Where will the CRC stored for CC2541 ImageA and ImageB in the generated Hex file

Hi JXS,

Thank you for your reply.

I have been through the OAD Guideline for CC254x and I had really good insight on the OAD Code. But still I have a doubt regarding the checkDL() function.

1. As per the "OAD_for_CC254x.pdf" section "14.7 Completion of OAD Process" it is mentioned:

The OAD Target will then invalidate its own image and reset so that the BIM can run the new image in-place.

e.g. we are running ImgA, and doing a OAD for ImgB, then from the above statement, will ImgA invalidate itself ?

If the above understanding is correct, then why in checkDL() function, we are setting the crc[0] as 0xFFFF, and writing it to the flash @ 0x4000 (the linker script for ImgA is not changed).

2. As per our understanding, checkDL() function, always returns false from the below statement:

return (crc[0] == crc[1]);

then from where will the device get reset, after a successful OAD.

It would be really great if you could help us understand checkDL() function in detail.

Thanks,

Sunny

I would appreciate if someone can throw some light on this.

Hi JXS,

Thank you for the reply.

But still, we have some confusion for the checkDL() function.

In the below code block of checkDL() function:

if (crc[1] == 0xFFFF)
{
crc[1] = crcCalcDL();

#if defined FEATURE_OAD_BIM // If download image is made to run in-place, enable it here.
uint16 addr = OAD_IMG_D_PAGE * OAD_FLASH_PAGE_MULT + OAD_IMG_CRC_OSET / HAL_FLASH_WORD_SIZE;
crc[0] = 0xFFFF;
HalFlashWrite(addr, (uint8 *)crc, 1);
HalFlashRead(OAD_IMG_D_PAGE, OAD_IMG_CRC_OSET, (uint8 *)crc, sizeof(crc));
#endif
}

return (crc[0] == crc[1]);

Here, for ImgA, we have the FEATURE_OAD_BIM preprocessor flag defined in our project.

The crc[1] which we believe is the Shadow CRC, is 0xFFFF when the ImgB is generated. We are checking the 3rd and 4th Byte of the generated .bin file, to confirm the same.

So we go inside the "if (crc[1] == 0xFFFF)" condition and proceed.

Now we calculate the shadow CRC and store it in crc[1] via the statement crc[1] = crcCalcDL().

Also the below statement:

uint16 addr = OAD_IMG_D_PAGE * OAD_FLASH_PAGE_MULT + OAD_IMG_CRC_OSET / HAL_FLASH_WORD_SIZE;

is evaluated as:

uint16 addr = 8 * 512 + 0 / 4;

So, addr = 4096.
The actual flash address is 4096 * HAL_FLASH_WORD_SIZE = 0x4000 (the location of CHECKSUM=0x4000-0x4001, for ImgB linker script)

After that, the crc[0] is invalidated by assigned 0xFFFF to it.
It is then written to location addr = 4096 (i.e. @0x4000), along with crc[1] having the newly calculated value via the below statement:

HalFlashWrite(addr, (uint8 *)crc, 1);

Then a HalFlashRead() is done from the OAD_IMG_D_PAGE (which is OAD_IMG_B_PAGE = 8).
The OAD_IMG_D_PAGE corresponds to the Flash Location 8 * HAL_FLASH_PAGE_SIZE = 0x4000, where we get crc[0] and crc[1], updated by HalFlashWrite() function.

Since crc[0] has already been made to 0xFFFF, and crc[1] has the calculated CRC updated by crcCalcDL() function, these two values will not match.

So, return (crc[0] == crc[1]) will return FALSE, and we will not get a RESET.

Please do let us know if we have misunderstood anything.

Thanking you,

Sunny

Hi JXS,

We tried to step through the debugger.

We are able to put breakpoints in the oadImgBlockWrite() function at the below line:

if (checkDL())

But we are not able to step-in the checkDL() function.

Also we tried to put break point in the checkDL() function, but we are not able to put break point in the checkDL() function.

We are able to point break points in other functions like crcCalcDL() function.

If possible, can you please try to put break point in the crcCalcDL() function, and help us out regarding the same.

Thanks,

Sunny