Channel tx_queue purging on disconnect was inconsistently handled
by the different channels: iso, l2cap, l2cap_br.
iso channels handled purging in the tx_data_pull hook.
l2cap and l2cap_br did the purging in channel delete functions
and did not expect tx_data_pull to be called for a disconnected
channel. Their data_pull functions could return a ptr to a
net_buf that was still on the tx_queue, which is problematic
when the conn tx_processor unrefs the returned buffer resulting
in multiple calls to the buf destroy function.
To make things consistent and correct, remove the code that tries
to purge tx_queues in the tx_processor and only do purging in
the channels themselves when they are deleted/disconnected.
Also refactor and clarify referencing of the net_buf returned
by tx_data_pull. It was confusing who had a reference and
when, which could vary depending on the length of the original
buffer. There are three cases: the buffer length is less
than the tx.mps, greater the mps but less than the mtu so
requiring segementation but not fragmentation, or greater than
both mps and mtu so requiring both segmentation and fragmentation.
The conn layer would increase the refcnt if the length was
greater than the mtu, but not have any awareness of whether
the net_buf was still on the tx_queue or not.
Now it is the tx_data_pull callbacks responsibitity to increment
the reference count if it is returning a pointer to a net_buf
that it is still keeping on the tx_queue for segmentation purposes.
The conn layer will now always transfer that reference into a
fragment view and not conditional it on the length relative to
the mtu, and always decrement the reference to the parent when
the fragment is destroyed.
So there is no risk of decrementing a reference to a net buf
that might still be on a tx_queue, which simplifies error
handling in particular.
Also add error handling paths for when asserts are not enabled.
Signed-off-by: Mike J. Chen <mjchen@google.com>
Fix peripheral role assertion during connection update and
simultaneous flash operations.
prepare_cb: Actual EVENT_OVERHEAD_START_US = 6149
This happens due to instant latency at connection update
where the ticks_at_expire was in the past.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Write operations are not constant time on nRF54Lx SoCs and
depend on the previous value present versus new value to be
written. Hence, perform no more than one iteration.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
In HFP PTS, cases run fail with the log 'Bluetooth Profile Descriptor
List UUID (0x111E) is missing'.
With this patch, the affected cases can pass correctly.
Signed-off-by: Jiawei Yang <jiawei.yang_1@nxp.com>
SonarQube reports: cast from 'const struct _snode *' to 'struct _snode *'
drops const qualifier. Fix it by removing the const.
Signed-off-by: Mark Wang <yichang.wang@nxp.com>
(cherry picked from commit 7982ee0df13b7f9e54595f799e32f38c464b0f9e)
1. bt_sdp_discover may be called in two tasks at the same time:
(1) If the 2 calls are for 2 acl conns,
the same session may be new allocated to two acl conns.
(2) If the 2 calls are for the same acl conn, the same session should be
used.
Use the bt_conn_index to map acl conn with session to fix it.
2. fix that bt_sdp_discover may use one session that is disconnecting,
put req to reqs_next and process it after disconnected.
3. remove memset of session, reset each members individually.
Signed-off-by: Mark Wang <yichang.wang@nxp.com>
(cherry picked from commit dcbe5d7dc5d72ce0d2915cf2b291cc84a916a536)
Utilize a code spell-checking tool to scan for and correct spelling errors
in `Kconfig` files within the `soc` and `subsys` directory.
Additionally, incorporates a fix recommended by the reviewer.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
Ensure that the BR/EDR L2CAP RX MTU is limited to the configured
BT_L2CAP_RX_MTU value during channel configuration.
This change prevent potential buffer overflow issues when receiving
data larger than the configured buffer size.
Signed-off-by: Jiawei Yang <jiawei.yang_1@nxp.com>
The MTU calculation in rfcomm_connected() was incorrectly adding the FCS
size instead of subtracting it.
This could lead to buffer overflows when sending data that exceeds
the actual available space.
Fix the calculation by properly subtracting both the RFCOMM header size and
the FCS size from the L2CAP MTU to get the correct RFCOMM session MTU.
Signed-off-by: Jiawei Yang <jiawei.yang_1@nxp.com>
Some of the Kconfig options did not reflect the actual
requirements from the TMAP spec. Relying on the AUDIO_TX
and AUDIO_RX was misleading, as they could be set by
broadcast configurations when the features required
unicast features.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Fix peripheral EVENT_OVERHEAD_START_US assertion due to
missing packet timer timestamp accumulation under single
timer use in nRF54Lx SoCs.
Relates to commit 373dc0db71 ("Bluetooth: Controller:
Single timer use Extended Advertising nRF54L support").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
There is an issue that the SCO connect cannot be un-referenced by HFP
AG when the SCO connection is broken if the SCO connect is not created
by HFP AG.
Register SCO connect change callback. And un-reference the SCO connect
in SCO disconnected callback.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
There are two issues found,
Issue 1, if the SCO connection has been established, the SCO connected
event will not be notified. Then the call that waiting for the SCO
connected event will not be updated. And the call will be rejected due
to the timeout.
Issue 2, the SCO con object will be cleared when SLC is disconnecting
and the SCO connection has been established. It causes the SCO connect
object will be `NULL` in the SCO disconnected callback.
Add a new field `sco_conn` to keep the SCO connect object. Manage the
SCO connection via the field `ag->sco_conn`.
When creating the SCO connection, keep the created SCO connect object
to `ag->sco_conn`. And if the `ag->sco_chan.sco` is `NULL`, it means
the SCO connection has been established. Just update the call that
waiting for the SCO connected event directly.
Increase the `ref` of SCO connect object and pass to the field
`ag->sco_conn` if the `ag->sco_conn` is `NULL`.
Decrease the `ref` of SCO connect object if the `ag->sco_conn` is not
`NULL` when disconnect the SCO connection or process SCO disconnected
event.
Signed-off-by: Lyle Zhu <lyle.zhu@nxp.com>
Commit removes unused host hci custom functionality
that was moved to mesh by misinterpretation of it.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
- Issue: There is a bus-fault while accessing empty userdata structure
pointer if application does not include any include service
userdata instance (which consist of UUID list of included service)
but service array has defined dummy entry for it assumed to be
overridden by app during initial flow.
- For example, the issue has happened in case of tmap-central sample
without "CONFIG_BT_OTS" support. there are some MCS attributes
dependent on OTS service because of that
"BT_GATT_INCLUDE_SERVICE(NULL)" entry is added as part of service
definition. The given entry does not have userdata handler defined
and is expecting to be overriden by the app if it will be included.
During "bt_mcs_init()" call, "mcs.attrs[i].user_data" is not
populated with any attr-instance pointer. This makes CPU to access
null-address during reading local-database include-service attribute
which was not provided by the app but the include-service entry was
added to the db.
- Fix: Adding condition to check if user-data has null address, and
returning back to avoid any hard-faults.
Signed-off-by: Nirav Agrawal <nirav.agrawal@nxp.com>
If the next ticker is too close then hop over it when
rescheduling a ticker with drift in slot window is used.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The command buffer was incorrectly encoded due to the recent change to H:4
encoding. Instead of fixing the header offset, just use the newly
introduced bt_hci_cmd_alloc() API, since that takes care of correctly
encoding the command header when it gets sent.
Signed-off-by: Johan Hedberg <johan.hedberg@silabs.com>
Introduce a new bt_hci_cmd_alloc() API which only allocates a command
buffer and reserves sufficient headroom for H:4 and command headers, but
doesn't actually encode any of those headers into the buffer. Additionally,
modify bt_hci_cmd_send() and bt_hci_cmd_send_sync() so that they handle
such buffers and perform the header encoding correctly. To avoid
duplication, bt_hci_cmd_send_sync() now builds upon (i.e. calls internally)
the bt_hci_cmd_send() API.
Signed-off-by: Johan Hedberg <johan.hedberg@silabs.com>
Fix incorrect aux offset value in the aux pointer in the
primary advertising channel PDUs when using single timer
tIFS switching feature in nRF54Lx.
End time was not captured when only Tx follow by Tx happens
in the primary advertising channels. End time needs to be
always captured when using single timer so that elapsed
packet timer value is correctly accumulated, for use to
setup subsequent radio transmission or reception inside the
same radio event.
Relates to commit 373dc0db71 ("Bluetooth: Controller:
Single timer use Extended Advertising nRF54L support").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Do not clear timer clear subscription until the radio event
is done. This is required for single timer switching when
periodic chains are to be received by LLL scheduling. Each
PDU in the train requires the timer to be cleared when the
parent PDU has been received such that reception of
auxiliary PDU's radio_tmr_start_us() can use the relative
offset since last PDU reception.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Use defines for Transmit/Receive and Inter-frame spacing
related Radio End events. This will ease some spaghetti of
conditional compilations for single timer and direction
finding support.
Direction Finding is not yet supported when using single
timer software tIFS switching.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fixup commit 031a17c271 ("Bluetooth: Controller: Align
DDPI usage with nRF54L power domains").
Align DDPI usage with nRF54L series power domains and
available DDPI configurations.
Use indices below 4 when needing to bridge with GPIO,
use indices below 8 when needing to bridge with MCU domain
like CCM and AAR; and use indices above 7 to bridge with
Peripheral domain.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
HAL_SW_SWITCH_TIMER_CLEAR_PPI is same as
HAL_SW_SWITCH_GROUP_TASK_ENABLE_PPI, hence the single PPI
suffice being disabled.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Bluetooth 6.1 has version number 0x0F and was released in May 2025.
This commit ensures this version number is properly decoded.
Going forward we may consider removing printing the version
number as a string because the feature set is more important
than the version number.
Signed-off-by: Rubin Gerritsen <rubin.gerritsen@nordicsemi.no>
The bt_le_ext_adv_info struct has been extended to also
contain the advertising and periodic advertising states.
Additionally, the function verifies the input to avoid
NULL pointer access, and the addr field is more
properly documented.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Adds a new abstract struct for unicast group that is
specific for CAP. The difference between this and the BAP
unicast group, is that the parameters are CAP streams and
thus ensuring that the streams in the group adhere to the
additional requirements that CAP has on top of BAP.
This also adds foreach functions for both CAP and BAP
to allow users to iterate on the streams in the
abstract groups.
Various samples, modules and tests have been updated
to use the CAP struct and API.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Core Specification 6.1 clarified exptected erro code in case peer
sending invalid Public Key. In case pairing is aborted during or
immediately after Public Key Exchange phase has completed (which is
the case here) expected reason code is set to "DHKey Check Failed".
This was affecting SM/CEN/KDU/BI-04-C and SM/PER/KDU/BI-04-C
qualification test cases.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
Fix up the bus fault by not violating the MAXPACKETSIZE
value range of NRF_CCM h/w peripheral.
Fix up relates to commit 920117922b ("Bluetooth:
Controller: nRF53x: Fix NRF_CCM MAXPACKETSIZE value").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix incorrect elapsed events value when LLL event prepare_cb
was invoked but was aborted before anchor point sync. This
caused premature supervision timeouts under applications
configured with CONFIG_BT_CTLR_EVENT_OVERHEAD_RESERVE_MAX=n
and CONFIG_BT_CTLR_PERIPHERAL_RESERVE_MAX=n.
Fixes commit 247037bd3e ("Bluetooth: Controller: Fix
incorrect elapsed events value").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
ticker_is_initialized() should only return true when the ticker is
running (triggered regularly). Users like nrf_flash_sync_is_required()
depend on this behavior.
When the bluetooth controller driver is closed, ll_deinit() calls
lll_deinit(), which stops the ticker from being triggered. Also
deinitialize the ticker to ensure that ticker_is_initialized() returns
false.
Signed-off-by: Marco Widmer <marco.widmer@bytesatwork.ch>
