T-Beam boot issue after being shut down from Android app #2364

IZ1IVA · 2023-03-17T16:06:02Z

Hi,

A T-Beam running firmware 2.1.2, after being shut down from Android app 2.1.1, will not boot up automatically when its USB port is plugged back into a power source. Pressing the POWER button does nothing, while pressing the USER button boots up the device.

This happens only when the device is shut down from the app, not from the POWER button.

Cheers,
Andrea

garthvh · 2023-03-17T16:14:18Z

This is not an issue on RAK devices after retesting

rpsainio · 2023-03-17T19:23:28Z

@IZ1IVA so, if you have USB-connected then you are able to run console and see what happens when you press USER-button. Most likely it will tell the wake-up reason. Might be a long Deep-Sleep allthough this is supposed to happen only on ESP32 platforms without AXP192

mverch67 · 2023-03-17T21:29:17Z

Same for TLora T3S3 !! Reboot from APP hangs and EVERY configuration change with reboot hangs. Need to press the reset button...

rpsainio · 2023-03-18T13:33:20Z

Maybe so, still at least I would like to see some console output. This application is quite talkative :)

karamo · 2023-03-18T22:27:38Z

T-Beam does shutdown with user-button, but suddenly reboots.

rpsainio · 2023-03-20T09:27:28Z

The original problem seems to be that TBeam goes to DeepSleep instead switching itself off.
INFO | ??:??:?? 272 From Radio onread
INFO | ??:??:?? 277 Shutting down from admin command
INFO | ??:??:?? 277 Shutting down
INFO | ??:??:?? 277 Entering deep sleep for 18880 seconds
INFO | ??:??:?? 277 Disable bluetooth
INFO | ??:??:?? 277 GPS prepare sleep!
INFO | ??:??:?? 277 GPS deep sleep!
INFO | ??:??:?? 277 Saving /prefs/db.proto
INFO | ??:??:?? 278 Saving /prefs/config.proto
INFO | ??:??:?? 278 Saving /prefs/module.proto
INFO | ??:??:?? 278 Saving /prefs/channels.proto
INFO | ??:??:?? 279 Setting GPS power=0

@caveman99 is this now a feature or bug? See: #2352 Besides it does not sleep forever

karamo · 2023-03-20T16:36:11Z

T-Beam does shutdown with user-button, but suddenly reboots.

here the DEBUG-LOG:

DEBUG | 16:30:20 151424 [Button] Long press start!
...
INFO  | 16:30:25 151429 [Button] Turning off screen
INFO  | 16:30:25 151429 [Button] Shutting down
INFO  | 16:30:25 151429 [Button] Entering deep sleep for 18880 seconds
INFO  | 16:30:25 151429 [Button] GPS prepare sleep!
INFO  | 16:30:25 151429 [Button] GPS deep sleep!
DEBUG | 16:30:25 151429 [Button] sx126x entering sleep mode (FIXME, don't keep config)
INFO  | 16:30:25 151429 [Button] Saving /prefs/db.proto
INFO  | 16:30:25 151430 [Button] Saving /prefs/config.proto
INFO  | 16:30:26 151430 [Button] Saving /prefs/module.proto
INFO  | 16:30:26 151430 [Button] Saving /prefs/channels.proto
INFO  | 16:30:26 151430 [Button] Setting GPS power=0
ets Jul 29 2019 12:21:46

rst:0x5 (DEEPSLEEP_RESET),boot:0x12 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:2
load:0x3fff0030,len:1184
load:0x40078000,len:13192
load:0x40080400,len:3028
entry 0x400805e4
E (71) esp_core_dump_flash: No core dump ðfÉÑ¥Ñ¥½¹�found!
E (71) esp_core_dump_flash: No core dump partition found!
[    12][D][esp32-hal-cpu.c:244] setCpuFrequencyMhz(): PLL: 480 / 2 = 240 Mhz, APB: 80000000 Hz
[   464][I][esp32-hal-psram.c:96] psramInit(): PSRAM enabled
”ÃNFO  | ??:??:?? 0 

//\ E S H T /\ S T / C
INFO  | ??:??:?? 0 Booted, wake cause 3 (boot count 4), reset_reason=reset

rpsainio · 2023-03-20T17:02:49Z

@karamo My guess for what happens here:

long press triggers Shutdown/DeepSleep <- this is the bug
TBeam configures USER-button as a wakeup source
TBeam goes DS
USER-button is still pressed
TBeam wakes up / reboots with DEEPSLEEP_RESET as reason

If you release the USER-button just when it is saving its config it will not wake up

karamo · 2023-03-20T17:43:01Z

Ohh, yes. If I release the Button when the display switched off ... stay off.
Press the User-Button ... wake up.

garthvh · 2023-03-22T16:14:45Z

This is part of this feature request

#2326

If a device is in deep sleep only GPIO will wake it back up (user button)

rpsainio · 2023-03-22T17:33:21Z

@garthvh but as you can read it wakes up immediately if the button is kept pressed while ESP prepares for deepsleep. Very hard to implment that FR if at all - STILL a bug
Besides TBeam has an AXP....

rpsainio · 2023-04-10T06:36:02Z

@caveman99 shall we open another issue for the fact that the deepsleep lasts only 18880 seconds, most likely on all ESP32 platoforms

garthvh · 2023-04-10T16:08:43Z

@caveman99 shall we open another issue for the fact that the deepsleep lasts only 18880 seconds, most likely on all ESP32 platoforms

There are no other logs with this number, only yours what value did you set for SDS?

rpsainio · 2023-04-10T18:32:55Z

@garthvh you might notice that there are two logs in this issue from me and another one from Karamo. They both show the same value 18880 seconds when the deep sleep is entered. So mine was most likely 3600 seconds and you have to ask Karamo for that other log. I think this value might be a result of wrong conversion or simply a programming error leading to this funny value. So, we must build some debugging to see where the value gets set up.

rpsainio · 2023-04-11T06:42:40Z

@garthvh So, the value is 3600 and I changed it to 7200 and deep sleep is still for 18880 seconds. I think there is no relation between the parameter and the sleeping time. Back to my original question: Shall we open another issue for this funny deepsleep-time ?

`
[risto@rsa46 nanovna-saver]$ meshtastic --port /dev/ttyACM0 --get power.sdsSecs
Connected to radio
power.sds_secs: 3600
Completed getting preferences
[risto@rsa46 nanovna-saver]$ meshtastic --port /dev/ttyACM0 --set power.sdsSecs 7200
Connected to radio
Set power.sds_secs to 7200
Writing modified preferences to device

INFO | ??:??:?? 36 [Button] Shutting down
INFO | ??:??:?? 36 [Button] Entering deep sleep for 18880 seconds
INFO | ??:??:?? 36 [Button] GPS prepare sleep!
INFO | ??:??:?? 36 [Button] GPS deep sleep!
INFO | ??:??:?? 36 [Button] Saving /prefs/db.proto
INFO | ??:??:?? 37 [Button] Saving /prefs/config.proto
INFO | ??:??:?? 37 [Button] Saving /prefs/module.proto
INFO | ??:??:?? 37 [Button] Saving /prefs/channels.proto
INFO | ??:??:?? 38 [Button] Setting GPS power=0
`

karamo · 2023-04-11T10:28:45Z

I am not a C/C++ specialist. But could this be the problem?

sleep.cpp line 204:

void doDeepSleep(uint64_t msecToWake)
    LOG_INFO("Entering deep sleep for %lu seconds\n", msecToWake / 1000);

All values are defined as UINT32, but in doDeepSleep is uint64_t and format %lu.

caveman99 · 2023-04-11T10:30:06Z

we're setting the device to sleep for portMAX_DELAY milliseconds, which actually is a magic number if INCLUDE_vTaskSuspend is defined, meaning 'sleep forever'. This is documented in the FreeRTOS spec. The printed number is wrong, the code is doing the right thing.

caveman99 · 2023-04-11T10:31:31Z

@karamo #define portMAX_DELAY ( TickType_t ) 0xffffffffUL

karamo · 2023-04-11T10:36:05Z

@caveman99 in freertosinc.h line 37:
#define portMAX_DELAY UINT32_MAX

@karamo

- fix wrong debug print - change shutdown logic for t-beam if PMU is detected - wait for 10 seconds instead of 5 for shutdown and resurrect screen brightness adjust for @karamo

rpsainio · 2023-04-11T11:11:09Z

@caveman99 Let me test this as I recall that it came back after 18880 seconds.... and did not sleep for ever.

caveman99 · 2023-04-11T11:11:31Z

please wait a bit, fixing compilation errors ...

caveman99 · 2023-04-11T11:15:15Z

@caveman99 in freertosinc.h line 37: #define portMAX_DELAY UINT32_MAX

this is not called if FreeRTOS is active. See a few line above that ...
// Include placeholder fake FreeRTOS defs

The real def is in .platformio\packages\framework-arduinoespressif32\tools\sdk\esp32\include\freertos\port\xtensa\include\freertos\portmacro.h

rpsainio · 2023-04-11T11:19:09Z

anyway - test runs now

garthvh · 2023-04-11T16:21:55Z

Makes sense that 18880 was a display error, it does not exist anywhere in all of GitHub other than this issue.

It also might make sense to set a SDS minimum value, anything less than 12 hours is essentially invalid and potentially dangerous, 24 hours is probably better. What would the real use case be for 2hours? If solar ran flat overnight you are just firing up a flat battery again without any charging being done by solar.

The tbeam behavior has been updated to not shut down with the user button again, the screen brightness "feature" is totally broken and really should probably go away again. I have honestly never engaged it on purpose.

karamo · 2023-04-11T18:19:21Z

@garthvh what do you mean? "18880 was a display error"
How can it be? Where does it came from?

garthvh · 2023-04-11T19:20:55Z

Bad math, neither of you set a timer and waited 5 and a half hours to see if it actually turned back on. I would reverse it, does anyone have any actual testing that indicates it was really going to wake from deep sleep after 5 hours?

rpsainio · 2023-04-11T19:23:41Z

@garthvh I just tested this scenario and it did not turn on after 5 hours - so seems to be a problem with number formatting

garthvh · 2023-04-13T05:00:36Z

Adding the screen brightness bug back is causing issues

tropho23 · 2023-04-13T05:08:43Z

New issue for this change: the primary problem is that it breaks user button-initiated shutdown, which is a critical feature for RAK-based devices. Firmware version 2.1.8 makes user button-initiated shutdown impossible for RAK devices.

In addition, this now doesn't even behave as intended; the change breaks screen function (via user button) of at least SH1106-based screens such as 1.3" OLEDs. When long-pressing the user button, the brightness indicator bar briefly appears and the screen flickers and sometimes turns off, or dims (random and uncontrollable). It takes a device reset to restore screen brightness and readability.

If this change resolved a T-Beam-specific issue, recommend keeping it only for T-Beam firmware and revert to the functionality & behavior from 2.1.7 for all other targets.

Alternately, simply remove the screen brightness dimming feature for all devices if it behaves unpredictably.

caveman99 · 2023-04-13T12:42:28Z

i am going to toss it again. this time for good :-) really more hassle than its worth.

tropho23 · 2023-04-13T13:03:24Z

Thanks! As Garth noted, I doubt anyone even uses this capability and a reboot resets to screen default brightness anyways. Dimming the screen would be a useful feature, if we were able to change a persistent brightness value that in the firmware/app settings somehow for low-light environments. I don't want to distract from other development so this is not a feature request, just stating that would be a better way to adjust brightness, assuming it's even possible to do it. Tony

…

________________________________ From: Thomas Göttgens ***@***.***> Sent: Thursday, April 13, 2023 08:42 To: meshtastic/firmware ***@***.***> Cc: tropho23 ***@***.***>; Comment ***@***.***> Subject: Re: [meshtastic/firmware] T-Beam boot issue after being shut down from Android app (Issue #2364) i am going to toss it again. this time for good :-) really more hassle than its worth. — Reply to this email directly, view it on GitHub<#2364 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AQ7GUPTGEA2YSSLUZIHQUDLXA7YD7ANCNFSM6AAAAAAV6XDNYA>. You are receiving this because you commented.Message ID: ***@***.***>

caveman99 · 2023-04-13T13:34:46Z

next firmware will (again) remove screen brightness.

andrekir transferred this issue from meshtastic/Meshtastic-Android Mar 17, 2023

garthvh closed this as completed Mar 22, 2023

caveman99 reopened this Apr 9, 2023

caveman99 added a commit that referenced this issue Apr 11, 2023

resolve #2364

7c4b267

- fix wrong debug print - change shutdown logic for t-beam if PMU is detected - wait for 10 seconds instead of 5 for shutdown and resurrect screen brightness adjust for @karamo

caveman99 closed this as completed in 29c13b5 Apr 11, 2023

garthvh reopened this Apr 13, 2023

caveman99 closed this as completed Apr 13, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

T-Beam boot issue after being shut down from Android app #2364

T-Beam boot issue after being shut down from Android app #2364

IZ1IVA commented Mar 17, 2023

garthvh commented Mar 17, 2023 •

edited

Loading

rpsainio commented Mar 17, 2023

mverch67 commented Mar 17, 2023

rpsainio commented Mar 18, 2023

karamo commented Mar 18, 2023

rpsainio commented Mar 20, 2023 •

edited

Loading

karamo commented Mar 20, 2023

rpsainio commented Mar 20, 2023

karamo commented Mar 20, 2023 •

edited

Loading

garthvh commented Mar 22, 2023

rpsainio commented Mar 22, 2023 •

edited

Loading

rpsainio commented Apr 10, 2023

garthvh commented Apr 10, 2023

rpsainio commented Apr 10, 2023

rpsainio commented Apr 11, 2023 •

edited

Loading

karamo commented Apr 11, 2023 •

edited

Loading

caveman99 commented Apr 11, 2023

caveman99 commented Apr 11, 2023

karamo commented Apr 11, 2023

rpsainio commented Apr 11, 2023

caveman99 commented Apr 11, 2023

caveman99 commented Apr 11, 2023

rpsainio commented Apr 11, 2023

garthvh commented Apr 11, 2023

karamo commented Apr 11, 2023 •

edited

Loading

garthvh commented Apr 11, 2023

rpsainio commented Apr 11, 2023

garthvh commented Apr 13, 2023

tropho23 commented Apr 13, 2023 •

edited

Loading

caveman99 commented Apr 13, 2023

tropho23 commented Apr 13, 2023 via email

caveman99 commented Apr 13, 2023

T-Beam boot issue after being shut down from Android app #2364

T-Beam boot issue after being shut down from Android app #2364

Comments

IZ1IVA commented Mar 17, 2023

garthvh commented Mar 17, 2023 • edited Loading

rpsainio commented Mar 17, 2023

mverch67 commented Mar 17, 2023

rpsainio commented Mar 18, 2023

karamo commented Mar 18, 2023

rpsainio commented Mar 20, 2023 • edited Loading

karamo commented Mar 20, 2023

rpsainio commented Mar 20, 2023

karamo commented Mar 20, 2023 • edited Loading

garthvh commented Mar 22, 2023

rpsainio commented Mar 22, 2023 • edited Loading

rpsainio commented Apr 10, 2023

garthvh commented Apr 10, 2023

rpsainio commented Apr 10, 2023

rpsainio commented Apr 11, 2023 • edited Loading

karamo commented Apr 11, 2023 • edited Loading

caveman99 commented Apr 11, 2023

caveman99 commented Apr 11, 2023

karamo commented Apr 11, 2023

rpsainio commented Apr 11, 2023

caveman99 commented Apr 11, 2023

caveman99 commented Apr 11, 2023

rpsainio commented Apr 11, 2023

garthvh commented Apr 11, 2023

karamo commented Apr 11, 2023 • edited Loading

garthvh commented Apr 11, 2023

rpsainio commented Apr 11, 2023

garthvh commented Apr 13, 2023

tropho23 commented Apr 13, 2023 • edited Loading

caveman99 commented Apr 13, 2023

tropho23 commented Apr 13, 2023 via email

caveman99 commented Apr 13, 2023

garthvh commented Mar 17, 2023 •

edited

Loading

rpsainio commented Mar 20, 2023 •

edited

Loading

karamo commented Mar 20, 2023 •

edited

Loading

rpsainio commented Mar 22, 2023 •

edited

Loading

rpsainio commented Apr 11, 2023 •

edited

Loading

karamo commented Apr 11, 2023 •

edited

Loading

karamo commented Apr 11, 2023 •

edited

Loading

tropho23 commented Apr 13, 2023 •

edited

Loading