[PATCH v5 3/3] efi: Add tee-based EFI variable driver

[Sumit Garg]

On Tue, 6 Jun 2023 at 12:28, Ard Biesheuvel <ardb at kernel.org> wrote:
>
> On Tue, 6 Jun 2023 at 08:52, Sumit Garg <sumit.garg at linaro.org> wrote:
> >
> > Hi Jan,
> >
> > On Tue, 6 Jun 2023 at 12:09, Jan Kiszka <jan.kiszka at siemens.com> wrote:
> > >
> > > On 26.05.23 03:07, Masahisa Kojima wrote:
> > > > When the flash is not owned by the non-secure world, accessing the EFI
> > > > variables is straightforward and done via EFI Runtime Variable Services.
> > > > In this case, critical variables for system integrity and security
> > > > are normally stored in the dedicated secure storage and only accessible
> > > > from the secure world.
> > > >
> > > > On the other hand, the small embedded devices don't have the special
> > > > dedicated secure storage. The eMMC device with an RPMB partition is
> > > > becoming more common, we can use an RPMB partition to store the
> > > > EFI Variables.
> > > >
> > > > The eMMC device is typically owned by the non-secure world(linux in
> > > > this case). There is an existing solution utilizing eMMC RPMB partition
> > > > for EFI Variables, it is implemented by interacting with
> > > > TEE(OP-TEE in this case), StandaloneMM(as EFI Variable Service Pseudo TA),
> > > > eMMC driver and tee-supplicant. The last piece is the tee-based
> > > > variable access driver to interact with TEE and StandaloneMM.
> > > >
> > > > So let's add the kernel functions needed.
> > > >
> > > > This feature is implemented as a kernel module.
> > > > StMM PTA has TA_FLAG_DEVICE_ENUM_SUPP flag when registered to OP-TEE
> > > > so that this tee_stmm_efi module is probed after tee-supplicant starts,
> > > > since "SetVariable" EFI Runtime Variable Service requires to
> > > > interact with tee-supplicant.
> > > >
> > > > Acked-by: Sumit Garg <sumit.garg at linaro.org>
> > > > Co-developed-by: Ilias Apalodimas <ilias.apalodimas at linaro.org>
> > > > Signed-off-by: Ilias Apalodimas <ilias.apalodimas at linaro.org>
> > > > Signed-off-by: Masahisa Kojima <masahisa.kojima at linaro.org>
> > > > ---
> > > >  drivers/firmware/efi/Kconfig                 |  15 +
> > > >  drivers/firmware/efi/Makefile                |   1 +
> > > >  drivers/firmware/efi/stmm/mm_communication.h | 236 +++++++
> > > >  drivers/firmware/efi/stmm/tee_stmm_efi.c     | 638 +++++++++++++++++++
> > > >  4 files changed, 890 insertions(+)
> > > >  create mode 100644 drivers/firmware/efi/stmm/mm_communication.h
> > > >  create mode 100644 drivers/firmware/efi/stmm/tee_stmm_efi.c
> > > >
> ...
> > >
> > > I think we have a probe ordering issue with this driver:
> > > efivarfs_fill_super() may be called before the TEE bus was probed, thus
> > > with the default efivar ops still registered. And that means
> > > efivar_supports_writes() will return false, and the fs declares itself
> > > as readonly. I've seen systemd mounting it r/o initialling, and you need
> > > to remount the fs to enable writability.
> > >
> > > Is there anything that could be done to re-order things reliably, probe
> > > the tee bus earlier etc.?
> >
> > This driver has a dependency on user-space daemon: tee-supplicant to
> > be running for RPMB access. So once you start that daemon the
> > corresponding device will be enumerated on the TEE bus and this driver
> > probe will be invoked. So I would suggest you to load this daemon very
> > early in the boot process or better to make it a part of initramfs.
> >
>
> That is not the point, really.
>
> If this dependency exists, the code should be aware of that, and made
> to work correctly in spite of it. Requiring a module to be part of
> initramfs is not a reasonable fix.

I am not sure if I followed you here. Until we have the tee-stmm-efi
module loaded we won't get corresponding EFI operations registered.
The key here is that the underlying OP-TEE device won't be available
until tee-supplicant (see enumeration with PTA_CMD_GET_DEVICES_SUPP)
is running.

Do you have any better ideas regarding how this should be handled?

>
> IIUC, this also means that the efivar ops are updated while there is
> already a client. This seems less than ideal as well

That's true. An ideal situation would be to allow in-kernel RPMB
access APIs for OP-TEE or other kernel drivers to use. With that we
should be able to remove this tee-supplicant dependency. I hope you
are already aware about the earlier efforts to add the RPMB subsystem
to the kernel. But with these real world use-cases emerging like EFI
variables and fTPM, we should be able to convince corresponding
subsystem maintainers.

-Sumit