From: Herbert Xu <herbert@gondor.hengli.com.au>
To: "Xin, Xiaohui" <xiaohui.xin@intel.com>
Cc: "Dong, Eddie" <eddie.dong@intel.com>,
Stephen Hemminger <shemminger@vyatta.com>,
"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
"kvm@vger.kernel.org" <kvm@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"mst@redhat.com" <mst@redhat.com>,
"mingo@elte.hu" <mingo@elte.hu>,
"davem@davemloft.net" <davem@davemloft.net>,
"jdike@linux.intel.com" <jdike@linux.intel.com>
Subject: Re: [RFC PATCH v7 01/19] Add a new structure for skb buffer from external.
Date: Sat, 3 Jul 2010 17:12:37 +0800 [thread overview]
Message-ID: <20100703091237.GA6632@gondor.apana.org.au> (raw)
In-Reply-To: <F2E9EB7348B8264F86B6AB8151CE2D7916B0129230@shsmsx502.ccr.corp.intel.com>
On Mon, Jun 28, 2010 at 05:56:07PM +0800, Xin, Xiaohui wrote:
>
> >OK, if I understand you correctly then I don't think have a
> >problem. With your current patch-set you have exactly the same
> >situation when the skb->data is reallocated as a kernel buffer.
>
> When will skb->data to be reallocated? May you point me the code path?
Anything that calls pskb_expand_head.
> >This is OK because as you correctly argue, it is a rare situation.
> >
> >With my proposal you will need to get this extra external buffer
> >in even less cases, because you'd only need to do it if the skb
> >head grows, which only happens if it becomes encapsulated.
> >So let me explain it in a bit more detail:
> >
> >Our packet starts out as a purely non-linear skb, i.e., skb->head
> >contains nothing and all the page frags come from the guest.
> >
> >During host processing we may pull data into skb->head but the
> >first frag will remain unless we pull all of it. If we did do
> >that then you would have a free external buffer anyway.
> >
> >Now in the common case the header may be modified or pulled, but
> >it very rarely grows. So you can just copy the header back into
> >the first frag just before we give it to the guest.
> >
> Since the data is still there, so recompute the page offset and size is ok, right?
Right, you just move the page offset back and increase the size.
However, to do this safely we'd need to have a way of knowing
whether the skb head has been modified.
It may well turn out to be just as effective to do something like
if (memcmp(skb->data, page frag head, skb_headlen))
memcpy(page frag head, skb->data, skb_headlen)
As the page frag head should be in cache since it would've been
used to populate skb->data.
It'd be good to run some benchmarks with this to see whether
adding a bit to sk_buff to avoid the memcmp is worth it or not.
Cheers,
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
next prev parent reply other threads:[~2010-07-03 9:12 UTC|newest]
Thread overview: 63+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-05 10:14 xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 02/19] Add a new struct for device to manipulate external buffer xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 03/19] Export 2 func for device to assign/deassign new strucure xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 04/19] Add a ndo_mp_port_prep pointer to net_device_ops xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 05/19] Add a function make external buffer owner to query capability xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 06/19] Add a function to indicate if device use external buffer xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 07/19] Add interface to get external buffers xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 08/19] Make __alloc_skb() to get external buffer xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 09/19] Ignore room skb_reserve() when device is using " xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 10/19] Don't do skb recycle, if device use " xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 11/19] Use callback to deal with skb_release_data() specially xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 12/19] Add a hook to intercept external buffers from NIC driver xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 13/19] To skip GRO if buffer is external currently xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 14/19] Add header file for mp device xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 15/19] Add basic funcs and ioctl to " xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 16/19] Manipulate external buffers in " xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 17/19] Export proto_ops to vhost-net driver xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 18/19] Add a kconfig entry and make entry for mp device xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 19/19] Provides multiple submits and asynchronous notifications xiaohui.xin
2010-06-05 10:14 ` [RFC PATCH v7 00/19] Provide a zero-copy method on KVM virtio-net xiaohui.xin
2010-06-05 14:56 ` [RFC PATCH v7 11/19] Use callback to deal with skb_release_data() specially Eric Dumazet
2010-06-09 7:30 ` Xin, Xiaohui
2010-06-05 14:53 ` [RFC PATCH v7 08/19] Make __alloc_skb() to get external buffer Eric Dumazet
2010-06-09 7:34 ` Xin, Xiaohui
2010-06-05 14:51 ` [RFC PATCH v7 03/19] Export 2 func for device to assign/deassign new strucure Eric Dumazet
2010-06-06 23:13 ` [RFC PATCH v7 01/19] Add a new structure for skb buffer from external Stephen Hemminger
2010-06-07 7:51 ` Andi Kleen
2010-06-07 8:17 ` Mitchell Erblich
2010-06-09 9:22 ` Xin, Xiaohui
2010-06-09 8:48 ` Xin, Xiaohui
2010-06-08 5:27 ` Herbert Xu
2010-06-09 9:54 ` Xin, Xiaohui
2010-06-11 5:21 ` Herbert Xu
2010-06-12 9:31 ` Xin, Xiaohui
2010-06-13 8:58 ` Xin, Xiaohui
2010-06-17 11:21 ` Herbert Xu
2010-06-18 5:26 ` Xin, Xiaohui
2010-06-18 5:59 ` Herbert Xu
2010-06-18 7:14 ` Xin, Xiaohui
2010-06-18 7:45 ` Herbert Xu
2010-06-20 10:06 ` Michael S. Tsirkin
2010-06-20 10:32 ` Herbert Xu
2010-06-20 10:39 ` Michael S. Tsirkin
2010-06-20 11:02 ` Herbert Xu
2010-06-20 11:11 ` Michael S. Tsirkin
2010-06-20 11:36 ` Herbert Xu
2010-06-20 11:47 ` Michael S. Tsirkin
2010-06-20 11:59 ` Herbert Xu
2010-06-20 12:48 ` Michael S. Tsirkin
2010-06-20 15:19 ` Ben Hutchings
2010-06-23 8:09 ` Dong, Eddie
2010-06-23 9:52 ` Herbert Xu
2010-06-23 10:05 ` Dong, Eddie
2010-06-24 10:08 ` Herbert Xu
2010-06-25 1:03 ` Dong, Eddie
2010-06-25 11:06 ` Michael S. Tsirkin
2010-06-27 6:14 ` Herbert Xu
2010-06-28 9:56 ` Xin, Xiaohui
2010-06-28 10:00 ` Michael S. Tsirkin
2010-07-03 9:12 ` Herbert Xu [this message]
2010-06-25 2:07 ` Xin, Xiaohui
2010-06-17 11:20 ` Herbert Xu
2010-06-09 8:29 ` Xin, Xiaohui
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=20100703091237.GA6632@gondor.apana.org.au \
--to=herbert@gondor.hengli.com.au \
--cc=davem@davemloft.net \
--cc=eddie.dong@intel.com \
--cc=jdike@linux.intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=mst@redhat.com \
--cc=netdev@vger.kernel.org \
--cc=shemminger@vyatta.com \
--cc=xiaohui.xin@intel.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
Powered by JetHome