Cancelling transfers on suspend contradicts Universal Serial Bus
Specification Revision 2.0, 9.1.1.6 Suspended:
* When suspended, the USB device maintains any internal status,
including its address and configuration.
The internal status definitely includes any pending USB transfers.
If there is a class that wants to cancel transfer on suspend, then the
cancel should be initiated by the class, not the device stack itself.
Update hal_nordic to a version that does not abort all endpoints at
suspend. It seems that aborting endpoints on suspend in nrfx driver was
the actual reason why transfers were canceled on suspend.
Remove transfer retriggering on resume from CDC ACM and Bluetooth class
implementations because transfers are no longer cancelled on suspend.
Other classes do not have any suspend related workarounds implemented.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
When we receive Node Reset message, we schedule work in the system
workqueue to call bt_mesh_reset(). In bt_mesh_reset() we call
bt_conn_disconnect() to disconnect proxy client. bt_conn_disconnect()
will put a buffer to a pool and call k_sem_take to let BT HCI TX thread
process this buffer. Because we stop scanner after disconnecting the
proxy, we can still receive a message at this point. Since BT RX thread
has higher priority than the system workqueue, if we receive a message
while in bt_mesh_reset(), it will be processed when we call
bt_conn_disconnect().
Stop scanner before resetting the mesh to avoid processing of a received
mesh message while in the process of resetting mesh.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
If, while storing app or net key, or changing cdb, we receive another
mesh message that affects setting of the same key that the mesh
currently stores, the new change won't be stored.
The settings subsystem API has rescheduling point inside. The mesh
settings are stored in the system workqueue and by default mesh messages
are processed from BT RX thread which has higher priority than the
system workqueue.
When the case above happens, a new change will be written to the same
cache entry. But this cache entry will be invalidated after leaving
settings API, which in turns will invalidate the new change:
- Receive Config AppKey Add message
cfg_srv.c -> app_keys.c: bt_mesh_app_key_add()
app_keys.c: update_app_key_settings()
app_keys.c: bt_mesh_settings_store_schedule()
- store_pending() in settings.c is scheduled
settings.c -> app_keys.c: bt_mesh_app_key_pending_store()
app_keys.c: store_app_key() called to store new app key
app_keys.c: settings_save_one() calls flash driver and sleeps
- Receive Config AppKey Delete message while in sleep,
which wakes up BT RX thread before returning to the system workqueue
cfg_srv.c -> app_keys.c: bt_mesh_app_key_del()
app_keys.c: update_app_key_settings()
app_keys.c: app_key_update_find() finds entry and returns it
app_keys.c: update->clear = 1
app_keys.c: bt_mesh_settings_store_schedule()
- returning back to bt_mesh_app_key_pending_store() from
settings_save_one()
app_keys.c: update->valid = 0
- the key won't be deleted next time store_pending() is scheduled.
This change moves entry invalidation before calling settings API so that
after returning from settings API, a new change won't be unintentionally
invalidated.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
Allow the streams to be paired when creating unicast group. This will
allow to reuse the same ISO for the paired streams.
Fixes: #51796
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
The img_mgmt_flash_area_id would add processing of
slot2_partition and slot3_partition if they only exist, even if
not used at all.
Signed-off-by: Dominik Ermel <dominik.ermel@nordicsemi.no>
Add BT_AUDIO_CONTEXT_TYPE_UNSPECIFIED as the default supported
context. This is primarily due to the requirement that it
shall always be supported, and the value 0 (previous
default) was invalid.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The call to set_supported_contexts in set_src_supported_contexts
gave the wrong value by reference.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Unit test project for bt_keys_update_usage().
This part of subsys/bluetooth/host/keys.c unit testing.
Signed-off-by: Ahmed Moheb <ahmed.moheb@nordicsemi.no>
A follow up to commit 1d7a077e11 - apparently the PPP interface should
not be brought entirely down internally, as this can break further
communication with the host. Because of that, reintroduce functionality
that used to be covered by net_if_carrier_down() (which basically
skips the L2 enable(false) call), but limited to PPP scope only.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Fixes#53786
The variable slot_plus_us is assumed to be a number in us unit.
To assign the maximum possible CTE length to this variable,
BT_HCI_LE_CTE_LEN_MAX has been used while this parameter is defined
in a 8us units (it's 0x14=20 that corresponds to 160us).
To convert this number to us, it's needed to be multiplied by 8.
Signed-off-by: Saleh Mehdikhani <saleh.mehdikhani@unikie.com>
The bt_has_register_param's preset_sync_support and independent_presets
make sense only if CONFIG_BT_HAS_PRESET_COUNT is non-zero meaning the
CONFIG_BT_HAS_PRESET_SUPPORT is enabled.
Otherwise, those parameters shall be skipped.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Add command to perform the Periodic Advertising Set
Info Transfer procedure, which transfers information
about a periodic advertising set to a connected device.
In essence this is just PAST but from the advertiser
instead of a 3rd device.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
In ascs.c we had a case where we assigned stream->conn
without taking the ref.
In bt_audio_stream_attach we did not check if stream->conn
was NULL before taking a reference, causing multiple calls
to ase_config to take multiple references.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The purpose of this commit is to facilitate future improvements to the
LPN and friendship feature. By being able to identify friendship related
messages on the advertising layer, it will be possible to treat these in
a specific manner so that message exchange between LPN and friend
devices can be conducted in a more power efficient manner.
This commit adds the following:
- A separate tag for friendship related messages
- A optional separate advertising set for friend related messages
Signed-off-by: Anders Storrø <anders.storro@nordicsemi.no>
Adds Kconfig option to enable tuning the Friend receive delay window.
Friend poll events can be made more efficient by compensating for the
time the non-ideal advertising latency otherwise would add to the receive
delay.
Signed-off-by: Anders Storrø <anders.storro@nordicsemi.no>
This changes some of the MCUmgr transport CMake select statements
to depends on, this is to align with other in-tree symbols that
do not pull in whole subsystems and instead only allow selection
if those subsystems are already enabled.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
- Adds mmc.c
- Edits sd.c to init and probe MMC
- Adds mmc init to sd_init
- Some functions from sdmmc.c should be in sd_ops because
they can be used by both sdmmc and mmc.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
In sd subsystem, made one header for all the init functions
instead of having a bunch of header files with one function.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
split reusable portions of SDMMC protocol code into sd_ops.c, so other
SD protocols can use these functions directly without compiling in the
SDMMC subsystem.
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
Add stub code for SDIO support, capable of verifying card responds to CMD5.
This commit also changes the architecture of the SDIO probe step to make
adding new protocol support more streamlined, and enable compiling out
support for undesired protocols.
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
This commit updates all in-tree code to use `CONFIG_CPP` instead of
`CONFIG_CPLUSPLUS`, which is now deprecated.
Signed-off-by: Stephanos Ioannidis <stephanos.ioannidis@nordicsemi.no>
This commit moves the files under `subsys/cpp` directory to the
`lib/cpp` directory because the C++ ABI runtime library and the
standard C++ library components are not a "subsystem" (aka. API) in
conventional sense and is better described as a "library."
Classifying the C++ ABI runtime library and the standard C++ library as
"libraries" instead of "subsystems" also better aligns with how the
existing C standard library implementation (`lib/libc`) is handled.
Signed-off-by: Stephanos Ioannidis <stephanos.ioannidis@nordicsemi.no>
Wi-Fi bands are regulated by a governing body depending on operating
country, add support for the user to provide a country of operation as a
hint to the Wi-Fi chipset.
Ideally if the chipset supports this is all handled internally, in that
case "get" is useful but for testing and other usecases add a "set" as
well, similar to "iw reg set" or "country_code=" configuration in
hostapd/wpa_supplicant in Linux world.
This add a new offload API operation "reg_domain" that can be used to
either get or set the regulatory information.
The validation is left to the underlying chipset, shell only does basic
validation, (XY/00).
This is just a regulatory hint to the chipset, there could be other
regulatory hints e.g., beacon that can override this configuration, so,
an additional option to force this setting despite other hints is also
given for testing purposes.
FYI, the standard database used is [1].
[1] - https://git.kernel.org/pub/scm/linux/kernel/git/sforshee/wireless-regdb.git/tree/db.txt
Signed-off-by: Krishna T <krishna.t@nordicsemi.no>
under certain conditions the current implementation did not maintain
the desired sort order.
Signed-off-by: Pascal Brogle <pascal.brogle@husqvarnagroup.com>
Add the ability to change supported contexts in PACS.
Enhance the context command in the shell module to make it configurable.
Signed-off-by: Szymon Czapracki <szymon.czapracki@codecoup.pl>
base name or name can contain up to two shorts.
(object id & object intstance or resource id & resource instance id)
Signed-off-by: Pascal Brogle <pascal.brogle@husqvarnagroup.com>
The purpose of the change was to have the Zephyr network stack
internally add the IGMP all systems 224.0.0.1 multicast address to a
multicast membership list so that multicast hash filter implementations
can add this address to the hash filter.
Fixes#53548
Signed-off-by: Chamira Perera <chamira.perera@audinate.com>
The call to bt_adv_reset_adv_pool is only required in the
broadcaster role, and not in the observer role, regardless
of the setting of CONFIG_BT_EXT_ADV
As the code was this call was also made for the observer.
Note: handling of the setting of CONFIG_BT_EXT_ADV
is already handled in the bt_adv_reset_adv_pool function
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
BOOT_MAGIC_SZ and BOOT_MAX_ALIGN were used in the header without
including bootutil/bootutil_public.h. This change remove the need of
the inclusion by making the dependency private.
Fixes#52095
Signed-off-by: Guillaume Lager <g.lager@innoseis.com>
Fix a bug in setting MAC address using net_if API, the API doesn't do a
memcpy but just stores the pointer and shell was passing stack pointer.
We can use dynamic allocation but freeing the memory for the MAC address
would be trickier, so, use the net management API and let the underlying
drivers figure out the MAC address memory management.
Signed-off-by: Krishna T <krishna.t@nordicsemi.no>
There is special handling done for resolved addresses to convert
them to "regular" addresses for the upper layers.
This commits adds two helper functions to check if they are
resolved, and if so, then properly copied.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
We only expose random/public address types to the upper layers.
This is done by checking if the address type of events are
resolved addresses, and if so, then we translate them to
public/random.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
As we have to provide LL addresses in big endian to userspace to be
POSIX compliant and we also do not want to reserve extra space for
such addresses, bff6a5cce5 introduced
a change that swaps address bytes in place in the packet before
returning the packet with LL address pointers to userspace.
Unfortunately a regression sneaked into the code base while doing
so: The byte swapping was duplicated when using 6LoWPAN compression
and the byte swapping caused decryption to fail in some cases,
see #53630. This commit fixes the problem.
Fixes: #53630
Signed-off-by: Florian Grandel <jerico.dev@gmail.com>
The L2 networking layer checks for return value from enable, but
Ethernet is not checking and always returns 0, so, relay the return
value from the Ethernet driver to networking stack.
This fixes the issue of interface start failing but interface still
being up.
Signed-off-by: Krishna T <krishna.t@nordicsemi.no>
This add storage abstraction to allow switch between different flash
APIs. This remove the erase command at updatehub core and move it to
storage init phase.
Signed-off-by: Gerson Fernando Budke <gerson.budke@ossystems.com.br>
