Software ESP32 restart ( ESP.restart()) doesn't seems to fully reset the hardware

[Rrobinet]

Hi,
With the following script I am using the Hardware Serial interfaces.
Practically the simplified script shown below, uses a Nextion LCD display that connects via the Hardware interface 2.
The script ((see below) writes and reads a Nextion HMI variable, through the hardware interface 2 (pin 16=RX, pin 17=TX baud rate 9600).
In a more elaborate script I am using the ESP.restart() call to reset the ESP32 (WEMOS LOLIN32) after reconfiguration, however I need to physically restart the ESP to let the hardware interface work again,
the software restart seems to leave the hardware interface in a unknown state.

The simplified script below writes and reads 5 times back the data stored by the Nextion MCU.

The result shows that everything is fine until the "software restart" occurs, only a physical restart or firmware reloads solves the problem

• !!!!! 1. Loading the script
  ⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮⸮ets Jun 8 2016 00:22:57
  
  rst:0x10 (RTCWDT_RTC_RESET),boot:0x13 (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:1
  load:0x3fff0018,len:4
  load:0x3fff001c,len:812
  load:0x40078000,len:0
  load:0x40078000,len:11392
  entry 0x40078a9c
  recvRetCommandFinished ok <-- Nextion Init is OK
  recvRetCommandFinished ok
  
  Doing Nextion Communication test <-- Communication test 1 is OK
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :319
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 2 is OK
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :267
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 3 is OK
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :697
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 4 is OK
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :779
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 5 is OK
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :29
  Nextion test is OK
  
  !!!! 2. ESP.restart()
  ets Jun 8 2016 00:22:57
  
  rst:0xc (SW_CPU_RESET),boot:0x13 (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:1
  load:0x3fff0018,len:4
  load:0x3fff001c,len:812
  load:0x40078000,len:0
  load:0x40078000,len:11392
  entry 0x40078a9c
  recvRetCommandFinished err <-- Nextion Init is NOK
  recvRetCommandFinished err
  
  Doing Nextion Communication test <-- Communication test 1 is NOK
  Now updating COMMV1 value
  recvRetCommandFinished err
  recvRetNumber err
  Nextion test FAILED
  653
  0
  
  Doing Nextion Communication test <-- Communication test 2 is NOK
  Now updating COMMV1 value
  recvRetCommandFinished err
  recvRetNumber err
  Nextion test FAILED
  352
  1073488868
  
  Doing Nextion Communication test <-- Communication test 3 is NOK
  Now updating COMMV1 value
  recvRetCommandFinished err
  recvRetNumber err
  Nextion test FAILED
  200
  1073488868
  
  Doing Nextion Communication test <-- Communication test 4 is NOK
  Now updating COMMV1 value
  recvRetCommandFinished err
  recvRetNumber err
  Nextion test FAILED
  58
  1073488868
  
  Doing Nextion Communication test <-- Communication test 5 is NOK
  Now updating COMMV1 value
  recvRetCommandFinished err
  recvRetNumber err
  Nextion test FAILED
  739
  1073488868
  
  !!! 3. Hardware reset
  rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
  ets Jun 8 2016 00:22:57
  
  rst:0x10 (RTCWDT_RTC_RESET),boot:0x13 (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:1
  load:0x3fff0018,len:4
  load:0x3fff001c,len:812
  load:0x40078000,len:0
  load:0x40078000,len:11392
  entry 0x40078a9c
  recvRetCommandFinished err
  recvRetCommandFinished ok <-- Nextion Init is OK
  
  Doing Nextion Communication test <-- Communication test 1 is OK
  Now updating COMMV1 value
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :976
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 2 is OK
  Now updating COMMV1 value
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :114
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 3 is OK
  Now updating COMMV1 value
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :761
  Nextion test is OK
  
  Doing Nextion Communication test <-- Communication test 4 is OK
  Now updating COMMV1 value
  Now updating COMMV1 value
  recvRetCommandFinished ok
  recvRetNumber :855
  Nextion test is OK
  .....

Code: Select all

#include "Nextion.h"                  // Include Nextion library
HardwareSerial Serial2(2);            // Default pins for Wemos Lolin32 are pin 16=RX, pin 17=TX baud rate 9600
long int nexPollTime;                 // USed to collected itermediate time  
NexVariable   COMMV1        =   NexVariable      (0,23,"COMMV1");         // Nextion Communication watchdog variable
#define DEBUGGING

void setup() {
     nexInit ();        // Init the nextion communication (typically Serial 2 interface  @9600 bps)
     delay (100);       // Settle down delay
}

void loop() {
  // Call 5 times the Nextion befor restart
  for (int i = 0; i <5; i++) 
  {  
    nexCommCheck (0);   // Check the Nextion connection immediately
    delay (1000);       // Wait 1 second
  }
   ESP.restart();       // Then restart the WeMos
  
}

void nexCommCheck (int long pollRate)
{
   if (millis()>nexPollTime+pollRate)
   {
#ifdef DEBUGGING 
  Serial.println ("\nDoing Nextion Communication test") ;
#endif      
      nexPollTime = millis();
      uint32_t setvalue = random (1000);
      uint32_t getvalue;
#ifdef DEBUGGING
  Serial.println ("Now updating COMMV1 value");
#endif
      COMMV1.setValue (setvalue);
      COMMV1.getValue (&getvalue);
      if (getvalue == setvalue)          // Nextion Write check
     {   
#ifdef DEBUGGING    
     Serial.println ("Nextion test is OK");
#endif 
     }
     else 
     {
#ifdef DEBUGGING
      Serial.println ("Nextion test FAILED"); 
      Serial.println (setvalue);
      Serial.println (getvalue);
#endif   
     }
   }
}

[vonnieda]

FYI: I've recently posted the same question, and no responses yet: https://esp32.com/viewtopic.php?f=2&t=4949

[Rrobinet]

This problem is solved (see https://github.com/espressif/arduino-esp32/issues/1189)
by applying the UART patch on esp323_hal_uart.c
replacing : void uartFlush(uart_t* uart) (on the latest https://github.com/espressif/arduino-esp32) by:

Code: Select all

void uartFlush(uart_t* uart)
{
    if(uart == NULL)
    {
    return;
    }

    UART_MUTEX_LOCK();
    while(uart->dev->status.rxfifo_cnt != 0 || (uart->dev->mem_rx_status.wr_addr != uart->dev->mem_rx_status.rd_addr))
    {
       READ_PERI_REG(UART_FIFO_REG(uart->num));
    }
UART_MUTEX_UNLOCK();
}

Robert

[tihomir.9a4gl]

The above modification does not help in my case, but...

I had the same issue and I read somewhere that Arduino HardwareSerial implementation has this issue and the ESP-IDF is fine. So, I tried to implement my HardwareSerial using ESP-IDF and this really works fine, you can do ESP.restart and UART2 works !

I pushed my implementation called ESP32Serial here https://github.com/9a4gl/ESP32Serial

[Rrobinet]

@tihomir.9a4gl
The issue(s) with the secondary UART interfaces is discussed on several places, and very confusing.
So as I understand the patch described above was a workaround for me, but not an actual solution.

Last week I have decided to verify if this problem was not solved by the latest version of the espressif-arduino library and as a matter of fact it is. So please try the last version to verify if this solves your problem too.

Now there is a undocumented feature; you don't (may) not specify

Code: Select all

HardwareSerial Serial2(2)

in the script or include the

Code: Select all

#include HardwareSerial.h

library , this is implicitly done in the arduino library.
Just specify

Code: Select all

Serial2.begin(115200)

for instance to set the baud rate for the 2nd interface for example
I tried my script above by simply remove

Code: Select all

//HardwareSerial Serial2(2);            // Default pins for Wemos Lolin32 are pin 16=RX, pin 17=TX baud rate 9600

which is working fine.
Note that the baud rate of the Serial2 is defined in the nextInit() function
Robert

[tihomir.9a4gl]

> Last week I have decided to verify if this problem was not solved by the latest version
> of the espressif-arduino library and as a matter of fact it is. So please try the last version to verify if this solves your problem too.

I am using atom + platformio, just noticed there is update for Espressif32 platform, 5 days old. New version is 1.2.0. So I upgraded, and I see news is that Serial2 does not need to be created, it is already created by Arduino framework. But bug is still there, Serial2 works for me when plugged to power supply. After ESP.restart() it get stuck when I try to use UART2 and complete platform is blocked. Using my ESP32Serial implementation it still works. So, I am not sure you are doing same as me or we are using different version of espressif-arduino. Anyone else still have issue described here ?

BR,
Tihomir

[Rrobinet]

thimor,

You are probably right, looking a the new release of the only difference with the esp32-hal-uart.c is that the patch is now included.
So if it was not working for you before, it should not work now.
However it fix my issue
Now with this patch there is a comment:

//Due to hardware issue, we can not use fifo_rst to reset uart fifo.
//See description about UART_TXFIFO_RST and UART_RXFIFO_RST in <<esp32_technical_reference_manual>> v2.6 or later

So it might be an hardware issue related to ESP32 versions

Robert