Re: [PATCH v3 3/5] media: i2c: add driver for the SK Hynix Hi-846 8M pixel camera

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Martin,

On Wed, Jun 02, 2021 at 02:00:11PM +0200, Martin Kepplinger wrote:
> Am Dienstag, dem 01.06.2021 um 03:14 +0300 schrieb Laurent Pinchart:
> > On Mon, May 31, 2021 at 02:07:35PM +0200, Martin Kepplinger wrote:
> > > The SK Hynix Hi-846 is a 1/4" 8M Pixel CMOS Image Sensor. It supports
> > > usual features like I2C control, CSI-2 for frame data, digital/analog
> > > gain control or test patterns.
> > > 
> > > This driver supports the 640x480, 1280x720 and 1632x1224 resolution
> > > modes. It supports runtime PM in order not to draw any unnecessary
> > > power.
> > > 
> > > The part is also called YACG4D0C9SHC and a datasheet can be found at
> > > https://product.skhynix.com/products/cis/cis.go
> > > 
> > > The large sets of partly undocumented register values are for example
> > > found when searching for the hi846_mipi_raw_Sensor.c Android driver.
> > 
> > A common story, unfortunately :-S
> > 
> > I've done an initial review, I'll likely have more comments on v4, but
> > you should have quite a few things to address already :-)
> > 
> > > Signed-off-by: Martin Kepplinger <martin.kepplinger@xxxxxxx>
> > > ---
> > >  MAINTAINERS                |    6 +
> > >  drivers/media/i2c/Kconfig  |   13 +
> > >  drivers/media/i2c/Makefile |    1 +
> > >  drivers/media/i2c/hi846.c  | 2138 ++++++++++++++++++++++++++++++++++++
> > >  4 files changed, 2158 insertions(+)
> > >  create mode 100644 drivers/media/i2c/hi846.c

[snip]

> Thank you, Laurent for that wonderful review. It made me rework/fix the
> power supply interface + sequencing in the driver (and even better
> understand how it's supplied on my board).
> 
> I want to take all your review into account for a next revision, except
> for the additional bits for the register definitions, that should
> encode the length, if that's ok. We can choose whether to write 1 or 2
> bytes at a given address and it just looks nice and simple to me as it
> is.

I won't push strongly, but in my experience it's error-prone, as it's
easy to select the incorrect number of bytes. That's what led me to
create this mechanism to bundle register addresses and sizes, it has
simplified my life when writing drivers. I think it should actually be
turned into a helper, possibly provided by regmap.

-- 
Regards,

Laurent Pinchart



[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux