MPLAB ICD3调试PIC32MX Timer产生计数错误值

谢谢你的回答。但你的解释是不正确的。原因是，当我开始执行代码时，我总是等待一段时间，所以在激活BP之前，有1个PPS脉冲（中断）通过。因此，不管我把BP放在哪里，都应该有相同的数字。此外，我做了一个小的代码变体，我用一个端口位作为外部脉冲计数器的门信号。当在使用第49行的BP时检查外部脉冲计数器值时，得到正确的值3999935，但是TMR23（GPSCKReC.UILASTMTMRCOUNT）给出不正确的值大约为3992390。当在第47行使用BP时，结果在TMR23和外部计数器（+/-1计数）上是完全相同的。



      以下为原文

    Thanks for the answer. But your interpretation is NOT correct. The reason is that when I start execute the code I always wait a while so a number of 1 PPS pulses (interrupts) passes before activating the BP. Therefore there should be the same numbers despite where I put the BP.
Furthermore, I did a slight variant of the code where I used a Port bit to act as a gate signal to an external pulse counter. When checking the external pulse counter value when using BP at line 49 I got the correct value 3 999 935 but the TMR23 (GpsClkRec.uiLastTmrCount) give an incorrect value of about 3 992 390. When using the BP at line 47 the result was exactly the same on the TMR23 and the external counter (+/- one count).

NKKZZMAN：PIC32 MX150 F128D在TXCON控制寄存器中没有中断选项位上的冻结外围设备。



      以下为原文

    @NKurzman: The PIC32MX150F128D doesn’t have the freeze peripherals on break option bit in the TXCON Control Register.

调试计时器生成计数错误值，我做了一个新版本的调试尝试并修改了中断代码如下：然后我做了两个测试：A）其中一个BP在执行程序之前被设置，b）在程序执行过程中设置BP。无计时器计数错误：iValueEK〔0〕399 99 34 IValueEK〔2〕399 99 34 IValueEK〔3〕399 99 34 IValueKe[ 5 ] 399 99 33 IValueKe[ 6 ] 399 99 34 IValueKe[ 7 ] 399 99 34 IValueKe[ 8 ] 399 99 34 IValueKe[ 9 ] 399 99 34 IValueKe[10 ] 0B B，BP在执行期间设置，给出计时器计数。错误：IValueVek [ 1 ] 399 934 IValueEK [ 2 ] 399 99 34 IValueKe[ 3 ] 399 99 34 IValueKe[ 4 ] 399 99 34 IValueKe[ 5 ] 399 99 34 IValueKe[ 6 ] 3992200和lt；-错误值（在此时间帧中的BP集）IValueEK [7 ] 399 99 33 IValueKe[8 ] 399 99 33 IValueKe[9 ] 399 99 34 IValueKe[ 10 ] 0SO显然设置了BP网络。在执行过程中动态（动态）干扰导致错误结果的计时器计数。在某种程度上，BP临时设置禁止计时器的计数。如果这种行为被认为是一个错误或不可以辩论。然而，知道是好的。对我来说，这是相当混乱，我花了不必要的时间来解决它。问候，博城，SM6FIE



      以下为原文

    Debugging timer generates count error values
I did a new version of my debug attempts and modified the interrupt code as below:
// External INT2 INTERRUPT ROUTINE ***************************************************

// IsrInt2, Timer interrupt routine, priority #7

void __ISR(_EXTERNAL_1_VECTOR, IPL7SOFT) Int1Handler(void)

{

  T2CONbits.ON = False;

  Nop // Let TMR stabilize

  Nop();

  GpsClkRec.uiLastTmrCount = TMR2;

  iValueVek[iIdx] = TMR2; // Save TMR23 values into vector

  iIdx++;

  if (iIdx > 24) {

    iIdx = 0;

  }

  if (iIdx == 10) {

    Nop(); // Set breakpoint HERE

  }

  Nop();

  Nop();

  TMR2 = 0;

  Nop();

  Nop();

  T2CONbits.ON = True;

  Nop();

  Nop();

  mxINT1ClearIntFlag();

}
 
Then I did two tests: A) One where the BP was set before executing the program, B) Set BP during execution of the program.


Case A, BP Set before execution gives, as expected no timer counting errors:
iValueVek[0]    3442455                                      
iValueVek[1]    3999934                                      
iValueVek[2]    3999934                                      
iValueVek[3]    3999934                                      
iValueVek[4]    3999933                                      
iValueVek[5]    3999933                                      
iValueVek[6]    3999934                                      
iValueVek[7]    3999934                                      
iValueVek[8]    3999934                                      
iValueVek[9]    3999934                                      
iValueVek[10]  0   
                     
Case B, BP set during execution gives, gives timer counting error:                                            
iValueVek[0]    960532                                        
iValueVek[1]    3999934                                      
iValueVek[2]    3999934                                      
iValueVek[3]    3999934                                      
iValueVek[4]    3999934                                      
iValueVek[5]    3999934                                      
iValueVek[6]    3992200    <-- WRONG Value (BP set in this time frame)
iValueVek[7]    3999933                                      
iValueVek[8]    3999933                                      
iValueVek[9]    3999934                                      
iValueVek[10]  0                                                   
                                                                                 
So obviously setting a BP interactively (dynamically) during execution interferes with the Timer counting that leads to an erroneous result. In some way the setting of the BP temporary disables the counting for the timer. If this behavior is considered a bug or not can be debated. However it is good to know. For me it was rather confusing and I spent unnecessary time to figure it out.
 
Regards, Bo, SM6FIE

也许我应该补充一下，第一个值（IValueEvk〔0〕）不是一个错误值。它具有不同的值的原因很简单，因为它取决于在1个PPS周期中的哪个地方开始执行程序。/博城，SM6FIE



      以下为原文

    Perhaps I should add that the first value (iValueVek[0]) is not an error value. The reason it has a different value is simple because it is depending on where in the 1 PPS cycle I start the execution of the program.
 
/Bo, SM6FIE

湿透了床，想到了问题，然后我突然意识到“NKurzman”终究会成为真正的“网络上的顶级枪手”。启动了我的开发系统，检查了我的项目选项。哎哟，在“冻结外设”下有一个选项取消定时器1，2，3，4，5。取消所有定时器的一个中提琴，现在它没有更多的定时器干扰时，动态设置BP。现在我知道了两件事。1）如何动态地（交互地）调试定时器和“NKurzman”确实是一个“PRO”。古德奈特，博城，SM6FIE。



      以下为原文

    Wet to bed, thinking of the problem and then I suddenly realized that “NKurzman” eventually could be a real “Top Gun Pro on the Net”.  Fired up my development system and did check my project options. Whoops, under “Freeze Peripherals” there is an options to deselect Timers 1,2,3,4,5. Deselected all timers an viola, now it works No more timer interference when dynamically setting BP. So now I know two things. 1) How too dynamically (interactively) debug timers and that “NKurzman” indeed is a “Pro”.
 
Goodnight, Bo, SM6FIE