Better understanding of fatal exception at epc1=0x4000228b #914
Comments
|
Is ESP power enough? It can be reason for strange errors with reading flash memory. |
|
@anakod Do you have an idea where this error is coming from? What is the reason for seeing such an error? |
|
I'm not sure, just have error like that once, but I didn't remember exact epc1 code so it can be different error also. But I think first what users should to do in cases like that - check power source :) If voltage drops on some operations - it's probably reason. |
|
I have similar problem too. This fatal exception appears with sample projects (e.g. MeteoControl, Basic_WebSkeletonApp) and my own projects, that compiled with Sming v3.0.1 (esp-open-sdk-1.5.4). My test hardware is NodeMCU DEV board. I'm not sure that the problem is the lack of power. I will continue testing. |
|
If error depends on Sming version then it shouldn't be related to hardware\power source. |
|
+can we get additional info from crash handler / stack? |
|
@ADiea my guess is that the issue is happening even before our application is loaded/started. I have seen this when the first level bootloader is not able to check/load the ROM file. An error message like this is shown: |
|
could it be caused by faulty flash chip (reached end of life) ? |
|
Hi, I'm having the same issue with my code, based on MQTT and HttpServer_ConfigNetwork: The problem, I think, is related with the size of the first bin file: 0x0000.bin, that is bigger than the expected by the configuration in the make file: Makefile-project.mk It spect to be less than 36864 bytes (0x9000 ), but with the new Sming version is 40960 bytes (0xA000). Sming 3.0 (esp-open-sdk 1.5.4 or 2.0.0) -- NOT WORKING Sming 2.0 --> WORKING Any idea? solution? |
|
BTW, can someone tell where these magical numbers come from? Thanks! |
|
@mtzfactory Take a look at the commits in this PR. Meanwhile I suggested a PR for esptool.py that gives warning or stops with an error if there is an overlap detected. |
|
Hi @slaff, I just follow your commits in the PR and didn't worked, I tried with different examples and even with the Basic_Blink example didn't succeed, when I connect to the serial port it prints some kind of random values, I think is reseting itself after a while... PD: To be sure I did it right, I cloned again the Sming repo and applied your instructions there... but same result... |
|
@anakod Thanks for noticing this. I guess I made a mistake somewhere... Have to check... |
@mtzfactory That is exactly what it does. It does not print any additional information on the serial and the only thing that should happen is your LED blinking. If you want to test an app that prints something in the console use our Basic_Serial example. |
|
Thanks! but I already knew that @slaff... I just modified it with a Serial.println every time blink function was called, just to test everything was correct, but it wasn't, even the led didn't blink... As the other examples, it prints some kind of random values, I think is reseting itself after a while... |
|
Hi, I just re-downloaded and compiled everything... and still get the same result... PD: I'm using the esptool.py with the PR mentioned above... And when I connect to the serial port I get this... constantly sending random data and reseting... |
You have only one board? Will be very good to test on different hardware also to be sure. |
|
Hi @anakod , Yes I have: an esp-01 (1Mbyte flash)... and I tried with the new sming version: 3.1.0. Sming 2.1.0, esp-01 (1Mbyte flash), Basic_Blink works perfectly. Sming 3.1.0, esp-01 (1Mbyte flash), Basic_Blink works perfectly. Sming 2.1.0, esp-01 (1Mbyte flash), HttpServer_ConfigNetwork works perfectly. Sming 3.1.0, esp-01 (1Mbyte flash), HttpServer_ConfigNetwork doesn't work. In both cases I get the following message: |
|
Please, paste the decoded stack trace here. |
|
Hi @slaff , Here you have: |
|
@mtzfactory Thanks. First make sure to have SPI_SIZE set to the correct size of the flashmem (what is your ESP device flashmen size by the way?). For example if you have 4 megabytes then set it to Test and write back with the results that you get. |
|
looks like it crashes when trying to mount spiff |
|
Hi @slaff, With the develop version it works perfectly... Yeey!!! Thanks a lot! |
|
The main cause for the fatal exception epc1=0x4..228b was overlapping ROMs. We proposed fix to estool.py team and it is now merged in the upstream version. Thanks to all who helped us track the problem. |
This issue is created to help us collect and understand the reason(s) for fatal exception at epc1=0x4000228b.
Reported problems:
See #857, #864.
Causes:
There is a PR for esptool.py that will detect overlapping roms and exit with an error.
If you have an idea what can cause this issue, please comment below. Hints for fixes, workarounds, etc are welcomed.
The text was updated successfully, but these errors were encountered: