netdev - Re: [PATCH 0/2] Get rid of ndo_xmit

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]

Date:	Wed, 27 Aug 2014 07:09:35 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Jesper Dangaard Brouer <brouer@...hat.com>
Cc:	Hannes Frederic Sowa <hannes@...essinduktion.org>,
	David Miller <davem@...emloft.net>, netdev@...r.kernel.org,
	therbert@...gle.com, jhs@...atatu.com, edumazet@...gle.com,
	jeffrey.t.kirsher@...el.com, rusty@...tcorp.com.au,
	dborkman@...hat.com
Subject: Re: [PATCH 0/2] Get rid of ndo_xmit_flush

On Wed, 2014-08-27 at 15:56 +0200, Jesper Dangaard Brouer wrote:

> I do worry a bit, if writing skb->xmit_more in a more cache cold
> scenarios could hurt us (as that cacheline seems to be read mostly),
> but this is not a problem right now because it will always be cache
> hot, as we always clear the entire SKB in alloc.
> 


Yes, but many of us use things like TCP stack and thousand of packets
sitting in a possibly complex qdisc hierarchy before hitting the
device ;)

In these situations, skb is dirtied at TX completion time, but not in
ndo_start_xmit() (where its only read)

In the (qdisc->dequeue() -> ndo_start_xmit()) stage, only the first
cache line of skb is typically dirtied.


--
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