Re: [PATCH v2 1/3] ipe: add the .ipe policy keyring

On Thu, Oct 29, 2026 at 12:30:38AM +0000, Manuel Huber via B4 Relay wrote:

From: Manuel Huber <redacted>

Add a dedicated .ipe keyring for IPE policy signature verification.
With this, userspace can provision its own keys into the keyring,
then restrict it and activate it for IPE policy signature verification.

Gate creation of the keyring behind a new by-default disabled
CONFIG_SECURITY_IPE_KEYRING Kconfig option. When created, the keyring
is empty and writable. The keyring is used as a fallback after the
existing policy trust sources only when it is created, non-empty, and
restricted with restrict_link_reject.

This enables scenarios in which a trusted userspace component generates
an IPE policy, and creates an ephemeral policy-signing key pair,
provisions its public certificate, and restricts the IPE keyring during
the early lifetime of a system. The trusted component can then sign the
IPE policy, destroy the private key, and load and activate the policy.
During that phase, the system is considered in a trusted state whereas
the system transitions into a possibly untrusted state afterwards,
running arbitrary user-provided applications. In such a scenario, the
trusted userspace component could reside on a dm-verity protected
rootfs disk, and applications would reside on additional dm-verity
protected disks. The latter disks would be dynamically attached and
their dm-verity root hashes and/or signatures not known a priori to
launching the system.

Not trying to muddy the water here, but I stared / thought about this
a while and wonder if I'm missing something below...

With this Kconfig approach, aren't we now saying "if you want to use IPE,
and you want to ever use this feature, you need to make sure you seal that
keyring or else someone can replace the IPE policy!"? That seems rather
easy to foobar, and means to me at most kernels (i.e. distros) won't have
this enabled which limits how easy this is to use. I know Fan wanted this
behind some Kconfig to make it controlled by build time, but does it make
sense to keep the command line option as well to really make this opt-in
for users who have it enabled at build time? Otherwise you really are gonna
have to craft a special kernel in my mind.

Likewise, but more nitpicky, I'm not sure the restrict_link_reject
requirement buys much over the chain restriction. restrict_link_reject
stops new keys being added, but pkcs7_validate_trust() walks certs
embedded in the signature, so whoever holds the private key of any cert
in .ipe can sign a new leaf, embed it, and have a policy accepted anyway
I believe. Under both restrictions that basically means "holders of an
enrolled private key and whoever they delegate to". If the concern is a
restriction like key_or_keyring:<id> pointing at a writable keyring,
that seems like a provisioner bug rather than a reason to rule out the
chain restriction. Am I missing another reason for the stricter check?
I do realize this is rather pedantic in argument and doesn't have a place
in the use case described, but I'm wondering if a future use case will
want such a thing.

Thanks,
Andrew