Support for using Link Time Optimization (LTO) when
building application with open source Bluetooth Low
Energy Controller. This reduces code space usage.
Current Memory Usage at the time of this PR (hci_ipc):
BT_CTLR_OPTIMIZE_FOR_SPEED:
Memory region Used Size Region Size %age Used
FLASH: 260112 B 256 KB 99.22%
RAM: 60136 B 64 KB 91.76%
SRAM1: 0 B 64 KB 0.00%
IDT_LIST: 0 B 32 KB 0.00%
BT_CTLR_OPTIMIZE_FOR_SIZE:
Memory region Used Size Region Size %age Used
FLASH: 232292 B 256 KB 88.61%
RAM: 60128 B 64 KB 91.75%
SRAM1: 0 B 64 KB 0.00%
IDT_LIST: 0 B 32 KB 0.00%
BT_CTLR_OPTIMIZE_FOR_APP_DEFAULT:
Memory region Used Size Region Size %age Used
FLASH: 232292 B 256 KB 88.61%
RAM: 60128 B 64 KB 91.75%
SRAM1: 0 B 64 KB 0.00%
IDT_LIST: 0 B 32 KB 0.00%
BT_CTLR_LTO:
Memory region Used Size Region Size %age Used
FLASH: 221484 B 256 KB 84.49%
RAM: 60120 B 64 KB 91.74%
SRAM1: 0 B 64 KB 0.00%
IDT_LIST: 0 B 32 KB 0.00%
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Remove legacy LL optimize for speed dependency required to
support encryption feature in Controller where crypto was
performed in highest priority ISRs to setup encryption in 3
radio events (this now requires 5 radio events in split LL).
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add a new option to split off the building of the HCI layer of the
controller. This layer is not needed by the controller unit tests, and
becomes problematic with the planned change to use devicetree for HCI
drivers.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
Ensure that in LLCP reference to node_rx is cleared when
retention is NOT used, to avoid corruption of node_rx later
re-allocated
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
In case a CPR is intiated but rejected due to CPR active on
other connection, rx nodes are leaked due to retained node not
being properly released.
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
The mic_state is a guarded field in event_done_extra where it was
accessed outside of its guard.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
For vendor datapaths that do not use a netbuffer to write SDUs
into, the callee of sdu_write currently has to keep track itself
of how much data has been written to the current SDU; This is wasteful
since ISOAL already keeps track of that. Add an sdu_written parameter
to the callback to inform the callee how much has been written to the
current SDU already so the callee can write using the correct offset
Signed-off-by: Troels Nilsson <trnn@demant.com>
This function was used to shortcut HCI for combined host + controller
builds. It doesn't provide much value and adds complexity to the HCI
driver interface, so just remove it. This means vendor-specific HCI
commands is now the only way for the host to access the same
information.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
Enabling vendor-specific extensions also implies support for vendor
events, so a separate Kconfig option for that is unnecessary.
One small additional thing this requires is the use of the
__maybe_unused annotation, since there's no-longer a single Kconfig
option that the controller hci.c can use to know that the vendor event
helper symbols are needed.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
The naming of these two options was problematic, since it's both of them
are about vendor extensions, even though one has _EXT in the name and
the other doesn't. Just merge one option into the other. This has a
slight overhead on the controller side of enabling some more vendor
features if BT_HCI_VS is enabled, but that should hopefully be
acceptable.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
Derive BT_CTLR_ISO_TX_BUFFER_SIZE from BT_ISO_TX_MTU to have
optimal Controller memory allocations.
BT_CTLR_ISO_TX_BUFFER_SIZE can be set lower than
BT_ISO_TX_MTU in which case upper layer can send fragmented
SDU to the Controller.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Rename struct bt_hci_iso_data_hdr to bt_hci_iso_sdu_hdr, and
struct bt_hci_iso_ts_data_hdr to bt_hci_iso_sdu_ts_hdr.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
In ull_disable, it is imperative that the callback is set up before a
second reference counter check, otherwise it may happen that an LLL done
event has already passed when the disable callback and semaphore is
assigned.
This causes the HCI thread to wait until timeout and assert after
ull_ticker_stop_with_mark.
For certain compilers, due to compiler optimizations, it can be seen
from the assembler code that the callback is assigned after the second
reference counter check.
By adding memory barriers, the code correctly reorders code to the
expected sequence.
Signed-off-by: Morten Priess <mtpr@oticon.com>
Implement LL_ASSERT_INFO1 and LL_ASSERT_INFO2 for triggering assertions
with parameter information provided for vendor core dump.
Adds Kconfig CONFIG_BT_CTLR_ASSERT_VENDOR by which the new LL asserts
map to BT_ASSERT_VND macros, which shall be implemented in the
debug_vendor_hal.h of the platform. If not enabled, LL_ASSERT_INFO will
map to existing BT_ASSERT_MSG with parameters printed.
Add use of LL_ASSERT_INFO2 where ull_ticker_stop_with_mark() result may
assert.
Signed-off-by: Morten Priess <mtpr@oticon.com>
When building ISO applications on a device that has the host
and controller on the same core, the application developer
shouldn't have to set controller specific configurations.
Signed-off-by: Rubin Gerritsen <rubin.gerritsen@nordicsemi.no>
Use reschedule margin as minimum ticks_slot when ticker with
reschedule expires. This will ensure not too many unreserved
tickers expire in close proximity that can lead of pile up
of CPU processing time and eventually causing timeout
callbacks to be delayed. This mitigate possible LL asserts
due to overhead in start of radio events.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Enable use of ticker slow window yield feature under mesh
usecase. This is make sure scan window either slides within
is window or skips to next interval so that it does not
lead to advertising event being skipped.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix short prepare preempt timeout start, such that if there
are any normal prepare before a short prepare in the
pipeline, then find the short prepare and start its preempt
timeout.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add prepare pipeline assertion checks to detect faults
like stopping preempt timeout without prior start. Also,
check if the preemptor param is correct when preemption
is performed. Check that the prepare pipeline is not
infinitely looping.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add assertion check for use of scan aux context so that
scan aux context is not assigned again while there is
already an existing LLL scheduling in use by the scan
context.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix BT_CTLR_EARLY_ABORT_PREVIOUS_PREPARE to setup new
shorter preempt timeout irrespective of whether there
is a previous prepare enqueued in the prepare pipeline.
Relates to commit d573951f0d ("Bluetooth: Controller:
Revert back early abort of previous prepare").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When building ISO applications on a device that has the host
and controller on the same core, the application developer
shouldn't have to set controller specific configurations.
Without this change the samples iso_receive and iso_broadcast
will fail to run on NRF52 series devices as the samples
try to set up two streams but the controller is configured
to support only one.
Controller unit tests that were previously only enabling the
controller specific ISO configurations now also enable the
top-level ISO configurations to ensure that the default
stream count is properly configured.
Signed-off-by: Rubin Gerritsen <rubin.gerritsen@nordicsemi.no>
Update the documentation of `BT_CTLR_ADV_DATA_CHAIN` Kconfig symbol to
say that it's only supported in *non-connectable* extended advertising.
Signed-off-by: Théo Battrel <theo.battrel@nordicsemi.no>
Ignore failure to Broadcast ISO remove data path when
terminating a BIG without a data path being setup.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When multiple advertisers overlap over time, the random
delay being applied need many ticker operation context.
This causes the ticker_update interface to return error
causing assertion in the Controller.
This is fixed by not applying random_delay value to the
overlapping advertiser that does not get to use the radio.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add aux offset population assertion check, to ensure the
aux offset calculation is completed before the primary
PDU is setup for transmission.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Change default to using advertising coding in ACL connection
when establishing connections on Coded PHY.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Add missing check for Coded PHY support when creating
extended advertising set.
Added a fallback in Controller for upper layers requesting
Coded PHY while Controller is implemented without Coded PHY.
In this case, 1M PHY will be used in extended advertising
for testing purposes.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Ignore LE 2M PHY being set in Initiating_PHYs of the
LE Extended Create Connection command.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
UBSan was showing the following error:
```
runtime error: left shift of 137 by 24 places cannot be represented in
type 'int'
```
Cast the value to `uint32_t` to make UBSan happy.
Signed-off-by: Théo Battrel <theo.battrel@nordicsemi.no>
Implement test version of the LE Create BIG command ll_big_test_create
(BT_HCI_OP_LE_CREATE_BIG_TEST) in the controller.
This enables use of CONFIG_BT_ISO_ADVANCED, which supports use of
the following BIG/QoS configuration parameters (otherwise derived):
- ISO_Interval
- NSE
- Max_PDU
- BN
- IRC
- PTO
Signed-off-by: Morten Priess <mtpr@oticon.com>
Fix regression in BT_RX_STACK_SIZE use due to
commit b91728619c ("Bluetooth: host: remove
`CONFIG_BT_RECV_BLOCKING`").
Profiling of Controller Rx thread has been mentioned in
commit 586ba9fd13 ("Bluetooth: Controller: Increase
BT_RX_STACK_SIZE for BT_HCI_RAW enabled").
Also, move the TX_STACK_SIZE value as default in Kconfig
from the explicit value being set in hci_ipc sample.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Generic node_rx_hdr struct contains data not used for done event, so this
pdu specific data is moved to node_rx_pdu to minimize memory usage for
MFIFO_done alloc.
This however calls for a considerable write-through replacing the generic
node_rx_hdr with node_rx_pdu when relevant and needed.
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Use CONFIG_BT_CTLR_CENTRAL_ISO instead of CONFIG_BT_CENTRAL in
declaring central CIS CREATE procedure context data. In some configs the
size of this data structure is bound by central CIS CREATE, so doing it
correctly will save memory in the procedure context allocation.
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
If central initiates incompatible procedure after having replied (with
_IND), peripheral fails to disconnect as spec'ed.
Fix by correctly setting the INCOMPAT flag to reserved on IND receipt to
enforce the disconnect.
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
This commit removes `bt_hci_evt_get_flags`, and the associated flag
symbols. These symbols no longer serve a purpose in the driver interface
after the removal of `bt_recv_prio` from the interface.
The implementation of `bt_hci_evt_get_flags` is distributed to its
still-existing use sites, `hci_core` and the `hci_driver` in the
controller.
The duplication of code is intentional. This is no longer part of any
interface, and the host-developers expect the host copy of the function
to change or dissapear entierly. This will allow the host implementation
to progress without fear of breaking the controller implementation.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
This config selects a variant of the HCI driver interface that spills
out host internals unto the drivers and even the Zephyr controller. It
will now be removed in favor of driver interfaces that hide the
internals of the host.
The new default is `CONFIG_BT_RECV_WORKQ_BT`.
Any references to the removed kconfig are refactored out.
Any out-of-tree driver using the removed interface can be easily adapted
by copying the following implementations into the driver as private
functions:
- `hci_driver.h:BT_HCI_EVT_FLAG_RECV_PRIO`
- `hci_driver.h:BT_HCI_EVT_FLAG_RECV`
- `hci_driver.h:bt_hci_evt_get_flags`
- `hci_raw.c:bt_recv_prio`
In combination these symbols function as a interface adapter. These
symbols will be removed in this PR in subsequent commits.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
In the simulated bsim boards there is no jittery delay
from a (D)PPI event until a TIMER starts.
The timer starts instantaneously so we do not need to
compensate for it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the HAL always when setting the TIMERs CC registers,
(also for coded Phy) to ensure things work properly for simulation.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Add HCI input parameter validation according to spec for passing
qualification tests.
- Conditional compile of LE Set PA Receive Enable feature
- Store BIGinfo encryption state and number of streams
- Do not report BIGinfo if phy is invalid or unsupported
- Return disallowed ll_read_iso_tx_sync uses sync recever handle
- Validate BIS indices in ll_big_sync_create
Signed-off-by: Morten Priess <mtpr@oticon.com>
