On 11/04/2013 05:10 PM, Vlad Yasevich wrote:

On 11/04/2013 07:11 AM, Daniel Borkmann wrote:

quoted

On 10/30/2013 03:29 PM, Vlad Yasevich wrote:

quoted

On 10/30/2013 06:50 AM, Daniel Borkmann wrote:

[...]

quoted

Daniel

Here is a follow-on idea that might help even more.
What if we put a pointer to skb_checksum_ops() in the skb
somewhere (I was thinking of skb_shinfo).  Then
skb_checksum can simply use the data from there.  This would
allow us to get rid of all the special cases in SCTP that do
checksumming.  We can just set it to partial, set up the right
fields and let HW or SW always do the right thing.

I need to think about this a bit. This would certainly have the
negative side-effect of a higher skb->truesize usage and thus
affecting memory accounting for everyone as we extend
skb_shared_info.

You are talking a single pointer here...  The alternative
is to do a per-protocol table.

Well, that is e.g. 8 more byte per skb, right? ;) I think if we
need these kind of changes in many more places than currently,
then this would make sense indeed.