From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 25 Jan 2001 15:36:43 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 25 Jan 2001 15:36:33 -0500 Received: from minus.inr.ac.ru ([193.233.7.97]:7184 "HELO ms2.inr.ac.ru") by vger.kernel.org with SMTP id ; Thu, 25 Jan 2001 15:36:22 -0500 From: kuznet@ms2.inr.ac.ru Message-Id: <200101252036.XAA10500@ms2.inr.ac.ru> Subject: Re: [UPDATE] Zerocopy patches, against 2.4.1-pre10 To: ionut@moisil.cs.columbia.edu (Ion Badulescu) Date: Thu, 25 Jan 2001 23:36:03 +0300 (MSK) Cc: linux-kernel@vger.kernel.org, andrewm@uow.EDU.AU In-Reply-To: <200101252028.f0PKSJR02124@moisil.dev.hydraweb.com> from "Ion Badulescu" at Jan 25, 1 12:28:19 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello! > Starfire card does, maybe the 3com is different. :-) 3com _is_ different. 8) I is not an issue, we do not make zerocopy on IP fragments. > Are we even bothering with the partial checksums at this point, or > are we falling back to CPU checksumming if the packet is fragmented? Of course, we are not bothering. > And, on a related note: what's involved in making a driver > zerocopy-aware? I haven't looked too closely to the current patch, > but I was thinking of playing with the starfire driver, since I > have all the chipset docs.. Nothing especially clever. Gather plus checksumming, as described in comment in linux/netdevice.h. Well, just look into one of existing drivers: better, sunhme. 3com and acenic are too stupid to learn something. Alexey - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/