Fix the order of setting the flag that indicates that
preempt timeout has been successfully requested versus
the request is waiting to be scheduled.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix ticker implementation to prefer to keep ticker node
started in case of race condition to start and stop.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Intent is to pass Ellisys ISOAL quality tests for framed TX scenarios
where the TX SDU includes a time stamp that is not based on the
controller's clock.
Changes:
-- Include controller's reception time as a separate field in the TX SDU
information
-- Include decision on whether SDU time stamp is valid and based on the
controller's clock
-- Arbitrate and select / compute time stamp for the SDU
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Changes:
-- Corrected reading and writing time offset in framed segment header to
account for endianness
-- Corrected bit positioning of time offset bit field in the segment
header structure definition
-- Fixed upstream Zephyr BSIM test build failure due to debug logging
39-bit variable
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Changes:
-- Disregards target event after segmenting the first TX SDU and instead
decide event based on the next payload and feasibility to fragment
for that event (time-offset > 0)
-- Corrected error SDU release logic when an error occurs while
expecting the start of a new SDU, to avoid releasing an SDU which
could be received in the next payload or event
-- Included additional debug logging
Signed-off-by: Nirosharn Amarasinghe <niag@demant.com>
Revert back the EVENT_OVERHEAD_END_US value to original
value in initial commit 5c3709c13f ("bluetooth:
controller: Add def's re. update to slot reservation calc").
40 us is a sufficient margin considering ~30 us being the
worst case Rx chain delay for S8 Coded PHY on nRF5x SoCs.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Convert the driver to a dt-based device. Also update Kconfig so that it
depends on the right compatible being enabled, initialize in POST_KERNEL
(APPLICATION should not be used for devices) and remove one redundant
include.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Fix missing lazy calculation when using ticker interface
required for Central CIS create.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix initialization of CIS lazy_active event count used when
first CIS event is active. Now initialization for first and
any subsequent CISes made active.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix some compiler instruction re-ordering. Mayfly code with
cpu_dmb() help avoid stalled memq_ull_rx processing when
rx_demux is to be executed using mayfly.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Use the HAL event clear functions to clear EVENTS
instead of accessing the registers directly.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the HAL event clear function to clear the RTC EVENT
instead of accessing the register directly.
This allows using an updated version of the RTC HW models
which generate level interrupts (as the real HW) and in
which if the EVEN register is not properly cleared,
the interrupts are kept high, resulting in the interrupt
handler reentering immediately after exiting.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
The CIS Central uses the algorithm described in BT Core 5.4 Vol 6,
Part G, Section 2.2 to calculate the Max_PDU value for framed mode.
However, HAP needs the Max_PDU to be calculated according to "Core
enhancement for ISOAL CR" coming with the "Atlanta" update. With this
update, the fragmentation is controlled via a parameter at CIG creation.
Enabling CONFIG_BT_CTLR_CONN_ISO_AVOID_SEGMENTATION will set the
ISO_Interval to 7.5 ms for a 10 ms framed CIG, and calculate Max_PDU to
45/65 for 16/24 kHz.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Use NRF_CCM HEADERMASK register to use correct Additional
Authentication Data (AAD) from ISO PDU header.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix timestamp calculation to use CIS reference anchor point
instead of incorrectly used CIG reference anchor point.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix incorrect use of Max_SDU_Size and SDU_Interval parameters in
ISO Transmit Test Mode code.
Fixes EBQ test LL/CIS/PER/BV-08-C.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Added optional support for read ISO link quality command.
Added link quality member in struct ll_iso_stream_hdr.
Signed-off-by: Mads Winther-Jensen <mdwt@demant.com>
If there are duplicate handles the Controller shall return the
error code Invalid HCI Command Parameters (0x12)
Fixes a failure in the EBQ test LE/AdvExt/Req-01 (part of the Ellisys
quality test suite)
Signed-off-by: Troels Nilsson <trnn@demant.com>
Fixes some incorrect assertions in ull_adv.c on connection establishment
Note that the usual checks using adv->lll.conn doesn't work here,
since ull_periph_setup() NULLs the lll's conn structure pointer
before stopping the ticker
Signed-off-by: Troels Nilsson <trnn@demant.com>
rp_enc_state_wait_ltk_reply() is executed in thread context, so
it is not allowed to move the encryption state machine forward.
Defer LTK reply handling to next prepare event by introducing a continue
state.
Update the unit test to reflect this, and remove the TODO that actually
said there was an issue in the first place.
Signed-off-by: Thomas Ebert Hansen <thoh@oticon.com>
Prevent compiler optimizations from changing the order of assignments in
the ticker API transactions. This could cause lost operation requests
and unexpected behavior under certain race conditions.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Adds a target RPA known unknown cache; it fulfills a similar function
to the existing peer RPA cache, used for scanning of directed
advertisements
Signed-off-by: Troels Nilsson <trnn@demant.com>
- Fix masking for no coded phy
- Allow ISO_Interval < SDU_Interval for framed mode
- Change BT_HCI_ERR_INSUFFICIENT_RESOURCES to
BT_HCI_ERR_CONN_LIMIT_EXCEEDED.
- Prevent starting same CIS twice
- Cancel an initiated CIS creation procedure if terminated before
sending CIS_IND.
- Implement canceling of local CC procedure. Respond to CIS_RSP with
REJECT, if canceled after CIS_REQ was sent.
- Introduce state CIG_STATE_INITIATING for central, to keep track of
initiating CIS connection, in transition between CONFIGURABLE and
ACTIVE.
Fixes EBQ test /HCI/CIS/BC-03-C.
Signed-off-by: Morten Priess <mtpr@oticon.com>
When using the JIT scheduler (CONFIG_BT_CTLR_JIT_SCHEDULING), doing
ticker updates which only change the slot reservation, will fail.
For compatibility, the updates to ull.ticks_slot are preserved, as some
link layer implementations use this to predict collisions.
Signed-off-by: Morten Priess <mtpr@oticon.com>
When executing a local CIS_TERMINATE procedure, the reason sent to the
host must be changed to match the perspective of the initiator.
According to BT Core 5.4 Vol 4, Part E, section 7.1.6, the disconnect
reason must be BT_HCI_ERR_LOCALHOST_TERM_CONN for a local disconnect.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Do not assert in case of race conditions or asymmetric CIS
create/release. There are cases where ll_conn_iso_stream_get_by_group
legally returns NULL, and this should just cause for-loop continuing to
next iteration.
Signed-off-by: Morten Priess <mtpr@oticon.com>
To keep track of requested/pending CIS LLL flushing, change 'flushed'
binary state to 'flush' with states NONE, PENDING and COMPLETE.
This enables CIS teardown to know that a CIS already has a pending LLL
flush, and ULL does not need to initiate it.
Signed-off-by: Morten Priess <mtpr@oticon.com>
While the maximum sizes were already correctly checked by the code, the
minimum sizes of the PDUs were not. This meant that PDUs smaller than
the minimum required length (typically 6 bytes for AdvA) were
incorrectly forwarded up to the Host.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
To prevent failure stopping a ticker node after starting it within the
same ticker_job execution (queued operations), defer the stopping until
after start. This will sequence the operations correctly and prevent
failures.
Signed-off-by: Morten Priess <mtpr@oticon.com>
For config CONFIG_BT_CTLR_CONN_ISO_RELIABILITY_POLICY, set
CIG_Sync_Delay fixed as high as possible, and calculate FT by iteration.
This favors utilizing as much as possible of the Max_Transport_latency,
and spreads out payloads over multiple CIS events (if necessary).
Add exit with cleanup as ll_cig_parameters_commit may fail after
creating new CIG/CIS instances.
Fixes issue #59605.
Fixes EBQ test HCI/CIS/BI-12-C.
Signed-off-by: Morten Priess <mtpr@oticon.com>
There are some calls to ull_adv_aux_pdu_set_clear that are set to 0
but the argument is actually unsigned. This PR replaces the '0'
with '0U'
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
First implementation had a math error where the offset would be the
remainder from division (modulo). It should have been an ISO interval
minus the remainder. Refactored to not use modulo.
Signed-off-by: Morten Priess <mtpr@oticon.com>
In certain scenarios, the semaphore sem_ticker_api_cb may be taken
during attempt to complete a synchronous ticker operation such as start
or stop, but is never released via ull_ticker_status_give. This could
happen if ticker temporarily runs out of ticker operation slots for any
ticker client.
The consequence of ULL stuck waiting in semaphore is anything from
allocation assertions to watchdog timeouts.
This commit also sets a timeout on ull_disable calls, which relies on
the disable callback being invoked from 'done'. Invocation of 'done' is
mandatory, and code shall not allow that not to happen, but to avoid
hiding the real cause, the semaphore now has a timeout which causes
assertion in the location the issue occurred.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Check mayfly_is_running() to determine use of blocking- or non-blocking
version of lll_csrand, instead of k_is_in_isr().
As mayflies run in ISR context for Nordic platform, change will have no
impact on Nordic configuration.
For vendors using META_IRQ mayflies, implementation will now prevent
blocking and reduce latencies and jitter in soft-realtime code.
Signed-off-by: Morten Priess <mtpr@oticon.com>
The separate done memq was introduced to handle the old LLCP
stalling processing of the rx queue; This is no longer an issue with
the new LLCP, so we can remove it
Signed-off-by: Troels Nilsson <trnn@demant.com>
These configs are very tied to the BT_LL_SW_SPLIT implementation,
so it makes sense that these are only visible when that link layer is
used.
For the ones that may be used by other controllers in the future,
a dependency has been added.
Signed-off-by: Rubin Gerritsen <rubin.gerritsen@nordicsemi.no>
Truncation of advertising data has to be done at a PDU
boundary; Including only part of a PDUs advertising data is
not allowed
Signed-off-by: Troels Nilsson <trnn@demant.com>
If the aux scanner already has CONFIG_BT_CTLR_SCAN_DATA_LEN_MAX
advertising data, there is no point in following an aux ptr -
instead flush the data and produce an incomplete report
Signed-off-by: Troels Nilsson <trnn@demant.com>
