Lopy4 Guru meditation error

robert-hh

@bjsi9 Does that also happen with the safe boot (P12 at 3.3v during reset)?

bjsi9

@robert-hh Hi, no error if i safe boot!
But when i remove the jumper, the errors comes back.
Also after reset.

robert-hh

@bjsi9 That means, like you suspekt, the crash is caused by something in main.py, or by the attempt of accessing main.py. Safe boot does not run boot.py or main.py.

bjsi9

@robert-hh Thanks, good point.
Here is fresh pictures of what is in the boot and main, that still gives me error:

robert-hh

@bjsi9 main.py looks definitely uncritical. I'm a little bit thinking about boot.py, because setting REPL to UART0 is the default. I have to check this evening back home what I have on my devices. But you could try to comment that out. During safe boot you will still have access to the file system via ftp. I would not use Pymakr for that basic debug stuff. It has its own complications.

bjsi9

@robert-hh Hi, the boot.py is the same code as on my SiPy. And that has no error.

bjsi9

I just updated my Sipy (which have been working for half year). And now i have the exsact same problem with Guru error.
(i dont know which firmware was on before the upgrade, i will try to find out)
Update: its 1.15.0.b1

robert-hh

@bjsi9 Looking into the sources, this is the code in the embedded _boot.py, which is executed on any boot, even the safe boot:

# _boot.py -- always run on boot-up, even during safe boot
import os
from machine import UART
os.dupterm(UART(0, 115200))

So this code is not required in boot.py, and you can drop it. Do the file copying with ftp. The youre sure about what you get.

robert-hh

@bjsi9 Duplicating the dupterm code in boot.py does not hurt, at least in my test with a FiPy and a WiPy.

bjsi9

Hi @robert-hh , still no luck.

The LoPy4 boots, and are ready for commands. But after some seconds it chrashes with the Guru error. Then it boots again, and chrashes again.. and so on.
Very sad story for my 2 Sipy's and 2 LoPy4's. But i'm sure we will find a solution at some point. Any suggestions are welcome :)

bjsi9

New update:
I dont care about my code anymore, i just want the LoPy4 to work as from factory.
I have tried with a new expansionboard, LoPy4 with erased filesystem and downgrading to 1.15.0.b1.
Still chrash chrash chrash every 2 seconds.
No luck after manually flash-reset, or FTP upload standard boot and main.

Safe boot dosent chrash.

robert-hh

@bjsi9 That's strange. I have a Lopy here running since 5 weeks in a row as Lora nanogateway, and some Fipy's for testing, but they never crashed.

When crashing, do your devices run code or are they just waiting in the REPL prompt?
Do you always connect the devices in the same way - same expansion board, same cable, same USB port?

bjsi9

@robert-hh Hi, yes its very strange. I have also some Sipy's for testing, never had any problems. But after upgrade they chrashes all the time, even when downgrading and flash-reset.

The device starts up, and waits in the prompt with: >>>>
After som seconds, it chrashes.
I have two expansionsboard, and tested with different USB ports also. Its the same cable i've used since i started working with pycom device the past year.

Remember: the safe boot works

robert-hh

@bjsi9 Looking at the code, this is hard to understand:
With safeboot, only _boot.py is executed.
with normal boot, _boot.py, boot.py, _main.py and main.py are run (in that order). That is mostly th whole difference. This is the section in the file mptask.c, which I'm talking about (line 282 ff):

    if (!safeboot) {
        // run boot.py
        int ret = pyexec_file("boot.py");
        if (ret & PYEXEC_FORCED_EXIT) {
            goto soft_reset_exit;
        }
        if (!ret) {
            // flash the system led
            mperror_signal_error();
        }
    }

    if (!safeboot) {
        // execute the frozen main first
        pyexec_frozen_module("_main.py");

        // run the main script from the current directory.
        if (pyexec_mode_kind == PYEXEC_MODE_FRIENDLY_REPL) {
            const char *main_py;
            if (MP_STATE_PORT(machine_config_main) == MP_OBJ_NULL) {
                main_py = "main.py";
            } else {
                main_py = mp_obj_str_get_str(MP_STATE_PORT(machine_config_main));
            }
            int ret = pyexec_file(main_py);
            if (ret & PYEXEC_FORCED_EXIT) {
                goto soft_reset_exit;
            }
            if (!ret) {
                // flash the system led
                mperror_signal_error();
            }
        }
    }

bjsi9

@robert-hh Okay, maybe _boot.py or _main.py is corrupted? What could i do about that, any suggestions? Thanks

robert-hh

@bjsi9 _boot.py should be OK, because that is executed always. _main.py is inside the frozen bytecode and comes as part of the firmware image. I have my doubts whether that is wrong.
Did you meanwhile try another (short) cable? Cables and especially connectors can wear.

bjsi9

@robert-hh Just tried another cable, no luck. Same chrashes, after erase file system and firmware update (done with the updater tool)
I know its a long shot, but i'm gonna try older version of the updater tool now.

bjsi9

No luck using updater tool 1.13.1.b5

bjsi9

@robert-hh Hi, quick question. sys, lib and cert folders are all empty when i view them in filezilla FTP (Show hidden files are on). Should these folders be empty?

robert-hh

@bjsi9 Yes. They can be empty and are empty on fresh installe devices. lib is used for storing python code of library style. It is also named in uos.path. But that is convention.

Explore Pybytes | Official Documentation | Report a Firmware Bug/Issue | GitHub

Lopy4 Guru meditation error

Pycom on Twitter