On 01/21/16 at 06:50P, Gleb Smirnoff wrote: > On Thu, Jan 21, 2016 at 06:43:10PM -0800, hiren panchasara wrote: > h> > h> > Provide new socket option TCP_CCALGOOPT, which stands for TCP > congestion > h> > h> > control algorithm options. The argument is variable length and > is opaque > h> > h> > to TCP, forwarded directly to the algorithm's ctl_output method. > h> > h> > > h> > h> > Provide new includes directory netinet/cc, where algorithm > specific > h> > h> > headers can be installed. > h> > h> > > h> > h> > The new API doesn't yet have any in tree consumers. > h> > h> > > h> > h> > The original code written by lstewart. > h> > h> > Reviewed by: rrs, emax > h> > h> > Sponsored by: Netflix > h> > h> > Differential Revision: https://reviews.freebsd.org/D711 > h> > h> > h> > h> Thanks Lawrence and Glebius for the patch. > h> > h> > h> > h> I know original plan was to have this in earlier but is it possible to > h> > h> still get this in 10.3 or would it be too soon? > h> > > h> > Do you have a module that is going to utilize it? > h> > h> Not right now. But having this in 10.3 can be useful. I don't want to > h> push. I can always carry this patch around if I need in 10. > > Yes, I'm quite reluctant to merge. First we need more modules using the API, > and then we can say ourselves that API is good and can go into stable. Since > when it wents into stable, we should no longer change it.
Fair enough. Cheers, Hiren
pgp5hDbCw48WM.pgp
Description: PGP signature
