Suggestions for improving initial draft for LP core interrupts on the esp32c6

[aarch64angel]

Hello!
I've put together a small proof of concept implementation of LP core interrupt handling on the esp32c6.
fork
It's not very sophisticated or well tested yet, I'm mainly just looking for architectural suggestions and or issues. It currently relies on the unstable 'riscv-interrupt-m' feature, that will probably have to be replaced by an intermediate naked asm function.
I tried to include references for every assumption that I made.
Again, any suggestion is welcome.

[MabezDev]

Thanks for looking into this! Very cool that you have this working already!

As you suggested, I would ditch the unstable riscv-interrupt-n feature, in favour of some asm and or naked functions (they're stable now, which is nice!). I would also tear our your manual interrupt handler storage code, and try to leveral the PACs, checkout the lp pac here: https://github.com/esp-rs/esp-pacs/blob/d1640337cd0df450f7d1302651fcf3792f5df166/esp32c6-lp/src/lib.rs#L34. I think it should be possible to create interrupt handling code in a very similar way to the HP cores, so I'd use the esp-hal interrupt module as some guidelines.

[aarch64angel]

Thanks for the feedback!
Unfortunately, I don't believe it is possible to implement it the same way as it is in the HP core. The c6's LP core only has a single interrupt vector, the ones you linked to in the PAC are still for the HP core, even if they are from LP peripherals. It's sad, because 16K of memory is already very limiting, but I'm not aware of a more efficient approach (I'm sure the actual code can still be made more efficient of course), at least not if we want separate handlers for different interrupts.
Technical Reference
ESP-IDF implementation
The alternative would be to let the user implement their own general handler and just provide tools to make that straight-forward. (or some middle ground with an included centralized handler that drivers can hook into, but the user could replace if they wanted to)
With regards to the unstable feature, I'll replace it with asm.