On Fri, 2017-01-27 at 15:46 +0000, David Laight wrote: > Reads code ... > My confusion is that the call is specifying the number of EXTRA bytes of > head/tail > room rather than the number of bytes needed.
And the fact that @size is changed in existing code (so not visible in patch diff) to size = SKB_WITH_OVERHEAD(ksize(data)); > I was thinking that pskb_expand_head(skb, 16, -16, ...) could be implemented > (mostly) with memmove(). Yes, might worth doing that if one day a caller tries that ;) In the meantime, not worth adding code that wont be reached. Thanks.
