Re: [PATCH 3/3] platform/x86: intel_speed_select_if: Remove hardcoded map size

[Andy Shevchenko]

On Tue, Sep 26, 2023 at 08:04:46AM -0700, srinivas pandruvada wrote:
> On Tue, 2023-09-26 at 16:16 +0300, Andy Shevchenko wrote:
> > On Mon, Sep 25, 2023 at 12:45:55PM -0700, Srinivas Pandruvada wrote:
> > > The driver is using 256 as the size while calling devm_ioremap().
> > > The
> > > maximum offset is already part of struct isst_mmio_range. Use the
> > > maximum offset (end field of the struct) plus 4 as the map size to
> > > remove
> > > hardcoded value of 256.

...

> > > +       punit_dev->mmio_range = (struct isst_mmio_range *) ent-
> > > >driver_data;
> > > +
> > > +       punit_dev->punit_mmio = devm_ioremap(&pdev->dev, base_addr,
> > > +                                            punit_dev-
> > > >mmio_range[1].end + sizeof(u32));
> > 
> > Can we rather fix the mmio_range driver data to have end be actually
> > not the
> > offset of the last dword? (Better maybe to keep length there.)
> > 
> We can. But that has to be separate patch on top as there are other
> places this range is used.

Still you can add a third member for now and then clean up it later as it's all
in one file.

...

> > With help of
> > 
> >         struct resource r;
> >         ...
> >         r = DEFINE_RES_MEM(base_addr, mmio_range.beg +
> > mmio_range.len);
> > 
> > you can switch to devm_ioremap_resource() API.
> What is the advantage of creating a resource and then call
> devm_ioremap_resource()?

It manages resource via global resource management and also prints an error
messages in case of errors.

-- 
With Best Regards,
Andy Shevchenko