Re: [PATCH v4 2/3] media: i2c: dw9719: Update DW9800K driver

From: Krzysztof Kozlowski

Date: Sat Nov 21 2026 - 05:46:14 EST


On Thu, Nov 19, 2026 at 10:51:24AM +0100, Griffin Kroah-Hartman wrote:
> The previous DW9800K driver code was created without access to a
> datasheet for the part, now that we have one we can clean up some
> misunderstandings and add some features for further expandability.
>
> More specifically:
>
> Previously we were blindly writing registers to their downstream values,
> Now we know what the correct default values of SAC and SACT, and write
> to the correct bits as specified from the datasheet.
>
> We found that the device does indeed have an ID, and add it to the
> switch statement when setting up the default device properties.
>
> The largest change would be the implementation of "Ringing," this is
> actually an internal switch bit used to differentiate whether the SAC
> (smart actuator control) modes are enabled. if the "ringing" bit is
> zero, then actuator modes DIRECT and LSC are available(based on the
> value written to the Mode register). If the bit is 1, SAC2-SAC5 are
> available.
>
> For the devices we tested this driver on, no new values are being
> written, but now the option for further use of the DW9800k is available
> to others if they wish to do so.
>
> Fixes: 083f7f0dce16 ("media: i2c: dw9719: Add DW9800K support")
> Signed-off-by: Griffin Kroah-Hartman <griffin.kroah@xxxxxxxxxxxxx>

Well, not your fault, but that's a nice example why register values
should not be the DT ABI. I am going to bookmark this patchset and use
it in future discussions when people insist on using hardware values as
the actual ABI :/

Best regards,
Krzysztof