On Wed, 2022-07-20 at 00:50 -0700, Christoph Hellwig wrote: > On Mon, Jul 18, 2022 at 04:01:56PM -0500, gjoyce@xxxxxxxxxxxxxxxxxx > wrote: > > From: Greg Joyce <gjoyce@xxxxxxxxxxxxxxxxxx> > > > > Platforms that have a permanent key store may provide unique > > platform dependent functions to read/write variables. The > > default (weak) functions return -EOPNOTSUPP unless overridden > > by architecture/platform versions. > > This is still lacking any useful implementation. It also seems to be > used in patch 3 before it actually is used. > > As the functionality seems optional I'd suggest to drop this patch > for > now and not call it from patch 3, and do a separate series later that > adds the infrastructure, at leat one useful backend and the caller. It's kind of a chicken and egg thing. I'd hoped to add the infrastructure and follow it up with another pseries specific patchset that provided platform specific implementations of those functions. But I can break it up as you suggest. I'll include your other comments as well as the keyring suggestion from Hannes.
