-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Tue, Jan 08, 2019 at 01:15:39AM +0000, 'awokd' via qubes-users wrote:
> Demi Obenour wrote on 1/7/19 3:16 PM:
> > Looking through the GPG CVE list, it appears that GPG has a fantastic
> > security record.  This seems to jus Most of the recent vulnerabilities have
> > been side-channel attacks.
> > 
> > Is it useful to use split-GPG with a hardware token to prevent side-channel
> > attacks?
> 
> I am far from a cryptographer, but IIRC those side channel attacks get the
> key by observing decryption leaks. So a hardware token wouldn't affect that
> either way, because once the key is unlocked it still gets processed the
> same.

Not really, if key lives on a hardware token, only it can perform
decryption/signing. So, if that hardware token is resistant against side
channel attacks, then split-GPG (or anything else) will not make it
worse.

> > Also, is it best to use one signing key per project one is working on?
> > 
> Again, not a crypto expert but if you're using the same development workflow
> for all projects, don't see much security gain from separate keys. If some
> demand a different, potentially less secure workflow, those might benefit
> from subkeys. Hopefully someone experienced has more insight!

There is one more thing: if you use a single key for multiple projects,
then it's harder to distinguish those projects based on cryptographic
proof. Which means code signed in one project could potentially be used
in another. 
An example: I have a qubes code signing key I use to sign my
qubes-related commits/tags. But I also contribute to other projects,
including also very simple patches, where I only fix one file and
definitely not review the whole repository. If I would use the same key
for both, then one could attack me like this:
1. Introduce a backdoor to some random software that I would likely
   contribute to (or even create new one specifically for this purpose).
2. Wait for me to contribute there (all kind of social engineering will
   help here).
3. Take my signed contribution and pretend the code belongs to qubes -
   this may is quite tricky, and probably require breaking into my github 
account
   (or github infrastructure) to place it under my (or QubesOS) account;
   but even without it, it would help in other attacks.

With separate keys (having project name in key comment) that attack
wouldn't work, or would require significantly more social engineering -
depending whether you attack a machine or a human.

You may also take into account security of development environment for
each project. If one depends on a lot of software without reliable
integrity verification method (or, say, a lot of NodeJS package ;) ),
then such environment would be significantly easier to compromise, and
so the key used there (even if not leaked, then used from there to
sign/decrypt anything).

- -- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEEhrpukzGPukRmQqkK24/THMrX1ywFAlw5PaQACgkQ24/THMrX
1yyotggAj6mbhIApmFSsajZ/Zjk1Lt49Lgnba5TXQDHgODwGp+i4QG3JqKVgHTma
QXvoTKsMZohuABe6wWiTxT/DvJJjUzpHAOEnj/XAzGm6mm8kJqZ/hih2pq7T7+qn
Oe+zOdLNPdS4olmLy/igw/V+CtjNhuWYKsSM7mCzSpRRIPGuG4IvhEX+WyHFDt6u
rMpCL2nNqRHcMo+Qve7/5e2IPnWFZPjDVsaeTiHpaAlFfzDVLUyg2qxGxamezuLo
fH6ZvUd1UOHntUCYWjeD7JpY05Y8P0dAPRsRlcW28eAKAeUy9cepQlLJeafRdYCo
b5e0pWhYe/DqZxMJKzVuSnJy2OpBeA==
=j4nW
-----END PGP SIGNATURE-----

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to qubes-users+unsubscr...@googlegroups.com.
To post to this group, send email to qubes-users@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/20190112010644.GB6577%40mail-itl.
For more options, visit https://groups.google.com/d/optout.

Reply via email to