Removed the LOG_WRN if the started or stopped callback
for streams are not set. The reasoning is that there is also
the group-wide started and stopped callbacks, and an application
may want to use those instead of the per-stream callbacks.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Removed the LOG_WRN if the started or stopped callback
for streams are not set. The reasoning is that there is also
the group-wide started and stopped callbacks, and an application
may want to use those instead of the per-stream callbacks.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The bt_bap_unicast_client_ep_qos function attempts to access
ep->iso->chan.iso->iso, but where some of the pointers may be
invalid. Add an additional check for these.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
In addition to aligning the destination address of all connection
objects during the identity resolution phase, we also need to update the
corresponding address field in the `struct bt_keys` structure for each
of these connection objects.
If we don't do this, `bt_keys_get_type` for that connection won't find
the `bt_keys` structure in the pool and will allocate a new one. This
will result in orphaned `bt_keys` structures in the pool.
This completes the change done in https://github.com/zephyrproject-rtos/zephyr/pull/61030/commits/637dabc3c9b018f47dee656d2ca0fe8f2fa21984
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
The event header is already accessed in `bt_recv_unsafe`. If length is
valid there, no point in checking it again in `hci_event` or
`hci_event_prio`.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
During debugging multiple.sh test, a race condition was observed where
advertiser state on Host and Controler was desynchronized.
Below is the sequence of actions led to this issue:
0. Host has queued few priority events at this moment.
1. Controller establishes a connection and stops the advertiser.
2. Controller sends to Host BT_HCI_EVT_LE_ENH_CONN_COMPLETE.
3. Host deprioritizes BT_HCI_EVT_LE_ENH_CONN_COMPLETE event and doesn't
process it immediately.
4. Controller sends to Host other priority events which are processed
immediately.
5. Host wants to disable the advertiser in order to update the Resolving
List (see `bt_id_add`).
6. Host sends to Controller BT_HCI_OP_LE_SET_ADV_ENABLE(Disable)
command.
7. Controller sends to Host BT_HCI_EVT_CMD_COMPLETE event on
BT_HCI_OP_LE_SET_ADV_ENABLE command.
8. Host prioritizes this BT_HCI_EVT_CMD_COMPLETE event, processes it
immediately and clears `BT_ADV_ENABLED` flag for the advertiser.
Note that Controller already has the advertiser disabled after
establishing connection on step 1.
9. Host sends few commands in a row and Controller sends
BT_HCI_EVT_CMD_COMPLETE events that are prioritized (see
`bt_id_add`).
10. Host wants to re-enable the advertiser after updating the Resolving
List.
11. Host sends to Controller BT_HCI_OP_LE_SET_ADV_ENABLE(Enable)
command.
12. Controller sends to Host BT_HCI_EVT_CMD_COMPLETE event on
BT_HCI_OP_LE_SET_ADV_ENABLE.
13. Host prioritizes this BT_HCI_EVT_CMD_COMPLETE, processes it
immediately and sets `BT_ADV_ENABLED` flag for the advertiser.
14. Controller has the advertiser enabled.
15. Host now processes the delayed BT_HCI_EVT_LE_ENH_CONN_COMPLETE
event.
16. Host resets the advertiser state, thinking the Controller's
advertiser is also stopped.
This commit prioritizes LE Connection Complete and LE Enhanced
Connection Complete events to drop advertiser state on Host immediately
once the corresponding event was received.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
- "bt_tbs_set_uri_scheme_list()" function does not setting URI for
GTBS instance. Using "inst_lookup_index()" fixes issue to find
registered GTBS instance for provided bearer_index.
- "bt_tbs_set_uri_scheme_list()" function overrides URI scheme list
with latest provided list. But the expectation is that it should
keep previously added URI schemes and only update new scheme which
are not part existing list. Code is updated to fix this.
- The same function uses net-buf to combine all URI schems for all
TBS bearer if current request is not for GTBS instance. Currently,
the code only uses combined URI scheme list for logging purpose and
it is not used to send for URI Scheme notification to provide
updated GTBS URI scheme to remote. Code is now updated to add
dynamic buffer to add all TBS URI scheme list for GTBS instance
and notify it similar to other functions to optimise memory usage.
Signed-off-by: Nirav Agrawal <nirav.agrawal@nxp.com>
If the chrc->uuid is somehow an invalid UUID, we log it and
terminate the discovery procedure.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Some TBS API functions attempted to take the mutex
of an instance before the NULL check.
Reorder the checks of the functions, and also modify function
to be more similar (using the same terms and structure).
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
There are two main issues found with using DLC TX thread,
Issue 1, the RAM consumption. Every DLC will have a dedicated thread
and thread stack.
Issue 2, the thread stack overflow issue. There is no way to strike a
balance between stack size and RAM consumption. Since the deep of call
stack is depended on the upper layer, the thread stack needs to set by
application. Due to the thread stack of DLC is dedicated, RAM
consumption is the product of the added value and the number of DLCs.
Use a TX worker to replace DLC TX thread.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add a Kconfig `BT_L2CAP_CONNLESS` to enable the feature.
Add an API `bt_l2cap_br_connless_register()` to register the monitor
to receive the unicast connectionless data.
Add an API `bt_l2cap_br_connless_unregister()` to unregister the
registered callbacks.
Add an API `bt_l2cap_br_connless_send()` to send unicast
connectionless data.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Implement the unicast to broadcast handover procedure,
as per the Bluetooth CAP specificiation.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Commit f499ea0 made it so that the compiler optimization flags are also
passed to the assembler. However, this breaks for the IAR toolchain
since it uses the GNU assembler which does not have the same
optimization flags as the IAR compiler.
This commit separates the compiler and assembler optimizations to make
it possible to specify appropriate flags in a toolchains corresponding
`compiler_flags.cmake`
Signed-off-by: Daniel Flodin <daniel.flodin@iar.com>
A recent change triggered a warning in LLVM that was promoted to error
when run with twister.
```shell
goep.c:102:3: error: label followed by a declaration is a C23 extension
[-Werror,-Wc23-extensions]
102 | int err;
```
Add a scope to the switch case so that the declaration has proper scope.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
After the #93703 PR was merged we get a possibility to remove _node field
from bt_conn_cb struct if the BT_CONN_DYNAMIC_CALLBACKS option is disabled.
The list conn_cbs can also be safely removed by more macro usage.
Signed-off-by: Radosław Koppel <r.koppel@k-el.com>
CONFIG_BT_CTLR_DATA_LENGTH_MAX should set its default value up to
and including the maximum value supported by the Bluetooth spec
(251 bytes).
Signed-off-by: Sam Friedman <sam@golioth.io>
According to the OBEX Version 1.5, the Name header could be a empty
string.
Update function `bt_obex_add_header_name` to support the case that the
length of name could be 0.
Update function `bt_obex_get_header_name` to support the case that the
name header is found but the name length is 0.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add a TLV array `tlvs` to keep the pending TLV data.
Update `app_param`, `auth_challenge`, and `auth_rsp` to set the one
TLV information each time. When the optional argument `last` set,
write the all pending TLVs to the tx buffer.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add a structure `struct bt_obex_tlv` to pass/save the one TLV info.
Use a array of `struct bt_obex_tlv` as the set of all TLV-triplets.
Update the following functions to support the feature, including
`bt_obex_add_header_app_param`, `bt_obex_add_header_auth_challenge`,
and `bt_obex_add_header_auth_rsp`.
Add a function `bt_obex_tlv_parse` to parse the encoded TLV-triplets.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
There are types of string can be added by using OBEX adding header
function bt_obex_add_header_*(). One is byte sequence. Another is null
terminated Unicode text.
Add a function bt_obex_string_is_valid() to check if the added string
is valid. And add a function bt_obex_unicode_is_valid() dedicated to
check Unicode string.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add commands for GOEP to test OBEX features.
Add command `alloc-buf` and `release-buf` to allocate and release TX
buffer.
Add command set `add-header` to add the OBEX headers to allocated TX
buffer.
Add command set `client` to send OBEX client requests with allocated
TX buffer.
Add command set `server` to send OBEX responses with allocated TX
buffer.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add commands for GOEP to test transport features.
Add commands for transport over RFCOMM, including `register-rfcomm`,
`connect-rfcomm`, and `disconnect-rfcomm`.
Add commands for transport over L2CAP, including `register-l2cap`,
`connect-l2cap`, and `disconnect-l2cap`.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
The MTU of the GOEP should be not less than 255. So, if the transport
is RFCOMM, the CONFIG_BT_BUF_ACL_RX_SIZE should be not less than 264.
It includes,
- 255 bytes for the minimum MTU of GOEP,
- 4 bytes for L2CAP Header,
- 5 bytes for RFCOMM header and FCS.
And if the transport is L2CAP, the CONFIG_BT_BUF_ACL_RX_SIZE should
be not less than 259.
It includes,
- 255 bytes for the minimum MTU of GOEP,
- 4 bytes for L2CAP Header.
Add Kconfig `BT_GOEP_RFCOMM_MTU` to configure the maximum size for
RFCOMM transport. The range of `BT_GOEP_RFCOMM_MTU` is
`[264, BT_RFCOMM_L2CAP_MTU]`. And the GOEP MTU via RFCOMM transport
should be in the range `[255, (BT_GOEP_RFCOMM_MTU-9)]`.
Add Kconfig `BT_GOEP_L2CAP_MTU` to configure the maximum size for
L2CAP transport. The range of `BT_GOEP_L2CAP_MTU` is
`[259, BT_BUF_ACL_RX_SIZE]`. And the GOEP MTU via L2CAP transport
should be in the range `[255, (BT_GOEP_L2CAP_MTU-4)]`.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Add a Kconfig BT_GOEP to control the GOEP feature.
Implement the GOEP protocol and transport, both for GOEP 1.1 and GOEP
2.x.
For GOEP transport, OBEX over RFCOMM, and OBEX over L2CAP are
supported.
For GOEP protocol, `put`, `get`, `abort`, `setpath`, and `action` are
supported. And only one operation can be processed at the same time.
The feature `Reliable Session` is unsupported.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
L2CAP channels will now, along with the ident, store the opcode of the
pending request. This commit expands the ident lookup function to also
compare received response types to this opcode, and will ignore
unsolicited responses.
Setting of idents for channels are moved after verification of buffer
allocation for the request to be sent. A TODO is added for improving
this functionality at a later time.
Signed-off-by: Håvard Reierstad <haavard.reierstad@nordicsemi.no>
This functionality is implemented in the HCI-facing side of the
controller and the ULL. LLL functionality is not implemented in this
commit.
Furthermore, this removes a dependency for LE Power Control, as the
features are related, but not necessarily required.
Signed-off-by: Tyler Joseph Huffman <tyhu@demant.com>
AICS gain mode does not prevent local nor remote operations from changing
the gain setting value. The gain setting value shall simply be ignored if
in the BT_AICS_MODE_AUTO_ONLY or BT_AICS_MODE_AUTO modes, but the local
application and remote devices can still modify the gain setting value.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This commit adds support for the frame space update feature
to the bluetooth host.
This is mainly just a wrapper around the frame space update
HCI command and event.
Signed-off-by: Sean Madigan <sean.madigan@nordicsemi.no>
Fix BT_LE_LOCAL_SUPPORTED_FEATURES_SIZE define, as this was
not working correctly without brackets around
CONFIG_BT_LE_MAX_LOCAL_SUPPORTED_FEATURE_PAGE.
Also fix CONFIG_BT_LE_MAX_LOCAL_SUPPORTED_FEATURE_PAGE help,
as this was talking about remote featues, where this only
applies for local features.
Also update for new CONFIG_BT_CONN_DYNAMIC_CALLBACKS kconfig
Signed-off-by: Sean Madigan <sean.madigan@nordicsemi.no>
The write_ct function calls bt_cts_send_notification and if
bt_cts_send_notification fails to send a notification (e.g.
if there are no clients subscribed), it would return an
error, and thus send a ATT error response to the client,
even if the value was actually written.
This commit changes the behavior so that the notification
does not affect the return value (as it is not directly
linked to the write itself), but rather just logs it if
it fails to send the notification.
The notification can failed to be send for 2 main reasons:
1) The application returns an error on fill_current_cts_time
2) bt_gatt_notify fails to notify due to missing buffers
The first we cannot do much about. For the latter we can
either just live with the missing notification, or implement
a retry if we run out of buffers. For this simple fix the
former was chosen, but can be revisited if it turns out
to be a significant problem.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Since the SDUs are now allocated by the application
and the host does not care about the SDU sizes, the check
is useless.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This commit allows to the user to control the use or not of dynamic
connection callbacks.
The stack is not using the dynamic connection callbacks so they are
only needed if the applications want to use it.
Setting it to true for backwards compatibility.
With dynamic connection callbacks
FLASH: 176600 B 512 KB 33.68%
RAM: 27428 B 64 KB 41.85%
IDT_LIST: 0 GB 32 KB 0.00%
Without dynamic connection callbacks
FLASH: 176376 B 512 KB 33.64%
RAM: 27428 B 64 KB 41.85%
IDT_LIST: 0 GB 32 KB 0.00%
Signed-off-by: Yago Fontoura do Rosario <yafo@demant.com>
This commit registers the connection callbacks using BT_CONN_CB_DEFINE,
which ensures they are set up at link time rather than runtime.
Signed-off-by: Yago Fontoura do Rosario <yafo@demant.com>
This commit registers the connection callbacks using BT_CONN_CB_DEFINE,
which ensures they are set up at link time rather than runtime.
Signed-off-by: Yago Fontoura do Rosario <yafo@demant.com>
Fix default value of BT_TICKER_REMAINDER_GET, it will be
selected based on enabled Controller features. This is not
needed by minimal peripheral role.
Fix depends on for BT_CTLR_CONN_RANDOM_FORCE as it is
applicable when more than one simultaneous connections is
supported. Again, this is not needed by minimal peripheral
role.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Check whether the connection response parameters both with and without
ECRED are within the valid ranges from the Bluetooth Core Specification
(part 3.A.4 v6.0). Changes validation checks in requests to match the
same pattern.
Signed-off-by: Håvard Reierstad <haavard.reierstad@nordicsemi.no>
The codec configuration operation for unicast (ASCS) contain
both a target PHY and a target latency value that was previously
just hardcoded by the client and unavailable by the server.
This commit adds them to the bt_audio_codec_cfg struct so that
applications can both set and get these values.
The default values used by BT_AUDIO_CODEC_CFG use the same
values as the client would previously set.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Adds a missing requirement from Core Spec V6.0 Vol 3.A chapters 10.1
and 10.2 to ignore L2CAP_FLOW_CONTROL_CREDIT_IND packets with the
credit value set to 0.
Matches existing credit-related functions by checking that the CID
is in the dynamic range (you can't add credits to fixed channels).
Signed-off-by: Håvard Reierstad <haavard.reierstad@nordicsemi.no>
