[Solved] Frequent core dump after 2.17

General
#1

Hi,

My esp8266 (https://github.com/michaelfung/sonoff-basic-openhab) starts to core dump frequently after built with Mongoose 2.17. It normally takes less than 24 hours after power up to occur.

init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":59102.161727}
init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":59402.161770}
init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":59702.161451}
init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":60002.161069}
init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":60302.160946}
init.js:215 Published:OK topic:sonoff_basic/a7f63d/state msg:{“sch_enable”:“ON”,“skip_once”:“OFF”,“oncount”:0,“relay_state”:“OFF”,“memory”:29668,"
uptime":60602.162017}
mac 684

HW WDT @ 0x40102880
A0: 0x40102880 A1: 0x3ffffbb0 A2: 0x00000008 A3: 0x00000000
A4: 0x00000000 A5: 0x60000000 A6: 0x00000003 A7: 0x3ffffae3
A8: 0x00000000 A9: 0x00000000 A10: 0x00000000 A11: 0x00000000
A12: 0x00000040 A13: 0x7fffffff A14: 0x00000002 A15: 0x00000140

(exc SP: 0x3ffff9f0)

— BEGIN CORE DUMP —
{“app”: “openhab_sonoff_basic”, “arch”: “esp8266”, “version”: “1.1_20200224r2”, “build_id”: “20200224-054313/p1-3-g06574e3-master-dirty”, “build_image”: “docker.io/mgos/esp8266-build:2.2.1-1.5.0-r5”,
“REGS”: {“addr”: 1073740320, “data”: "
gCgQQLD7/z8IAAAAAAAAAAAAAAAAAABgAwAAAOP6/z8AAAAAAAAAAAAAAAAAAAAAQAAAAP///38CAAAAQAEAAIAoEEAEAAAAAAAAAAAAAAAAAAAAMAAAAA==", “crc32”: 3906734821},
“DRAM”: {“addr”: 1073643520, “data”: "
voEhQAAAAAAQgP4/AAAAAAIAAAD8gv4/ZIP+P8yD/j8AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACK2iBAAQAAADDNIUBkEP8/AAAAAJQc/z+8D/8/AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

I wonder if anyone having the same issue after Mongoose updated to 2.17?

#2

AFAIK, HW WDT means a hardware watchdog timeout.

If you use local build and have mos console running, mos will analyze the core dump and provide you information where the exception occured.

#3

The log I posted is captured by a Pi, so I cannot use mos… Anyway, I guess I found the cause of the problem. There is a flaw in the Sonoff power supply circuit and this flaw is allergic to 2.17. This issue does not happen when 3.3v is supplied from another source. Thanks!