DaveBerkeley

It is a bit of a hack. I added a --dual-bank switch on the command line (but not to --help) and use a table to look up the sectors to...

I've removed the --dual-bank option and added code to read the optcr reg for stm32f76x devices. Now in https://github.com/DaveBerkeley/stlink/tree/f7_integrate branch. I have tested it on a Nucleo F767 board. Will...

I don't know how to run the tests that it is failing. If I can find out how to do that, then I can have a go at making sure...

I have a colorlight_i9 platform file. My PMOD connectors are wired with the rows reversed to this one. Have I got them the wrong way round? Should I change my...

> There are other expansion boards waiting for a sane design pattern before the maintainer is willing to merge them I'd love to see this expansion board use inheritance because...

That looks like a good approach. It reflects the physical wiring. For litex-boards I added a board="i9" parameter to the platform and target files. Then took a deep copy of...

Zephyr seems to use a similar method of adding initialisation functions. https://github.com/zephyrproject-rtos/zephyr/blob/main/include/zephyr/init.h#L85-L90

The priority (order) of boot methods could be passed as a define. This also allows a decoupling of having a device and booting from it. Currently boot_sequence() is hard coded...

Thanks, I noticed that in the linker.ld script. It serves a slightly different purpose to what I'm suggesting, but uses a very similar technique. The __attribute__((constructor, priority)) technique allows the...

It would not add any dependency on C++. The .init_array would be empty if not used. The only overhead would be the call to __libc_init_array() which would , in that...