1 |
Brian G Peterson writes: |
2 |
|
3 |
> I assume that you intend to 'blow the whistle' because |
4 |
> you are incapable or unwilling to submit a patch for the |
5 |
> issue yourself? |
6 |
|
7 |
If you read the recent messages carefully, you'll find that |
8 |
I have tried _numerous_ times to provide details how to |
9 |
remedy the situation. |
10 |
|
11 |
|
12 |
> I agree that there is a lot of room for improvement in |
13 |
> the portage security system. |
14 |
|
15 |
Then why don't we stop discussing what I know or don't know, |
16 |
do or won't do, and talk about a solution? The vast majority |
17 |
of text posted in this thread is concerned with all kinds of |
18 |
things BUT finding a good, technical solution to a |
19 |
vulnerability that _does_ exist. |
20 |
|
21 |
Generating a signed hash list of all files is really not |
22 |
that difficult. It would solve the problem in a matter of |
23 |
hours for those who are concerned about it, and it would |
24 |
probably set things in motion for a better solution to be |
25 |
developed that solves the problem for all users as well as |
26 |
possible. |
27 |
|
28 |
So why is the Gentoo team so incredibly reluctant to do |
29 |
anything about it? |
30 |
|
31 |
Again: |
32 |
|
33 |
(1) Configure your main site to update the portage tree |
34 |
from CVS in a time interval that's sufficient large to |
35 |
allow for the hash list to be generated. Someone else |
36 |
already suggested once an hour. I can't say what is |
37 |
appropriate since I don't know your setup. |
38 |
|
39 |
(2) Calculate hashes for all files in the /usr/portage |
40 |
hierarchy. One could probably use a trivial Makefile to |
41 |
generate hashes incrementally, even, to ease the load |
42 |
on the machine. |
43 |
|
44 |
(3) Sign the hash file with a GPG key. That means that |
45 |
either someone has to enter the pass phrase manually, |
46 |
or you'll have to set up a pass phrase agent, or you'll |
47 |
have to use a key without a password at all. |
48 |
|
49 |
Everything but the first solution is sub-optimal but |
50 |
still a _lot_ better than what we have now. If someone |
51 |
manages to compromise the main site, we all have far |
52 |
greater problems than a lost secret key, so even _if_ |
53 |
the pass phrase is empty we still gain security. |
54 |
|
55 |
(4) Distribute the signed hash file with the portage tree. |
56 |
|
57 |
(5) Provide scripts that verify the integrity of the tree |
58 |
after an emerge sync _before_ any other code is run |
59 |
that has been obtained from the network. |
60 |
|
61 |
(6) Make the matching public key available on the key |
62 |
servers, on the web site, and every other place that |
63 |
you can think about. Give an expiry date of, say 3 |
64 |
months to make clear that this is an intermediate |
65 |
solution that will change. |
66 |
|
67 |
(7) Get as many people to sign the key as possible to |
68 |
properly authenticate it. |
69 |
|
70 |
(8) Write a security advisory that educates the users about |
71 |
the problem. |
72 |
|
73 |
Peter |
74 |
|
75 |
|
76 |
-- |
77 |
gentoo-security@g.o mailing list |