ESP SoC Reset sequences

[MabezDev]

Builds on top of #2151
Supercededes #1965
Closes #1963.

I still notice some oddities, where the probe/dtm becomes unresponsive, particularly after flashing once. I usually then need to physically unplug the board and replug for flashing to work again. Not sure what's causing that atm, but this PR should still bring us a step closer.

Draft until all chips are supported.

[bugadani]

Better, but still not quite right: this PR only works once per power cycle, just as OP said. However, for me pressing Rst works, too, no need to unplug.

❯ probe-rs run --chip esp32c3 target\riscv32imc-esp-espidf\debug\c3-idf-demo
      Erasing ✔ [00:00:02] [############################] 576.00 KiB/576.00 KiB @ 261.30 KiB/s (eta 0s )
  Programming ✔ [00:00:04] [#############################] 249.81 KiB/249.81 KiB @ 58.07 KiB/s (eta 0s )
    Finished in 6.735s
 WARN probe_rs::util::rtt: No RTT header info was present in the ELF file. Does your firmware run RTT?


❯ probe-rs run --chip esp32c3 target\riscv32imc-esp-espidf\debug\c3-idf-demo
 WARN probe_rs::session: Could not clear all hardware breakpoints: A RISC-V specific error occurred.

Caused by:
    Timeout during DMI access.
 WARN probe_rs::session: Failed to deconfigure device during shutdown: A RISC-V specific error occurred.

Caused by:
    Timeout during DMI access.
Error: Connecting to the chip was unsuccessful.

Caused by:
    0: A RISC-V specific error occurred.
    1: Timeout during DMI access.

[bugadani]

Please keep bugging me to work on this, these need to be custom flash algo functions and not sequences.

Also I don't think we need the clone here any more, but I'm not 100% sure.

[MabezDev]

So I have found the issue, and a work around, but it's not clear to me if we can do better.

The IDF application re-enables watchdogs during its application, when you Ctl+C, the application halts (stop petting the wdg) and the board resets after some period. It's possible to begin reflashing before the wdg reset if you're fast enough and reflashing works perfectly.

The workaround is to pass --catch-reset which will leave the core in a halted state on wdg reset, but this is just a workaround. Probe-rs doesn't seem to be able to coax the DTM out of whatever state it's in when the wdg resets the chip and the DTM thinks the core should be halted, its like the chip and DTM get out of sync. Not sure what I can try here to improve that, any ideas?