1 |
The OpenPGP smartcard standard, and the Nitrokey Pro smartcards that |
2 |
are being distributed to Gentoo developers, do not support having a |
3 |
separate primary/signing key for keys that are generated on the cards. |
4 |
As a result they can only be used in accordance with our current |
5 |
requirements if the keys are generated in software, which places them |
6 |
at a higher risk of compromise. |
7 |
|
8 |
The intent of the separate primary key is to reduce the risk of it |
9 |
being compromised by keeping it offline. However, if it were |
10 |
generated on a smartcard it would be exclusively be maintained |
11 |
offline, so it is counterproductive to require that it be generated |
12 |
online and then recommend that it be kept offline after this. |
13 |
Additionally this key needs to be brought back online anytime the key |
14 |
expiration is updated, which is at least annually. |
15 |
|
16 |
I believe this is creating a false sense of security, and that |
17 |
developers should be encouraged to generate new keys using their |
18 |
smartcards, so that these keys are never stored outside the protected |
19 |
hardware. By default gpg creates revocation certificates at this time |
20 |
as well. If a key is lost it can still be revoked, and of course the |
21 |
gpg fingerprint associated with the developer can be changed in any |
22 |
case and is the de-facto root of our trust system. These failure |
23 |
modes really are no different from an offline primary key that is |
24 |
separate from the signing keys. |
25 |
|
26 |
The revision adds an exception for keys generated on a smartcard. It |
27 |
does not prohibit those who wish to have separate keys from doing so, |
28 |
though doing this with card-generated keys requires having two |
29 |
smartcards. |
30 |
|
31 |
An obvious objection I can see is that nobody will be able to tell |
32 |
whether any particular key was generated on a smartcard or not. While |
33 |
this is true, it is also the case that there is no way to verify that |
34 |
a primary key is being stored offline, or used on a non-compromised |
35 |
PC, or that it even has a passphrase set. Unless we want to use some |
36 |
kind of hardware key that supports remote attestation we're going to |
37 |
have to trust developers to be security conscious, and IMO making it |
38 |
easy to generate keys on smartcards actually will facilitate security. |
39 |
This change actually makes the more secure mode of operation simpler |
40 |
than the less secure one (unless the primary key is kept online, in |
41 |
which case it provides no benefit). |
42 |
|
43 |
Full version online at: |
44 |
https://gist.github.com/rich0/7d11e2297d8b8a5586997baec2a99e30 |
45 |
|
46 |
Patch follows: |
47 |
|
48 |
|
49 |
diff --git a/glep-0063-v3.rst b/glep-0063-v3.rst |
50 |
index 5895873..86e5fd9 100644 |
51 |
--- a/glep-0063-v3.rst |
52 |
+++ b/glep-0063-v3.rst |
53 |
@@ -12,6 +12,12 @@ OpenPGP key management policies for the Gentoo |
54 |
Linux distribution. |
55 |
Changes |
56 |
======= |
57 |
|
58 |
+v3 |
59 |
+ The requirement to have a separate signing and primary key was removed |
60 |
+ in the case of keys generated/stored on smartcards, to encourage the use |
61 |
+ of these keys, and acknowledging that the main use case for a separate |
62 |
+ primary key is largely fulfilled by having all the keys stay offline. |
63 |
+ |
64 |
v2 |
65 |
The distinct minimal and recommended expirations have been replaced |
66 |
by a single requirement. The rules have been simplified to use |
67 |
@@ -69,7 +75,8 @@ not be used to commit. |
68 |
at least 256-bit. All subkey self-signatures must use this digest. |
69 |
|
70 |
2. Signing subkey that is different from the primary key, and does not |
71 |
- have any other capabilities enabled. |
72 |
+ have any other capabilities enabled. This requirement does not apply |
73 |
+ if the primary key was generated on a smartcard. |
74 |
|
75 |
3. Primary key and the signing subkey are both of type EITHER: |
76 |
|
77 |
|
78 |
-- |
79 |
Rich |