svn commit: r338257 - head/sys/netinet6

[Bjoern A. Zeeb]

On 23 Aug 2018, at 16:54, Bjoern A. Zeeb wrote:

> Author: bz
> Date: Thu Aug 23 16:54:22 2018
> New Revision: 338257
> URL: https://svnweb.freebsd.org/changeset/base/338257
>
> Log:
>   MFp4 bz_ipv6_fast:
>
>   Migrate udp6_send() v4mapped code to udp6_output() saving us a 
> re-lock and
>   further simplifying the address-family handling code by eliminating
>   AF_INET checks and almost all v4mapped handling right after the 
> start
>   as cases could actually not happen anymore.
>
>   Rework output path locking similar to UDP4 allowing for better
>   parallelism (see r222488, and later versions).
>
>   Sponsored by: The FreeBSD Foundation (2012)
>   Sponsored by: iXsystems (2012)
>   Differential Revision:	https://reviews.freebsd.org/D3721
>
> Modified:
>   head/sys/netinet6/udp6_usrreq.c

Yes, this was from a perforce branch (originally) but people might 
remember seeing it on net at .  Not that anyone but me really cared so far.

In case you wonder what the difference once was, there was a local 
benchmark case here showing the difference:
https://people.freebsd.org/~bz/bench/20120408-diff_pps-pps.html   (there 
are some more on those pages or in my old BSDCan presentation).

This latest update also includes the epoch(9) changes and should finally 
get UDP6 on par with UDP4 and has addressed all but one comment from the 
old review.

Happy UDP-encaping or QUICing or ..


/bz