Re: [PATCH v5 2/2] Add "shutdown" to "struct class".

[Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>]

On Thu, Jun 01, 2017 at 09:55:20AM -0700, Josh Zimmerman wrote:
> On Thu, Jun 1, 2017 at 9:39 AM, Jason Gunthorpe
> <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org> wrote:
> > On Thu, Jun 01, 2017 at 09:33:43AM -0700, Josh Zimmerman wrote:
> >
> >> -               if (dev->bus && dev->bus->shutdown) {
> >> +               if (dev->class && dev->class->shutdown) {
> >> +                       if (initcall_debug)
> >> +                               dev_info(dev, "shutdown\n");
> >> +                       dev->class->shutdown(dev);
> >> +               } else if (dev->bus && dev->bus->shutdown) {
> >>                         if (initcall_debug)
> >>                                 dev_info(dev, "shutdown\n");
> >>                         dev->bus->shutdown(dev);
> >
> > Looking at this closer, now you definately have to change the TPM
> > patch to call through to the other shutdown methods. We can say
> > current TPM drivers have no driver->shutdown, but we cannot be sure
> > about the bus->shutdown, so may as well call both from tpm's
> > class->shutdown.

> Why can't we be sure? Could bus->shutdown methods be registered
> outside of the drivers/char/tpm/ tree?

Yes.

> > I would say this should be done after the tpm2_shutdown completes as
> > lower level shutdowns could remove register access.
> 
> This doesn't necessarily work.  The spec states that tpm2_shutdown
> must be the last command issued for an orderly shutdown, so if we
> call

After tpm2_shutdown is called ops must be null'd then we call down the
rest of the shutdown methods which can no longer issue TPM commands
due to ops being NULL.

Jason

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot