From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Garzik Subject: Re: [git patches] net driver fixes Date: Sun, 23 Dec 2007 01:42:14 -0500 Message-ID: <476E0346.1030902@garzik.org> References: <20071223053314.GA31966@havoc.gtf.org> <20071223062731.GP8181@ftp.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Andrew Morton , Linus Torvalds , netdev@vger.kernel.org, LKML To: Al Viro Return-path: Received: from srv5.dvmed.net ([207.36.208.214]:41687 "EHLO mail.dvmed.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750709AbXLWGmT (ORCPT ); Sun, 23 Dec 2007 01:42:19 -0500 In-Reply-To: <20071223062731.GP8181@ftp.linux.org.uk> Sender: netdev-owner@vger.kernel.org List-ID: Al Viro wrote: > On Sun, Dec 23, 2007 at 12:33:14AM -0500, Jeff Garzik wrote: >> A couple [minorly] notable wireless bug fixes, and plenty of viro fixes >> for obscure issues :) > > Heh... FWIW, forcedeth patch (sent your way about two weeks ago) also > belongs in the same set. If you need a resend - tell... I applied it to #upstream (2.6.25) since forcedeth is not on any big-endian platforms AFAIK. Is it .24 material for some obvious reason, which I missed? :) > There's another pile in drivers/net/wireless, but that's for linville to > forward when he gets around to it. Pure annotation patches belong to > past-2.6.24 merge. > > I also have starfire and epic100 fixes, but that'll have to wait until > I get around to putting the cards into sparc box (mcast breakage for > starfire and full-driver one for epic100; since nobody had cared for > the latter since 2.3.late, well...) I have an epic100 card too if you need it (though it sounds like you have something testable). > I think I'll have an ipg fix for you tomorrow, but I want to RTFM first > to make sure that it makes sense. And there are several interesting > issues in atl1, netxen and cxgb3, but those will have to wait for when > I get around to asking maintainers just what the hell did they mean those > to do. > > FWIW, drivers/net is fairly noise-free wrt sparse endianness warnings > in my tree; the main exceptions are prism54 (oid_mgt.c and the nightmares > it pulls) and skfp (AIX-shared vendor driver; 'nuff said, IMO). Awesome :) > BTW, if you still have any documentation for xircom_cb from your fighting > tulip-related stuff, it would be welcome - there are some oddities with > rx ring handling (assuming that we care about that driver at all and it's > not on the way out, that is). xircom_tulip_cb should probably be deleted, since it is the crappier of the two drivers for the same hardware (xircom_cb being the other one). xircom_cb _the driver_ is pretty odd. It is less like tulip than it should be, actually. There are several things that could have been done to improve the throughput/etc. of the driver, but it was more important at the time to simply find a driver that always worked. IIRC its RX filtering was broken, implying the need to enable promisc mode just to receive packets normally. Jeff