From owner-freebsd-current Tue Feb 20 18:59:16 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id SAA27524 for current-outgoing; Tue, 20 Feb 1996 18:59:16 -0800 (PST) Received: from Root.COM (implode.Root.COM [198.145.90.17]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id SAA27519 Tue, 20 Feb 1996 18:59:12 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by Root.COM (8.6.12/8.6.5) with SMTP id SAA04978; Tue, 20 Feb 1996 18:59:13 -0800 Message-Id: <199602210259.SAA04978@Root.COM> X-Authentication-Warning: implode.Root.COM: Host localhost didn't use HELO protocol To: Naoki Hamada cc: andreas@knobel.gun.de, hackers@FreeBSD.org, current@FreeBSD.org Subject: Re: mbuf enhancement patch In-reply-to: Your message of "Wed, 21 Feb 1996 11:46:44 +0900." <199602210246.LAA18404@sirius.sbl.cl.nec.co.jp> From: David Greenman Reply-To: davidg@Root.COM Date: Tue, 20 Feb 1996 18:59:13 -0800 Sender: owner-current@FreeBSD.org Precedence: bulk >>>I found the ep driver always keeps some mbuf's in its pool. Is this >>>because mbuf allocation is too expensive for boards which equip small >>>receive buffer? If this is the case, some improvement (not mine :-) is >>>desirable. >> >> I think that's what the author thought, but the FIFO on the 3c509 should be >>sufficiently large enough to not need the extra 1% of speed that having the >>private pool gets you. Our malloc implementation is quite efficient, actually. > >The old 3c509 has 2k bytes RX FIFO. Is this large enough? Yes, but a bit tight. If the driver were properly written, large packets would be put in mbuf clusters which are allocated out of a private pool and should be as fast as the pool that the driver is maintaining. I haven't looked at the driver source in any detail...I've instead decided to rewrite it at some point in the future, but haven't had the time + enough interest yet. -DG David Greenman Core-team/Principal Architect, The FreeBSD Project