Re: [PATCH] tpm: ibmvtpm: Call tpm2_sessions_init() to initialize session support

[Jarkko Sakkinen]

On Wed Jul 3, 2024 at 2:57 AM EEST, Jarkko Sakkinen wrote:
> On Wed, 2024-07-03 at 02:48 +0300, Jarkko Sakkinen wrote:
> > On Mon, 2024-07-01 at 15:14 -0400, Stefan Berger wrote:
> > > Applying it is probably the better path forward than restricting HMAC to 
> > > x86_64 now and enabling it on a per-architecture basis afterwards ...
> > 
> > Why is this here and not in the associated patch?
> > 
> > Any, what argue against is already done for v6.10.
> > 
> > The actual bug needs to be fixed before anything
> > else.
> > 
> > I can look at the patch when in August (back from
> > holiday) but please response to the correct patch
> > next time, thanks.
>
> Next steps forward:
>
> 1  Comment out sessions_init().
> 2. See what happens on x86 in QEMU.
> 3. All errors were some sort size errors, so look into failing
>    sites and fixup the use of hmac shenanigans.

For anything "fast" or "quick" I think this really the only
possible sane thing to do right now:

https://lore.kernel.org/linux-integrity/20240703003033.19057-1-jarkko@xxxxxxxxxx/T/#u

There's also bunch of other drivers than tpm_ibmvtpm so better
to limit it to known good drivers.

I can take at the actual issue in August and will review any
possible patches then. This one I'll send after my current PR
for TPM has been merged.

BR, Jarkko