On 16/09/2022 12:03, Sakari Ailus wrote:
And even if you have small differences in the registers you'll need to
write there, you can still differentiate between the sensors based on the
compatible string.
I don't have strong opinion on the grey areas though. Still if the register
set is exactly the same, then the driver should also be the same.
Right now we have
- An imx412 driver that works on imx577 unmodified on Qcom hardware
- A Nvidia driver modified by Leopard imaging that ignores the chip id
and uses the same init sequence.
This driver is called "imx477" and I can verify it works with
imx412 and imx577.
The code for this driver modifies the original out of tree driver they
had and stops validating the CHIP_ID
So I think we can take it as read it works with imx412, imx477 and
imx577 - I've verified the first and last is the case.
We know the upstream driver works with the Intel platform and I've
tested/used it on Qcom with minimal change, so I'm happy to stand over
listing both imx412 and imx577.
Its pretty clear the init sequence works for imx412, imx477 and imx577
so, it feels to me like the right thing to do is to add in the
compatible strings and if/when we get better chip specific data say for
higher resolutions, add that resolution init sequence in when the compat
matches.
---
bod
