Re: [git patches] libata updates, GPG signed (but see admin notes)

From: Ingo Molnar <mingo@elte.hu>
Date: Oct 31, 2011, 08:40 UTC
Message-ID: <20111031084048.GA11807@elte.hu>
In-Reply-To: <CA+55aFx1NGWfNJAKDTvZfsHDDKiEtS4t4RydSgHurBeyGPyhXg@mail.gmail.com>

* Linus Torvalds <torvalds@linux-foundation.org> wrote:

> That said, even the "BEGIN PGP SIGNED MESSAGE" things are a massive 
> pain in the butt. We need to automate this some sane way, both for 
> the sender and for the recipient.

The most practical form would be if Git supported such oneliner pull requests:

 git pull git://foo.com bar.branch                           \
  --pull-sha1 0acf00014bcfd71090c3b0d43c98e970108064e4       \
  --gpg-by: "Ingo Molnar <mingo@kernel.org>"                 \
  --gpg-sig: 8a6f134afd1d212fe21345

maintainers could just paste them into a shell and it would abort if it's not trusted. The maintainer verifies the visible, 'Ingo Molnar' bit. The 8a6f134afd1d212fe21345 is a signed-by-Ingo-Molnar version of this content:

    git://foo.com bar.branch 0acf00014bcfd71090c3b0d43c98e970108064e4

And Git would verify that what ends up being pulled is indeed 0acf00014bcfd and also verifies that it was signed by me.

[ If we are extra diligent/paranoid then beyond the sha1 we might 
  even GPG sign the shortlog, or even the full raw log of all commits 
  leading to the sha1: this introduces some Git shortlog and patch 
  formatting version dependency though.

  Git could also double check foo.com's DNS coherency, or check it 
  against a known-trusted whitelist of domain names specified in the 
  maintainer's .gitconfig, as an extra layer. ]

Doing it in this form would remove all the mail formatting madness - one could paste such a pull request into a shell straight away, from HTML email, from text email, from MIME email, etc.

In fact i would trust such a Git based solution far more than any opaque, invisible tool that claims to have checked a signature with cooperation of my mail client (ha!).

The only somewhat non-obvious bit is that Git should be *very* careful about its key ID and signature parsing strategy, to protect against social engineering attacks.

For example neither this:

  --gpg-by: "Ingo Molnar <mingo@kernal.org>"

nor this:

  --pgp-by: "Ingo Molnar <mingo@kernel.org>"

malicious pull request should slip through in any fashion:

 - Git should only use keys that are in your ring of trust - not pull 
   keys from the public keyring automatically and just check 
   coherency of the pull request or such. [I'm sure people will be 
   tempted to have such a feature - but that temptation should be 
   resisted.]

 - Git should abort the moment it sees an unknown option

Thanks,

	Ingo