Re: [PATCH v3 2/3] media: i2c: Add GT97xx VCM driver

[Kieran Bingham <kieran.bingham@ideasonboard.com>]

Hi Zhi,

Also - Cc: Dongchun Zhu <dongchun.zhu@mediatek.com> who is listed as the
DW9768 VCM driver author...

Quoting Zhi Mao (毛智) (2024-06-12 12:13:40)
> Hi Angelo,
> 
> Thanks for your review.
> 
> On Wed, 2024-06-12 at 09:07 +0200, AngeloGioacchino Del Regno wrote:
> > Il 12/06/24 03:20, Zhi Mao ha scritto:
> > > Add a V4L2 sub-device driver for Giantec GT97xx VCM.
> > > 
> > > Signed-off-by: Zhi Mao <zhi.mao@mediatek.com>
> > 
> > Hello Zhi,
> > 
> > I fail to see why would you need to upstream this new driver instead
> > of
> > simply adding the IC_INFO_REG to the already existing (and more
> > featureful)
> > dw9768 driver, which also seems to support the Giantec GT9769 VCM.

Even more so especially as
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/i2c/dw9768.c
already directly supports the compatible strings added by this driver -
surely we don't want multiple (near identical) drivers matching the same
compatible string?

> > 
> 
> Our project uses Giantec VCM hardware. 
> For detailed vendor information, please visit: (
> https://en.giantec-semi.com/yqmd/164). 
> The VCM datasheet we are referencing is provided by Giantec. 
> Currently, the relationship between Giantec VCM and Dongwoon VCM is
> unclear, but Dongwoon seems to be another manufacturer of VCM
> hardware. 
> 
> From the perspective of software driver development and maintenance, it
> makes sense for each vendor's hardware should have its own software
> driver.

Personally, I don't think so. If two vendors make identical parts, we
shouldn't have two identical drivers.

I still have plans to refactor VCM drivers if I get some spare-time(tm)
as almost each driver does the same identical task. They're all just
copies of the boilerplate.  That seems like something we should reduce,
not increase.

--
Kieran


> So, I upstream a new VCM driver for Giantec.
> 
> > Cheers,
> > Angelo