Instead of relying only on the `tx` and `rx` qos pointers,
we extract further information from the
bt_hci_evt_le_cis_established event to properly determine if we
can actually send or receive data.
This is useful to help determine which data paths to setup,
and whether to reject requests to send.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
If CONFIG_SHELL_STATS is disabled, shell->stats is NULL and
must not be dereferenced. Guard against it.
Fixes#44089
Signed-off-by: Alexej Rempel <Alexej.Rempel@de.eckerle-gruppe.com>
Remove call to bt_audio_stream_iso_listen which
calls bt_iso_server_register for the unicast client.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The bluetooth ISO shell module now supports the newly
added central and peripheral ISO configs.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Move a few unicast functions around so that we reduce the number
of #if defined guards. No code has changed, but a single
prototype has been added.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Add two new Kconfigs: BT_ISO_CENTRAL and BT_ISO_PERIPHERAL
that is used to do central or peripheral only builds,
similar to the BT_CENTRAL or BT_PERIPHERAL Kconfigs.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
output of ztest is not easy to ready with many testsuites running in
sequence, make it a bit more readable.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
There is a race condition in a multicore system that happens when the
idle thread in a CPU checks if the state was forced, if not it will
call the policy manager. If a secondary core forces a state after that
this point the value returned by the policy will be rewritten.
Another case is, if a state is forced while a CPU is sleeping,
when this CPU resumes, the forced bit is cleared and the forced state
is never be used.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
In pm_system_suspend there is a goto label that is supposed to be used
to do cleanup before exit the function but it is not doing anything
right now. Just remove it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This extends client implementation with Read Presets Request operation
support.
The implementation can maintain up to BT_L2CAP_TX_BUF_COUNT control
point requests so that user could perform/queue another Control Point
Operation if one is ongoing. E.g. preset can be selected from Read Preset
Response notification context or perform multiple operations if EATT
enabled.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
The periodic CTE_REQ disable command, requested by Host, may be locked
until connection is dropped due to missing CTE_RSP from peer device.
That is caused by implementation of CTE_REQ disable and CTE_REQ
control procedure handling.
The procedure is marked as active when CTE request was send to peer
device. It is marked as inactive after completion of the procedure.
That caused locking of CTE disable on a semaphore.
The BT 5.3 Core Spec, Vol 4, Part E, section 7.8.85 says the HCI_LE_-
Connection_CTE_Request_Enable should be considered active on a conne-
ction from when Host successfully issues the command with Enable=0x1
until a command is issued with Enable=0x0 or single LLCP CTE request
has finished (CTE_Request_Interval=0x0). Also there is a clarification
from BT SIG that the command with Enable=0x0 does not affect any
initiated LLCP CTE request. That means Controller is allowed to finish
already started procedure and it is not allowed to start new LLCP CTE
request procedure after completion of the command with Enable=0x0.
Taking that into account, there is no need to synchronize ULL and LLL
in regard of disable the LLCP CTE request while the procedure is
pending. Controller is free to complete the procedure or terminate it.
The change removes all code related with cte_req.is_active, disable
callback and waiting of ULL for LLL to finish the LLCP CTE request.
The ULL will complete the HCI_LE_Connection_Request_Enable with
Enable=0x0 immediately. In case the procedure is disabled in before
the response arrives, then further processing of the response is
dropped and the procedure context released.
The context is not released by the code responsible for disable
handling, to have single place where it is done.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
There is an error in the condition that checks if new CTE_REQ should
be started after end of connection event. The condition verifies if
counter req_expire is set to zero. Then new CTE_REQ is started
irrespectively to CTE_REQ being disabled.
req_interval is used to store information if the CTE_REQ is:
- periodic, then value doesn't equal zero,
- single shot or disabled, then value equals zero.
The condition should verify if the req_interval is not zero and
req_expire is not zero. The second part of the if condition is
required to avoid starting next CTE_REQ until last one has been
completed.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
Some of control procedures has set ctx->rx_opcode in rp_comm_tx
function to values that state there is an expected PDU.
This is not always true. In case of procedures that do not expect
any response from remote device, the ctx->rx_opcode should be set
to PDU_DATA_LLCTRL_TYPE_UNUSED.
In the worst cases scenario, when the same control procedure
is executed locally and remotely, correct response PDU may
cause an assertion in ull_cp_rx. It could happen because
of wrong ctx->rx_opcode value. A packet with opcode that is
set in remote and local control procedure context rx_opcode
will be treated as expected value for both. That is a situation
that cannot happen.
The commit changes the assignments of rx_opcode to fix the problem.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
If the echo flag is disabled the cmd_buff isn't printed and the
characters in it must not be counted in this function.
Signed-off-by: Gerhard Jörges <joerges@metratec.com>
Adds a check that number of records to be encoded does not exceed the
maximum limit configured through Kconfig.
Signed-off-by: Veijo Pesonen <veijo.pesonen@nordicsemi.no>
This commit fixes the build errors for the testcases in
the tests/bluetooth/init folder
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
The tests/bluetooth/ctrl_user_ext test fails due to
compilation errors, which are fixed with this PR
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
Make use of the status field, reported by TCP, in the socket receive
callback. This allows to differentiate a graceful connection shutdown
from actual errors at TCP level (transmission timeout or RST received).
In case of error reported from TCP layer, set a new SOCK_ERROR flag on
the socket, and store the error code in the net_context user_data.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Make use of the status argument in the recv_cb() callback function -
instead of blindly reporting ECONNRESET whenever TCP context is
dereferenced, indicate whether an actual error condition happened (by
setting respective errno value) or a graceful shutdown took place (by
setting status to 0).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Previous check in the if-statement would never allow to send last
segment if msg->len + 2 == MTU * x.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
Added macro to test that a value falls within a defined range, including
the provided end points.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Get actual keys from references when PSA crypto is enabled.
A more secure method should be implemented once 802.15.4 platforms
support other than clear text keys.
Signed-off-by: Eduardo Montoya <eduardo.montoya@nordicsemi.no>
Remove NET_TCP_BACKLOG_SIZE from KConfig and from test,
because it's not present anymore in current version of TCP stack.
Signed-off-by: Andrey Dodonov <Andrey.Dodonov@endress.com>
This allows an application to inspect a mcumgr img upload command to
provide additional information for acceptance or rejection of it, and
makes the previous private version compare function public so that
application code can call it.
Signed-off-by: Jamie McCrae <jamie.mccrae@lairdconnect.com>
Enable ZTEST_DO_THE_SHUFFLE to shuffle the order tests are ran.
Additional configs ZTEST_DO_THE_SHUFFLE_SUITE_REPEAT_COUNT
ZTEST_DO_THE_SHUFFLE_TEST_REPEAT_COUNT specify the number of times
the test or suite is executed.
Signed-off-by: Al Semjonovs <asemjonovs@google.com>
This allows an application to inspect a mcumgr file access command and
either allow it or deny it with a result code.
Signed-off-by: Jamie McCrae <jamie.mccrae@lairdconnect.com>
After introducing SO_SNDBUF socket option, a possible deadlock situation
slipped into the TCP implementation. The scenario for the deadlock:
* application thread tries to send some data, it enters
net_context_send() which locks the context mutex,
* internal context_sendto() blocks on a TX packet allocation, if the
TX pool is empty rescheduling takes place,
* now, if at the same time some incoming packet has arrived (ACK for
example), TCP stack enters tcp_in() function from a different
thread. The function locks the TCP connection mutex, and tries to
obtain the SNDBUF option value. net_context_get_option() tries to
lock the context mutex, but it is already held by the transmitting
thread, so the receiver thread blocks
* when TX packet is available again, the transmitting thread unblocks
and tries to pass the packet down to TCP stack. net_tcp_queue_data()
is called which attempts to lock the TCP connection mutex, but it is
already held by the receiving thread. Both threads are in a deadlock
now with no chance to recover.
Fix this, by obtaining the SNDBUF option value in tcp_in() before
locking the TCP connection mutex.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Thread network makes no use of Solicited-node multicast addresses,
thereby do no create them on the interface to save multicast address
entries for important ones.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
There is a risk of deadlock in case net_if APIs are used from within
net_mgmt handlers as both module APIs are protected with their own
mutexes.
The scenario observed with OpenThread happend when
NET_EVENT_IPV6_ADDR_ADD/NET_EVENT_IPV6_MADDR_ADD events were processed.
The net_mgmt mutex is locked when both, an event handler is being
processed (from a separate net_mgmt thread) and when an event is raised
(for example when a new address is added on an interface). In case a
net_mgmt handler tried to use some mutex-protected net_if API, we could
end up in a deadlock situation - the net_mgmt would wait for the net_if
mutex to release, while some other thread (in this case main during
initialization) could wait within some net_if function, pending on
net_mgmt mutex to be released to notify the event.
Fix this, by preventing net_if APIs from being used from within OT
net_mgmt handlers.
Additionally, simplify the net_mgmt handlers logic, by making use of
additional info provided with an event. Instead of blindy assuming that
recently added address was the last on the list (which might not always
be the case, if addresses are added/removed dynamically), read the
actual address being added from the net_mgmt_event_callback structure.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The audio stream receive callback now contains a
recv_info struct, which contain crucial information
such as timestamps and packet validity.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The values represent an enum, and it makes sense for the
enum values to follow the enum type name, so an
_DIR_ infix was added to the values.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The PAC_TYPE_UNUSED would be breaking
enum rules, as it is setting and comparing
a value outside the enum range.
Instead we check if the `dir` field has been
set to SINK or SOURCE.
The reason why this still works, is that
we memset the struct unicast_client_pac
causing `dir` to become 0. This still
does not really follow the rules of enums, but
it is the best we can do without adding another
value to determine if a struct unicast_client_pac
is unused or not, without adding another value to
public enum struct.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Many functions and struct fields had the directory/type
value, but named in different ways and stored in different ways.
This change updates all uses of it to use the same name
and type.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The enum bt_audio_pac_type does not fully
represent the what the value is used for.
The typical use of it is not for just published
audio capabilities (PAC), but rather describes the
type, or direction, of audio endpoints.
For the unicast client, the type/direction is
relative to the unicast server.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The unicast_client_ep_init took a ID value that was
always 0. Moved the initialization to the value to
the function, instead of an argument. The value
is only used for the unicast server, so it can
probably be removed from the endpoint
struct for unicast clients only. To be optimized later.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The endpoint.type field was unused as it was
always LOCAL for the unicast server and always
REMOTE for the unicast client, and the few places
where we actually checked it, we already had a
check for acl->role which gave the same value.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The PSM of the first channel in the connection request is used, but the
PSM value is used for all of them on the receiving side.
Signed-off-by: Herman Berget <herman.berget@nordicsemi.no>
There were some errors in the detection of ECRED connection collisions,
so the retry was triggered incorrectly.
The number of channels requested in the retry was wrong in some cases,
using the number of channels from the received request instead of the
sent request.
Signed-off-by: Herman Berget <herman.berget@nordicsemi.no>
This embeds the log message source IDs inside the origin unit
as module IDs in Sys-T messages. This allows Sys-T message
parsers to see where the log messages are coming from.
This is enabled by default if using Sys-T catalog messages as
the collateral XML file contains the information to interpret
the module ID.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Devices using the MCUX FlexSPI flash driver erase their flash too
slowly, causing the usb stack to freeze, and USB DFU utilities to
report a timeout. Enable IMG_ERASE_PROGRESSIVELY for SOCs using the
MCUX FlexSPI driver to prevent this error.
Fixes#45359
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
SDMMC busy wait timeout was incorrectly waiting for
CONFIG_SD_DATA_TIMEOUT microseconds, but should be waiting for
CONFIG_SD_DATA_TIMEOUT milliseconds. Multiply wait value by 1000.
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
