Re: [PATCH v3 1/6] net: pad skb data and shinfo as a whole rather than individually

[Eric Dumazet <eric.dumazet@gmail.com>]

Le mercredi 25 janvier 2012 à 12:27 +0000, Ian Campbell a écrit :
> This reduces the minimum overhead required for this allocation such that the
> shinfo can be grown in the following patch without overflowing 2048 bytes for a
> 1500 byte frame.
> 
> Reducing this overhead while also growing the shinfo means that sometimes the
> tail end of the data can end up in the same cache line as the beginning of the
> shinfo. Specifically in the case of the 64 byte cache lines on a 64 bit system
> the first 8 bytes of shinfo can overlap the tail cacheline of the data. In many
> cases the allocation slop means that there is no overlap.
> 
> Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: Eric Dumazet <eric.dumazet@gmail.com>
> ---

Hmm... you missed build_skb() and all places we use
SKB_DATA_ALIGN(sizeof(struct skb_shared_info))

(for example in some drivers)

If you want to see possible performance impact of your changes, see
commit e52fcb2462ac (bnx2x: uses build_skb() in receive path), and
expect a drop from 820.000 pps to 720.000 pps