fix: device sleep (part 1)

[mverch67]

This PR fixes some issues introduced in #3251 (e.g. router/repeater/powersave wakeup issues) and does a few debug cosmetics:

The function enableLoraInterrupt() assumes that all LORA/DIO1/CS/RESET GPIOs are RTC GPIOs which is not true (e.g. only RTC GPIOs can be used for ext0).
As a result there are error messages in the log and all devices where the DIO1 pin is not an RTC GPIO will fail to wake up from sleep when a LoRa packet is received (one packet is received though, but processed on the next scheduled wakeup).

E (221030) RTCIO: rtc_gpio_pulldown_en(129): RTCIO number error
E (72312) RTCIO: rtc_gpio_pullup_en(109): RTCIO number error
E (72313) RTCIO: rtc_gpio_pullup_en(109): RTCIO number error

This PR replaces the RTC GPIO functions by generic functions which check for RTC GPIO internally and does a check for the validity of RTC GPIOs for DIO1.

[mverch67]

Tested with tbeam, t3s3, heltec v3 and heltec wireless paper. Now all wake up from router mode. In case you still experience issues in sleep mode (e.g unresponsive, noise) you'll have to do a factory reset, there may be some wrong setting.

[mverch67]

As the PR re-introduces GPIO wakeup source for all hardware devices that don't have RTC GPIO DIO1 some code changes from #3521 had to be reverted in the same fashion as in #2412 (to avoid screen turn on with every single packet).

[thebentern]

Good call. Technically we have targets like the T-Beam that have RF95 macros for the SX1276 and also SX126X ones, so runtime discrimination probably heads off some potential issues there hopefully.

[mverch67]

yes, I've tested that already. Working as expected.