NON-OS 1.5.4 (16_05_20) PWM bug

[eriksl]

If you use PWM on this version, the ESP will restart in about 15 minutes (exact time will vary) due to "hardware watchdog timeout". If you switch to one version earlier (ESP8266_NONOS_SDK_V1.5.3_16_04_18) the problem is gone.

I am very grateful for the IRAM cleanup btw. Gained about 4 Kbytes from 1.5.2 to 1.5.3.

[ESP_Faye]

Hi,

Can you provide your test code ? We will have a try.

Thanks for your interest in ESP8266 !

[eriksl]

That's a bit difficult, because it's part of a huge project (image is > 256 kbytes, mostly actual code). It also includes an alternative boot loader (rboot) which makes OTA actually possible in this case. It's here if you still want to try: http://github.com/eriksl/esp8266-universal-io-bridge

But, the PWM code like it is now, has been working fine in several SDK releases, for at least half a year, maybe longer. Only with the last update of the SDK the crashing started. When I go back one SDK version (or more), recompile+link everything, the problem is completely gone, exactly the same code and the same flow.

[eriksl]

It shouldn't be too hard to reproduce. Connect PWM channels to io 14 and io 15, activate one of them (or both) and wait.

[pratik]

Hello,

If you use only GPIO14 for PWM, does it still cause a watchdog reset after some time?

I'll see if I can reproduce the issue. What hardware are you using?

[eriksl]

Hello Pradik.

I have configured IO 14 and IO 15 for PWM but in practise only IO 14 is used. So both IO 14 and IO 15 are in the gpio_info_t struct array, but only IO 14 has a duty cycle > 0 normally.

I can reconfigure, so IO 15 isn't included into the struct array at all, so the struct array only has one member -> IO 14? Should that make a difference?

I have been using up to four PWM channels in the past and that was never a problem, btw.

[pratik]

Hello,

I could not reproduce the problem.
I'd suggest that you use an SDK version released just before the latest one available for stability reaons in end products.
I'll let you know if I find a solution to this or if I am able to reproduce the issue.

One thing you could try is wipe the flash clean and reprogram init_data and blank sectors and firmware. Sometimes that sorts some unexplained problems.

[eriksl]

This is not "production" stuff, and for the moment it's not a problem because I can fall back to the older SDK version.

But in future SDK versions, I really hope this will get fixed, so I can make use of new features.

[rudi]

This is not "production" stuff, and for the moment it's not a problem because I can fall back to the older SDK version.

But in future SDK versions, I really hope this will get fixed, so I can make use of new features.

hi erik

not sure - my unthink vision
15 min = 15 * 60 = 900 / 6.5 ns ( 0,0000000065 ) = 138461538461,5385 ( clk )
duty? ( 10 ) / 10
freq? ( 10 ) / 10
intval? ( 3 ) / 3
total 15 min ..= diff 3 sec ..

try to use/add

Code: Select all


while(1) {

// try this..
system_soft_wdt_feed();

// your pwm call
}


perhabs it helps

best wishes
rudi

[ESP_Yuhao]

Hi Erik!
I'm so sorry that ESP engineer could not reproduce the problem, either.

Code: Select all

/*
 * ESPRSSIF MIT License
 *
 * Copyright (c) 2015 <ESPRESSIF SYSTEMS (SHANGHAI) PTE LTD>
 *
 * Permission is hereby granted for use on ESPRESSIF SYSTEMS ESP8266 only, in which case,
 * it is free of charge, to any person obtaining a copy of this software and associated
 * documentation files (the "Software"), to deal in the Software without restriction, including
 * without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the Software is furnished
 * to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all copies or
 * substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
 * FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
 * COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
 * IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 *
 */

#include "esp_common.h"
#include "../sample_lib/wifi/wifi_test.h"
#include "uart.h"

#include "gpio_test.h"


/*Definition of GPIO PIN params, for GPIO initialization*/

#define PWM_4_OUT_IO_MUX PERIPHS_IO_MUX_MTDI_U
#define PWM_4_OUT_IO_NUM 12
#define PWM_4_OUT_IO_FUNC  FUNC_GPIO12

#define PWM_1_OUT_IO_MUX PERIPHS_IO_MUX_MTDO_U
#define PWM_1_OUT_IO_NUM 15
#define PWM_1_OUT_IO_FUNC  FUNC_GPIO15

#define PWM_2_OUT_IO_MUX PERIPHS_IO_MUX_GPIO4_U
#define PWM_2_OUT_IO_NUM 4
#define PWM_2_OUT_IO_FUNC  FUNC_GPIO4

#define PWM_3_OUT_IO_MUX PERIPHS_IO_MUX_MTMS_U
#define PWM_3_OUT_IO_NUM 14
#define PWM_3_OUT_IO_FUNC  FUNC_GPIO14

//#define PWM_4_OUT_IO_MUX PERIPHS_IO_MUX_GPIO5_U
//#define PWM_4_OUT_IO_NUM 5
//#define PWM_4_OUT_IO_FUNC  FUNC_GPIO5


uint32 io_info[][3] = {
    {PWM_1_OUT_IO_MUX,PWM_1_OUT_IO_FUNC,PWM_1_OUT_IO_NUM},
    {PWM_2_OUT_IO_MUX,PWM_2_OUT_IO_FUNC,PWM_2_OUT_IO_NUM},
    {PWM_3_OUT_IO_MUX,PWM_3_OUT_IO_FUNC,PWM_3_OUT_IO_NUM},
    {PWM_4_OUT_IO_MUX,PWM_4_OUT_IO_FUNC,PWM_4_OUT_IO_NUM}
};

void user_init(void)
{
   uint32 duty[3]={0,0,0},i;


      pwm_init(1000,duty,4,io_info);//1000us
      pwm_set_duty(20,0);/*20us */
      pwm_set_duty(10,1);/*10us This is IO4*/
      pwm_set_duty(15,2);/*15us */
      pwm_set_duty(5,3);/*5us */
      pwm_start();
   

}