ASSERTION FAIL [err == 0] and Error writing variable 1 (err: -116)

[Zihao]

Now I am using the https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/samples/bluetooth/direction_finding_central/README.html and https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/samples/bluetooth/direction_finding_peripheral/README.html for a direction finding project.
I am using the nRF52833DK and an antenna array designed by Nordic. The SDK I used is nRF Connect SDK v2.2.0.

-------------------------------------------------------------------------------------------------------

Let's just talk about the RX(central) device in this case.

I need a sample function for the nRF52833DK. The function should reboot the DK 15 seconds after scanning starts if no connection is established. Before rebooting the DK, I need to update a variable value in NVS.

The code should look like:

#define STORAGE_PARTITION_LABEL storage
#define NVS_ID_COUNT 1
#define NVS_ID_AOA_AOD_SETTING 2
#define NVS_ID_INIT_ALL_SETTINGS 3

static struct nvs_fs fs;

static int my_nvs_init(void)
{
    struct flash_pages_info info;
    int err;

    const struct device *flash_dev = DEVICE_DT_GET(DT_CHOSEN(zephyr_flash_controller));

    if (!device_is_ready(flash_dev)) {
        printk("Flash device %s is not ready!\n", flash_dev->name);
        return -ENODEV;
    }

    err = flash_get_page_info_by_offs(flash_dev, FLASH_AREA_OFFSET(STORAGE_PARTITION_LABEL), &info);
    if (err) {
        printk("Unable to get page info (err: %d)\n", err);
        return err;
    }

    fs.offset = FLASH_AREA_OFFSET(STORAGE_PARTITION_LABEL);
    fs.sector_size = info.size;
    fs.sector_count = 3;

    err = nvs_init(&fs, flash_dev->name);
    if (err) {
        printk("Flash initialization failed\n");
        return err;
    }

    return 0;
}

bool AOA_AOD_SETTING_reinit(void){
	printk("AOA_AOD_SETTING reinit...\n");
	AOA_AOD_SETTING = AOA_AOD_START;
	int err = nvs_write(&fs, NVS_ID_AOA_AOD_SETTING, &AOA_AOD_SETTING, len);
	if (err < 0) {
        printk("Error writing variable (err: %d)\n", err);
        return false;
    }
	else{
		printk("AOA_AOD_SETTING reinit success.\n");
	}
	return true;
}

static void scan_timeout(struct k_timer *dummy)
{
    printk("Scan timeout, no connection established. reinit...\n");
	if(AOA_AOD_SETTING_reinit()){

		sys_reboot(SYS_REBOOT_MODEL);
	}
}

void main(void){
    	err = my_nvs_init();
		if (err) {
			printk("Cannot initialize NVS!\n");
			return err;
		}
}

However, I encountered the following error: Error writing variable 1 (err: -116).

I understand that error -116 indicates a certificate/key issue through this issue. However, I have not configured any certificates or keys. Therefore, I suspect this issue might be related to the scanning process. The scan_timeout callback function is triggered while the central device is still scanning. Interestingly, I can successfully write the variable when the central device is connected.

To address this, I added the following:

bool AOA_AOD_SETTING_reinit(void){
	printk("AOA_AOD_SETTING reinit...\n");
	AOA_AOD_SETTING = AOA_AOD_START;
	int err = nvs_write(&fs, NVS_ID_AOA_AOD_SETTING, &AOA_AOD_SETTING, len);
	if (err < 0) {
        printk("Error writing variable 1 (err: %d)\n", err);
        return false;
    }
	else{
		printk("AOA_AOD_SETTING reinit success.\n");
	}
	return true;
}

static void scan_timeout(struct k_timer *dummy)
{
    printk("Scan timeout, no connection established. reinit...\n");
	int err = bt_le_scan_stop();
	k_sleep(K_MSEC(100));
	if (err) {
		printk("Stop LE scan failed (err %d)\n", err);
	}

	if(AOA_AOD_SETTING_reinit()){
		k_sleep(K_MSEC(100));
		sys_reboot(SYS_REBOOT_MODEL);
	}
}

However, I got:

Scan timeout, no connection established. reinit...
ASSERTION FAIL [err == 0] @ WEST_TOPDIR/zephyr/subsys/bluetooth/host/hci_core.c:329
	k_sem_take failed with err -11
[00:00:15.299,468] <err> os: esf_dump: r0/a1:  0x00000003  r1/a2:  0x00000000  r2/a3:  0x00000001
[00:00:15.299,499] <err> os: esf_dump: r3/a4:  0x00020bf5 r12/ip:  0x00000000 r14/lr:  0x00007593
[00:00:15.299,530] <err> os: esf_dump:  xpsr:  0x41000021
[00:00:15.299,530] <err> os: esf_dump: Faulting instruction address (r15/pc): 0x0000759e
[00:00:15.299,560] <err> os: z_fatal_error: >>> ZEPHYR FATAL ERROR 3: Kernel oops on CPU 0
[00:00:15.299,591] <err> os: z_fatal_error: Fault during interrupt handling

[00:00:15.299,621] <err> os: z_fatal_error: Current thread: 0x200016b0 (unknown)
[00:00:15.601,745] <err> fatal_error: k_sys_fatal_error_handler: Resetting system
*** Booting Zephyr OS build v3.2.99-ncs1 ***

The cases about error -11 found in Devzone do not address my issue.. For example.

error -11 (EAGAIN) means "No more contexts." So, why can't I stop the scanning? I have added some k_sleep calls, but they didn't help.

Here are the steps I have tried. The main issue remains: why can't I update a variable in NVS?

Thank you! Any information would be appreciated.

[Elfving]

Hello,

I understand that error -116 indicates a certificate/key issue through this issue. However, I have not configured any certificates or keys. Therefore, I suspect this issue might be related to the scanning process. The scan_timeout callback function is triggered while the central device is still scanning. Interestingly, I can successfully write the variable when the central device is connected.

You can find what all of these error codes mean in errno.h. For instance here. -116 allways means means "Connection timed out". In the context of the issue you linked to, that typically stems from certificate

or key issues, though I do not think that is the issue here. 

My immediate idea on this issue is that it is related to the time-sensitive nature of this procedure. You should make sure that you do other things (ie. write to NVS) in a workqueue so that it won't hinder the connection.

Regards,

Elfving

[Zihao]

Sure, thank you. That's probably because interrupt operations are not recommended to occur simultaneously with NVS write operations. I think there might be a proper solution that allows NCS to handle this time-sensitive procedure well. After directly removing the assert statement, my program seems to be running normally.