Re: RFC: (re-)binding the VFIO platform driver to a platform device

[Scott Wood <scottwood@freescale.com>]

On Wed, 2013-10-09 at 12:16 -0700, gregkh@linuxfoundation.org wrote:
> On Wed, Oct 09, 2013 at 07:02:25PM +0000, Yoder Stuart-B08248 wrote:
> > Have been thinking about this issue some more.  As Scott mentioned,
> > 'wildcard' matching for a driver can be fairly done in the platform
> > bus driver.  We could add a new flag to the platform driver struct:
> > 
> > diff --git a/drivers/base/platform.c b/drivers/base/platform.c
> > index 4f8bef3..4d6cf14 100644
> > --- a/drivers/base/platform.c
> > +++ b/drivers/base/platform.c
> > @@ -727,6 +727,10 @@ static int platform_match(struct device *dev, struct device_driver *drv)
> >         struct platform_device *pdev = to_platform_device(dev);
> >         struct platform_driver *pdrv = to_platform_driver(drv);
> > 
> > +       /* the driver matches any device */
> > +       if (pdrv->match_any)
> > +               return 1;
> > +
> >         /* Attempt an OF style match first */
> >         if (of_driver_match_device(dev, drv))
> >                 return 1;
> > 
> > However, the more problematic issue is that a bus driver has no way to 
> > differentiate from an explicit bind request via sysfs and a bind that
> > happened through bus probing.
> 
> That was by design, nice to see I implemented it properly :)
> 
> > I think something like the new flag in the snippet below would enable the platform
> > bus to support platform drivers that only bind on explicit request:
> > 
> > diff --git a/drivers/base/bus.c b/drivers/base/bus.c
> > index 4c289ab..daf6d24 100644
> > --- a/drivers/base/bus.c
> > +++ b/drivers/base/bus.c
> > @@ -201,7 +201,7 @@ static ssize_t bind_store(struct device_driver *drv, const char *buf,
> >         int err = -ENODEV;
> > 
> >         dev = bus_find_device_by_name(bus, NULL, buf);
> > -       if (dev && dev->driver == NULL && driver_match_device(drv, dev)) {
> > +       if (dev && dev->driver == NULL && driver_match_device(drv, dev, 1)) {
> 
> Magic flags are the spawn of your favorite anti-$DIETY.  I'm never going
> to accept that, sorry.
> 
> If you really want to do something "special" for the platform bus, then
> do it only for the platform bus.  But even then, you'll find me arguing
> that you really don't want to do it at all, sorry.

It's not (or shouldn't be) special for the platform bus.  The "don't
bind by default" flag (note that I am *not* referring to the above code
snippet) would be useful for VFIO PCI as well, as it would replace the
hacky and racy usage of new_id.

> I'm still yet to be convinced this is even an issue at all, but maybe
> that's just the jetlag talking...

If this is because you think we should use new_id, we just had a
discussion in this thread about why that's no good.

-Scott