From mboxrd@z Thu Jan 1 00:00:00 1970 From: "J. Bruce Fields" Subject: Re: [GIT] Networking Date: Fri, 11 Mar 2011 15:48:23 -0500 Message-ID: <20110311204823.GB7906@fieldses.org> References: <20110310.153444.115930379.davem@davemloft.net> <20110310.155556.48513201.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , akpm@linux-foundation.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: Linus Torvalds Return-path: Received: from fieldses.org ([174.143.236.118]:44099 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754172Ab1CKUs0 (ORCPT ); Fri, 11 Mar 2011 15:48:26 -0500 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Mar 10, 2011 at 04:29:30PM -0800, Linus Torvalds wrote: > On Thu, Mar 10, 2011 at 3:55 PM, David Miller w= rote: > > I should have put: > > > > =C2=A0 =C2=A0 =C2=A0 =C2=A0Merge to get commit 8909c9ad8ff03611c9c9= 6c9a92656213e4bb495b > > =C2=A0 =C2=A0 =C2=A0 =C2=A0("net: don't allow CAP_NET_ADMIN to load= non-netdev kernel modules") > > =C2=A0 =C2=A0 =C2=A0 =C2=A0so that we can add Stephen Hemminger's f= ix to handle ip6 tunnels > > =C2=A0 =C2=A0 =C2=A0 =C2=A0as well, which uses the MODULE_ALIAS_NET= DEV() macro created by > > =C2=A0 =C2=A0 =C2=A0 =C2=A0that change. >=20 > Yeah, that would have explained it. That said, if you are merging for > something like that, may I suggest actually starting off with >=20 > git merge 8909c9ad8ff03611c9c96c9a92656213e4bb495b >=20 > that then actually makes the history itself also show the relationshi= p > (you'd still have to write the commit message explaining why, By the way, I occasionally wonder whether it would make sense to make a habit of committing bugfixes on top of the commit that introduced the bug (at least in cases where there *is* a single commit that introduced the bug). As with the above, it'd make the history a little more self-documenting= =2E It might simplify life for backporters. (In theory, they could do merges instead of a cherry-picks if they wanted to.) The set of "bad" commits would be described by "fix^...fix". But then I had some mental image if your saying "WTF?" the first time I send you a post-rc1 pull request that looks like an octopus merge of a dozen little 1- or 2- commit branches based all over the place. I dunno, would it be annoying? --b.