CCS/TMS320F28379D: Memory Allocation display and memory allocation configured in cmd file.

Hi,

Looks like the ebss section is small enough to fit in LS5 and GS0. There are no variables allocated in the other GSRAMs, hence you see them empty.

If needed, you can combine them and have it as a single memory block.

For example,

RAMGS0 : origin = 0x00C000, length = 0x001000
RAMGS1 : origin = 0x00D000, length = 0x001000

This can be combined as a single block

RAMGS01 : origin = 0x00C000, length = 0x002000

You can combine n number of RAMs as long as the addresses are contiguous.

By default, we keep them as separate memory blocks, in case application wants to use them for different purposes.

Regards,

Veena

Veena,

Thank you for your answer,

I resolved my issue in the .cmd file, 

I have another issue:

I am using freeRTOS, the PC jumps to the interrupt void ILLEGAL_ISR(void) after executing the _portRESTORE_FIRST_CONTEXT in porASM.asm.

_portRESTORE_FIRST_CONTEXT
; Restore stack pointer from new task control block.
MOVL XAR0, #_pxCurrentTCB
MOVL XAR0, *XAR0
MOVL XAR0, *XAR0
MOV @SP, AR0

; Restore XAR4 and RPC from saved task stack.
; and return to main task function.
; SP should be set to stack start plus 2 before LRETR.
SUBB SP, #28
POP XAR4
SUBB SP, #14
POP RPC
SUBB SP, #10
LRETR

I am not very familiar with TI assembly code but, I could dig into details if you could help me.

I am using the TMS320f28379D launch xl for the tests,

I modified the cmd file, you could find it in the project, you don't need to link anything, all resources are in the project.

Before this modification, the RTOS runs well.

Thank you in advance,

S.Tarik 

univcadiayyad-my.sharepoint.com/.../EeO_RCHXm6xDlJ6KQtT5mcsBVOcDf7YCL03T66XUCLjRQQ

Hi S.Tarik,

So, the issue happened after you updated the cmd file. Is that right?

Can you check the value of RPC register just before LRETR instruction. This should hold the address of the first task.

Regards,

Veena

Hi Veena,

Veena Kamath said:

So, the issue happened after you updated the cmd file. Is that right?

Yes,

The RPC = 0x0 befor LRETR!

Hi S.Tarik,

Looks like your task stack is not updated as expected.

I hope this is on the CPU1 side. The GSRAMs are shared between CPU1 and CPU2 and I believe the ownership of all the GSRAMs are assigned properly. By default GSRAMs are owned by CPU1.

Regards,

Veena 

Veena Kamath said:

Looks like your task stack is not updated as expected.

Venna,

How you could pretend that this is a Stack issue?

I change the .ess section from GSRAMs to LSRAMs and it seems to work well. One section was not enough to fit, I concatenate many LSRAMs to get proper compiled and it run well.

I have only the CPU1 activated, I could understand why this couldn't work with GSRAM?

Thanks

Hi,

The task stacks typically contains the return address of the task  which will be used by the task restore routine. 

I am not sure why allocating them in GSRAMs is not working. Can you create global variable, allocate to GSRAM, initialize the  variable in main and check if you can see the value in the memory?

Also, please configure the GSRAM ownership to CPU1 incase not done. 

Regards,.

Veena

Hello Veena,

I have to use the LSRAM for .ess variables and it works well,

sorry for the delayed replay, I don't have any idea why this is not working using GSRAM

Thank you