-- Handle time calculations across the point at which the controller's
time information from the controller's clock will wrap (in the
microsecond range)
-- Use wrapped time in all ISO-AL time calculations
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Adding support for allowing user extension code to support deferring
anchor point moves. Refactored LLCP only.
Note: This is NOT supported by LEGACY LLCP impl.
KConfig'd by BT_CTLR_USER_CPR_ANCHOR_POINT_MOVE
Modified ll_conn_update API, to accommodate passing offsets
Modified CPR/CU context data structure
Modified CPR/CU procedure to handle anchor point move wait state
Fixed up unit tests, adding anchor point move cases to CPR/CU test
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Using a macro instead of repeated identical code for validation
Use macro for pdu length field encoding
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Adding controller support for updating SCA in ACL and ISO peripherals
Adding HCI API support for le_request_peer_sca and complete event.
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Check LE event mask state as well as host controlled feature mask state
and reject CIS request accordingly.
Release pre-allocated ISO resources on rejection of request
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Program CIG ticker with window widening drift. Introduce vendor specific
conversion macros allowing sub-microsecond resolution in the
accumulation of window widening drift per interval.
Calculation of window_widening_max_us is done for NSE<3, and must be
adjusted in the LLL before use, if the first CIS in the CIG has a
NSE>=3. In that case window_widening_max_us shall be limited to one sub
interval.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Functions related to string manipulation that were defined in
`common/log.h` has been moved to the `common/bt_str.h` file and their
implementation in `common/bt_str.c`.
Files that were using those functions has been updated consequently.
Signed-off-by: Théo Battrel <theo.battrel@nordicsemi.no>
Fix ticker slot reserved check to handle deferred execution
of ticker worker where ticks_elapsed is greater than
ticks_slot_previous. In which case, each ticker expiring be
checked for overlap with previous slot reservation.
This fixes:
ASSERTION FAIL [0] @ WEST_TOPDIR/zephyr/subsys/bluetooth/
controller/ll_sw/nordic/lll/lll.c:476
lll_preempt_calc: Actual EVENT_OVERHEAD_START_US = 3906
Flash erase operation exposed the assertion, CPU being
halted deferred the ticker_worker collision resolution and
ticker expiry was not individually checked for overlap with
previous expired ticker reservation.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix missing first connection event due to first connection
event ticks_slot of central and peripheral overlapping with
the initiator window ticks_slot and advertising ticks_slot,
respectively.
Use ticker_stop_abs interface so that when the ticker is
stopped the ticks_slot_previous is truncated to the actual
ticks used as requested by the supplied absolute ticks.
Relates to commit 8a294a62d4 ("Bluetooth: controller:
legacy: Fix missing first conn event").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix ticker instance ticks_current initialization to be
acquired from the RTC counter value. The RTC counter may
have run in previous use of ticker before bt_disable hence
keep the instance ticks_current equal to RTC counter value
when ticker_init() is called again.
Relates to commit 4349a475a8 ("Bluetooth: Controller: Add
deinit() infrastructure").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
In order to avoid using multiple sources of truth for the platfom's
endianness, convert the in-tree code to use the (BIG|LITTLE)_ENDIAN
Kconfig variables exclusively, instead of the compiler's
__BYTE_ORDER__.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
According to specification, when executing a remote procedure the
rx parameters used need to be updated when the REQ from the peer
is received. Currently this is done only when an ACK is
received on the RSP PDU send out.
This commit updates the RX parameters upon receipt of the REQ PDU
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
Macro that was defined in the `common/log.h` has been moved to the newly
created `common/assert.h` file.
Files that were using those macro has been updated consequently. Also, the
include of `common/log.h` has been removed of files that were actually not
using any logging methods.
Some `#include "common/log.h"` have been removed. Those were required
before when including `hal/debug.h`. With this change, `hal/debug.h` no
longer has this requirement, because the replacement, `common/assert.h`
includes what it uses.
Signed-off-by: Théo Battrel <theo.battrel@nordicsemi.no>
Change automated searching for files using "IRQ_CONNECT()" API not
including <zephyr/irq.h>.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Adds proper support for having more than one PDU pool for the advertising
PDUs (to save memory)
- Introduces lll_adv_aux_data_init, lll_adv_aux_scan_rsp_alloc and
lll_adv_sync_data_init to indicate if the PDU is a primary/legacy PDU
or secondary/auxillary PDU
- Scan response PDUs are initialized later, since the correct pool to draw
from depends on whether the advertising is extended or legacy
Signed-off-by: Troels Nilsson <trnn@demant.com>
There is access to the procedure context after a potential release
of the context, which (in theory) can lead to incorrect behaviour
There is no good way of testing this with a unittest without adding
specal debug code to the ull_llcp_local module
After code-inspection no other location has been found with
potential access after context release
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
According to Core 5.3 Vol 4, Part E, section 7.8.82 slot_durations,
switch_pattern_len and ant_ids are used only for AoA and do not affect
reception of AoD CTE.
To improve interoperability checking of correctness of these parameters
is not required. It will not affect lower link layer in case the AoA
CTE reception feature is not enabled in Kconfig.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
The sys* ops like sys_clear_bit are indirectly included via arch CPU
header. Other stuff like find_msb_set end up included via this header as
well.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
- Account for the added random delay when updating remaining duration
- Re-worked JIT re-scheduling update a bit to keep the random delay
information and remaining duration correct
Signed-off-by: Troels Nilsson <trnn@demant.com>
- Remaining duration is now in us instead of ticks to avoid
overflow when ticks have higher resolution
- First advertising event should not reduce remain duration
Signed-off-by: Troels Nilsson <trnn@demant.com>
Add EVENT_OVERHEAD_START_US to ticks_anchor when enabling
advertising, so the first advertising event is not late due
to calculations for aux packets etc.
Signed-off-by: Troels Nilsson <trnn@demant.com>
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Tune the EVENT_OVERHEAD_START_US for typical usecases on
nRF52833 SoC using the central_gatt_write and
peripheral_gatt_write samples.
Manually run the samples for over 30 mins without any
assertions and use the actual profiled value that would
otherwise be printed on assertion.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When ticks_to_expire is in the past, i.e. ticks_to_expire is
0, perform an explicit trigger of ticker worker instead of
setting up the compare which delays the trigger by the
minimum ticks offset margin. The changes reduce latencies
of the O(n) ticker_resolve_collision by upto the minimum
ticks offset margin (3 for nRF SoCs) for each resolution.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When an event's prepare is delayed due to ULL scheduling or
ISR latencies, check and have a verbose assertion that logs
the actual EVENT_OVERHEAD_START_US required.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Conditional compile ticker interfaces like ticker_update
which are not required when individual state or role samples
are build.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix the BIG event counter calculation from stored payload
counter. Event latency should not be multiplied by burst
count.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix missing includes and rearrange header includes to be able
to build a full feature set Controller.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Set the correct number of Ctrl Tx buffers used for the new
implementation, so that the ACK-fifo gets the correct size
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
Conditionally compile the legacy control procedure defined
control Tx buffers that is not required in the new control
procedure implementation.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Adds support for handling fragmented advertising data over HCI when
CONFIG_BT_CTLR_ADV_AUX_PDU_LINK is not enabled
- Added support for appending advertising data to ull_adv_aux_hdr_set_clear
and ull_adv_aux_pdu_set_clear via ULL_ADV_PDU_HDR_FIELD_AD_DATA_APPEND
- Updated ll_adv_aux_ad_data_set and ll_adv_aux_sr_data_set to
handle fragmentation ops without CONFIG_BT_CTLR_ADV_AUX_PDU_LINK
Signed-off-by: Troels Nilsson <trnn@demant.com>
Remove redundant assertion check, and minor indentation fix
to correct the alignment of function call.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Enable ticker ticks slow window extensions implementation
when broadcaster role is enabled. This feature moves the
primary channel advertising event within the specified
window when overlapping with other roles.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Implemented:
-- Combining segments from multiple SDUs in a single PDU for TX
-- Release of PDU based on remaining capacity below threshold
-- Release of any PDUs in production on receiving a timeout for the
upcoming event
-- Release of PDUs in production if source if destroyed
-- Synchronization flag to exclude multiple contexts in ISO-AL
-- New Kconfig BT_CTLR_ISO_TX_SEG_PLAYLOAD_MIN to allow definition of
minimum TX segment size
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Implements:
-- SDU fragments can be buffered in the ISO-AL so that the correct size
of the SDU can be computed such that it can be released for inclusion
in the first fragment's header
-- Changed SDU emit interface to allow the same structure to be used to
buffer as well as release fragments
-- New configuration BT_CTLR_ISO_RX_SDU_BUFFERS to specify the number of
SDU buffers for each sink
(Buffering disabled when BT_CTLR_ISO_RX_SDU_BUFFERS=0)
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Implemented:
-- Storing information required to service HCI_LE_Read_ISO_TX_Sync
request
-- Implemented interface to retrieve information
-- Cleared ISO-AL source and sink state on deallocation
-- Renamed cig_ref_point to grp_ref_point as it serves a common purpose
in both CIS and BIS
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Improvement for IAL/CIS/FRA/PER/BI-01-C & IAL/CIS/FRA/PER/BI-02-C
-- Test appears to fail as the transmitted PDU is packed with less data
than is indicated in the segment headers.
-- Guards were included to verify that sufficient data exists in the PDU
before extracting the segment.
General
-- Updated timestamps in error SDUs to better approximate the expected
timestamps if time information is missing.
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
In case a connection update complete event reports an error, the
parameters reported should be the ones active on the connection and not
the ones requested
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
LE Write Suggested Default Data Length and LE Set Data Length commands
are suggestions from host and should be validated only as per HCI
specification regarding internal setting of LLCP.
LLCP is allowed to use other values if needed.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
The memory used for Tx buffers and procedure context must
be aligned on a word boundary, otherwise it is possible
to get unaligned memory access
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
Fix radio_tmr_start_us to be soft realtime, if the requested
start in microseconds is in the past, setup the next
earliest expiry. This is to prevent radio not being started
and not let the LLL hanging. When setup to receive auxiliary
PDU in LLL scheduling, is_aux_sched is set and if radio does
not start and the window is requested to be pre-empted, then
it is blocked from being preempted due to the is_aux_sched
flag remaining to be set. This stalls the scanning
infinitely until explicit scan disable.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix missing is_aux_sched flag being reset before resuming to
scan for new advertising PDU in the same scan window after
having received an earlier primary channel PDU and auxiliary
channel PDU using LLL scheduling.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix receiving spurious overlapping ADV_IND and SCAN_IND PDU
when LLL state expects to receive SCAN_RSP PDU. Explicitly
check that these PDUs are not processed when in active
scanning and waiting for SCAN_RSP PDUs.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
