Re: [PATCH RESEND] dt-bindings: trivial-devices: Document SPI measurement on LWE boards

[Conor Dooley]

On Wed, Mar 26, 2025 at 10:27:02PM +0100, Lukasz Majewski wrote:
> Hi Conor,
> 
> > On Wed, Mar 26, 2025 at 06:42:40PM +0100, Lukasz Majewski wrote:
> > > Hi Conor,
> > >   
> > > > On Wed, Mar 26, 2025 at 03:09:30PM +0100, Lukasz Majewski wrote:  
> > > > > The measurement device on Liebherr's (LWE) boards is used to
> > > > > monitor the overall state of the device. It does have SPI
> > > > > interface to communicate with Linux host via spidev driver.
> > > > > Document the SPI DT binding as trivial SPI device.
> > > > > 
> > > > > Signed-off-by: Lukasz Majewski <lukma@xxxxxxx>
> > > > > ---    
> > > > 
> > > > You should not do a resend with no explanation as to why.
> > > > Additionally, I would like to know why my review on the original
> > > > patch was ignored:
> > > > https://lore.kernel.org/all/20250225-despair-rural-dc10216005f4@spud/#t
> > > >   
> > > 
> > > I've made a mistake, as I've used the lwe prefix, which is the
> > > different branch office for Liebherr.
> > > 
> > > As we discussed last time - it would be better to use the already
> > > present 'lwn' vendor prefix as several other boards from this
> > > company use it (display5, bk4, xea, btt3).
> > > 
> > > And this was apparent, after I've resent the patches. My mistake.
> > > 
> > > Regarding the comment - on xea, btt the binding would be used, as
> > > those two boards (based on imx287) are using it.
> > > 
> > > Hence, single "trivial device" would be OK.
> > > 
> > > The v2 of this patch has the proper 'lwn,btt' binding.  
> > 
> > I'm sorry, I don't understand how this excuses using the same binding
> > for different devices.
> 
> Ok, so maybe I will just explain how things are on those devices and we
> can find some solution.
> 
> So we do have two devices - based on imx287: XEA (rev 1,2) and BTT (rev
> 0,1,2).
> 
> We do have a measurement device connected to SPI (on both above
> devices). This device has a protocol, which is fully served in user
> space (just the /dev/spidevX.Y is required for it).
> 
> Hence the 'lwn,btt' can be used for all those measurement devices.

You mentioned in the previous version of this that the devices used a
"similar approach", not the same approach. Now you're saying that it is
the same. Which is it?
If a kernel driver was handling this, rather than userspace via spidev,
would it have to differentiate between xea and btt devices?

Attachment: signature.asc
Description: PGP signature