[PATCH v2 2/5] drm/rockchip: dw_hdmi_qp: Improve error handling with dev_err_probe()

[Laurent Pinchart]

On Thu, Oct 09, 2025 at 10:24:25AM +0300, Cristian Ciocaltea wrote:
> On 10/6/25 7:26 PM, Laurent Pinchart wrote:
> > On Mon, Oct 06, 2025 at 05:37:23PM +0300, Cristian Ciocaltea wrote:
> >> On 10/6/25 3:02 PM, Dmitry Baryshkov wrote:
> >>> On Mon, Oct 06, 2025 at 02:55:38AM +0300, Laurent Pinchart wrote:
> >>>> From: Cristian Ciocaltea <cristian.ciocaltea at collabora.com>
> >>>>
> >>>> The error handling in dw_hdmi_qp_rockchip_bind() is quite inconsistent,
> >>>> i.e. in some cases the error code is not included in the message, while
> >>>> in some other cases there is no check for -EPROBE_DEFER.
> >>>>
> >>>> Since this is part of the probe path, address the aforementioned issues
> >>>> by switching to dev_err_probe(), which also reduces the code a bit.
> >>>>
> >>>> Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea at collabora.com>
> >>>> Signed-off-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> >>>> ---
> >>>>  .../gpu/drm/rockchip/dw_hdmi_qp-rockchip.c    | 62 +++++++------------
> >>>>  1 file changed, 24 insertions(+), 38 deletions(-)
> >>>>
> >>>> diff --git a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> >>>> index 7d531b6f4c09..4e7794aa2dde 100644
> >>>> --- a/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> >>>> +++ b/drivers/gpu/drm/rockchip/dw_hdmi_qp-rockchip.c
> >>>> @@ -457,10 +457,8 @@ static int dw_hdmi_qp_rockchip_bind(struct device *dev, struct device *master,
> >>>>  		return -ENODEV;
> >>>>  
> >>>>  	if (!cfg->ctrl_ops || !cfg->ctrl_ops->io_init ||
> >>>> -	    !cfg->ctrl_ops->irq_callback || !cfg->ctrl_ops->hardirq_callback) {
> >>>> -		dev_err(dev, "Missing platform ctrl ops\n");
> >>>> -		return -ENODEV;
> >>>> -	}
> >>>> +	    !cfg->ctrl_ops->irq_callback || !cfg->ctrl_ops->hardirq_callback)
> >>>> +		return dev_err_probe(dev, -ENODEV, "Missing platform ctrl ops\n");
> >>>
> >>> This only makes sense for the purpose of unification.
> >>
> >> Right, as mentioned in the commit description, the intention was to ensure
> >> consistent error handling across the probe path rather than limiting the scope
> >> to -EPROBE_DEFER exclusively.
> > 
> > Should I revert this change in v3 or keep it ? I see value in
> > unification, but I don't mind either way. Dmitry, what's your preference
> > ?
> 
> I missed to point out this patch has been also sent a while ago as part of
> another series [1] which should be ready for merging.  It'd be great if there's
> no need to revert any changes, otherwise we need to keep those in sync.
> 
> Regardless, I'll let you know if that gets applied first, allowing us to drop
> this one after rebasing.

Thanks for the information. I'm happy to merge 3/5 on top of that
series. The DT binding change in 1/5 can be merged separately.

> [1] https://lore.kernel.org/all/20250903-rk3588-hdmi-cec-v4-3-fa25163c4b08@collabora.com/

-- 
Regards,

Laurent Pinchart