-
Notifications
You must be signed in to change notification settings - Fork 9
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
v2.0.6 Issues #40
Comments
Hi Matthias, could you record the log? Would be really helpful. You can connect an USB-UART adapter to the "debug" connector on the back of device. Home Buttons TX -> UART adapter RX, GND -> GND. The baud is 115200. Please also try the v2.0.5. I pulled it due to a bootloader compatibility issue, but your devices have the correct bootloader, so it should work. |
How to connect the debug pins? Soldering seems to be hard since there is the display glued to the other side. Looking at the holes supposed for the debug connectors, it looks like there is some metal at the back (backside of the display?) so just plugging some cables loosely will result in a short circuit I guess. Any suggestions? Could also not find anything in the docs. |
It's the back side of display, but it's not metal. Just plug some cables in loosely. |
Will try!
Flashed back using OTA to 2.0.5, same restart problem. Flashed back using OTA to 2.0.4, which was previously working without the restart problem, now the restart problem is there, too. Then I used esptool to flash 2.0.6 again, hoping it makes a difference... so, it makes, but it doesn't seem to but up anymore :( Still can get into the flash mode, so if you have an idea on how to get it back working... |
Some additional info on that: Seems like my mistake was to do erase_flash before writing the new firmware. I guess I lost whatever was flashed to |
Please try to get log on another device, I'm really interested in what could be wrong if even reverting to v2.0.4 didn't fix the issue. |
Serial (I guess) is 2301-112. Thanks for your help! Debug log: Will try later this day. |
Here are the full flash images for both v2.0.4 and v2.0.6. Yo can flash them with: |
Thanks!!! Device is alive again, but still has the restart issue. Below some debug output:
I'm a bit confused about the |
Thanks for the log! Have you tried the v2.0.4 again? Is it working? You've said previously:
So you can get to web setup and update page without crashing? It's not obvious to me what's the problem. The backtrace decoder is giving me this:
Can you please record another log with v2.0.6_debug build? It produces more detailed output. It's on the release page, you can use OTA to install it. |
Thanks for investigating and helping me!
Working in the sense of: I flashed it successfully and the device starts up: yes. Without this restart issue? no.
Yep.
Quite interesting. The restart problem does not occur with the debug firmware. debug log
|
Just to confirm: I just flashed back to the non-debug v2.0.6, the restart problem is back again. Also I can see a huge difference in speed: With the debug firmware, after a button press it takes a short while until the LED lights up (presumably due to the amount of debug output to be sent prior LED is switched on). Anyway, the event is then sent close in time (1-2 seconds) to MQTT and the LED is off again. With non-debug firmware, LED light turns on comparatively fast, then it remains on for a quite long while (significantly longer than with the debug firmware...) and then a reboot occurs. Wild guess: Some race conditions? Is there any action relying on an already active WiFi connection, but it might not be completely set up at that time? And waiting for the logs to be sent with the debug firmware, WiFi has enough time to initialize? |
Initially, did you upgrade from v2.0.4 to v2.0.6 or was it from v2.0.3? The debug is slower because of additional output and because there's no compiler optimisations. That's normal. |
If I remember correctly, it was v2.0.3 -> v2.0.4 -> v2.0.6.
I know, just mentioned this because suspecting a race condition. |
Should I be able to downgrade from 2.0.6 to v2.0.3? Trying any flashing with OTA after full image recovery (regardless if v2.0.4 or v2.0.6): Website looses connection, but after a couple of seconds is back. Same version, nothing changed on the display. First flashing the full image you provided (regardless if 2.0.4 oder 2.0.6) and then flashing any other image to |
I think, beside the WiFi issue, there is an issue with the full image you sent me. I just tried to downgraded one of the other devices, where I never used a full image. Successfully downgraded to v2.0.3 using OTA, the WiFi issue is gone for this other device. For the first device, I'm not able to flash anything beside the full images you provided. Using OTA or esptool to flash a normal release does not work anymore. |
Downgrading to v2.0.3 from full image v2.0.4 & v2.0.6 will not work because the bootloader is different. On your other devices that were not flashed with full images v2.0.4 or later it's working, because they have the right bootloader from before. The old v2.0.3 bootloader is working with newer versions, new bootloader is not working with older versions. The documentation is missing one fact: the flash has two app partitions. One at 0x10000 an the other at 0x150000. OTA writes to the inactive one and changes a switch to boot from it. So if you previously used OTA and later use esptool you might not flash to the active slot - the new image won't boot. To be sure you must flash the same image to both addresses. Can you please try this update image. It's the same as v2.0.6 just with lower log level set. I think the log output of the arduino Wi-Fi library is causing this issue. The log will be mostly empty since it also disables output from my code. homebuttons_fw_v2.0.7-beta.bin.zip First flash the full image v2.0.6 and then this one with OTA or esptool. I confirmed this update steps on my device and I see no reason why it would be any different on yours. |
Bootloader issues: Understood, thanks for clarification. Thanks for the v2.0.7. update. v2.0.6 full (ESP) and then v2.0.7-beta OTA works, button press causes LED to go on... and to go off, without reboot :) Home Assistant is triggered, so... looks very good! Thanks for all the effort! |
Great, thanks for testing. I'll need to find another solution since it's still nice to have some log output, but at least we know where the issue is. |
Sure, let me know if I should test something else or can be helpful otherwise. |
espressif/arduino-esp32#7959 is solved now. If you provide me with an update I can test if the problem is gone with that patch. |
This is solved in v2.1.0 by disabling the problematic logs, while still having the useful log output. The fixes from arduino core will be implemented when they are included in a PlatformIO platform release. @MatthiasLohr please test the v2.1.0. |
Seems to be working! Thanks! |
Hi,
I have 3 Home Button devices and all of them experience the same issues after upgrading to v2.0.6:
@nplan, could you take a look please?
Best regards
Matthias
The text was updated successfully, but these errors were encountered: