Gather enable from background C program crash PowerPMAC

[george.kontogiorgos]

Hello,

I am unsuccessful trying to trigger data gather from a C background program. I wrote the following program (capp1.c) to debug the problem:

/*For more information see notes.txt in the Documentation folder */
#include <gplib.h>

#define _PPScriptMode_		// for enum mode, replace this with #define _EnumMode_
#include "../../Include/pp_proj.h"

int main(void)
{
	InitLibrary();  // Required for accessing Power PMAC library

	pshm->Gather.Enable=2;
	
	CloseLibrary();
	return 0;
}

Then I built, downloaded and save the project. After that I restarted the controller with $$$ on IDE terminal and run my background program via SSH connection (./capp1.out). The controller then suddenly  stops responding on IDE and the controller`s watchdog LED lights up.

When I first do some gather via IDE terminal (Configure everhing, Gather.Enable=2 and Gather.Enable=0) , the background program starts working (it sets correctly the enable register and gather the desired data).

I have the hypothesis that the commands issued from terminal do some initialization on gather feature but I do not know how to track it. This idea of initialization comes to me when I realized that when I do Gather.Enable=2 on C program the Gather.Index does not automatically erase as it happens when the same command is issued from IDE terminal. So I manually erase the buffer and reset index as follows:

/*For more information see notes.txt in the Documentation folder */
#include <gplib.h>

#define _PPScriptMode_		// for enum mode, replace this with #define _EnumMode_
#include "../../Include/pp_proj.h"

int main(void)
{
	InitLibrary();  // Required for accessing Power PMAC library

	//Put your code here

  	// Disable any running gather
	pshm->Gather.Enable=0;
  
	// Set sample counter to zero
	pshm->Gather.Samples=0;

  	// Erase buffer with gpascii command (Index points to buffer's position 0 after clear)
	GetResponse("clear gather", gpascii_response, 100, 1);

	// Start gather
	pshm->Gather.Enable=2;

	CloseLibrary();
	return 0;
}

 Another point is that I could successfully enable the gather at first time after reset by using:

GetResponse("Gather.Enable=2", gpascii_response, 100, 1);

Instead of:

pshm->Gather.Enable=2;

I  tryed on two different controller's architecture:

Type: POWER PMAC BRICK

CPU: PowerPC, AMP86xxx

Firmware: 2.5.4.0

And

Type: POWER PMAC BRICK

CPU: arm,LS1021A

Firmware: 2.6.0.0

Thanks in advance,

George

Edited February 4, 2022 by george.kontogiorgos

[Omron Forums Support]

I can start gathering by running a C program with the following line after using the plot tool to initialize settings.

pshm->Gather.Enable=2;

I have a feeling the issue lies in the setup of the "Gather." variables usually configured by the plot tool. Can you try using the tool to set these up as a test? You should be able to close it before your actual gathering and keep the settings.

[george.kontogiorgos]

Hi Eric,

As your suggestion I did some tests as follows.

Test 1

I issued the folowing commands (on that order) on the terminal:

$$$
Gather.Items=1
Gather.Addr[0]=Sys.ServoCount.a
Gather.Period=1
Gather.MaxSamples=100000

And then I ran my C Background Program with the following content

/*For more information see notes.txt in the Documentation folder */
#include <gplib.h>   

#define _PPScriptMode_		// for enum mode, replace this with #define _EnumMode_
#include "../../Include/pp_proj.h"

int main(void)
{
	InitLibrary();  // Required for accessing Power PMAC library

	pshm->Gather.Enable=2;				//Put your code here

	CloseLibrary();
	return 0;
}

And then the PowerPMAC crashed (the red arrow 😞

Test 2

I reseted the PowerPMAC with $$$, opened the plot window and wait it initialize. After that I closed the plot window and set my gather parameters (Just to be the same as the last test since plot window sets Gather.Period and Gather.MaxSamples when it opens).

Then I ran my C Background Program (the same as Test 1) and everything goes fine.

Conclusions from Test 1 and Test 2

Opening the plot window before using pshm->Gather.Enable=2 on the C Backgroud Program works fine (the PowerPMAC does not crash by enabling gather on a C program). I suspected the Gather.Enable=1 followed by a Gather.Enable=0 that is on plot window Gather Settings area. That was difference from my configuration from terminal and configuration from Plot window. This lead me to Test 3.

Test 3

I did exactly the same as Test 1 but adding Gather.Enable=1 followed by a Gather.Enable=0 on Initial gather configuration on IDE Terminal:

$$$
Gather.Items=1
Gather.Addr[0]=Sys.ServoCount.a
Gather.Period=1
Gather.MaxSamples=100000
Gather.Enable=1
Gather.Enable=0

Then I ran the  C Background Program and the PowerPMAC did not crash.

Conclusions from Test 3

PowerPMAC Test 3 lead me to the conclusion that setting Gather.Enable=1 or Gather.Enable=2 (value 2 I tested earlier on my first post of this thread) run some initialization on Gather. structure. So I did Test 4 to know if setting Gather.Enable=0 and run the C Background Program crash the PowerPMAC.

Test 4

I reseted the PowerPMAC controller and then issued the following commands into the IDE Terminal

Gather.Items=1
Gather.Addr[0]=Sys.ServoCount.a
Gather.Period=1
Gather.MaxSamples=100000
Gather.Enable=0

Then I ran the  C Background Program and the PowerPMAC crashed.

Conclusions from Test 4

Setting the same value (Gather.Enable=0) to register and running the C program crash the controller. The last test, the test 5, I just tryed the last option for enabling gather (Gather.Enable=3).

Test 5

I reseted the PowerPMAC controller and then issued the following commands into the IDE Terminal

Gather.Items=1
Gather.Addr[0]=Sys.ServoCount.a
Gather.Period=1
Gather.MaxSamples=100000
Gather.Enable=3
Gather.Enable=0 // Stop rotary gather

Then I ran the  C Background Program and the PowerPMAC did not crashed.

Conclusion

Changing Gather.Enable value from 0 to something different from IDE terminal may initialize something inner the controller that allow pshm-Gather.Enable works fine on a C Background Program. This is what Plot window does when it initializes (we can see on Gather registers changes). Obviously I do not know in fact what the Plot window does and maybe it do some inner controller initialization.

 

Thanks,

George

Edited February 11, 2022 by george.kontogiorgos

[george.kontogiorgos]

15 minutes ago, george.kontogiorgos said:

And then the PowerPMAC crashed (the red arrow 😞

The sad Emoji was a mistake made by a set of special characters toghether, I'm sorry about it.

[Omron Forums Support]

I believe that the act of setting Gather.Enable from the script environment applies the settings in Gather. variables. Setting Gather.Enable to 2 or 3 from C code after changing these settings, but before changing the value in the script environment, will crash PMAC.

The "Command" command can be used from C to run code in the script environment after changing these settings.

Command("Gather.Enable=2");

 

[george.kontogiorgos]

Hi Eric,

16 hours ago, Eric Hotchkiss said:

The "Command" command can be used from C to run code in the script environment after changing these settings.

Command("Gather.Enable=2");

Thanks for the tip. I was using to circumvent the problem on my API:

GetResponse("Gather.PhaseEnable=2", gpascii_response, 100, 1);

But maybe your command is faster than mine since GetResponse has to manipulate the return from the command.

I was thinking in just initialize Gather. structure on the begining of my program with something like that

// Initialize Gather. struture
Command("Gather.Enable=0");
Command("Gather.Enable=1");
Command("Gather.Enable=0");

// Use gather structure directly
pshm->Gather.Enable=2;

Thanks,

George

[george.kontogiorgos]

Hello,

On 2/14/2022 at 8:31 PM, Eric Hotchkiss said:

The "Command" command can be used from C to run code in the script environment after changing these settings.

Command("Gather.Enable=2");

I would like to use it on CaptCompISR routine, is it possible?  What is the best way of enabling gather from interruption?

[Omron Forums Support]

The CaptCompISR should be as short as is possible. The best method would be to set a flag (such as a script global variable), that can be reacted to by a background Script/C PLC.

Here is an example Script PLC.

GLOBAL MyGatherFlag
OPEN PLC StartGatherPLC
LOCAL MyGatherLatch
IF(MyGatherFlag == 1)
{
    IF(MyGatherLatch == 0)
    {
        Gather.Enable = 2
        MyGatherLatch = 1
    }
}
ELSE
{
    IF(MyGatherLatch == 1)
    {
        Gather.Enable = 0
        MyGatherLatch = 0
    }
}
CLOSE

 

[george.kontogiorgos]

Hi Eric!

Thanks for replying me.

12 hours ago, Eric Hotchkiss said:

The CaptCompISR should be as short as is possible.

Yes, I do know that this is a good pratice, thanks for the advice. I think it would not be a problem if the CaptCompISR just enable the gather.

12 hours ago, Eric Hotchkiss said:

The best method would be to set a flag (such as a script global variable), that can be reacted to by a background Script/C PLC.

I had this idea but I was trying to avoid any delay between the interruption and enabling the gather.