On 2014-04-18, Henning Brauer <lists-openbsdt...@bsws.de> wrote: > so, what are we doing with this now? > I still want to hide in_cksum_phdr() and kill in_cksum_addword() so that > nobody ever uses that sh*t again. > yes, sk loses is half-baked cksum offload support with this, as > discussed before. > as naddy pointed out there are (at least) two private copies of > in_cksum_phdr in other drivers, that is a seperate issue in my book if > it is an issue at all, that's not an exposed API
And I would prefer all three drivers to handle this in a consistent way. I don't care too much if they all have private copies, make use of in_cksum_phdr() etc., or lose this "half-baked cksum offload support" altogether. Well, having three private copies would be kind of stupid. -- Christian "naddy" Weisgerber na...@mips.inka.de