[PATCH] drm/bridge: adv7511: Fix cec clock EPROBE_DEFER handling

[Vincent Whitchurch]

On Mon, Apr 06, 2020 at 02:58:17AM +0200, Laurent Pinchart wrote:
> On Tue, Mar 31, 2020 at 04:16:29PM +0200, Vincent Whitchurch wrote:
> >  int adv7511_cec_init(struct device *dev, struct adv7511 *adv7511)
> >  {
> >  	unsigned int offset = adv7511->type == ADV7533 ?
> >  						ADV7533_REG_CEC_OFFSET : 0;
> > -	int ret = adv7511_cec_parse_dt(dev, adv7511);
> > +	int ret;
> >  
> > -	if (ret)
> > -		goto err_cec_parse_dt;
> > +	if (!adv7511->cec_clk)
> > +		goto err_cec_no_clock;
> > +
> > +	clk_prepare_enable(adv7511->cec_clk);
> > +	adv7511->cec_clk_freq = clk_get_rate(adv7511->cec_clk);
> >  
> >  	adv7511->cec_adap = cec_allocate_adapter(&adv7511_cec_adap_ops,
> >  		adv7511, dev_name(dev), CEC_CAP_DEFAULTS, ADV7511_MAX_ADDRS);
> > @@ -342,8 +331,11 @@ int adv7511_cec_init(struct device *dev, struct adv7511 *adv7511)
> >  err_cec_alloc:
> >  	dev_info(dev, "Initializing CEC failed with error %d, disabling CEC\n",
> >  		 ret);
> > -err_cec_parse_dt:
> > +	clk_disable_unprepare(adv7511->cec_clk);
> > +	/* Ensure that adv7511_remove() doesn't attempt to disable it again. */
> 
> Would it make sense to call devm_clk_put() here to already release the
> clock ?

I've just sent out a v2 with this added.  Thank you for the review!