Add POWMAN_PASSWORD_BITS to pico_bootsel_via_double_reset.

[Gadgetoid]

Fix pico_bootsel_via_double_reset for RP2350.

Was borrowing this code for a USB MSC user boot toggle. Found it didn't work, delved into the bootloader and found:

https://github.com/raspberrypi/pico-bootrom-rp2350/blob/fd6104450fa8f55c11c0c9b54dbc69a27537130f/src/main/arm/varm_boot_path.c#L774

[kilograham]

I'm assuming this is correct (can't see any docs which say which regs need the password ;-) )

[Gadgetoid]

Section 6.4 from the good book for posterity:

All registers with address offsets up to and including 0x000000ac are password protected. Therefore, the following writeable registers are unprotected and have 32-bit write access:

• POWMAN_SCRATCH0 → POWMAN_SCRATCH7
• POWMAN_SCRATCH0 → POWMAN_SCRATCH7
• POWMAN_BOOT0 → POWMAN_BOOT3
• POWMAN_BOOT0 → POWMAN_BOOT3
• POWMAN_INTR
• POWMAN_INTE
• POWMAN_INTF

Doesn’t help that this section lands on a page boundary and it looks like a formatting error leading into another list of registers right below...

edit: of course I can’t judge ‘cos I looked at the bootrom code and only skimmed the docs until just now 🤣

[lurch]

Maybe it'd be better to use powman_set_bits ? 🤔

[Gadgetoid]

Maybe it'd be better to use powman_set_bits ? 🤔

Agreed! And updated.

Edit: I flew too close to the sun 😭

[lurch]

CI is failing with e.g.

/usr/lib/gcc/arm-none-eabi/12.2.1/../../../arm-none-eabi/bin/ld: cannot find -lhardware_powman_headers: No such file or directory
/usr/lib/gcc/arm-none-eabi/12.2.1/../../../arm-none-eabi/bin/ld: cannot find -lhardware_powman: No such file or directory
collect2: error: ld returned 1 exit status
gmake[2]: *** [test/kitchen_sink/CMakeFiles/kitchen_sink_printf_compiler_no_shortcircuit.dir/build.make:1256: test/kitchen_sink/kitchen_sink_printf_compiler_no_shortcircuit.elf] Error 1
gmake[1]: *** [CMakeFiles/Makefile2:2710: test/kitchen_sink/CMakeFiles/kitchen_sink_printf_compiler_no_shortcircuit.dir/all] Error 2
gmake[1]: *** Waiting for unfinished jobs....

[kilograham]

yeah lets revert that last change, as it obviously breaks RP2040, and i don't know if it is worth having to do a platform test in the CMake and bazel

[lurch]

yeah lets revert that last change, as it obviously breaks RP2040, and i don't know if it is worth having to do a platform test in the CMake and bazel

Sorry @Gadgetoid for my duff suggestion 🤦‍♂️

[Gadgetoid]

Sorry @Gadgetoid for my duff suggestion 🤦‍♂️

It was a good and reasonable suggestions, but SDK says no 😆

Reverted and push'd!

[kilograham]

yeah it was a good suggestion, and probably hit back in the day too!