Capture image from imx219 sensor on i.MX8MM

[Dave Stevenson]

Hi Frieder

On Mon, 6 Feb 2023 at 15:41, Frieder Schrempf
<frieder.schrempf at kontron.de> wrote:
>
> Hi Laurent,
>
> On 06.02.23 16:31, Laurent Pinchart wrote:
> > Hi Frieder,
> >
> > On Mon, Feb 06, 2023 at 03:24:41PM +0100, Frieder Schrempf wrote:
> >> Hi,
> >>
> >> I'm trying to get a camera stream on our i.MX8MM board up and running
> >> with Linux 6.1. I'm using the RPi v2.1 camera module (imx219) as sensor.
> >>
> >> I basically copied the devicetree setup from Tim's overlay at
> >> arch/arm64/boot/dts/freescale/imx8mm-venice-gw72xx-0x-imx219.dts.
> >>
> >> With all the drivers enabled the devices (imx219, imx7-media-csi,
> >> imx-mipi-csis) seem to probe just fine and I see /dev/video0 and
> >> /dev/media0.
> >>
> >> For the v4l subdevices I would expect to see one for the sensor and one
> >> for the CSI bridge. But only the latter is there (see below).
> >>
> >> Is this correct? Am I missing something? How can I setup/enable the
> >> pipeline/stream?
> >
> > Your expectations are correct, but the result isn't. You should see the
> > camera sensor in the media graph. Has the imx219 been probed by the
> > driver ? Did probe succeed ?
>
> Yes, the imx219_probe() works just fine and returns with code 0. I also
> see v4l2_async_register_subdev_sensor() and the underlying code being
> called.
>
> So I really don't understand why I'm missing the sensor subdev. Any
> ideas how to continue debugging?

It'd be worth checking
/sys/kernel/debug/v4l2-async/pending_async_subdevices to see if the
v4l2_async framework is waiting on any (sub)devices.

  Dave