Relationship between MULTI_STACK and KERNEL_TYPE?

[srl100]

I have an OIL file that includes the following OS section:

Code: Select all

	OS myOs {
		EE_OPT = "DEBUG";
		EE_OPT = "NODEPS";	//Don't do a separate dependency file pass
		EE_OPT = "NOSRC";	//Don't create intermediate .src files

		CPU_DATA = PIC30 {
			APP_SRC = "code.c";
			MULTI_STACK = FALSE;
			ICD2 = TRUE;
		};

		MCU_DATA = PIC30 {
			MODEL = PIC33FJ256MC710;
		};
		
		BOARD_DATA = EE_FLEX {
			USELEDS = TRUE;
		};
		
		KERNEL_TYPE = ECC1;
		
		STATUS = EXTENDED;
	};

I get the following errors when I build my application:

Code: Select all

.\libee.a(ee_evwait.o)(.text+0x84): In function `EE_oo_WaitEvent':
C:\Evidence\evidence\eclipse\plugins\COAAB5~1.201\ee_base\pkg\kernel\oo\src\ee_evwait.c:210: undefined reference to `EE_hal_stkchange'
.\libee.a(ee_evwait.o)(.text+0xaa):C:\Evidence\evidence\eclipse\plugins\COAAB5~1.201\ee_base\pkg\kernel\oo\src\ee_evwait.c:224: undefined reference to `EE_hal_stkchange'
.\libee.a(ee_evset.o)(.text+0xa6): In function `EE_oo_SetEvent':
C:\Evidence\evidence\eclipse\plugins\COAAB5~1.201\ee_base\pkg\kernel\oo\src\ee_evset.c:240: undefined reference to `EE_hal_stkchange'
.\libee.a(ee_ulockres.o)(.text+0x96): In function `EE_oo_ReleaseResource':
C:\Evidence\evidence\eclipse\plugins\COAAB5~1.201\ee_base\pkg\kernel\oo\src\ee_ulockres.c:238: undefined reference to `EE_hal_stkchange'
.\libee.a(ee_activate.o)(.text+0x96): In function `EE_oo_ActivateTask':
C:\Evidence\evidence\eclipse\plugins\COAAB5~1.201\ee_base\pkg\kernel\oo\src\ee_activate.c:208: undefined reference to `EE_hal_stkchange'
make: *** [pic30.cof] Error 1

Is there a relationship between the KERNEL_TYPE and MULTI_STACK that I am missing? What are the legal combinations for the OSEK conformance classes?

[paolo.gai]

BCC1 and BCC2 work both with single and multistack.

ECC1 and ECC2 work only with MULTISTACK, becasue these conformance classes have the support for the blocking primitive WaitEvent!

PJ

[srl100]

Thanks for the info - RT-Druid wasn't complaining, so I thought that it was a valid combination. I'll update my OIL file and try again!