On Thu, Oct 18, 2012 at 05:39:59PM +0300, Tero Kristo wrote:
> On Thu, 2012-10-18 at 16:53 +0300, Felipe Balbi wrote:
> > hi,
> >
> > On Thu, Oct 18, 2012 at 03:18:04PM +0300, Tero Kristo wrote:
> > > > > +static int __init omap4430_phy_power_down(void)
> > > > > +{
> > > > > + void __iomem *ctrl_base;
> > > > > +
> > > > > + if (!cpu_is_omap44xx())
> > > > > + return 0;
> > > > > +
> > > > > + ctrl_base = ioremap(OMAP443X_SCM_BASE, SZ_1K);
> > > > > + if (!ctrl_base) {
> > > > > + pr_err("control module ioremap failed\n");
> > > > > + return -ENOMEM;
> > > > > + }
> > > > > +
> > > > > + /* Power down the phy */
> > > > > + __raw_writel(PHY_PD, ctrl_base + CONTROL_DEV_CONF);
> > > > > +
> > > > > + iounmap(ctrl_base);
> > > > > +
> > > > > + return 0;
> > > > > +}
> > > > > +early_initcall(omap4430_phy_power_down);
> > > > > +#endif
> > > >
> > > > I think you could do it even if the driver is enabled.
> > >
> > > Actually not at least now, it looks like the driver is not controlling
> > > this bit at all, so the driver would fail if we do this.
> >
> > then we can consider that a bug in the driver. Kishon, I thought you had
> > added SCM address space to PHY driver for this particular reason until
> > we get SCM driver, wasn't it ??
>
> Yes, I would say its a bug in the driver. However we need this disable
> mechanism for the case where we don't have the driver also (which is the
> default config for omap.)
sure, of course. But I'd like to see it unconditionally done, meaning
that there needs to be a counterpart in the driver to make sure you can
run this unconditionally during early phases of boot up ;-)
--
balbi
Attachment:
signature.asc
Description: Digital signature
