On Thu, Jul 04, 2024 at 11:20:53PM +0200, Stefano Brivio wrote:On Thu, 4 Jul 2024 14:58:28 +1000 David Gibson <david(a)gibson.dropbear.id.au> wrote:Ah! That makes sense once you know.Make the salient points about these various arrays clearer with renames: * udp_l2_iov_sock and udp[46]_l2_mh_sock don't really have anything to do with L2.The original idea behind that 'l2' there was to have the type of destination in the name first, and then the source ('sock').On the other hand, they're actually clearer this way.Right, I think this is more obvious out the gate.Oops, corrected.They are, however, specific to receiving not sending not receiving.I failed to decrypt this one. :)Easy fix, and on consideration I prefer "splice_to" as well. Changed. -- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibsonRename to udp_iov_recv and udp[46]_mh_recv. * udp[46]_l2_iov_tap is redundant - "tap" implies L2 and vice versa. Rename to udp[46]_l2_iov * udp[46]_localname are (for now) pre-populated with the locan address but the more salient point is that these are the destination address for the splice arrays. Rename to udp[46]_splicetoVery slight preference (but not worth a lot of changes, in case): udp[46]_splice_to. To me it's not immediately obvious those are two words otherwise.