Re: [PATCH] media: i2c: imx258: Parse and register properties

From: Jacopo Mondi
Date: Tue Jan 03 2023 - 12:17:03 EST


Hi Robert

On Tue, Jan 03, 2023 at 03:11:44PM +0100, Robert Mader wrote:
> On 02.01.23 15:06, Jacopo Mondi wrote:
> > Hi Robert
> >
> > On Sun, Dec 25, 2022 at 04:42:34PM +0100, Robert Mader wrote:
> > > Analogous to e.g. the imx219. This enables propagating
> > > V4L2_CID_CAMERA_SENSOR_ROTATION values so that libcamera
> > > can detect the correct rotation from the device tree
> > > and propagate it further to e.g. Pipewire.
> > >
> > > Signed-off-by: Robert Mader <robert.mader@xxxxxxxxxxxxx>
> > > ---
> > > drivers/media/i2c/imx258.c | 13 ++++++++++++-
> > > 1 file changed, 12 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/media/i2c/imx258.c b/drivers/media/i2c/imx258.c
> > > index eab5fc1ee2f7..85819043d1e3 100644
> > > --- a/drivers/media/i2c/imx258.c
> > > +++ b/drivers/media/i2c/imx258.c
> > > @@ -9,6 +9,7 @@
> > > #include <linux/pm_runtime.h>
> > > #include <media/v4l2-ctrls.h>
> > > #include <media/v4l2-device.h>
> > > +#include <media/v4l2-fwnode.h>
> > > #include <asm/unaligned.h>
> > >
> > > #define IMX258_REG_VALUE_08BIT 1
> > > @@ -1149,6 +1150,7 @@ static int imx258_init_controls(struct imx258 *imx258)
> > > {
> > > struct i2c_client *client = v4l2_get_subdevdata(&imx258->sd);
> > > struct v4l2_ctrl_handler *ctrl_hdlr;
> > > + struct v4l2_fwnode_device_properties props;
> > Might be nicer to move this one line up
>
>  Can you say what's your reasoning? I personally slightly prefer
> alphabetical order, but no strong opinion :)
>

I've often been instructed to try to respect the inverse-xmas-tree
order when declaring variables, if possible. I now realize it's a sort
of cargo cult, as the rule is not written anywhere, so I can't ask
you to comply with what seems to be a personal preference :)

> > > s64 vblank_def;
> > > s64 vblank_min;
> > > s64 pixel_rate_min;

And anyway this breaks the rule already, so up to you, really

> > > @@ -1156,7 +1158,7 @@ static int imx258_init_controls(struct imx258 *imx258)
> > > int ret;
> > >
> > > ctrl_hdlr = &imx258->ctrl_handler;
> > > - ret = v4l2_ctrl_handler_init(ctrl_hdlr, 8);
> > > + ret = v4l2_ctrl_handler_init(ctrl_hdlr, 10);
> > I count 9 controls being registered before this patch, not 8. Do I
> > count them right ?
> >
> > If that's case, as v4l2_ctrl_new_fwnode_properties()
> > can register up to two controls (V4L2_CID_ROTATION and
> > V4L2_CID_ORIENTATION) I would pre-reserve 11 controls not 10 to avoid
> > relocations.
>
> Indeed, looks like bumping this was forgotten in
> c6f9d67e2ac625e331f6a7f5715d2f809ff0a922
>

Sorry, not your fault but since you're here... You can either make a
patch that fixes the existing number of controls first then apply
this one on top, or address the issue here, maybe with a small mention in
the commit message along the lines of:

While at it, reserve space for 3 additional controls even if
v4l2_ctrl_new_fwnode_properties() can only register 2 of
them, to fix the existing implementation which reserve space for 8
controls but actually registers 9.

Thanks
j

> > > if (ret)
> > > return ret;
> > >
> > > @@ -1232,6 +1234,15 @@ static int imx258_init_controls(struct imx258 *imx258)
> > > goto error;
> > > }
> > >
> > > + ret = v4l2_fwnode_device_parse(&client->dev, &props);
> > > + if (ret)
> > > + goto error;
> > > +
> > > + ret = v4l2_ctrl_new_fwnode_properties(ctrl_hdlr, &imx258_ctrl_ops,
> > > + &props);
> > > + if (ret)
> > > + goto error;
> > > +
> > The rest looks good to me!
> >
> > Thanks
> > j
> Thanks!
>
> --
> Robert Mader
> Consultant Software Developer
>
> Collabora Ltd.
> Platinum Building, St John's Innovation Park, Cambridge CB4 0DS, UK
> Registered in England & Wales, no. 5513718
>