> -----Original Message----- > From: Neal Cardwell [mailto:ncardw...@google.com] > Sent: Monday, April 3, 2017 2:59 AM > To: gfree.w...@foxmail.com > Cc: David Miller <da...@davemloft.net>; Alexey Kuznetsov > <kuz...@ms2.inr.ac.ru>; James Morris <jmor...@namei.org>; Patrick McHardy > <ka...@trash.net>; Netdev <netdev@vger.kernel.org>; Gao Feng > <f...@ikuai8.com> > Subject: Re: [PATCH net-next 1/1] net: tcp: Define the TCP_MAX_WSCALE > instead of literal number 14 > > On Sat, Apr 1, 2017 at 12:14 AM, <gfree.w...@foxmail.com> wrote: > > From: Gao Feng <f...@ikuai8.com> > > > > Define one new macro TCP_MAX_WSCALE instead of literal number '14', > > and use U16_MAX instead of 65535 as the max value of TCP window. > > There is another minor change, use rounddown(space, mss) instead of > > (space / mss) * mss; > > > > Signed-off-by: Gao Feng <f...@ikuai8.com> > > --- > > Looks OK to me. > > IMHO this is a nice direction, since there is a proposal to increase the > maximum > shift factor from 14 to 15: > https://tools.ietf.org/html/draft-nishida-tcpm-maxwin-03
Do you mean enlarge the wscale to 15 now ? > > Though there are a few spots where a literal 14 appears: > > tcp_input.c:3763: net_info_ratelimited("%s: Illegal window scaling > value %d >14 received\n", > > tcp_output.c:238: * See RFC1323 for an explanation of > the limit to 14 Ok, I will update these two spots. Regards Feng > > > neal