Fix missing synchronization when stopping primary and
auxiliary PDU scheduling.
Added implementation to ensure primary and auxiliary event
count match when terminating extended advertising on maximum
number of events or on duration.
Fixes#37571.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add implementation to release Auxiliary Context on failure
to receive auxiliary chain being in LLL scheduling.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Flush the scanned advertising PDUs to HCI while receiving
the scan response PDU when Extended Active scanning.
This reverts earlier implementation to remember the
advertising PDUs.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add support for Extended Active Scanning and generation of
Extended Advertising Report with Scan Response event type.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Reverting Extended scanning with LLL scheduling as this is
causing unresolvable merge conflict with the commits that
is implementing the Extended Active Scanning support.
Subsequent commit add the revert implementation back again.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix a race condition that caused the Controller to deadlock
waiting for a semaphore that be given when LLL events have
been disabled.
ULL reference count is checked in thread context to decide
if LLL events are pending, but the reference count can be
decremented by the ULL execution context which prevents the
set `disabled_cb` function not being called due to no
pending event to produce the done events.
Fixed by re-checking if reference count is zero and avoid
waiting to take the semaphore.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The new JIT scheduler does not have slot reservation, which means that
the ticker extension feature for automatically re-scheduling a colliding
non-anchored event, e.g. ADV, cannot be used.
This implementaion reacts to ADV envent done with result ABORTED or
TOO_LATE, and in those cases attempts to re-schedule the ADV event again
within the 10 ms pertubation window.
As the original scheduling, the re-scheduling is randomized, so there
is no absolute predictability as to how many attempts will be made. The
advertiser will attempt with randomly delayed re-schdules until the
window is exhausted.
If re-scheduling is unsuccessful, the weight of the ADV event is
increased, improving it's chances of success in the next event.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Fix HCI Reset Command from asserting when there is pending
connection creation requested on coded PHY.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Implement the functions required to calculate the SubEvent 1
and SubEvent n mapped channel indices.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Reserve a minimum node rx of 2 that can happen when local
central initiated PHY Update reserves 2 node rx, one for PHY
Update complete and another for Data Length Update complete
notification. Otherwise, a peripheral only needs 1
additional node rx to generate Data Length Update complete
when PHY Update completes.
Relates to #36381.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Replace the post decrement in loop's conditional into
explicit decrement inside the loop so as to avoid
decrementing the maximum count without enqueueing free rx
buffers into the free rx buffer MFIFO.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Increase the ticker operations count in ULL high context
when Extended Advertising Connection Establishment on Coded
PHY is supported.
This fixes assertion in Controller when initiating a
connection on Coded PHY, wherein two scan instance ticker,
one window stop ticker, and a new connection instance
ticker operations needs to be enqueued.
Relates to commit a6b8eba7c5 ("Bluetooth: controller:
Implement disabling the other PHY initiator").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When an ACL connection with active CISes terminates, inject CIS/CIG
teardown to ensure CIS is stopped before ACL disconnection completes.
This includes stopping CIG ticker when last CIS has stopped.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Added Kconfig option to configure the maximum Broadcast and
Synchronized Receiver ISO PDU length.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The amount of needed done events changes when using the "must expire"
logic.
Make the amount vendor configurable (via ull_vendor.h)
Defaults to the original amount of 3.
Signed-off-by: Thomas Ebert Hansen <thoh@oticon.com>
Added function ull_conn_iso_resume_ticker_start to enable LLL ISO
peripheral/central to start the one-shot resume ticker.
At timeout the common lll_resume handler is called, and based on the
LLL state, the peripheral/central is able to resume the CIG event.
The resume acts exactly like a normal event resume from the prepare
pipeline, with the exception that the LLL must provide specific
lll_event instances.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Change use of dedicated memory pool for linked list nodes for
node_rx_iq_report to common mem_link_rx. Former solution had
a drawback. Released link nodes may be enqueued to wrong memory
pool. E.g. link related with nopde_rx_iq_report went to common
link memory pool, whereas link nodes from common pool were enqueued
to dedicated list.
The solution was working because links have the same memory layout,
just different memory pools they originated from.
The problem may occur if one of those link memory pools is reset.
Then the same link may be used by node_rx and node_rx_iq_report
at the same time, causing controller failure.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
To enable or disable CTE sampling in periodic advertising scanner LLL
implementation lll_sync is extended by lll_df_sync member.
The lll_df_sync is a double buffer for lll_df_sync_cfg.
It will be used by LL thread to enable or disable CTE sampling and
provide configuration for sampling CTE in AoA mode.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
Add handling of Direction Finding node_rx_iq_report type in ULL.
This is required to correctly deliver IQ samples collected by
LLL during receive of PDU including CTE.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
CTE sampling enable or disable requests should be provided to LLL.
Double buffer of lll_df_sync_cfg is added to lll_sync for that purpose
Thanks to that:
- there is no need to synchronize LLL with LL thread
- LLL has always access to valid configuration (new or former)
The configuration is stored in double buffer as instances
of lll_df_sync_cfg, not as pointers. This is the only
difference between double buffer pattern used to store adverising PDU.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
Check whether the preempt event matches with the head of the
pipeline before aborting the currently active event.
This is required to avoid preemption of events that became
active due to done event and there has been a race in
stopping the preempt ticker.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Refactor and simplify the bluetooth buffer configurations to improve the
easy of configurations and eliminate invalid ones.
By moving configurations out of host and controller specific
configurations and into a common one it becomes easier to configure
the host and controller separately as the same configurations can be
used as would be for a combined build.
All HCI configurations are now given exluding the matching HCI header,
which eases the configuration as the application don't have to know the
different header sizes.
The BT_RX_BUF_LEN is split into ACL and Event, as well as the suprising
use of Command size.
BT_L2CAP_RX_MTU is removed as the stack does not support reassembling of
HCI ACL data to larger L2CAP PDUs. The application will have to set
ACL RX size and account for the L2CAP PDU header itself.
BT_EATT_RX_MTU was removed as it is only used for setting a different
default value for another option which leads to the stuck kconfig symbol
problem.
The configurations can be updated according to the table below:
** New configuration | ** Old configuration
All configurations
BT_BUF_ACL_RX_SIZE | BT_L2CAP_RX_MTU + 4
BT_BUF_ACL_RX_SIZE | BT_RX_BUF_LEN - 4
BT_BUF_EVT_RX_SIZE | BT_RX_BUF_LEN - 2
BT_BUF_CMD_TX_SIZE | BT_RX_BUF_LEN - 3
BT_BUF_CMD_TX_COUNT | BT_HCI_CMD_COUNT
BT_BUF_EVT_RX_COUNT | BT_RX_BUF_COUNT
BT_BUF_ACL_RX_COUNT | BT_RX_BUF_COUNT
BT_BUF_ACL_RX_COUNT | BT_ACL_RX_COUNT
BT_BUF_EVT_DISCARDABLE_SIZE | BT_DISCARDABLE_BUF_SIZE - 2
BT_BUF_EVT_DISCARDABLE_COUNT | BT_DISCARDABLE_BUF_COUNT
Controller-build
BT_BUF_ACL_TX_SIZE | BT_CTLR_TX_BUFFERS_SIZE
BT_BUF_ACL_TX_COUNT | BT_CTLR_TX_BUFFER
HCI-bridge
BT_BUF_ACL_TX_SIZE | BT_HCI_ACL_DATA_SIZE
BT_BUF_ACL_TX_COUNT | 6
Fixed invalid configurations setting either BT_L2CAP_RX_MTU or
BT_CTLR_DATA_LENGTH_MAX larger than BT_RX_BUF_LEN could lead to buffer
overruns.
Fix advertising report max data length calculation.
This always used the BT_DISCARDABLE_BUF_SIZE macro but this feature
can be turned off and advertising reports will be allocated from the RX
buffer in that case. Also controller-build does not have this buffer
(in hci_raw.c). Also the wrong HCI header was used in the calculation,
HCI event header should have been used instead of HCI ACL header.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix ULL reference count decrement to be performed before the
done events are handled, so that correct reference count is
used to determine if events are pending in the pipeline.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When connection is initiated in one of either 1M or Coded
PHY initiating scan instance then the other scanning
instance's scheduling and memory allocation needs to be
cleaned up.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Reserve an additional node rx buffer when Extended Initiator
is supported as the received ADV_EXT_IND PDU is being
buffered in the auxiliary channel scanning instance and is
only released/flushed in the done event of the initiating
auxiliary channel radio event.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Avoid using sizeof to access ULL and LLL struct members.
Based on the alignment requirements of structures, due to
padding between structure members, use of sizeof of previous
struct member to access next struct member is incorrect.
Continue to use explicitly stored parent pointer to access
ULL context. Combine event header and ULL header so that
the parent pointer point directly to the combined ULL
struct.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Consolidate Kconfigs for ISO roles to make compilation guards more
concise. Just introduces aliases, does not change functionality.
Signed-off-by: Wolfgang Puffitsch <wopu@demant.com>
First design towards ISO adaptation layer, this PR introduces
data-structures and framework for Rx unframed PDUs (BT RX ingress).
Two callbacks are defined for the SDU production (BT RX egress), one for
SDU allocation as well as a callback for emitting a reassembled SDU.
Signed-off-by: Asger Munk Nielsen <asmk@oticon.com>
Priority in the legacy stack has been unused and "work in progress" for
some time. With this commit, the priority passing/handling is cleaned
up, preparing for the new JIT scheduling priority handling.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Move the LLL done event handling from ULL to LLL, this
reduces CPU utilization and reduces overhead between radio
events.
Fixes#21993.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
There is a race between the pipeline enqueue by the ULL
context when processing the LLL done event, and the
pipeline queue by the ticker timeout for a new radio
event. This caused an already enqueued event prepare to be
placed after the newly arrived request to enqueue a prepare
by ticker timeout.
Fix this race by placing the new prepare in the pipeline
instead of starting the event, when ULL is still processing
the LLL done event.
As ticker timeout and ULL processing of LLL done event are
in the same ULL_HIGH context, the pipeline handling is safe
with the order of prepares maintained correctly.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Make bluetooth.init tests for nRF platform pass again by adding ULL/LLL
stub functions, and fixing minor compile issues.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Fix flash driver sync regression introduced due to changed
default ULL_HIGH in the commit 30634334a8 ("Bluetooth:
controller: Fix ULL_HIGH priority to be lower than LLL").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
LL_CIS_IND starts a ticker for the created CIG, using the event_counter
and offset provided. Ticker generates callbacks to
lll_peripheral_iso_prepare. Event done with ISO (extra) type is demuxed
and done handled for CIG including ticker update with drift
compensation.
TODO: Handle multiple CISes as well as pause/resume and scheduling
latency.
Signed-off-by: Morten Priess <mtpr@oticon.com>
When Extended Advertising terminated due to duration or
maximum number of events, the auxiliary PDU scheduling
is now correctly stopped.
Fixes#31254.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Initial work in progress implementation of Create Connection
Cancel for Extended connection initiation.
Adds implementation to teardown connection initiated at ULL
layer and gracefully release allocated resources.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Remove inclusion of vendor specific internal LLL include
files in ULL source code.
Prefix `lll/` include file path to correctly include vendor
defined types and function implementation.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Use `#if defined(...)` instead of `#if IS_ENABLED(..)` when
conditional compilation of definitions in source code.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When ticker job is disabled inside radio events then all
advertising, scanning, and slave latency cancel ticker
operations will be deferred, requiring increased ticker
thread context operation queue count.
Relates to #32430.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Consistently use the internal BT_CTLR_ADV_SET value in the
controller's implementation.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
In order to avoid stalling LLCP procedures blocking done event
processing, done events are now put in a separate memory queue,
and processed by own mayfly function.
This ensures prepare/done refcount balancing and prevents assertion
due to lack of done events.
Signed-off-by: Morten Priess <mtpr@oticon.com>
