On Tue, Nov 24, 2015 at 10:37 AM, David Miller [off-list ref] wrote:

From: Hannes Frederic Sowa <redacted>
Date: Tue, 24 Nov 2015 18:43:35 +0100

quoted

On Tue, Nov 24, 2015, at 18:32, Tom Herbert wrote:

quoted

As you said this in only feedback and nobody is forcing anyone to do
anything. But encouraging HW vendors to provide generic mechanisms so
that your users can use whatever protocol they want is the exact
_opposite_ of punishing users, this is very much a pro-user direction.

Some users will suffer worse performance in case we don't correctly set
ip_summed for a specific protocol before we do the copy operations from
user space into skbs but if they are always done in the driver.

Your concern presumes that looking backwards is as important as looking
forward.

We want to simplify things _and_ move away from protocol specific
csums, and if some old crufty hardware based systems pay some performance
cost for this I say so be it.

So this is not a valid argument against Tom's changes in my mind.

And for that matter these arguments have nothing to do with these UDP
encapsulation patches at all, they seem to be directed to the patches
to eliminate NETIF_F_IP{V6}_CSUM so please post on that thread.