From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wang Chen Subject: Re: [PATCH] net-driver: Drivers don't set IFF_* flag [Was: [PATCH 3/3] netdevice: order of synchronization of IFF_PROMISC and IFF_ALLMULTI] Date: Tue, 24 Jun 2008 13:39:19 +0800 Message-ID: <48608887.6000102@cn.fujitsu.com> References: <4857B6DC.5020805@trash.net> <48587295.40705@cn.fujitsu.com> <48587854.8050400@pobox.com> <485BC7BA.60406@cn.fujitsu.com> <485F8332.6010203@trash.net> <485FA635.4090903@cn.fujitsu.com> <485FA962.5060704@trash.net> <485FB6D1.7000604@cn.fujitsu.com> <485FB8B1.8020000@trash.net> <4860478B.2080109@cn.fujitsu.com> <20080624051029.GB19434@colo.lackof.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Patrick McHardy , Jeff Garzik , Alan Cox , "David S. Miller" , NETDEV , kyle@parisc-linux.org To: Grant Grundler Return-path: Received: from cn.fujitsu.com ([222.73.24.84]:60895 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751531AbYFXFn0 (ORCPT ); Tue, 24 Jun 2008 01:43:26 -0400 In-Reply-To: <20080624051029.GB19434@colo.lackof.org> Sender: netdev-owner@vger.kernel.org List-ID: Grant Grundler said the following on 2008-6-24 13:10: > On Tue, Jun 24, 2008 at 09:02:03AM +0800, Wang Chen wrote: > ... >> So I want to remove "clear promisc" feature by ioctl of this driver >> instead of adding it to feature-removal-schedule. >> "set promisc" feature by ioctl can be fixed temporarily and should be >> added to feature-removal-schedule. >> Fortunately most of the features of this driver's ioctl are for >> developer testing. >> >> This week I want to wait for tulip driver maintainer's confirmation. > > Confirmation of what? > Apologies for not being able to follow the whole conversation. > Sorry for that I cc-ed you in the mid of discussion. Patrick and me were discussing about that driver should not change device's promisc mode, because it's not a boolean state, but a refcnt to decide whether IFF_PROMISC should be set or not. In de4x5_ioctl(), user can simply set IFF_PROMISC flag and that can cause upper layer protocol broken. Patrick suggest to use dev_change_flags() to change IFF_PROMISC flag and add the feature to feature-removal-schedule. My suggestion is that we remove IFF_PROMISC changing feature from de4x5_ioctl() now. > If you want to remove a developer debug feature, please just submit > a patch like usual to kyle and myself. > Is there any user space binary use these features? case DE4X5_SET_PROM: case DE4X5_CLR_PROM: If not, I will send a patch to remove them.