| 1 |
On Sat, 07 Mar 2015 18:31:44 -0800 |
| 2 |
Zac Medico <zmedico@g.o> wrote: |
| 3 |
|
| 4 |
> On 03/07/2015 05:24 PM, Brian Dolbec wrote: |
| 5 |
> > On Sat, 07 Mar 2015 15:26:26 -0800 |
| 6 |
> > Zac Medico <zmedico@g.o> wrote: |
| 7 |
> > |
| 8 |
> >> On 03/06/2015 09:50 AM, Mark Kubacki wrote: |
| 9 |
> >>> We're on the same side here. |
| 10 |
> >>> |
| 11 |
> >>> Do we have numbers showing the ratio "portage used with defaults" |
| 12 |
> >>> vs. where "[webrsync-gpg] is described in many hardening guides |
| 13 |
> >>> for gentoo and widely used among the security conscious" applies? |
| 14 |
> >>> |
| 15 |
> >>> DNS not being encrypted is just painting the whole picture. Point |
| 16 |
> >>> is, the default is that "emerge --sync" results in a transfer |
| 17 |
> >>> using RSYNC (or http). |
| 18 |
> >>> |
| 19 |
> >>> And by default you cannot compare the result with any |
| 20 |
> >>> authoritative source. |
| 21 |
> >>> |
| 22 |
> >> |
| 23 |
> >> Ideally, we can rely on security mechanisms built into git [1], |
| 24 |
> >> possibly involving signed commits. |
| 25 |
> >> |
| 26 |
> >> [1] https://github.com/gentoo/gentoo-portage-rsync-mirror |
| 27 |
> > |
| 28 |
> > |
| 29 |
> > Actually, git makes it nearly impossible to get the correct info |
| 30 |
> > required to re-process the git commit signature. It is all embedded |
| 31 |
> > into the commit itself, but not in a way for gpg to be able to |
| 32 |
> > re-verify it. Git relies on the issuer's gpg verification status |
| 33 |
> > which it records in it's data before the push. So, without some |
| 34 |
> > major hacking on git data, it is not viable for routine |
| 35 |
> > verification purposes. |
| 36 |
> |
| 37 |
> Maybe we can use the relatively recently added git verify-commit |
| 38 |
> command [1]. If we run that command with --verbose, then we can use |
| 39 |
> that to verify that that a commit was signed with a trusted key. |
| 40 |
> Shouldn't verification of the HEAD commit be enough to vouch for the |
| 41 |
> integrity of the entire tree? |
| 42 |
> |
| 43 |
> [1] |
| 44 |
> https://github.com/git/git/commit/d07b00b7f31d0cb6675b4bdba269ce7f087ddd24 |
| 45 |
|
| 46 |
GREAT! So, same as for push --sign, we just need to configure a |
| 47 |
replacement gpg command for git that uses gkeys instead of gpg |
| 48 |
directly. |
| 49 |
|
| 50 |
That will simplify things over the previous. |
| 51 |
|
| 52 |
A replacement gpg command is needed because of the way gkeys stores the |
| 53 |
developer keys in separate keyrings. gkeys determines which keyring to |
| 54 |
use and runs gpg with the correct settings. It is also capable of |
| 55 |
searching is db for the correct keyring to verify against as well. |
| 56 |
|
| 57 |
-- |
| 58 |
Brian Dolbec <dolsen> |
Archives