nrf52 watchdog (attempt #2)#8670
Conversation
* nrf52: add watchdog Main thread only for now. * bump framework-arduinoadafruitnrf52 to pick up new wdt support * clang-format the new parts of main-nrf52.cpp --------- Co-authored-by: Ben Meadors <benmmeadors@gmail.com> (cherry picked from commit 8395429)
The 90 seconds wdt timeout matches the esp32 wdt timeout.
|
@markusgritsch can you see if this resolves #8468 for you? |
|
awesome, i'll set a couple to test out |
|
@thebentern Let's merge this so we can test it a bit more; I do like the rework here, great job @SebKuzminsky |
|
I've been running it with no issues at all |
|
Following the merge I've tested so far my Wio Tracker L1 Pro and Heltec T114; was able to change settings successfully. |
|
Yay, thanks folks! :) |
|
On the NRF52840 Seeed XIAO, the system is functioning as expected. I am wondering if it is possible to log the reboot reason during boot, specifically if there is a mechanism in the hardware that preserves the trigger for the reboot. |
|
I'm getting T114 nodes that are off(deep sleep) and they will randomly reboot and power on by themselves. It's very random and sometimes it can take a full day or two to do this. |
I have two T114 running 2.7.16 powered by a battery, which never sporadically reboot. They also never turn on themselves from deep sleep. Do you have anything connected to the USB connector? My devices turn themselves on from deep sleep if I connect them to power on the USB port. |
|
@SebKuzminsky between @HarukiToreda's comment and now #8959, do we need to investigate? |
|
I think it's unlikely that @HarukiToreda's comment indicates a problem with the new watchdog, given the large and variable time. But it's not impossible. I like @FFAMax's suggestion to somehow log/record the reboot reason, but i don't know how to do that. #8959 is reported against 2.7.15, which does not include the nrf watchdog (but does include some other nrf deep sleep changes), so i think this also does not suggest a wdt problem. It's possible the reported version number is wrong and it's really somewhere between 2.7.15 and 2.7.16, which could potentially include the nrf watchdog, i'll ask for more info in that issue. |
…ake-2 nrf52 watchdog (attempt exploiteers#2)
This PR differs from the disastrous #8485 in two ways:
I would appreciate a sanity test from @Xaositek and @HarukiToreda