to use napi_struct independent of struct net_device Roland Dreier
2007-10-09 22:51 ` [ofa-general] " Roland Dreier
2007-10-09 23:17 ` [PATCH 1/4] IPoIB: Fix unused variable warning David Miller
2007-10-10 0:32 ` Jeff Garzik
2007-10-10 0:47 ` [ofa-general] " Jeff Garzik
-- strict thread matches above, loose matches on Subject: below --
2007-10-08 13:17 [PATCH 2/3][NET_BATCH] net core use batching jamal
2007-10-09 3:09 ` [ofa-general] " Krishna Kumar2
2007-10-09 13:10 ` jamal
2007-09-14 9:00 [PATCH 0/10 REV5] Implement skb batching and support in IPoIB/E1000 Krishna Kumar
2007-09-16 23:17 ` [ofa-general] " David Miller
2007-09-17 0:29 ` jamal
2007-09-23 17:53 ` [PATCHES] TX batching jamal
2007-09-23 17:56 ` [ofa-general] [PATCH 1/4] [NET_SCHED] explict hold dev tx lock jamal
2007-09-23 17:58 ` [ofa-general] [PATCH 2/4] [NET_BATCH] Introduce batching interface jamal
2007-09-23 18:00 ` [PATCH 3/4][NET_BATCH] net core use batching jamal
2007-09-30 18:52 ` [ofa-general] [PATCH 2/3][NET_BATCH] " jamal
2007-10-01 4:11 ` Bill Fink
2007-10-01 13:30 ` jamal
2007-10-02 4:25 ` [ofa-general] " Bill Fink
2007-10-02 13:20 ` jamal
2007-10-03 5:29 ` [ofa-general] " Bill Fink
2007-10-01 10:42 ` Patrick McHardy
2007-10-01 13:21 ` jamal
2007-10-08 5:03 ` Krishna Kumar2
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=20071010.022550.21928751.davem@davemloft.net \
--to=davem@davemloft.net \
--cc=Robert.Olsson@data.slu.se \
--cc=andi@firstfloor.org \
--cc=gaagaan@gmail.com \
--cc=general@lists.openfabrics.org \
--cc=hadi@cyberus.ca \
--cc=herbert@gondor.apana.org.au \
--cc=jagana@us.ibm.com \
--cc=jeff@garzik.org \
--cc=johnpol@2ka.mipt.ru \
--cc=kaber@trash.net \
--cc=mcarlson@broadcom.com \
--cc=mchan@broadcom.com \
--cc=netdev@vger.kernel.org \
--cc=peter.p.waskiewicz.jr@intel.com \
--cc=randy.dunlap@oracle.com \
--cc=rdreier@cisco.com \
--cc=shemminger@linux-foundation.org \
--cc=sri@us.ibm.com \
--cc=tgraf@suug.ch \
/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.