FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

Jared — Fri, 03 Mar 2023 09:13:55 GMT

Hi,

[quote user="ellamoss"]This led me back to my event handler, where I APP_ERROR_CHECK(p_fds_evt->result). This result is 0x8609, FDS_ERR_NOT_FOUND.[/quote]

And the event ID fds_evt_id_t id for that p_fds_evt was a FDS_EVT_UPDATE?

[quote user="ellamoss"]Do you know why fds_record_update would lead to this error? Immediately before calling flash_update, I verify that the key exists by calling my flash_read function, which returns 0 (meaning the file exists - right?).[/quote]

If fds_record_open() returns 0 then the record has been found in the fds module.

Can you try do it in one function as in the fds example? I suspect that the there is some scope issue in that the descriptor isn't set correctly between the two function calls.

regards

Jared

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

ellamoss — Thu, 02 Mar 2023 22:40:47 GMT

That makes sense. When I open it up in the debugger, I put a print statement after that fds_record_update call but I don't hit it - it jumps to app_error_weak.c, and that's where I see the NRF_FAULT_ID_SDK_ERROR code.

So as far as I can tell, this is triggered by write_enqueue in fds.c. I put in a bunch of print statements (pausing with the debugger always yields an exception, as you mentioned), and found the culprit to be the queue_start() command in write_enqueue, in fds.c.

So I stepped into that function, used print statements again to determine it came from event_send(&evt).

This led me back to my event handler, where I APP_ERROR_CHECK(p_fds_evt->result). This result is 0x8609, FDS_ERR_NOT_FOUND.

Do you know why fds_record_update would lead to this error? Immediately before calling flash_update, I verify that the key exists by calling my flash_read function, which returns 0 (meaning the file exists - right?).

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

Jared — Thu, 02 Mar 2023 07:56:06 GMT

This is the return values from fds_record_update:

none of them correspond to NRF_FAULT_ID_SDK_ERROR.

Note that trying to halt the CPU by pausing it with the debugger will always result in a exception when it's resumed, if the application is using the softdevice.

Can you share a screenshot that shows the return value from fds_record_update()?

regards

Jared

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

ellamoss — Wed, 01 Mar 2023 18:48:03 GMT

Okay, I think I'm getting closer. But now when I try to do a flash_update (ie, fds_record_update in the first code block), it errors out to the app_error_fault handler. I'm getting an error code 0x4001, which seems to correspond to "NRF_FAULT_ID_SDK_ERROR". Do you know why this might be happening, or how I can troubleshoot it?

Thanks,
Ella

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

Jared — Tue, 28 Feb 2023 09:43:46 GMT

Hi,

Yes, that is exactly how I would implement it.

regards

Jared

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

ellamoss — Mon, 27 Feb 2023 22:43:16 GMT

Hi,

The error happens during calls to flash_read; from there, it seems to be coming from a call to fds_record_open() (line 137 of my first code block). So do I need to write code to wait on reads until I get the FDS_EVT_WRITE event? Maybe a volatile flag that I wait for being set before doing any reads?

Thanks,
Ella

RE: FDS with SoftDevice reporting FDS_ERR_CRC_CHECK_FAILED on read

Jared — Mon, 27 Feb 2023 14:15:26 GMT

Hi,

Exactly which FDS function is returning FDS_ERR_CRC_CHECK_FAILED?

Typical reason for getting FDS_ERR_CRC_CHECK_FAILED is trying to read the record before the fds module has completed writing it and returned with FDS_EVT_WRITE.

regards

Jared