From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934013Ab1JEAby (ORCPT ); Tue, 4 Oct 2011 20:31:54 -0400 Received: from out3.smtp.messagingengine.com ([66.111.4.27]:41356 "EHLO out3.smtp.messagingengine.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933803Ab1JEAbw (ORCPT ); Tue, 4 Oct 2011 20:31:52 -0400 X-Sasl-enc: 2qXrGyRGdcjntkg7dO1MjVtMV9Nv0FcQ4nacLzF5TnXQ 1317774432 Date: Tue, 4 Oct 2011 21:27:07 -0300 From: Henrique de Moraes Holschuh To: Heiko Carstens Cc: Jiri Kosina , "H. Peter Anvin" , Linux Kernel Mailing List Subject: Re: kernel.org status: establishing a PGP web of trust Message-ID: <20111005002707.GA20062@khazad-dum.debian.net> References: <4E8655CD.90107@zytor.com> <20111004125134.GB2803@osiris.boeblingen.de.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20111004125134.GB2803@osiris.boeblingen.de.ibm.com> X-GPG-Fingerprint: 1024D/1CDB0FE3 5422 5C61 F6B7 06FB 7E04 3738 EE25 DE3F 1CDB 0FE3 User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 04 Oct 2011, Heiko Carstens wrote: > > I have a question here. In case people are 'reasonably certain' that the > > old key has never been jeoparadized, why are they required to create a new > > key? You do that when your key is either old, or weaker than the requested keys. In the kernel.org case, if your old key is not RSA/4096... If your old key is already RSA/4096, there are some reasons to create a new key: 1. "Just in case the old key was compromised", in which case you should revoke the old key soon. 2. To use different keys for different purposes. 3. To have a backup key (when proper use of subkeys is not enough to protect the main key) -- you better think about this carefully: how safe can you make the backup key storage, and _how fast_ you will notice should someone manage to access it? 4. When prepararing for a key rollover in the near future (in which case the keys are cross-signed, and you keep collecting signatures on BOTH keys until the old one is decomissioned). > > (if the old key would have been compromised, the attacker could as well > > generate a new key and sign it with the old key himself, so I fail to see > > any benefit of this PGP excercise). Yes, he could. The signature from the old key is used to assert that you indeed have control of the old private key, but it cannot assert that you are the only party that has control of the old private key. > > It doesn't make too much sense to force people to live with two different > > personalities in this "PGP web of trust" world just for the sake of > > kernel.org, does it? > > Also same question here. And as far as I can tell nobody has given an > answer yet. That depends, you might want to have separate keys for separate purposes (which only makes sense if key usage will be different enough to cause different key exposure to risks). In any case, if you'll have different keys, it is good practice to use the UID comment field to disclose the purpose of the key. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh