Re: [PATCH 0/3] arm64/virt: Add Arm CCA measurement register support

[Yeoreum Yun <yeoreum.yun@arm.com>]

On Tue, Sep 29, 2026 at 04:04:50PM -0300, Jason Gunthorpe wrote:
> On Tue, Sep 29, 2026 at 05:06:02PM +0100, Yeoreum Yun wrote:
> > Hi Jason,
> > 
> > > On Wed, Apr 22, 2026 at 08:57:00AM +0000, Sami Mujawar wrote:
> > > 
> > > > > So this is tying it to the same FW event log that TPM uses.
> > > > > 
> > > > > I think that strengthens my point this should all be uninform. TPM
> > > > > drivers are directly exposing the event log today, but I guess that
> > > > > needs generalization if non-TPM drivers are going to present it as
> > > > > well.
> > > > > 
> > > > > How do you imagine getting and manipulating the EFI event log to use
> > > > > with this?
> > > >
> > > > The event logs from UEFI will be handed off to the OS using CCEL
> > > > ACPI table. The CCEL table spec update can be seen at
> > > > https://github.com/tianocore/edk2/issues/11384
> > > 
> > > I ment from linux userspace. event log is well establihsed in the aCPI
> > > side for TPM and in Linux today it is only exposed to userspace
> > > through TPM.
> > > 
> > > This is not TPM, so how do you intend to give the event log to
> > > userspace?
> > 
> > The event log would be exported via /sys/firmware/acpi/tables/data/CCEL
> > with binary format just as TPM event log is also exported with binary
> > format via binary_bios_measurements.
> > 
> > Instead of Sami, I'll post v2 which rebased on for-next/smccc-bus in
> > arm64 tree.
> 
> Okay, Jiri is close to posting the "attestation subsystem" which this
> needs to plug into instead of creating a bunch more sysfs files.
> 
> The CCEL log needs to come out through that interface.

Agree. Thanks for sharing!

-- 
Sincerely,
Yeoreum Yun