Hi,

quoted

Nope.

Are you really questioning my picture? Ridiculous ... see [0]

No, I mean some boards of this model have SoC silkscreen RK3528 and
others have RK3528A. The same is true for another Hinlink H28K SBC.

quoted

I'm sort of impressed on with which conviction you continue to claim
plain wrong things: [1], [2], [3].

If you spend a few minutes running mainline u-boot or BSP kernel
on your RK3528 board before blaming me:

I can't see the point here: I don't think it matters wether you or I
actually have this version on any of our boards. It exists: you may like it
or not.

This picture was actually a reply to your claim "... so we have never seen
the silk screen printed with RK3528 ... "  in your mail dated 2025/07/19.

Initially my only question was, why we don't use "rockchip,rk3528a" as
compatible when working on boards where the silkscreens says exactly that.
It was obviously copied from vendor and now it's "too late", "too
complicated" or something, idk and I'm fine with it.

BL31:
INFO:    rk_otp_init finish!
INFO:    RK3528 SoC (0x101)

mainline u-boot:
------
U-Boot 2025.07-...

Model: Generic RK3528
SoC:   RK3528A
------

BSP kernel:
[    0.768514] rockchip-cpuinfo cpuinfo: SoC            : 35281000
[    0.768990] rockchip-cpuinfo cpuinfo: Serial         : ...

quoted

I'm fine if upstream decides not to care. But it is and remains wrong
to claim that the other version does not exist

Unless Rockchip says they fused the wrong OTP during production.
Regardless of the SoC silkscreen, the chip type on OTP is the same,
so how does Rockchip distinguish these chips?

Please read the rest of my previous reply where I sent code locations 
where and how they do.

--
2.25.1