• State Accepted Answer
  • +1 person also asked this people also asked this
  • Locked Locked
  • Replies 3 replies
  • Subscribers 351 subscribers
  • Views 7527 views
  • Users 0 members are here
2025 survey

We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.

Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Data Abort on read, although write can be executed without any abort.

r4c00n

I am a begginer to arm and I have a problem with understanding Data Abort that I get when reading from memory (ATCM)

when I read from atcm. next pc jumps to 0x10, so Its Data Abort.

However, If I write to the same address. then I can read from it without data abort.

I disabled MPU using code from Cortex R TRM.

Now I do not know the reason of this behaviour.

1. I thought this was platform-related , but as It's TCM, and SOC TRM referred to Cortex R TRM, I started looking for a clue in cortex R - related resources. (please forgive me if it appears to be not cortex-r-related)

2. I am missing knowledge about what can cause this error, especially, that the memory is finally accessible (after write) and that I disabled the MPU.

I am using one core of Cortex R5, in split mode. ATCM is 64KB, and example address that caouses this is 0x2a08 (As far as I know, Its in proper range, ATCM starts in 0x0)

There are many more addresses where I get the same result, but most of them are ok. (I think the ones ending with 8 make this problem)

The address is used for stack, pop instruction is used to get data from the address to r0, but I think this should not matter (?)

Please help me find a reason of this behaviour.

Parents
  • +1

    Cortex R5 profile has a Data fault status register and Auxiliary data fault status register (refer section 4.2 from ARM DDI 0460D). Dump them in the your D-abort handler to get more information about the Data abort that you are seeing. Also, Is ECC check enabled on ATCM ? if yes, then are you also initializing it by pre-loading the TCM (refer preloading TCMs section from ARM DDI 0460D) ? Because, ECC check is done on the read operation and are ECC calculation is done on the write operation. A read operation from ECC protected TCM with an uninitialized ECC code will always give you Data abort unless you have ECC correction enabled and also If it's a non-fatal ecc error. 

Reply
  • +1

    Cortex R5 profile has a Data fault status register and Auxiliary data fault status register (refer section 4.2 from ARM DDI 0460D). Dump them in the your D-abort handler to get more information about the Data abort that you are seeing. Also, Is ECC check enabled on ATCM ? if yes, then are you also initializing it by pre-loading the TCM (refer preloading TCMs section from ARM DDI 0460D) ? Because, ECC check is done on the read operation and are ECC calculation is done on the write operation. A read operation from ECC protected TCM with an uninitialized ECC code will always give you Data abort unless you have ECC correction enabled and also If it's a non-fatal ecc error. 

Children
No data