Is it possible to queue sd_ble_gatts_hvx calls ?

RE: Is it possible to queue sd_ble_gatts_hvx calls ?

x_arrange — Sat, 18 Jan 2020 12:06:45 GMT

To sum up, here is the problem:

No matter what I send (indication/notification) peer receives same data each time but data length differs.

Look slike sd_ble_gatts_hvx just ignores content of params.p_data

RE: Is it possible to queue sd_ble_gatts_hvx calls ?

x_arrange — Fri, 17 Jan 2020 13:37:22 GMT

Thanks, that's what I was afraid of..

I thought I could use notification instead but apparently I can't.

Event though I call sd_ble_gatts_value_set before sd_ble_gatts_hvx each time, but peer (GATT client) receive the same value eache time

RE: Is it possible to queue sd_ble_gatts_hvx calls ?

Hung Bui — Fri, 17 Jan 2020 13:21:56 GMT

I see. You would need to wait for the confirmation of the peer after you send the first indication (BLE_GATTS_EVT_HVC event) before you can send the next indication. See here.

RE: Is it possible to queue sd_ble_gatts_hvx calls ?

x_arrange — Fri, 17 Jan 2020 13:05:41 GMT

Hi,

I send write command from GATT client (Android app) to GATT server (nRF52).

Currently I can send only one indication. When I send the next one right after the first one I get NRF_ERROR_BUSY

I have only one characteristic so there is no background work.

RE: Is it possible to queue sd_ble_gatts_hvx calls ?

Hung Bui — Fri, 17 Jan 2020 12:43:16 GMT

Hi,

Could you let me know which write event have you received ? Was it a write request or it's a write command ?

If it's a write request and it hasn't been replied by the server then the server is not allowed to send an indication, if I understand the spec correctly. So one indication/write request at a time.

Could you try to test writing a write command and check if you can send the indication ?
There is a chance that there is another characteristic is sending an indication or a write request, a sniffer trace would reveal if that's the case.