Likeliness of 368b7eb also impacting ESP32-S3? (IDFGH-13282)

espressif / esp-idf

Espressif IoT Development Framework. Official development framework for Espressif SoCs.

Apache License 2.0

13.7k stars 7.3k forks source link

Likeliness of 368b7eb also impacting ESP32-S3? (IDFGH-13282) #14212

Open JVKran opened 3 months ago

JVKran commented 3 months ago

Answers checklist.

[x] I have read the documentation ESP-IDF Programming Guide and the issue is not addressed there.
[X] I have updated my IDF branch (master or release) to the latest version and checked that the issue is present there.
[X] I have searched the issue tracker for a similar issue and not found a similar issue.

General issue report

How likely would it be that the fixes listed over here also apply to the ESP32-S3? We're currently having issues where we think the Wi-Fi stack fails to deinitialize (due to some reason) which is blocking the ESP32-S3 from entering deep-sleep. We've now updated to V5.1.4 (i.c.w. ESP-ADF V2.6) and are going to test if this doesn't happen anymore.

esp-wzh commented 3 months ago

This fix is not for esp32s3, could you describe the detailed phenomenon when the error occurs?

JVKran commented 3 months ago

We had an issue with application logic which didn't manage the case when there wasn't a wifi connection, but we also hadn't fired up the AP yet. This resulted in the Wi-Fi stack not being deinitialized when entering deep-sleep. Sometimes, deep-sleep wasn't entered but the main-loop appeared like it was no longer running. Some other tasks in the background were still active. The Wi-Fi connection sometimes was also still active.

esp-wzh commented 3 months ago

Some other tasks in the background were still active. The Wi-Fi connection sometimes was also still active.

This situation should not happen. The software will ensure that the process of entering sleep is single-threaded. The core requesting sleep will enter the critical section and stop other cores, and then prepare for sleep. Even if entering sleep fails, esp_deep_sleep_start function will not return but trigger the watchdog.