lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

Vidar Berg — Wed, 17 Jul 2024 12:27:50 GMT

I'm glad to hear that it works now after you increased the number of adv. sets. Thanks for the updates. Regarding the recover command, this should ideally be a one time operation. It disables the APPROTECT mechanism on the chip (i.e., readback protection). However, certain actions can cause it to become enabled again. For instance, if you erase the firmware on the netcore or appcore followed by a reset, as the firmware (in the startup files) is responsible for performing the software unlocking of the protection on boot. A crash very early in the boot process can also lead to the same symptoms.

It may also be worth adding that a --recover or --chiperase command will erase the whole chip, including any bonding information that may be stored in the settings partition. In that case, you will need to delete the bond on the PC and repeat the pairing procedure.

https://docs.nordicsemi.com/bundle/ncs-2.7.0/page/nrf/security/ap_protect.html

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Wed, 17 Jul 2024 04:48:44 GMT

I find that I have to do recover to get the board flashed, I just do it every time.

I'm wondering if this affects the bond/pairing information.
-- runners.nrfjprog: Recovering and erasing flash memory for both the network and application cores.
Recovering device. This operation might take 30s.
Erasing user code and UICR flash areas.
Writing image to disable ap protect.
Recovering device. This operation might take 30s.
Erasing user code and UICR flash areas.

I set the CONFIG_BT_CTLR_ADV_SET to 6 and haven't seen 0x2036 since.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Tue, 16 Jul 2024 19:25:11 GMT

I marked the ticket as closed. But;

[00:00:05.356,994] <wrn> bt_hci_core: opcode 0x2036 status 0x07
Failed to create a connectable advertising set hids_mouse (err -5)

I still get

SDC_HCI_OPCODE_CMD_LE_SET_EXT_ADV_PARAMS = 0x2036,

0x07 Memory Capacity Exceeded

after multiple connection attempts. Aside from putting things in a work queue or upping the CONFIG_BT_CTLR_ADV_SET any other suggestions?

It pairs fine and works on first pair. It's hard to understand why the mouse is disconnecting and reconnecting.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Tue, 16 Jul 2024 15:29:19 GMT

The hids_mouse also does the workqueue to decouple the advertising and I pulled that in. I also attempted to wrap my other calls that were from adv_multi sample in a workqueue and did not see a difference.

Let me know if you have any other ideas. I will keep shuffling code...
-----
Everything looks to be working now from nrf connect. Any ideas on why I would not be seeing the device on my windows PC since changing it to bt_le_ext_adv_start?
----
Nordic_HIDS_mouse shows up and it connects and shortly after it tells me driver error. I will keep trying...if you know of a way to debug the interaction between the mouse and the windows pc. lmk
---
Working! Had to get the Kconfig correct; you can't do windows mouse without
BT_HIDS_SECURITY_ENABLED

I had troubles with

BT_DIRECTED_ADVERTISING

I disabled that; with it I could not see the device. Regardless, thank you for your help Vidar.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

Vidar Berg — Mon, 15 Jul 2024 14:32:29 GMT

Thank you.

[quote user="tldr"]The hids mouse does not exist in this project yet, I will be attempting to do that. I'm surprised by the way in which it beats up the connection; if the service doesn't exist, you'd think it would back off at some point. It looks like a denial of service it connects so many times. [/quote]

I agree, it seems like it would make sense for the PC to back off after a few attempts. The link is probably terminated by the PC when the peripheral reports "pin or key missing" in response to the encryption request. Since it is HID over GATT, which is natively supported by most operating systems, the re-connections in the background will be handled by the OS and not through an app/program. However, it is possible to make the peripheral ignore connection requests from "unknown" devices by using the Bluetooth Filter Accept List feature.

[quote user="tldr"][00:00:41.568,359] <wrn> bt_hci_core: opcode 0x2039 status 0x0d
00> [00:00:41.575,408] <err> bt_adv: Failed to start advertiser
00> Failed to create a connectable advertising set (err -12)[/quote]

It may be that the controller does not have the resources to start advertiser at the time of the request. I'm not sure where in your code you are starting the advertiser here, but please try to offload this task to the system workqueue to see if that fixes the problem.

Here's an example where advertisment start task is offloaded to the workqueue: https://github.com/nrfconnect/sdk-nrf/blob/b3293095a1374d9fa4a7480fb5d145760155596e/samples/bluetooth/peripheral_hr_coded/src/main.c#L82

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Mon, 15 Jul 2024 06:35:31 GMT

Trying to advertise the peripheral_hids_mouse via advertising_set_create

[00:00:41.568,359] <wrn> bt_hci_core: opcode 0x2039 status 0x0d
00> [00:00:41.575,408] <err> bt_adv: Failed to start advertiser
00> Failed to create a connectable advertising set (err -12)

devzone.nordicsemi.com/.../insufficient-resources-warning-when-connecting
> The status states BT_HCI_ERR_INSUFFICIENT_RESOURCES, and from what I've found, the opcode 0x2039 stands for BT_HCI_OP_LE_SET_EXT_ADV_ENABLE

I've up'd the CONFIG_BT_CTLR_ADV_SET and I think that is correct, any suggestions appreciated.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Mon, 15 Jul 2024 04:17:35 GMT

Impressed you were able to discern that from the information given. The machine I am using has an intel bt chipset and this DK I have had paired as a mouse using peripheral_hids_mouse. I tried the other DK board I have and it didn't exhibit the behavior either so that makes sense.

The hids mouse does not exist in this project yet, I will be attempting to do that. I'm surprised by the way in which it beats up the connection; if the service doesn't exist, you'd think it would back off at some point. It looks like a denial of service it connects so many times.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

Vidar Berg — Fri, 12 Jul 2024 06:21:00 GMT

It seems like you have a central device within BLE range that keeps connecting and disconnecting with your nRF5340 DK, and it's probably filtering on the BLE address since you are not seeing these connection attempts being made when using the Thingy.

The OUI in the public address of the Bluetooth central is registered to Intel, so I guess the connections are made from a nearby PC with an Intel BT/Wifi chipset.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Thu, 11 Jul 2024 17:35:08 GMT

The connect/disconnect is not happening on the thingy53. It only happens on the nrf5340DK.
I do not have special conf or overlay files for the dk. If the code helps, send me an email from my profile address and I will provide.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Thu, 11 Jul 2024 16:39:55 GMT

Disconnected: A4:F9:33:82:A2:FC (public) (reason 19)
Connected A4:F9:33:82:A2:FC (public)
Advertiser[2] 0x20001278 connected conn 0x20001408
Disconnected: A4:F9:33:82:A2:FC (public) (reason 19)

That mac address does not show in the nrf scanner; not sure what else to try.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Thu, 11 Jul 2024 15:18:47 GMT

Spot on Vidar.

CONFIG_SEGGER_RTT_BUFFER_SIZE_UP=4096

Gives me the whole output! That's a huge problem solved for me. Why the difference in behavior from cold start and wake sense? The 1024 default might be changed if normal output causes RTT users to get cut off.

>Do you have any BLE security enabled in your application?

I have exactly what is coming from multiple_adv_sets, which I believe does nothing with security. I was wondering if its from nrf connect or some other device I have. There is no other peer; no other device is scanning and I'm not using nrfconnect when the connections and disconnections occur.

> I was wondering where and when you saw the Kconfig warning above.
The warning came from the tooltip hover provided in vscode that shows you the criteria for the flag. I was reaching...

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

Vidar Berg — Thu, 11 Jul 2024 14:21:16 GMT

[quote user="tldr"]The first time the device turns on the output still terminates; wake sensing the device works as expected.
Any ideas on how to go about debugging that? I need to do some more testing...I keep seeing connected and disconnected events with (reason 19) happening; more to look into but I'm happy to see the services. lmk if you've seen this before.[/quote]

Disconnect reason 19 corresponds to BT_HCI_ERR_REMOTE_USER_TERM_CONN and indicates that the connection was terminated by the peer device. Do you have any BLE security enabled in your application? Reason for asking is that the peer device may in some cases terminate the connection if it fails to secure the connection.

Regarding the missing log messages, are you printing logs over RTT or UART. If using RTT, it may help to increase the 'CONFIG_SEGGER_RTT_BUFFER_SIZE_UP' size. The main() thread does not yield to the log processing thread during initilization which may cause the log buffer to become full.

[quote user=""]CONFIG_BT_PERIPHERAL was assigned the value y, but got the value n. Missing dependencies:
BT_RPC_STACK || (BT_HCI && BT)[/quote]

I was wondering where and when you saw the Kconfig warning above. I was not sure what could have caused this as both BT_HCI and BT should be enabled in the application.

CONFIG_BT_RPC_STACK should only be enabled if you want to run the whole BT stack on the netcore: https://docs.nordicsemi.com/bundle/ncs-2.7.0/page/nrf/device_guides/nrf53/features_nrf53.html

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

tldr — Thu, 11 Jul 2024 13:50:41 GMT

Thank you for teaching me how to fish and making me a meal.

I was able to get the the additional beacon and connectable service working by setting:

CONFIG_BT_CTLR_ADV_SET=3 in prj.conf in sysbuild/ipc_radio

.
The first time the device turns on the output still terminates; wake sensing the device works as expected.
Any ideas on how to go about debugging that? I need to do some more testing...I keep seeing connected and disconnected events with (reason 19) happening; more to look into another fishing expedition for a code but I'm happy to see the services. lmk if you've seen this before.

>Did you also include the Kconfig.sysbuild file which enables the NRF_DEFAULT_IPC_RADIO symbol? And did you see the Kconfig warning while building the ipc_radio or the main application?

Yes, I included the Kconfig.sysbuild and the symbol is enabled. I did not see a warning and I don't see one in your screenshot.

RE: lookup meaning of bt_hci_core: opcode 0x2036 status 0x07

Vidar Berg — Thu, 11 Jul 2024 13:29:51 GMT

Hello,

The Bluetooth specification does specify the opcodes for all HCI commands (except vendor specific ones), the problem is that the value is split into fields named OCF and OGF, so the opcode value from the error log is not directly searchable.

I use to search for the opcode in the SDK tree instead:

The status code can be found in the specification. See Vol. 1, Part F, Chapter 2.

This error may, for instance, be raised if you try to add more advertisement sets than what has been allocated by the CONFIG_BT_CTLR_ADV_SET symbol in the 'ipc_radio' build. You can confirm the configurations set for the ipc_radio build by checking the generated .config file in build/ipc_radio/zephyr.

[quote user=""]

and I'm now trying to use multiple_adv_sets to provide the beacon and connectable service.

I've pulled all the settings from the prj.conf in multiple_adv_sets. I'm using sysbuild and I have the ipc_radio prj.conf defined.
The original existing broadcast/observer code is working even though this bt_hci_core has crashed/failed to initialize.

CONFIG_BT_PERIPHERAL was assigned the value y, but got the value n. Missing dependencies:
BT_RPC_STACK || (BT_HCI && BT)

[/quote]

Did you also include the Kconfig.sysbuild file which enables the NRF_DEFAULT_IPC_RADIO symbol? And did you see the Kconfig warning while building the ipc_radio or the main application?

Best regards,

Vidar