-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Tue, Apr 02, 2019 at 01:20:54PM +1100, haaber wrote:
> > On Tue, Apr 02, 2019 at 07:19:46AM +1100, haaber wrote:
> > >
> > > So do I understand that correctly: if I have, say, a debian-XYZ AppVM on
> > > clearnet it will check if the corresponding template needs an update,
> > > unless I de-activate the qubes-update-check service? Thank you
> >
> > Yes
> >
>
> Oups ! To me, one of the points of using tor as upgrade-transport-layer
> seems to me to render "aimed attacks" on *my* machine much harder. Is
> that a misconception?
> Assuming that 'yes', an attacker would typically see clearnet apt-update
> preceding a tor-based upgrade -- and could be made a reasonable guess
> *who* is upgrading (I don't think there are millions of qubes copies
> running, right?). This opens a (admittedly) small, probability-based
> attack surface, that comes only with small gain, if ever. Do you agree?
The updates _check_ only needs to download repository metadata, not
actual packages. Qubes based on a template do that from time to time,
using own network connection and report if there are any updates
available.
When you actually download and install those updates (over Tor) in the
template is up to you, it isn't immediately after checking if something
is available, so time based correlation isn't really an issue here.
- --
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/THMrX1ywFAlyjhOoACgkQ24/THMrX
1yzrVgf/cpAa8ZF7aw1UUkMVW3L+YndBFVOmH0vG1XZ1ppQ3RqG/5OpZnG+eSaQV
l2iyMMWpSDKY6niHEEhXIHBGO17ABmZcybvMe8jGtovm6e+kwRa1ef1yarSI3aLL
W2IcAFoo2XYRVpO+/sGWFD0WHNdIzqcVVNK5o45MKnJPgb+ZQ3+Wg7h9nbU3NCMh
zTlUHjW59gGgx1IKtylc69IM/zgBxKysfrC6SuTRTid2YGpUNfqyMR+oj+FEa2W9
VMoySbjOUnAxrOydvFyUL8vTZ/w1rDNpGAoWyUBcCoUmpDW9ZdfCCYuO1l2fWbE6
SZexjBIGsEzKbDfm2dD9HQT4VPicbQ==
=bswd
-----END PGP SIGNATURE-----