On Mon, 23 Jan 2023 at 09:59, Johan Hovold <johan@xxxxxxxxxx> wrote:
>
> On Mon, Jan 23, 2023 at 09:39:41AM +0100, Ard Biesheuvel wrote:
> > On Mon, 23 Jan 2023 at 09:19, Johan Hovold <johan+linaro@xxxxxxxxxx> wrote:
> > >
> > > The efivars sysfs interface was removed by commit 0f5b2c69a4cb ("efi:
> > > vars: Remove deprecated 'efivars' sysfs interface").
> > >
> > > Remove also the corresponding sysfs ABI documentation.
> > >
> > > Signed-off-by: Johan Hovold <johan+linaro@xxxxxxxxxx>
> > > ---
> > >
> > > Changes in v2
> > > - drop reference in gsmi sysfs documentation
> > > - drop reference in efivarfs.rst (kernel test robot)
> > >
> >
> > Ugh. So there is a remaining implementation of that interface. That is
> > a bit disappointing, tbh.
>
> No, you removed the implementation in the commit mentioned above. The
> Google SMI driver only provides a efivars "backend" but the interface
> was shared. The driver continues to work with efivarfs.
>
Ugh. So as far as I can tell, this interface is still being used
internally at Google.
> > So for now, let's disregard this patch, and I will check internally
> > whether or not that sysfs gsmi interface is actually used. If it is,
> > the docs should be kept but updated to clarify that it only describes
> > gsmi sysfs. Otherwise, we can drop the whole thing, including the gsmi
> > sysfs pieces themselves.
>
> So you'd need to bring back the sysfs implementation and make it Google
> SMI specific if it's still needed by someone. I don't think we want to
> do that if it can be avoided.
>
Indeed.
