[CFT/Review] net byte order for AF_INET

[Adrian Chadd]

On 12 October 2012 08:47, Gleb Smirnoff <glebius at freebsd.org> wrote:
> On Fri, Oct 12, 2012 at 04:46:40PM +0400, Gleb Smirnoff wrote:
> T>   Latest version of patch for further review and testing
> T> Changelog:
> T>  - Fixed TCP checksums
> T>  - Added comment about raw sockets byte ordering.
> T>  - More explicit htons(0), when assigning ip_off field.

I've just eyeballed the patch again:

* You've patched SCTP and IGMP - have you done any SCTP and IGMP testing at all?
* This kind of stuff almost begs for some kind of automated test suite
for testing IPv4, IPv6, TCP/UDP/ICMP, IGMP, SCTP, all the tunneling
stuff - is there anything out there like this? I know of the IPv6 test
suites that exist; what about being able to regression test the other
stuff?

Also whilst I'm nitpicking - do you think there's any performance
issues that may creep up? Remember that "performance issues" to me
don't necessarily mean "on a current generation intel", but mean "all
those cache starved ARM/MIPS/PPC/Atom boards out there that aren't
natively in network byte order." Making everything use network byte
order throughout the stack is nice for read-only packet work and nice
for cache-happy i386s, but what about the rest of the world?

(Don't get me wrong, I think this tidy-up is very nice and maybe quite
needed, I just wonder what other unknown magic is hiding behind the
existing code..)



Adrian