Developing a simple program - CCS V3.3

indiantuktuk said:

Data verification failed at address 0x3FFC0.

One suggestion I have is starting with the peripheral examples that come with the C/C++ header files and peripheral examples.  Only the specific run from XINTF example requires external memory on zone 7. 

http://focus.ti.com/docs/toolsw/folders/print/sprc097.html

But for now, back to the problem at hand:

If MP/MC is high at reset, then zone 7 of the XINTF is enabled.  If you have external RAM on your board connected to zone 7 you should be able to load the code without the error.  If you have RAM but sitll get the error make sure you have reset the device after changing MP/MC to a high signal.   This signal is only sampled at reset.   If you still can't load it then it could be an issue with the CCS memory map - the region may be defined as read only instead of read/write. 

If you don't have external RAM on your board then you won't be able to load this address through the external interface.  In this case set MP/MC low at reset, so the boot ROM is present.  In this case you cannot load the address 0x3FFFC0 since it is in the boot ROM so you can remap the section or tell the tools to not use it or load it.

In the linker command file (.cmd)  you can add the section type type = DSECT so .reset won't be loaded/used.

i.e.       .reset   : > RESET,   PAGE = 0, type = DSECT

This is described in the assembler users guide (SPRU513)

The same issue likely exists with the allocation of other sections - for example probably the VECTOR section.

-Lori

Thanks Lori for the info.

My custom board has IS61WV6416BLL-12TLI external memory.

Based on my understanding I would do the following

1)  Power ON the F2812 processor.

2)  Change MP/MC to a high signal.  At the moment it is neither high or low.

3)  Reset the F2812 processor by using the reset pin provided in the custom board

4)  Implement the peripheral examples as in SPRC097.  Which is not done till now.

5)  Start developing the Simple program as in the tutorial

6)  If the problem exists change the linker command as suggested by you.

I would do the above and report back.  If my understanding is wrong please suggest.

I appreciate your support.

Thx

Lori Heustess said:

One suggestion I have is starting with the peripheral examples that come with the C/C++ header files and peripheral examples.  Only the specific run from XINTF example requires external memory on zone 7

Dear Mr.Lori,

I was started with the peripheral examples from SPRC097.  The example I tried was run_from_xintf.

1)  Pulled the XMP_MCn to high and then powered on the board.

2)  configured for "boot to H0" operation.

3)  Started the code composer studio V3.3

4)  Removed the F2812.gel file

5)  File -> Open the Example_281xRunFromXintf.gel file

6)  Debug-->Connect.

7)  The disassembly window opened.  Its amazing the window is blank with    ---------   -----------  -----------    in all rows.

could you please review and let me know if I am missing any steps in the above.  Please suggest.

Thx & rdgs

Brandon,

Thanks for the guidance.  I retraced the steps as mentioned below.

1)  File -> Open the Example_281xRunFromXintf.gel file
2)  Debug-->Connect.

The GEL callback “OnReset(0)” for the target “F2812 XDS 510USB emulator/cpu_0” is still processing.  The function could contain an infinite loop.
Do you wish to pause the processing of this function and continue.
Yes/no

Selected : No

4)  The disassembly file is still blank.
5)  Opened GEL->DSP281x Run from XINTF->Load_and_build_project
6)  error observed :Encountered a problem loading file:\debug\Example_281xRunFromXintf.out.  Could not open the file
 7)  GEL->DSP281x Run from XINTF->Load_Code
Error observed "Function “F28112_XINTF_Map()” not found

Shut down the CCS and after a gap again started CCS to check if above errors would be repeated,

a)  loaded the Example_281xRunFromXintf.GEL file
b)  GEL->DSP281x Run from XINTF->Load_and_build_project

Amazing the program loaded and disassembly window opened.  The message at the bottom indicates Halted :S/W breakpoint.

Oh what a relief.

Noted the watch window; Debug---> closed; removed the GEL file and closed down CCS.
------------------------------------------
Query :-

1)  does the above example file indicate that I am able to access the external memory IS61WV6416BLL-12TLI. 
2)  Any more tests needs to be done to confirm the run from XINTF?.
3)  Why I got the errors in my first run?.
-------------------------------------------
Encouraged I went back to my earlier tutorial CCS - L1.

Started CCS, File --> load --> volume.gel
Project-> open-> volume1.pjt
Project--> rebuild all
The build shows 0 erros, 4 warnings, 0 remarks

Message 1
Data verification failed at address 0x8688.  Please verify target memory and memory map.
Pressed OK

Message 2
A data verification error occurred, file load failed.

Not sure what's the problem this time around.  could you please review and suggest.

Thx & rdgs

Hi,

Started CCS V3.3,File --> load --> volume.gel
Project-> open-> volume1.pjt
Project--> rebuild all
The build shows 0 erros, 4 warnings, 0 remarks

Message 1
Data verification failed at address 0x8688.  Please verify target memory and memory map.
Pressed OK

Message 2
A data verification error occurred, file load failed.

Not sure what's the problem.  Please suggest.

Settings :- XMP/MCn :- High & Boot to H0

Thx & rdgs

This particular region of the memory map is in the L0 SARAM Secure RAM which needs to be unlocked via the CSM.  It seems that perhaps the CSM is not being unlocked to allow access to this region.  Which GEL file are you using at this particular moment in time when running the Volume1 example?

Thanks Brandon for the response.

I am using the F2812.GEL file.

Please suggest.

Thx

BrandonAzbell said:

This particular region of the memory map is in the L0 SARAM Secure RAM which needs to be unlocked via the CSM.  It seems that perhaps the CSM is not being unlocked to allow access to this region.

As Brad suggests, the gel files that come with the peripheral examples have a CSM unlock function.  It will be automatically called when you reset the device if it is in the "OnReset()" function of the GEL file.

In addition to this you can keep the CSM unlocked during debug by opening a memory window to the password location (0x3F7FF8-0x3F7FFF on 281x).  When you first open the window it may read back all 0's.  Make sure to refresh the window and then it should read back 0xFFFF's if it is unlocked (assuming the password is not programmed).  If it reads 0x0000's then it is locked with a password.  By keeping the memory window open CCS will re-read the password and unlock the CSM after each reset.

Regards,

-Lori

Hi,

I am new to CCS and would wish to avoid CSM at the moment as I am scared that it could get locked. 

Is there a safer option to overcome this. 

I wonder if I could reset the board with the reset pin provided on the controller board as the first line in F2812.gel has the "OnReset()" function?. 

Does this help.

I have a query : For the first time when I POWERED ON the board the jumpers were are not placed does this lock the processor by any chance?.

thx

indiantuktuk said:

I am new to CCS and would wish to avoid CSM at the moment as I am scared that it could get locked. 

Is there a safer option to overcome this. 

The only thing that can lock the processor is to program the CSM password or not power the flash (ie leave VDD3VFL unconnected).   If you leave the power pin unconnected then the processor cannot read the password so it is by default locked.  Just power the pin and the processor will again be able to read the password.  

After a reset the CSM will "lock" but it isn't really secure if the password is erased (0xFFFF's) because it can be unlocked by simply reading the password locations.   Either the GEL file can do this read or you can open a memory window as I suggested.

indiantuktuk said:

I wonder if I could reset the board with the reset pin provided on the controller board as the first line in F2812.gel has the "OnReset()" function?. 

Either a reset with the pin or a debugger reset will "lock" the CSM.  But if the password is not programmed it will unlock as soon as code composer reads the password locations.  Does OnReset() call something like "CSMUnlock()"  ? (name might be slightly different).

indiantuktuk said:

I have a query : For the first time when I POWERED ON the board the jumpers were are not placed does this lock the processor by any chance?.

No.    

-Lori

Hi,

The function to unlock CSM is

 OnReset(int nErrorCode)

{

    Unlock_CSM();

}

I did the memory map (view ---> memory) at the locations suggested and the results are as given below

Location           Values

0x3F7FF8        0xFFFF           0xFFFF           0xFFFF

0x3F7FF9        0xFFFF           0xFFFF           0xFFFF

0x3F7FFA       0xFFFF           0xFFFF           0xFFFF

0x3F7FFB       0xFFFF           0xFFFF           0xFFFF          

0x3F7FFC       0xFFFF           0xFFFF           0xFFFF

0x3F7FFD       0xFFFF           0xFFFF           0xFFFF

0x3F7FFE       0xFFFF           0xFFFF           0x736C

0x3F7FFF       0xFFFF           0x736C            0x0B5C

I am not sure why the couple of values for 0x3F7FFE & 0x3F7FFF are not 0xFFFF.  What amazes me is the value 0x736C was 0x636C earlier in the day.  Am I missing something.  I did not give any password since I do not know how to do it at the moment.

Could you please review and suggest how to sort this.

thx

indiantuktuk said:

I did the memory map (view ---> memory) at the locations suggested and the results are as given below

 

Location           Values

 

0x3F7FF8        0xFFFF           0xFFFF           0xFFFF

0x3F7FF9        0xFFFF           0xFFFF           0xFFFF

0x3F7FFA       0xFFFF           0xFFFF           0xFFFF

0x3F7FFB       0xFFFF           0xFFFF           0xFFFF          

0x3F7FFC       0xFFFF           0xFFFF           0xFFFF

0x3F7FFD       0xFFFF           0xFFFF           0xFFFF

0x3F7FFE       0xFFFF           0xFFFF           0x736C

0x3F7FFF       0xFFFF           0x736C            0x0B5C

I'm not exactly sure about your data.  The password is 128 bits and the data has 3x that. 

If you are seeing non "FFFF" password locations and didn't do anything to unlock the part then I suspect these are really part of the memory space following the flash.  If any of the password locations were not "FFFF" then the CSM would stay locked and you would read all "0000" unless you knew the password and could unlock it.

Can you double check the address locations against the value again? 

-Lori

Thanks Lori for the response.

Could you confirm the method to check the values as given below.

View--> memory and the

bottom of the memory window there are couple of drop down options such Hex.32 bit C style; TI-style;64 bit float; 64 bit exponential float and many more.  However there is no 128 bit option. 

The other drop down has DATA, IO and program as option.

Could you please suggest how to look for the same.

thx

8304.reg_mem_map.doc

Hi Lori,

I checked the memory map again.  I think I need to ignore data from the second column since the number of columns keep increasing as I enlarge the window.  Please confirm my understanding.  If I ignore the additional columns the values would look as below.  The reading is with Format chosen as "16bit Hex - C style" and page field as "Data" selected.

Location           Values

0x3F7FF8        0xFFFF

0x3F7FF9        0xFFFF

0x3F7FFA       0xFFFF

0x3F7FFB       0xFFFF                 

0x3F7FFC       0xFFFF

0x3F7FFD       0xFFFF

0x3F7FFE       0xFFFF

0x3F7FFF       0xFFFF          

Please also find the image of memory map at one of the location.  Could you please review and suggest the next course of action.

thx