FDS Implementation

RE: FDS Implementation

Somesh Burkule — Tue, 19 May 2020 06:33:47 GMT

Thanks Jared for help.

I have implemented successfully fds operations with 2 file id's and 6 records in nrf52. not much difficult as much I was thinking.

RE: FDS Implementation

Jared — Thu, 07 May 2020 06:58:24 GMT

Hi,

[quote user="Somesh Burkule"]what this queue useful for? how we can define this queue size depend on records?[/quote]

From documentation:

The size of the internal queue of FDS operations. Increase the size if there are many users, or if your application will queue many operations at a time without waiting for the previous operations to complete. In general, you should increase the queue size if you frequently receive FDS_ERR_NO_SPACE_IN_QUEUES errors.

The FDS module is an asynchronous module, and uses a queue for scheduling operations. The size of the queue determines therefore how many write/read etc operations you can schedule at once.

regards

Jared

RE: FDS Implementation

Somesh Burkule — Thu, 07 May 2020 05:16:34 GMT

Hi Jared,

I am using one file id and 5 record key's. while writing to file, giving error to 5th record. then I increased fds internal queue size from 4 to 8. then it worked properly.

what this queue useful for? how we can define this queue size depend on records?

I need to add one more file id and record key. what value should I use for queue?

RE: FDS Implementation

Jared — Tue, 05 May 2020 13:52:48 GMT

From documentation:

fds_record_find (): This function finds the first record in a file that has the given record key. To search for the next record with the same key in the file, call the function again and supply the same fds_find_token_t structure to resume searching from the last record that was found.

RE: FDS Implementation

Somesh Burkule — Tue, 05 May 2020 13:48:53 GMT

What if I use file I'd and same record key?

RE: FDS Implementation

Jared — Tue, 05 May 2020 12:47:53 GMT

You don't need to pass the Record id when you want to access a record since this is something that will be handled by the module. Only Record Key and File Id is necessary.

RE: FDS Implementation

Somesh Burkule — Tue, 05 May 2020 12:44:30 GMT

[quote userid="73165" url="~/f/nordic-q-a/60947/fds-implementation/247848"]You use record KEY and file ID when you want to access a Record. [/quote]

Thanks Jared.

So we can use fds without using Record Id?

RE: FDS Implementation

Jared — Mon, 04 May 2020 12:08:44 GMT

Hi.

I actually do not agree, I think the specification explains is pretty well:

Record key

16 bits

Key that can be used to find the record. The value FDS_RECORD_KEY_DIRTY (0x0000) is reserved by the system to flag records that have been invalidated. See Restrictions on keys and IDs for further restrictions.

File ID

16 bits

ID of the file that the record is associated with. The value FDS_FILE_ID_INVALID (0xFFFF) is used by the system to identify records that have not been written correctly. See Restrictions on keys and IDs for further restrictions.

Record ID

32 bits

Unique identifier of the record.

So a specific Record will be stored in a file with a file ID. A file can have several records with the same record KEY but different record ID. Record ID is therefore unique, record KEY and file ID however is not. Which means that you can have several records stored with the same file ID and record KEY but with a unique record ID. You use record KEY and file ID when you want to access a Record.

Documentation.

regards

Jared