Hi!


> > > 1) Make sure there will be no regressions,
> > 
> > Well, all I have running recent kernels is N900. If ccp branch works
> > for you on N9, that's probably as much testing as we can get.
> > 
> > > 2) clean things up in the omap3isp; which resources are needed and when
> > > (e.g. regulators, PHY configuration) isn't clear at the moment and
> > > 
> > > 2) have one driver using the implementation.
> > > 
> > > At least 1) is needed. I think a number of framework patches could be
> > > mergeable before 2) and 3) are done. I can prepare a set later this week.
> > > But even that'd be likely for 4.14, not 4.13.
> > 
> > Yep, it is too late for v4.13 now. But getting stuff ready for v4.14
> > would be good.
...
> > @@ -302,13 +303,16 @@ int omap3isp_csiphy_acquire(struct isp_csiphy *phy)
> >  	if (rval < 0)
> >  		goto done;
> >  
> > -	rval = csiphy_set_power(phy, ISPCSI2_PHY_CFG_PWR_CMD_ON);
> > -	if (rval) {
> > -		regulator_disable(phy->vdd);
> > -		goto done;
> > +	if (phy->isp->revision == ISP_REVISION_15_0) {
> 
> Shouldn't you make the related changes to omap3isp_csiphy_release() as
> well?
> 
> Other than that the patch looks good to me.

Ah, yes, that needs to be fixed. Thanks for review.

I'll refresh the series. I believe we now have everything neccessary
to have useful driver for 4.14. Series is still based on 4.12-rc3, I
can rebase it when there's better base.

Best regards,

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html