Closed oxbown closed 7 months ago
Thanks for building and testing this. Anyway for the support of new hardware I would like to avoid the definitions of new environments for each combination but instead going with the approach of defining building blocks and adding them to the build config (build.yaml) so that you can select the necessary pieces in the build UI. So the support for the port ABC would need to become a "KIT" definition - like https://github.com/wellenvogel/esp32-nmea2000/blob/e1044821a7b61de3e1c393986262f44952bffddd/lib/hardware/GwHardware.h#L159. And we would need some config options for the build UI. Basically I would assume to have serial/can/i2c as potential grove users. Most probably it should be iic2 and serial2.
After some long time I now integrated the portABC into the master and into the build service. It would be really helpful if you could test the build service at https://www.wellenvogel.net/software/esp32/cibuild.html?tag=testci20240321 As I currently do not have a port ABC I cannot really test it (just was able to check the logs...).
Sure, I´ll test it! Feedback will come.
@wellenvogel Testing the actual revision i found the same ussie as here #61 Stack canary triggered. The setup is: AtomS3+CAN+GPS+RS485+ENV3
17:36:41.699` -> E (2370) esp_core_dump_flash: Core dump flash config is corrupted! CRC=0x7bd5c66f instead of 0x0 17:36:41.732 -> Rebooting... 17:36:41.732 -> ESP-ROM:esp32s3-20210327 17:36:41.732 -> Build:Mar 27 2021 17:36:41.732 -> rst:0xc (RTC_SW_CPU_RST),boot:0x2b (SPI_FAST_FLASH_BOOT) 17:36:41.732 -> Saved PC:0x40377870 17:36:41.732 -> SPIWP:0xee 17:36:41.732 -> mode:DIO, clock div:1 17:36:41.732 -> load:0x3fce3808,len:0x44c 17:36:41.732 -> load:0x403c9700,len:0xbe4 17:36:41.732 -> load:0x403cc700,len:0x2a38 17:36:41.732 -> entry 0x403c98d4 17:36:41.937 -> E (231) esp_core_dump_flash: No core dump partition found! 17:36:41.937 -> E (231) esp_core_dump_flash: No core dump partition found! 17:36:41.970 -> [ 246][E][Preferences.cpp:50] begin(): nvs_open failed: NOT_FOUND 17:36:42.220 -> LOG:400:GWWIFI: AP auto shutdown disabled (0s) 17:36:42.220 -> GWSERIAL:505:initialized IIC 1 at sda=5,scl=6 17:36:42.220 -> GWSERIAL:506:wifi AP pass: esp32nmea2k 17:36:42.220 -> GWSERIAL:506:admin pass: esp32admin 17:36:44.000 -> Guru Meditation Error: Core 0 panic'ed (Unhandled debug exception). 17:36:44.000 -> Debug exception reason: Stack canary watchpoint triggered (iicTask) 17:36:44.000 -> Core 0 register dump: 17:36:44.000 -> PC : 0x40380ce7 PS : 0x00060b36 A0 : 0x80383944 A1 : 0x3fcc4fe0
17:36:44.000 -> A2 : 0x3fce9f14 A3 : 0xb33fffff A4 : 0x0000abab A5 : 0x00060b23
17:36:44.000 -> A6 : 0x00060b20 A7 : 0x0000cdcd A8 : 0xb33fffff A9 : 0xffffffff
17:36:44.035 -> A10 : 0x3fcc53e0 A11 : 0x00000000 A12 : 0x7ae147ae A13 : 0x404cee14
17:36:44.035 -> A14 : 0x02ce9f14 A15 : 0x00ffffff SAR : 0x0000000a EXCCAUSE: 0x00000001
17:36:44.035 -> EXCVADDR: 0x00000000 LBEG : 0x400556d5 LEND : 0x400556e5 LCOUNT : 0xffffffff
17:36:44.035 -> 17:36:44.035 -> 17:36:44.035 -> Backtrace: 0x40380ce4:0x3fcc4fe0 0x40383941:0x3fcc5020 0x40383bad:0x3fcc5040 0x403780a9:0x3fcc5060 0x403780bd:0x3fcc5090 0x403780ea:0x3fcc50b0 0x40383f91:0x3fcc50d0 0x420bb2f6:0x3fcc50f0 0x420b53f2:0x3fcc5190 0x420b2ad5:0x3fcc54a0 0x42032eee:0x3fcc5560 0x4202b22f:0x3fcc55d0 0x4202874e:0x3fcc5660 0x42006bc6:0x3fcc56b0 0x42006db6:0x3fcc5710 0x4203037e:0x3fcc5730 0x420aa4db:0x3fcc5760 0x420aa73d:0x3fcc58c0 0x420c7765:0x3fcc5910 0x420a8f65:0x3fcc5930 0x420c568d:0x3fcc59d0
This is the log of the online build:
GIT_SHA="bfc6af1f89bf7a0a931e9b82734fe2f201d50a76" PLATFORMIO_BUILD_FLAGS=" -DM5_PORTABC -DSERIAL_GROOVE_485_A -DM5_GPS_UNIT_B -DM5_ENV3_C -DM5_CANUNIT -DGWRELEASEVERSION=testci20240321" pio run -e m5stack-atoms3-generic
Really interesting. Was running nearly the same set up on an atom lite without issues. So I have to increase the stack size...
Maybe you could give it a second try with testci20240322 - iic task stack size now at 4000.
Maybe you could give it a second try with testci20240322 - iic task stack size now at 4000.
Actual testci20240322 tag not compile in circle ci, maybe the last commit break something.
ahhh, should have been testing also with S3...
ok new tag: testci20240323
Build OK! Seems to work well on the esp32s3
Many thanks for your awesome work!
Two new build versions for Atom and AtomS3 with port ABC Commented #58 Also encreased i2c tasksize #61