Re: [PATCH 3/4] dt-bindings: mtd: macronix,mx25l12833f: add SPI-NOR chip

[Erez <erezgeva2@gmail.com>]

On Fri, 28 Jun 2024 at 19:45, Michael Walle <mwalle@kernel.org> wrote:
>
> > What is the kernel policy regarding obsolete flash chips?
> > Macronix annonce on end of life of mx25l12805d in 2010.
>
> Which doesn't mean there are no boards using it. But EOL in 2010
> might be a strong argument to remove it. But I don't see the need
> for it, yet.

Fair enough. That means I was in the right direction.
Just need a better description for the patch :-)
I'll update for version 2.

>
> > > Also please have a look at
> > > https://docs.kernel.org/driver-api/mtd/spi-nor.html
> >
> > The new mx25l12833f supports SFDP, the obsolete mx25l12805d does not.
> > I did manage to read the SFDP, though I do not have a copy of it (I do
> > not have the hardware any more).
>
> So how would you test newer versions of this series?

I develop the OTP callbacks with a mx25l12805d for a company using it.
I am no longer in contact with them.

The testing can be done with any Macronix SPI-NOR with OTP.
I did not check, but I guess most of their chips have an OTP using the
same opcodes/methods.

At least mx25l12805d and mx25l12833f use the same.
I did not add mx25l12833f as I did not use it and as it uses
asymmetric OTP (the MTD supports symmetric OTP).

>
> > To my best knowledge SFDP table do not contain information on OTP,
>
> That's true, but we need it to detect flashes during runtime in the
> future. Imagine sometime later, there's a third flash with this
> exact ID, then we need to have the SFDP to figure out if there are
> any differences between yours and the new one.

Sure, as I said the mx25l12833f does provide SFDP.
In patch 4, I removed the size property to force the driver to read SFDP.
I was not sure about the flags though.
I did not see any difference compared to using the setting of
mx25l12805d without SFDP.
Yet my focus of work was the OTP.

Erez

>
> -michael