Timer pull request

[Tjeerdie]

Adding on timer to make channel 1..4 IRQ callbacks possible

The timer.c and timer.h are modified such that you can now attach
interrupt callbacks to all channels of the hardware timers. This feature
should be fully backwards compatible with the previous implementation.

The addition is done to make an arduino core implementation of
HardwareTimers possible (Next step). That would be inline with what other
arduino stm32 cores offer.

This PR starts implementation of the following bugs/features
See issue #146

[Tjeerdie]

Example ino to check function

//
/*
 * Hardware timer interrupt 
 * Overflow example and channel example
 *
 * Setup the timers and call interrupt routines
 *
 * This example code is in the public domain.
 */
 
volatile int counterCH1;
volatile int counterCH2;
volatile int counterCH3;
volatile int counterCH4;
volatile int counterOC;
volatile int timertrig_channel;
volatile int channel_nr;
stimer_t _timer;

extern void IRQcallback_TIM2OC(stimer_t *Timer, uint32_t channel)
{
    counterOC++;
//    digitalWrite(PA6, !digitalRead(PA6));
//    Serial.println(__channel);
    
}

extern void IRQcallback_TIM2CH1()
{
    counterCH1++;
    digitalWrite(PA6, !digitalRead(PA6));
//    Serial.println(__channel);
    
}
extern void IRQcallback_TIM2CH2()
{
    counterCH2++;
    digitalWrite(PA6, !digitalRead(PA6));
//    Serial.println(__channel);
    
}
extern void IRQcallback_TIM2CH3()
{
    counterCH3++;
    digitalWrite(PA6, !digitalRead(PA6));
//    Serial.println(__channel);
    
}
extern void IRQcallback_TIM2CH4()
{
    counterCH4++;
    digitalWrite(PA6, !digitalRead(PA6));
//    Serial.println(__channel);
    
}

extern void IRQcallback_TIM2(stimer_t *Timer)
{
    timertrig_channel++;
    digitalWrite(PA6, !digitalRead(PA6));
    
}


void setup() {
  counterCH1 = 0;
  counterCH2 = 0;
  counterCH3 = 0;
  counterCH4 = 0;
  counterOC = 0;

  //begin serial for debug info.
  Serial.begin(115200);

  //attach timer to struct
  _timer.timer = TIM2;

  //led/pin output to see result
  //results can only be seen with logic analyzer/oscilloscoop by monitoring this pin
  //Result is showing the different phases the intrupts occur at. 
  pinMode(PA6, OUTPUT);

  Serial.println("--- Start timer init ---");
      
  TimerHandleInit(&_timer, 2000, 168);
  attachIntHandle(&_timer, IRQcallback_TIM2);

  TimerPulseInit(&_timer, 2000, 500, IRQcallback_TIM2OC);
  attachIntHandleOC(&_timer, IRQcallback_TIM2CH1, 1, 500);
  attachIntHandleOC(&_timer, IRQcallback_TIM2CH2, 2, 250);
  attachIntHandleOC(&_timer, IRQcallback_TIM2CH3, 3, 880);
  attachIntHandleOC(&_timer, IRQcallback_TIM2CH4, 4, 100);

  Serial.println("--- Finish timer init ---");
  

}

void loop() {
  // put your main code here, to run repeatedly:
          Serial.print("Counter_1: ");
          Serial.println(counterCH1);
          Serial.print("Counter_2: ");
          Serial.println(counterCH2);
          Serial.print("Counter_3: ");
          Serial.println(counterCH3);
          Serial.print("Counter_4: ");
          Serial.println(counterCH4);
          Serial.print("Channel_OC: ");
          Serial.println(counterOC);
          Serial.print("Channel: ");
          Serial.println(timertrig_channel);
           
          delay(1000);

}

[fpistm]

Thanks @Tjeerdie
I will review this soon and maybe the last part of the implementations. 😉
About astyle issue, I guess you use astyle without format config.
It is available here: https://github.com/stm32duino/Arduino_Core_STM32/blob/master/CI/astyle/.astylerc

[Tjeerdie]

When i get home i try to fix the Astyle issue.

[Tjeerdie]

@fpistm Did you already start with a HardwareTimer implementation or is there nothing implemented yet? I was considering using the STM32GERNERIC core implementation as a starting point for this. Because it is also based on the ST HAL. But i do not want to do redundant work if not necessary.

[fpistm]

@Tjeerdie
In fact not really, as I would study all current implementation and request from user.
I guess the one from STM32Generic is a good starting point if I well remember.
I will try to check soon, but I think I could not give you a feedback before the next week.

[Tjeerdie]

@fpistm Any news if this is going to be accepted like it is? Or does it need more work before that?

[fpistm]

@Tjeerdie Sorry I didn't have time to check.
Will try to review it this week but I could not not guarantee it as I'm very busy on several stuff. 😭

[noisymime]

This would be a great addition. Currently using the stimer arrangement, there's no way at all to utilise any compare channel except 1, which makes the implementation very limited.

The PR allows use of these extra channels without any breakage to current use.

The only alternative to getting the equivalent functionality to this PR would be to move the stm32generic HardwareTimer implementation over.

[BennehBoy]

Pr's are welcome :)

…

On Thu, 7 Mar 2019, 10:48 Josh Stewart, ***@***.***> wrote: This would be a great addition. Currently using the stimer arrangement, there's no way at all to utilise any compare channel except 1, which makes the implementation very limited. The PR allows use of these extra channels without any breakage to current use. The only alternative would be to move the stn32generic HardwareTimer implementation over. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#447 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AX0eXuwCPwDb-kJQfqo2LJscqRgLEZiPks5vUO58gaJpZM4bBk7T> .

[fpistm]

First review:

• Several commits needs to be squash and commit messages to review following contributing guidelines.
  Ex: 2 commits:

[Timer] Add get/set prescaler register function
[Timer] Add channels support

• Rebase on top of the master
• Some enhancement could be performed to avoid lot of return or factorize code

[fpistm]

@Tjeerdie I will try to provide a branch updated this afternoon, then you will be able to test and give feedback --> done in #473

[Tjeerdie]

@fpistm i will check this as soon as i have time. Hopefully this will be tomorrow.

[fpistm]

Replace by #473