Commit adds destruction of the persisted in PSA ITS key if
mesh does not own it (zero bit in the bitmap of persisted keys).
This is not standard mesh behavior, but might happen
if something happens between removing key data in mesh and
in the crypto library (for example power off in bettwen).
Previously, mesh wasn't able to import key with gotten stuck
key id. The current fix reproduces more robust behavior.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Fix Max Chunk reporting during Block_Information_Get:
If friendship is established, then max possible chunk size for transfer
is reported according to friend's queue size.
ChunkSize = (FndQ sz * Bytes per seg (12)) - Opcode(1) - Chunk_Num (2)-
8 byte MIC (max)
= (FndQ sz * 12) - 11
This allows all segments of Chunk_Transfer messages to fit in a friend
queue. If this is not done, friend queue keeps overflowing
systematically causing unstable PULL_MODE transfers.
Fix Chunk requests:
Adapt number of requested chunks according to current friend queue size.
If friend queue is too small, at least 1 chunk is requested.
Signed-off-by: Omkar Kulkarni <omkar.kulkarni@nordicsemi.no>
Commit removes unused host hci custom functionality
that was moved to mesh by misinterpretation of it.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Commit changes BT_MESH_DFU_FWID_MAXLEN option to be
compliant with BLE Mesh specification.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Commit fixes bug when advertisements were disabled and stopped
during suspending without unreferencing allocated advertisements.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
The va.h file used both uintX and bool without including
the appropriate header files, which can cause a compile issue
if the file is included in the wrong order.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Bluetooth had two public types with similar name _bt_gatt_ccc and
bt_gatt_ccc, but for absolutely different purposes.
That caused misunderstanding of relationship of them and cases
where to use which one.
Commit changes name of _bt_gatt_ccc to more suitable by usage and
improves documentation of it.
Additionally, it changes name of BT_GATT_CCC_INITIALIZER
to correspond the type name.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Private Node Identity advertisement on a subnet should stop as soon as
the network is removed.
Signed-off-by: alperen sener <alperen.sener@nordicsemi.no>
The publication period is defined by Steps and Resolution.
At BLE Mesh protocol level, the period is encoded in a byte
where first LSB 6 bits correspond to PerSteps and the last
two to PerRes. There is an issue on how the code at
`subsys/bluetooth/mesh/shell/cfg.c:model_pub_set` is encoding
these two values into the publication period byte.
This is commit fixes issue #87780
Signed-off-by: Raúl Gotor <raulgotor@gmail.com>
Commit fixes bug when corresponding group ID was never
incremented for independent corresponding relationships.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Commit removes device key candidate if provisionee failed
during device key refresh procedure or mesh reset has been
called in between. Otherwise, device key candidate gets
stuck in the internal trusted storage of crypto library.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
When the board has been provisioned and the board is rebooted
it was not possible to connect to the board through the proxy
servcie. This fixes that the Mesh Proxy servive is restarted
after the board is rebooted.
Signed-off-by: Ingar Kulbrandstad <ingar.kulbrandstad@nordicsemi.no>
Previously, PB-GATT unprovisioned advs would continue being sent if a
device is provisioned or in the process of being provisioned (an active
provisioning link) over PB-ADV.
This commit introduces the `link_cancel` callback for the `prov_bearer`
struct. When a provisioning link is established over a prov bearer, all
other (active) prov bearers will be suspended through this callback, and
re-enabled when the link is closed. An exception is made for PB-REMOTE,
which can utilize the other bearers.
The call to `bt_mesh_pb_gatt_srv_disable` in settings.c is removed. This
is called in `mesh_commit`, and is not needed as this is only used
during boot, at which point the Mesh provisioning service has not yet
been enabled/registered, meaning the function does nothing.
Signed-off-by: Håvard Reierstad <haavard.reierstad@nordicsemi.no>
When the GATT server is enabled the proxy advertiser
server need to be stopped. This solved a bug where the
Mesh Proxy Service is not restared.
This fix is done to fix several failing PTS test that
was failing when CONFIG_BT_MESH_ADV_EXT_GATT_SEPARATE
was enabled. Calling the stop_proxy_adv was left out
in a prevouis commit when doing code cleanup.
Signed-off-by: Ingar Kulbrandstad <ingar.kulbrandstad@nordicsemi.no>
The Generic Provisioning protocol requires that a transaction must be
acknowledged before the next Generic Provisioning PDU is processed
(section 5.3.3, MshPRT_v1.1).
In the current PB-ADV implementation, if there is an ongoing transaction
(link.tx.adv points to a valid advertisement), any received Generic
Provisioning Start and Continuation PDUs are ignored.
This creates a scenario where a remote device misses the Transaction
Acknowledgment PDU and retransmits the last PDUs, but they are ignored
by the local device because the local device has already started sending
the next Provisioning PDU.
In particular, in the observed failure, the Provisioner missed the
Generic Transaction Acknowledgment PDU during the transmission of the
Provisioning Data PDU.
This commit modifies the processing logic for Generic Provisioning Start
and Continuation PDUs by checking if the received PDU has already been
received. If so, it attempts to re-send the Generic Transaction
Acknowledgment PDU.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
This fixes an issue where a subsequent Config Model Subscription Add
message with the same virtual address UUID would result in a Config
Model Subscription Status message with the Address field incorrectly set
to 0.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
This updates the help text for the Kconfig option BT_MESH_WORKQ_SYS to
take into account the change made in PR #84282 which causes the host to
no longer return -ENOBUFS. Since the mesh will now instead block the
work queue, a note has been added about a potential consequence of this.
Signed-off-by: Ludvig Jordet <ludvig.jordet@nordicsemi.no>
Commit removes weak attribute and renames some functions
in crypto_psa.c since there is no centralized distribution
of the PSA key ID in bsim tests and no necessity to
reimplement native mesh approach.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
We need to make sure that within Zephyr different users of the PSA APIs
don't interfere with each other because of using the same numerical IDs
for persistent assets.
This takes care of the PSA key IDs when using persistent keys through
the PSA Crypto API.
See the comments in `<zephyr/psa/key_ids.h>` for more information.
This removes the recently-introduced Kconfig options that allowed changing
the base IDs subsystems were using for their persistent keys.
Signed-off-by: Tomi Fontanilles <tomi.fontanilles@nordicsemi.no>
The BLE acronym is not an official description of Bluetooth
LE, and the Bluetooth SIG only ever refers to it as Bluetooth
Low Energy or Bluetooth LE, so Zephyr should as well.
This commit does not change any board or vendor specific
documentation, and the term BLE may still be used in those.
It will be up to the vendors to update it if they want,
since many of them are using the term BLE in their
products.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Commit removes the double cdb node entry storing
into the settings if device was reprovisioned over
nppi interface.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
Remove START_PENDING flag, and replace with
bt_mesh_adv_is_empty_by_tag
not only make adv_ext.c more readable, but also enhancement
In previous implementation, after proxy started, will check
this flag, and when the flag is true, will direct stop proxy
and re-enter schedule, but has one problem, this flag set true,
but buf will be empty, such as when sending mesh relay messages
will for-each every sets, until find not-active, but also set
previous set START_PENDING flags, so will cause proxy advertising
started->stop->started.
Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
The first why introduce ADV_FLAG_PROXY_START
see here(https://github.com/zephyrproject-rtos/zephyr/pull/58826)
The main core change is to mark ADV_FLAG_PROXY as moving to the last
enabled proxy instead of before.
But introducing the ADV_FLAG_PROXY_START flag seems meaningless.
Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
Explicitly cast Kconfig option value to avoid truncation issue reported
by Coverity.
Fixes#84697
Coverity-CID: 487737
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
This commit checks that ATT MTU value returned by `bt_gatt_get_mtu` is
greater or equal to 3 to prevent integer overflow.
Fixes#84693
Coverity-CID: 487743
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
Ensure that ~(link.rx.seg) & SEG_NVAL doesn't result in 0.
Fixes#84804
Coverity-CID: 393090
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
The commit fixes issue where flash_area_flatten has been used where
code was only supposed to erase devices by hardware requirement prior
to write, by replacing the call with flash_area_erase and supporting
logic to select proper path.
There have been following Kconfig options added:
- CONFIG_BT_MESH_BLOB_IO_FLASH_WITHOUT_ERASE
- CONFIG_BT_MESH_BLOB_IO_FLASH_WITH_ERASE
that are available for user depending on devices in the system and allow
to turn off paths that are not used by BLOB IO; for example if user
never writes to device with erase CONFIG_BT_MESH_BLOB_IO_FLASH_WITH_ERASE
will disable the path.
Both Kconfig options are y by default and enable all paths.
Signed-off-by: Dominik Ermel <dominik.ermel@nordicsemi.no>
MBEDTLS_AES_ROM_TABLES let Mbed TLS use precomputed AES tables instead
of generating them at runtime and placing in RAM. It clearly helps
reducing the RAM footprint, but at the same time it increases the ROM
one. In order to let the final user decide which kind of optimization
they prefer, we change the "select" statement with "imply".
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
According to MshDFUv1.0, section 6.2.3.5, the Firmware Distribution
Receivers List message is not limited by the size of the receivers list:
```
The number of selected entries shall be limited by the following:
- The number of entries shall not exceed the value of the Entries Limit
field from the Firmware Distribution Receivers Get message.
- The number of entries shall not cause the message payload to exceed
the maximum Access PDU size.
```
Thus, this assertion is incorrect as it doesn't allow to have the
receivers list bigger than number of maximal Access PDU size.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
Commit:
- adds dependency of the mbedtls psa usage on secure storage
- removes PSA ITS emulator and enables usage of
the secure storage in ble mesh bsim tests
- enables secure storage in all ble mesh and related samples
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
This change aims to eliminate the dependency on `ctx_shell` in
the Bluetooth `mesh/shell/*`, making the code more maintainable.
Replaced `shell_*` functions that depended on `ctx_shell` with
the appropriate `bt_shell_*` functions.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
By using `SETTINGS_STATIC_HANDLER_DEFINE_WITH_CPRIO` we can ensure that
BT subsystems always get loaded after BT, and BT Mesh after both of
them.
This solves the host having to register a GATT service in a delayed
manner, as we are sure to now register the GATT service after sc_commit
sets `SC_LOAD`.
Signed-off-by: Kyra Lengfeld <kyra.lengfeld@nordicsemi.no>
uuid needs to be keept in pb_gatt_cli server context until the client
is connected to server, otherwise, since we only kept the pointer from
the net_buffer, any incoming unprovisioned beacon before the connection
is established may overwrite uuid.
Signed-off-by: alperen sener <alperen.sener@nordicsemi.no>
The Health Client model supports periodic publication and thus requires
the update callback to be initialized.
This commit implements a callback inside the health client model and
adds a wrapper to it so that the user can fill up the msg buffer.
This way, the Config Server will always accept
Config Model Publication Set message with periodic publication
parameters. If the users callback returns an error or is not
implemented, the periodic publication will be aborted.
Signed-off-by: Stine Åkredalen <stine.akredalen@nordicsemi.no>
Once proxy gatt is enabled, due to the nature of gatt advertisements
being continuous, the controller will reject starting the scanner after
enabling proxy gatt. As such we should start the scanner first.
Currently the order works as we are using k_work to postpone the proxy
gatt service registration, because services can't be registered between
initializing BT and loading the settings.
Once we start using the new settings priority feature, we are sure to
envoke enabling proxy gatt after loading the settings, so the delayed
enabling can and should be removed.
This preventetive fix has been tested by running relevant bsim tests
after rearranging in which order the settings of the subsystems are
loaded via `h_commit`.
Signed-off-by: Kyra Lengfeld <kyra.lengfeld@nordicsemi.no>
To signal to the mesh extended advertiser that a proxy advertisement is
started, `bt_mesh_adv_gatt_send()` is called in combination with
setting the `ADV_FLAG_PROXY` flag in adv_ext.c. This ensures that it
won't try to start the advertiser again, which would result in the
controller rejecting it.
Setting this flag outside adv_ext.c is not viable, and as such calling
`bt_mesh_adv_gatt_send()` should also be limited to the advertising
source code.
As it stand now, once we utilize the new commit priority for `h_commit`
calls such a recall of `bt_mesh_adv_gatt_send()` with following
rejection by the controller will happen.
Within the `bt_mesh_resume()` function we can confidently assume the
extended advertiser has already been started. As such we can call
`bt_mesh_adv_gatt_update()` instead of `bt_mesh_adv_gatt_send()`.
This change has been tested with both running relevant bsim tests after
rearranging in which order the settings of the subsystems are loaded via
`h_commit`, as well as a modified light switch sample, adding
`bt_mesh_resume/suspend` calls on buttons.
Signed-off-by: Kyra Lengfeld <kyra.lengfeld@nordicsemi.no>
Referring to MshDFU_v1.0 Sections 6.1.1, 6.2.1 and 7.1.1 model
descriptions: DFU/DFD server/clients extend BLOB Transfer root models
and DFD server requires Firmware Update Client on the same element. For
this reason we need to make sure that those main models or root models
exist on the same element. And also firmware update client can not be
forced to be in the first element.
For all model extention call return the error code in case of an error.
Signed-off-by: alperen sener <alperen.sener@nordicsemi.no>
Deprecate BT_CTLR, and add a new HAS_BT_CTLR as a virtual option which
specific users (like BT_LL_SW_SPLIT) select. This also means that we can
remove all places that were forcefully enabling the BT_CTLR option, and
instead we now depend on devicetree to get some local LL HCI driver
enabled which in turn also enables the HAS_BT_CTLR option.
Signed-off-by: Johan Hedberg <johan.hedberg@silabs.com>
