Re: [PATCH 0/2] lockdown: Prevent booting with lockdown enabled but IOMMU disabled

From: Nicolas Bouchinet

Date: Tue Nov 17 2026 - 03:50:13 EST


On Mon, Nov 16, 2026 at 02:29:01PM +0100, Nicolas Bouchinet wrote:
> >
> > How would you feel about at least blocking the user from turning off IOMMU
> > (if it's present on the system) via the kernel command line while lockdown
> > was enabled?
> >
> > Thinking about the case where most distros enable lockdown by default, this
> > will strenghten it's use without much downside that I can see.
> >
> > At least on a properly configured system that contains an IOMMU it would
> > prevent a user from subverting lockdown with a malicious device during the
> > boot sequence.
>
> I do agree with you, it could be a great addition. With Lockdown enabled
> as an early LSM (CONFIG_SECURITY_LOCKDOWN_LSM_EARLY), this might be
> feasible before boot parameter parsing. Lockdown already locks unsafe
> module parameters.
>
> If I remember correctly on distro's that have a Lockdown patch that
> enables it if secure boot is enabled, it should also be called before
> early param parsing.
>
> But taking this threat into account also brings other questions.
> i.e, a physical attacker with root access on a system where Lockdown is
> enabled and with a bootchain configured with shim+MOK is able to change
> kernel cmdline, generate a MOK and sign a compromised kernel
> binary. Thus defeating lockdown purpose.
>
> In order to take the physical attacker into account, a securely
> configured bootchain where the kernel image can't be tampered with and
> where an attacker is not able to inject it owns certificates/hash in
> db/MOK is mandatory so that Lockdown makes sense.

To be perfectly clear, my intention was not to suggest including a
physical attacker in the Lockdown threat model, but rather to provide
examples of how such an attacker could bypass Lockdown through other
exploitation paths.