Re: [PATCH 06/11] media: adv748x-csi2: Implement enum_mbus_codes

["Niklas Söderlund" <niklas.soderlund+renesas@ragnatech.se>]

On 2024-05-06 10:42:32 +0200, Jacopo Mondi wrote:
> Hi Laurent
> 
> On Mon, May 06, 2024 at 11:38:41AM GMT, Laurent Pinchart wrote:
> > On Mon, May 06, 2024 at 10:10:00AM +0200, Jacopo Mondi wrote:
> > > On Mon, May 06, 2024 at 12:07:40AM GMT, Laurent Pinchart wrote:
> > > > On Fri, May 03, 2024 at 05:51:21PM +0200, Jacopo Mondi wrote:
> > > > > Define a list of supported mbus codes for the TXA and TXB CSI-2
> > > > > transmitters and implement the enum_mbus_code operation.
> > > >
> > > > The commit message should explain why, not just what. Explaining why the
> > >
> > > Should I explain why the driver has to implement enum_mbus_codes ?
> >
> > You can just note it's mandatory to implement ?
> >
> > > > formats for TXA and TXB differ would also be useful.
> > >
> > > ok
> > >
> > > > > Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> > > > > ---
> > > > >  drivers/media/i2c/adv748x/adv748x-csi2.c | 35 ++++++++++++++++++++++++
> > > > >  1 file changed, 35 insertions(+)
> > > > >
> > > > > diff --git a/drivers/media/i2c/adv748x/adv748x-csi2.c b/drivers/media/i2c/adv748x/adv748x-csi2.c
> > > > > index 9da7f6742a2b..219417b319a6 100644
> > > > > --- a/drivers/media/i2c/adv748x/adv748x-csi2.c
> > > > > +++ b/drivers/media/i2c/adv748x/adv748x-csi2.c
> > > > > @@ -21,6 +21,18 @@ static const struct v4l2_mbus_framefmt adv748x_csi2_default_fmt = {
> > > > >  	.field = V4L2_FIELD_NONE,
> > > > >  };
> > > > >
> > > > > +static const unsigned int adv748x_csi2_txa_fmts[] = {
> > > > > +	MEDIA_BUS_FMT_YUYV8_1X16,
> > > > > +	MEDIA_BUS_FMT_YUYV10_1X20,
> > > >
> > > > CSI-2 uses UYVY, not YUYV.
> > >
> > > As this a recurring comment in the series (rightfully, I'm not
> > > questioning that) how does it work with existing test scripts assuming
> > > YUYV ? The same question could be asked about the issue Niklas had: is
> > > changing what pad an operation is allowed to be called on legit ?
> > >
> > > My answer would be yes, otherwise we'll be forever stuck, but I would
> > > like to check, especially with Niklas which maintains vin-tests.
> >
> > As usual, changes of behaviour are only regressions if someone complains
> > about them. We can keep backward-compatibility, but it would then be
> > nice to also support the right media bus codes, and to add a comment
> > that clearly indicates which ones are for backward-compatibility.
> >
> > That being said, given that userspace should use UYVY, and given that
> > the driver currently accepts UYVY, it should at least be enumerated. I
> > would skip enumeration of the media bus codes that we keep for
> > backward-compatibility, even if we accept them at runtime.

I'm OK just dropping them and update the use-cases, better get this 
right for the future. If we find use-cases we can't update, keeping 
accepting the formats but not enumerating them seems like a good option.

> >
> 
> What about the change to the pad on which s_fmt is allowed on R-Car
> CSI-2 ?
> 
> > > > > +	MEDIA_BUS_FMT_RGB565_1X16,
> > > > > +	MEDIA_BUS_FMT_RGB666_1X18,
> > > > > +	MEDIA_BUS_FMT_RGB888_1X24,
> > > > > +};
> > > > > +
> > > > > +static const unsigned int adv748x_csi2_txb_fmts[] = {
> > > > > +	MEDIA_BUS_FMT_YUYV8_1X16,
> > > > > +};
> > > > > +
> > > > >  int adv748x_csi2_set_virtual_channel(struct adv748x_csi2 *tx, unsigned int vc)
> > > > >  {
> > > > >  	return tx_write(tx, ADV748X_CSI_VC_REF, vc << ADV748X_CSI_VC_REF_SHIFT);
> > > > > @@ -146,6 +158,28 @@ static const struct v4l2_subdev_video_ops adv748x_csi2_video_ops = {
> > > > >   * But we must support setting the pad formats for format propagation.
> > > > >   */
> > > > >
> > > > > +static int adv748x_csi2_enum_mbus_code(struct v4l2_subdev *sd,
> > > > > +				       struct v4l2_subdev_state *sd_state,
> > > > > +				       struct v4l2_subdev_mbus_code_enum *code)
> > > > > +{
> > > > > +	struct adv748x_csi2 *tx = adv748x_sd_to_csi2(sd);
> > > > > +	const unsigned int *codes = is_txa(tx) ?
> > > > > +				    adv748x_csi2_txa_fmts :
> > > > > +				    adv748x_csi2_txb_fmts;
> > > > > +	size_t num_fmts = is_txa(tx) ? ARRAY_SIZE(adv748x_csi2_txa_fmts)
> > > > > +				     : ARRAY_SIZE(adv748x_csi2_txb_fmts);
> > > > > +
> > > > > +	if (code->pad != ADV748X_CSI2_SOURCE)
> > > > > +		return -EINVAL;
> > > > > +
> > > > > +	if (code->index >= num_fmts)
> > > > > +		return -EINVAL;
> > > > > +
> > > > > +	code->code = codes[code->index];
> > > > > +
> > > > > +	return 0;
> > > > > +}
> > > > > +
> > > > >  static struct v4l2_mbus_framefmt *
> > > > >  adv748x_csi2_get_pad_format(struct v4l2_subdev *sd,
> > > > >  			    struct v4l2_subdev_state *sd_state,
> > > > > @@ -235,6 +269,7 @@ static int adv748x_csi2_get_mbus_config(struct v4l2_subdev *sd, unsigned int pad
> > > > >  }
> > > > >
> > > > >  static const struct v4l2_subdev_pad_ops adv748x_csi2_pad_ops = {
> > > > > +	.enum_mbus_code = adv748x_csi2_enum_mbus_code,
> > > > >  	.get_fmt = adv748x_csi2_get_format,
> > > > >  	.set_fmt = adv748x_csi2_set_format,
> > > > >  	.get_mbus_config = adv748x_csi2_get_mbus_config,
> >
> > --
> > Regards,
> >
> > Laurent Pinchart

-- 
Kind Regards,
Niklas Söderlund