RP2350 A and B variant

[elral]

The new RP2350 is available in variant A and B. The A variant has the same number of pins as the RP2040 where the B variant has 40 pins (and at least the analog pins on another pins).
For now there is one generic variant for the RP2350. I am wondering if it would make sense to mark this variant as A and setup another one for variant B.

Yesterday I moved our project from platformIO to the Arduino IDE to test it with the RP2350. For what I could test everything was working fine (except AnaologIn as I am using an PGA2350, but will test it with a Pico2 again).
Really great job!! Thanks a lot for your effort!

[earlephilhower]

They're both the same die, so it should just be a matter of adding the new GPIO_HI_* registers when they're used explicitly (only IRQ masking in SPI/Ethernet IIRC) and adjusting the digital/analogWrite/Read to support >31 GPIOs.

[earlephilhower]

Actually, scratch that 😞 . They do have a strapping and wiring difference which makes them incompatible. May need completely new libpico rebuild just for it. For example, ADC will be wired up to upper GPIOs which will differ -A vs -B

[PontusO]

Maybe move the low level analog stuff into the bsp instead. Seems over the top to have a new lib simply for the adc.

[earlephilhower]

TBH I'm not sure that this is the only code/user-visible difference, it's just the first one I noticed. The SDK sometimes pulls in lots of dependencies so it may not be possible, but it's worth a look.

In the worst case end users won't know about it, there will just be one more pass in tools/libpico and things will "just work."

[elral]

I set up a new board this morning to test it out. After copying the generic one I just extended the pins and defined the new analog pins.
For what I can see everything was working, but for analog pins just tested A0/1 on Pin 40/41 and not all 8 analog pins.