Skip to content
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

Xtensa compiler (esp8266): Declaration of constructor does not match implementation #11

Closed
nerk opened this issue Jan 2, 2016 · 18 comments

Comments

@nerk
Copy link

nerk commented Jan 2, 2016

Compiling TaskScheduler.h using the xtensa-lx106-elf-g++ compiler (esp8266) fails. Apparently, it can not match the declaration of the constructor with its implementation because of mixing boolean and bool for parameter aEnable.

public:
   Task(unsigned long aInterval=0, long aIterations=0, void (*aCallback)()=NULL, Scheduler*  aScheduler=NULL, boolean aEnable=false, bool (*aOnEnable)()=NULL, void (*aOnDisable)()=NULL);

needs to become

public:
   Task(unsigned long aInterval=0, long aIterations=0, void (*aCallback)()=NULL, Scheduler* aScheduler=NULL, bool aEnable=false, bool (*aOnEnable)()=NULL, void (*aOnDisable)()=NULL);
@arkhipenko
Copy link
Owner

I have not yet tried to use TaskScheduler with esp8266. Very cool. Will fix the type mismatch as soon as I get back to my laptop.
Please do tell if it works. I have the chip, just never got around to compiling fo it.
I suspect (strongly) that power management (IDLE sleep) might not work, and I am not sure how well current version will play with the WFL.
Would be great if you could share your experience. Thanks.

@arkhipenko
Copy link
Owner

Could you please test v2.0.1?

@alw1746
Copy link

alw1746 commented Jan 3, 2016

Fyi,I did a quick-n-dirty private port a few months back on v1.8.2 and it did work very well. I noted:

-the task constructor declaration bug which you've now fixed.
-examples 1-5 worked with some minor recode.
-example 6(power management) worked,can't remember the counts.
-not example 7 (WDT not available in ESP Arduino).
-example 8 worked.

A diff of my code showed I made a declaration at top:

#ifndef _ESP8266
#include <avr/sleep.h>
#include <avr/power.h>
#endif

and a conditional test for ESP8266 in Scheduler::execute()

#ifdef _TASK_SLEEP_ON_IDLE_RUN
  if (idleRun && iAllowSleep) {
#ifdef _ESP8266
    const int sleepTimeMs = 1;
    ESP.deepSleep(sleepTimeMs*1000);
#else
    set_sleep_mode(SLEEP_MODE_IDLE);
    sleep_enable();
      /* Now enter sleep mode. */
      sleep_mode();
      
      /* The program will continue from here after the timer timeout ~1 ms */
      sleep_disable(); /* First thing to do is disable sleep. */
#endif

@arkhipenko
Copy link
Owner

Great. Thanks for sharing. I will incorporate your updates in the next version.
I do recommend switching to the latest v2.x - I have made significant optimization of the execute loop finding a few (minor) bugs in the process.
Plus v2 now supports task prioritization.

Enjoy! Glad you find TaskScheduler suitable for your projects.

@nerk
Copy link
Author

nerk commented Jan 3, 2016

Thanks a lot to you for this great project! This code is exactly what I was looking for.
I only tested the most basic things and they seem to work as expected, Obviously, idle sleep cannot work with the current code base and ESP8266. @alw1746 seems to have solved most issues.

I will try again once the other changes are in place. Cool stuff! 👍

@arkhipenko
Copy link
Owner

Thank you. Always good to know your work is useful for someone. :)

@arkhipenko
Copy link
Owner

@alw1746
Alex, it seems WDT might be available for ESP. Please check this:
https://github.com/sandeepmistry/esp8266-Arduino/blob/master/esp8266com/esp8266/cores/esp8266/Esp.h

@arkhipenko
Copy link
Owner

@alw1746
Alex, another note:
I don't think this code would work:

const int sleepTimeMs = 1;
ESP.deepSleep(sleepTimeMs*1000);

According to many articles (this for example: http://tinker.yeoman.com.au/2015/03/08/reducing-esp8266-power-consumption-using-deep-sleep/) esp8266 does a full reset after wake up, which would restart the sketch - not the behavior execute() function needs.

Instead suggest this:

Add wifi sleep mode setter in the allowSleep() method:

#ifdef _TASK_SLEEP_ON_IDLE_RUN
void Scheduler::allowSleep(bool aState) { 
    iAllowSleep = aState; 

#ifdef ARDUINO_ARCH_ESP8266
    wifi_set_sleep_type( iAllowSleep ? LIGHT_SLEEP_T : NONE_SLEEP_T );
#endif

}
#endif

and this to the execute() method:

#ifdef _TASK_SLEEP_ON_IDLE_RUN
    if (idleRun && iAllowSleep) {

#ifdef ARDUINO_ARCH_AVR    // Could be used only for AVR-based boards. 
      set_sleep_mode(SLEEP_MODE_IDLE);
      sleep_enable();
      /* Now enter sleep mode. */
      sleep_mode();

      /* The program will continue from here after the timer timeout ~1 ms */
      sleep_disable(); /* First thing to do is disable sleep. */
#endif

#ifdef ARDUINO_ARCH_ESP8266
// to do: find suitable sleep function for esp8266
        delay(985);        // ESP8266 implementation of delay() uses timers and yield
#endif
    }
#endif

    return (idleRun);
}

@alw1746
Copy link

alw1746 commented Jan 6, 2016

hmm..in that case I don't understand. I ran example 6 and this is what I got,
Scenario 1): With #define _TASK_SLEEP_ON_IDLE_RUN enabled
Start
c1=497878
c2=1001
Scenario 2): With #define _TASK_SLEEP_ON_IDLE_RUN disabled (commented out)
Start
c1=828914
c2=1001
C1 in scenario 2) is higher than in scenario 1) because processor is put to sleep for 1), but not for 2). But the count is only half so it's not a big diff, plus I did not see a reset. Perhaps ESP.deepsleep isn't working as I thought? Let me reflash my ESP8266 with the latest Arduino release and taskscheduler and I'll report back.

@arkhipenko
Copy link
Owner

@alw1746
Alex, Comparing your results to the 16Mhz Uno:

for the following tasks:

Task c(10, -1, &Count, &ts);
Task t(10000, 1, NULL, &ts, true, &tOn, &tOff);

c1 should be "around" 10000 (with the IDLE RUN SLEEP enabled), so clearly it does not sleep.
c2 should always be 1000 (10000/10)

Without IDLE RUN SLEEP for UNO at 16 MHz c1=529783 (v1.9.0: 551947)

I suppose ESP8266 is running at 80 MHz, but is probably doing a lot more in the background for WFL, so straight extrapolation (~550000 / 16 * 80 = 2750000) will not be accurate. 828914 "feels" low, but then again, I am not sure what else CPU is scheduled to do (probably a lot).

I guess it is time for me to pull my esp8266 out and start testing :)

PLEASE CHECK THE TESTING BRANCH WITH RECENT UPDATES FOR ESP8266!!

@arkhipenko
Copy link
Owner

@alw1746
Hi Alex, just wondering if you had a moment to test. Not urgent, just curious to know how it works with ESP8266... Thanks.

@alw1746
Copy link

alw1746 commented Jan 12, 2016

sorry mate been away so haven't had a chance to make blinky lights. But I'm downloading the latest ESP8266 Arduino build now so will let you know in the next day or so...

@arkhipenko
Copy link
Owner

@alw1746 no problem! Thanks for doing this actually. Please use the testing branch for this.

@alw1746
Copy link

alw1746 commented Jan 13, 2016

Had some problems with my ESP8266 Arduino v2.0.0 download so I tested Taskscheduler 2.02 (testing branch) against ESP8266 v1.65 on my board already. Attached test results. All worked apart from Example 7 (WDT) which I'll test again once I get 2.0.0 sorted out. Some minor mods needed, see the test results.
TaskschedulerTestResults_ESP8266.txt

@alw1746
Copy link

alw1746 commented Jan 13, 2016

Running with ESP8266 Arduino 2.0.0 now but can't get Example 7 to work. Replaced the timeout setting code with,

//reset watchdog
wdt_reset();
ESP.wdtEnable(WDTO_2S);
//Enable global interrupts

But below ISR failed to compile as WDT_vect is not exposed in Esp.h.

ISR(WDT_vect)

@arkhipenko
Copy link
Owner

@alw1746
Alex, thanks for testing. Good news!

Let me look into WDT a little closer.

And I still didn't put the chip to sleep per se, just delayed by 1 ms in case IDLE SLEEP is activated, so a bit of "cheating" on my end...
But it is entirely possible that ESP8266 does not support sleep the same way AVRs do, so maybe this is the only way.

To be investigated further....

@arkhipenko
Copy link
Owner

@alw1746
Addressed all of your changes in the main code of the examples for the next release, except for the WDT. Will have to look into that and possible sleep next.

@arkhipenko
Copy link
Owner

Just pushed the update with examples changes into 'testing' branch.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants