netdev - Re: [PATCH] net: increase skb->users instead of skb

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1292510518.2883.207.camel@edumazet-laptop>
Date:	Thu, 16 Dec 2010 15:41:58 +0100
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Junchang Wang <junchangwang@...il.com>
Cc:	Changli Gao <xiaosuo@...il.com>,
	"David S. Miller" <davem@...emloft.net>,
	Tom Herbert <therbert@...gle.com>,
	Jiri Pirko <jpirko@...hat.com>,
	Fenghua Yu <fenghua.yu@...el.com>,
	Xinan Tang <xinan.tang@...el.com>, netdev@...r.kernel.org
Subject: Re: [PATCH] net: increase skb->users instead of skb_clone()

Le jeudi 16 décembre 2010 à 22:31 +0800, Junchang Wang a écrit :
> On Thu, Dec 16, 2010 at 10:18 PM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> >
> > Yes, and no.
> >
> > Consider the case you have one receiver.
> >
> > Packet given after Changli patch wont be shared, so packet_rcv wont
> > clone it : Thats a win. Only one skb_clone() done instead of two.
> >
> > Consider case with 2 receivers :
> >
> > First time we call packet_rcv, packet is shared (because we call
> > deliver_skb(), so packet_rcv clones it. Normal situation, we really need
> > to clone it.
> 
> Got it. Thanks.
> 
> >
> > Second time, we give a non shared packet : Thats a win over previous
> > situation.
> >
> But, if we have N receivers, we get only the last one win - the first N-1 will
> call deliver_skb().
> 

Yes, but you want to, because each receiver has to make a private copy
of the skb.

The big win is that if packet if filtered out (not accepted by the
socket filter), you end with no extra skb_clone() at all.

Say you have 8 receivers, with a filter matching some hash/cpu, and only
one af_packet socket will take the message.

Before patch : 8 skb_clones()

After patch : one skb_clone()

If I undertood patch intent ;)



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html