Gentoo Archives: gentoo-dev

From: Jesse Nelson <yoda@××××××.com>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] 2004.1 will not include a secure portage.
Date: Wed, 24 Mar 2004 00:46:53
Message-Id: 20040324004651.GA10327@obi.f00bar.com
In Reply to: Re: [gentoo-dev] 2004.1 will not include a secure portage. by Chris Bainbridge
1 Back when the zynot guys did thier thing i laid out a system for dist/pacakage signing on the wiki, but that has gone away.
2 I have thrown up a quick htmlized version of that *and removed the zynot refs for those still sore bout that* and put it up
3
4 http://f00bar.com/files/ds.html
5
6 heres the quick overview:
7 * GPG signatures for builds, patches, distfiles and sources from cvs or svn repositories
8 * Auditing of distfiles/ebuilds for inclusion in the "secure or stable" branch
9 * Options for securing the transmission channel for trees and distfiles
10 * Revocation of certificates
11 * Policy for dealing with poison mirrors,distfiles, and ebuilds
12 * Policy for accetping new developer/maintainer keys ( see http://www.debian.org/devel/join/nm-step2 debians docs on this)
13 * Web of trust and how it could apply to p2p and other areas
14
15 the zynot forums had a good thread bout this as well @ p2p stuff and thats linked from the top of this html page.
16
17
18 * Chris Bainbridge (c.j.bainbridge@×××××.uk) wrote:
19 > Date: Wed, 24 Mar 2004 00:09:20 +0000
20 > From: Chris Bainbridge <c.j.bainbridge@×××××.uk>
21 > To: gentoo-dev@l.g.o
22 > User-Agent: KMail/1.6.1
23 > X-Spam-Status: No, hits=0.0 required=5.0 tests=none autolearn=no version=2.63
24 > Subject: Re: [gentoo-dev] 2004.1 will not include a secure portage.
25 >
26 > I think the idea of a central key controlling everything is bad - this means
27 > one person is ultimately responsible for the portage tree, and compromising
28 > this will allow access to everything. It would be better if every gentoo
29 > developer had a gpg key. Each package in the portage tree would then have
30 > a .gpg file which lists signatures for the package digest which contains
31 > hashes of each ebuild, files/*, and downloaded distfiles, and a permissions
32 > section listing who has access to make modifications to this dir (such as
33 > writing new ebuilds). People who already have access to a package are then
34 > free to grant access to others merely by inserting their public key into this
35 > file. emerge sync would have to be modified so that it checks signatures
36 > before files are updated.
37 >
38 > Important packages may require multiple signatures before a file is installed
39 > - this is to eliminate the possibility that a compromise of a single gentoo
40 > developer will hand root access to every gentoo installed system. At the
41 > moment, every developer is a point of cvs write access from which an attacker
42 > could root many gentoo installations.
43 >
44 > The key downloads, checking, revocation etc. would be handled by the existing
45 > gpg keyserver infrastructure (eg. keyserver.net). There is no need for an all
46 > powerful gentoo key, or even distribution system. Simply have emerge call gpg
47 > to do everything.
48 >
49 > This only really requires changes to the emerge sync process and a developer
50 > script to check, sign, and post changes. Everything else can be handed off to
51 > gpg. It would also enable some more exciting distribution methods like RSS
52 > channels listing new signed files in portage along with a p2p backend to
53 > fetch them, automatic security updates, etc.
54 >
55 > On Tuesday 23 March 2004 13:12, Paul de Vrieze wrote:
56 >
57 > > Ok, the biggest problem is an idea for how the actual checking should
58 > > function. (Signing is straightforward).
59 > >
60 > > I'll take a first shot at describing a key chain idea. Please shoot at it
61 > > and try to find the holes. But remember it needs to stay workable. First
62 > > the premises:
63 >
64 > --
65 > gentoo-dev@g.o mailing list
66 >
67 >
68
69 --
70 gentoo-dev@g.o mailing list