Use critical_section for Peripherals::take.

[reitermarkus]

• cortex_m Add implementation for critical-section 1.0 cortex-m#447
• msp430 Provide critical_section implementation. msp430#13
• riscv Add critical-section 1.0 implementation, fix multicore unsoundness. riscv#110
• xtensa_lx Correct Xtensa critical sections esp-rs/xtensa-lx#20, critical_section implementations & esp_backtrace esp-rs/esp-hal#151
• mips_mcu Provide critical_section implementation. kiffie/pic32-rs#5

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @therealprof (or someone else) soon.

Please see the contribution instructions for more information.

[Emilgardis]

lgtm, this is great :D

do we need another reviewer maybe?

[adamgreig]

I'm a bit concerned about what this means for platforms that don't have critical-section support yet. I know msp430 is working on it but having some codegen size issues, and I don't know anything about mips_mcu. It might be worth waiting a little longer for those architectures to get support before releasing svd2rust that uses it?

Having it as an optional dependency/feature seems useful though and does mean those platforms don't need immediate critical-section support, even end-users could add simple interrupt-based implementations to use take() if needed.

[reitermarkus]

It might be worth waiting a little longer

I don't think it's worth waiting. Other architectures will have to implement these only once they upgrade svd2rust anyways. I have opened issues for the affected architectures to give them a heads up though.

[adamgreig]

Oh, thanks, I didn't see the edit to the first comment. SGTM to me then.

[Emilgardis]

A mention to take not being provided if critical-section is not enabled could probably be added here

[therealprof]

LGTM,too. Thanks a lot @reitermarkus for taking this on.

bors r+

[bors]

Build succeeded:

• CI
• Rustfmt