RE: Persistent storage manager writing problem for one block

asma — Thu, 12 Feb 2015 07:24:24 GMT

Thanks Bill. it works.i have posted another question at this link related to pstorage update. can you please check it. https://devzone.nordicsemi.com/question/29547/pstorage-update-problem-call-back-is-not-getting-called/

RE: Persistent storage manager writing problem for one block

Bill Siever — Mon, 09 Feb 2015 13:05:15 GMT

Hi (updated post. I replaced my earlier answer with the following)

It looks like both the pstorage_clear and the pstorage_store are queued. You should wait for each to complete before proceeding to the next since they are working on the same block. That is, do:

    pstorage_block_identifier_get(&handle, 0, &block_0_handle);

    pstorage_wait_handle = block_0_handle.block_id;        // Specify which pstorage handle to wait for

    pstorage_wait_flag = 1;                                // Set the wait flag.
    pstorage_clear(&block_0_handle, 16);                   // Clear 16 bytes
    while(pstorage_wait_flag)  power_manage();             // Wait for clear

    pstorage_wait_flag = 1;                                // Set the wait flag. 
    pstorage_store(&block_0_handle, source_data_0, 16, 0); // Write to flash
    while(pstorage_wait_flag) power_manage();              // Wait for store

    pstorage_load(dest_data_0, &block_0_handle, 16, 0);    // Read from flash,

You should probably also change:

static uint8_t pstorage_wait_flag = 0;

to:

 static volatile uint8_t pstorage_wait_flag = 0;

In your earlier version:

clear was queued
write was queued
you wait for a command on the block to complete. You get an acknowledgement when the CLEAR completes.
you do the load, which loads the cleared values. The write is still in the queue.

I believe when you used the additional breakpoint the extra delay allowed the write to complete.