From owner-freebsd-net Thu Aug 27 11:28:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA25843 for freebsd-net-outgoing; Thu, 27 Aug 1998 11:28:39 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from khavrinen.lcs.mit.edu (khavrinen.lcs.mit.edu [18.24.4.193]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA25822 for ; Thu, 27 Aug 1998 11:28:28 -0700 (PDT) (envelope-from wollman@khavrinen.lcs.mit.edu) Received: (from wollman@localhost) by khavrinen.lcs.mit.edu (8.9.1/8.9.1) id OAA10118; Thu, 27 Aug 1998 14:27:34 -0400 (EDT) (envelope-from wollman) Date: Thu, 27 Aug 1998 14:27:34 -0400 (EDT) From: Garrett Wollman Message-Id: <199808271827.OAA10118@khavrinen.lcs.mit.edu> To: Julian Elischer Cc: net@FreeBSD.ORG Subject: Re: Next big network patch: specialized sosend for TCP In-Reply-To: References: <199808270155.VAA07365@khavrinen.lcs.mit.edu> Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org < said: > On Wed, 26 Aug 1998, Garrett Wollman wrote: > can be rewritten to: >> + if (resid >= MINCLSIZE) { >> + MCLGET(m, M_WAIT); >> + if (m->m_flags & M_EXT) >> + mlen = MCLBYTES; >> + } >> + len = min(min(mlen, resid), space); There's a lot more work to be done in here. For example, one of the things I'd like to see is whether sucking the data in MSS-sized chunks helps or hurts us. I think we may also see a win by aligning short packets against the end of the mbuf, so that when TCP m_prepend()s its headers, it doesn't need to allocate another buffer. (This is still really klugey; hopefully I'll get a chance to finish my UDP work which is much more fully-formed.) -GAWollman -- Garrett A. Wollman | O Siem / We are all family / O Siem / We're all the same wollman@lcs.mit.edu | O Siem / The fires of freedom Opinions not those of| Dance in the burning flame MIT, LCS, CRS, or NSA| - Susan Aglukark and Chad Irschick To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message