esp32c3 interrupt implementation

[zdima]

@aykevl this is only interrupt implementation for esp32c3.
No exported APIs except enum values of HandlerID type to initialize chip's modules.
Tested with pin src/examples/pininterrupt/pininterrupt.go (not checked in)

[aykevl]

I've been thinking about this PR and how to implement interrupts correctly on the ESP32-C3.
I think it would actually make more sense to make the Interrupt type be the 31 interrupt vectors. It would then work as follows:

• Code that wants to use an interrupt can do so using interrupt.New, for example interrupt.New(5, func(interrupt.Interrupt) { ... }).
• It will have to manually map the interrupt source to the interrupt vector, for example esp.INTERRUPT_CORE0.GPIO_INTERRUPT_PRO_MAP.Set(5).
• The handleInterrupt logic now becomes a lot simpler. It can now directly call callInterruptHandler with the interrupt number from mcause.

It is still possible to map multiple interrupt sources to a single interrupt vector: interrupt.New can be used for multiple handlers on a single interrupt number. For example:

// configuring GPIO:
intr := interrupt.New(5, func(interrupt.Interrupt) {
    status := esp.GPIO.STATUS.Get()
    if status != 0 {
        // handle GPIO interrupt
    }
})
intr.Enable()
esp.INTERRUPT_CORE0.GPIO_INTERRUPT_PRO_MAP.Set(5)

// configuring UART:
intr := interrupt.New(5, func(i interrupt.Interrupt) {
    flag := esp.UART0.INT_ST.Get()
    if flag != 0 {
        // handle UART interrupt
    }
})
intr.Enable()
esp.INTERRUPT_CORE0.UART0_INTERRUPT_PRO_MAP.Set(5)

(These two don't need to be on the same interrupt vector, I'm only showing that it is possible).

What do you think?

[zdima]

I didn't know you can call interrupt.New with the same id. I never get response from you on how does it work in the slack PM :(. If I knew, perhaps the implementation will be much simpler.
So how does it work if make a call interrupt.New(5, func(interrupt.Interrupt) {}) twice?

[zdima]

@aykevl not sure if you get notification on my comments above

[aykevl]

Looks good to me, thank you for all the work!

[aykevl]

CI is failing, but I think that's an unrelated issue.

[aykevl]

Oh now I see. You have made the PR against the espnet branch. Can you rebase it against the dev branch?

[zdima]

I could, but don’t we need other changes you made like stdio, pico lib and FreeRTOS?

…

On Oct 22, 2021, at 6:59 PM, Ayke ***@***.***> wrote: Oh now I see. You have made the PR against the espnet branch. Can you rebase it against the dev branch? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2170 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACI2FGP4NA36ES4CPIN3L3UIHUDVANCNFSM5F2XORKA>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[aykevl]

I could, but don’t we need other changes you made like stdio, pico lib and FreeRTOS?

Yes, but by merging this in the dev branch it will become available to everybody, not just those working from the espnet branch. That's why I also made a few other PRs like #2135, #2136, and #2143. In general I prefer working as much as possible directly from the dev branch for independent features.
After the PR is merged, it can easily be added to the espnet branch too.

[zdima]

this is merged with PR #2200