Re: [PATCH 1/2] [media] imx214: Add imx214 camera sensor driver

[Sakari Ailus]

Hi, Ricardo!

On Wed, Sep 26, 2018 at 08:06:55AM +0200, Ricardo Ribalda Delgado wrote:
> Hi Sakari
> On Wed, Sep 26, 2018 at 7:26 AM Ricardo Ribalda Delgado
> <ricardo.ribalda@xxxxxxxxx> wrote:
> >
> > Hello Sakari
> > On Mon, Sep 24, 2018 at 10:32 PM Sakari Ailus
> > <sakari.ailus@xxxxxxxxxxxxxxx> wrote:
> > >
> > > Hi Ricardo,
> > >
> > > On Fri, Sep 21, 2018 at 12:15:55PM +0200, Ricardo Ribalda Delgado wrote:
> > > ...
> > > > > > +static struct reg_8 mode_1920x1080[];
> > > > > > +static struct reg_8 mode_4096x2304[];
> > > > >
> > > > > Const. Could you rearrange the bits to avoid the forward declarations?
> > > > Const done, but I prefer to keep the forward declaration. Otherwise
> > > > the long tables will "hide" the mode declaration.
> > >
> > > Well, I guess the long tables do "hide" a bunch of other stuff, too. :-)
> > > But... I agree there's no trivial way around those tables either.
> > >
> > > It appears I'm not the only one who's commented on the matter of the
> > > forward declaration.
> >
> > Ok, I will change it, Eppur si muove  ;)
> > >
> > > ...
> > >
> > > > > > +static int imx214_probe(struct i2c_client *client)
> > > > > > +{
> > > > > > +     struct device *dev = &client->dev;
> > > > > > +     struct imx214 *imx214;
> > > > > > +     struct fwnode_handle *endpoint;
> > > > > > +     int ret;
> > > > > > +     static const s64 link_freq[] = {
> > > > > > +             (IMX214_DEFAULT_PIXEL_RATE * 10LL) / 8,
> > > > >
> > > > > You should check the link frequency matches with that from the firmware.
> > > >
> > > > I am not sure what you mean here sorry.
> > >
> > > The system firmware holds safe frequencies for the CSI-2 bus on that
> > > particular system; you should check that the register lists are valid for
> > > that frequency.
> > >
> 
> Can you point me to a driver that does this? Just for the example. Thanks

The imx319 driver:

<URL:https://patchwork.linuxtv.org/patch/52233/>

> > I am with you here. The i2c controller is a different driver but is
> > integrated with camss. Checkout
> > https://patchwork.kernel.org/patch/10569961/ I am interacting with
> > Todor and Vinod to enable the i2c port indepently with camss. At least
> > now it does not kill the port after an i2c timeout :)

Nice!!

-- 
Cheers,

Sakari Ailus
sakari.ailus@xxxxxxxxxxxxxxx