Joerg Sonnenberger wrote:

quoted text
> so_setsockopt:
> Shouldn't valsize be the normal size_t? Does socklen_t really make sense
> here? 

I vaguely remember there was a reason for this socklen_t... maybe the 
original FreeBSD changes. Anyway, I'm thinking it should be unsigned
int, actually, to fit the len parameter to sys_setsockopt() and
sys_getsockopt(), no?

quoted text
> The val pointer should be const.

Fixed.

quoted text
> sockopt_ensure_writeable --> I don't like the name.

I don't mind the naming. Bring up any name you'd like, and if people
agree, I'll change.

quoted text
> If you can now also push the socket options down into ipv4/ipv6 and the
> patch can go without the sockopt_setmbuf/getmbuf, that would be very
> nice :-)

Yes. Unfortunately, some code is *very* hairy about that kinda stuff so
I'm trying to avoid it in the first pass (see, for example, netiso).

quoted text
> Without a full review, this looks like a huge improvement.

Great to hear! :)

Thanks,

-e.