Date: Mon, 15 Jun 1998 12:08:42 -0400 (EDT) From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> To: Marc Slemko <marcs@znep.com> Cc: freebsd-net@FreeBSD.ORG Subject: Re: Fwd: NetBSD network code improvements Message-ID: <199806151608.MAA25075@khavrinen.lcs.mit.edu> In-Reply-To: <Pine.BSF.3.95.980502132010.28278t-100000@alive.znep.com> References: <199805020641.XAA17167@implode.root.com> <Pine.BSF.3.95.980502132010.28278t-100000@alive.znep.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Following up on some more really old messages in my INBOX.... <<On Sat, 2 May 1998 13:30:50 -0600 (MDT), Marc Slemko <marcs@znep.com> said: > ie. keep minimal state for TIME_WAIT, optimize the timeout handling to > check less frequently for TIME_WAIT timers, etc. From what I have seen, > this can make a big difference on a system with a lot of TIME_WAITs. What we've discussed in the past is slightly changing the memory allocation in Justin's ``new'' (only a year old or so) timeout code so that we can statically allocate timeout structures as a part of the PCB. The ``callout wheel'' design makes it fairly cheap to change timers on the fly, provided we don't get shot down by memory allocation -- in fact, if I remember the discussion correctly, this is the original application which motivated the development of the ``callout wheel'' design. (All of the TCP timers could be reimplemented to use system timers -- the only reason the networking code currently doesn't is because system timeouts were too expensive on a VAX-11/780.) -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
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806151608.MAA25075>