Only the first one requires it. Actually drivers know that already and
handle the frags list correctly.
In case ethernet has to run along with 15.4 on the same SoC, this will
optimize things quite a bit knowing that biggest ethernet frame will be
forcefully split in as many 128 bytes frags as necessary.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
A typo in update_attrs() was setting every observer to a PMIN of 0.
This meant we could send observer data as often as the process was
called. This is out of spec as the default minimum is 10 seconds.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
The hop limit value in net_pkt was not updated according to
received IPv6 header.
Fixes#8182
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The LwM2M engine thread is used for various periodic triggers.
None of these are in a critical path that requires super sensitive
timing and the current K_PRIO_COOP(7) setting was causing the
Bluetooth RX thread to have to wait too long for certain actions
to complete.
Let's lower the priority to -1 (effectively) to eliminate these
conflicts.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Fix file exist error within settings_subsys_init in order to correctly
reload existing settings when CONFIG_SETTINGS_FS is enabled.
Signed-off-by: Daniele Biagetti <daniele.biagetti@cblelectronics.com>
When att_disconnected is called a thread may be waiting for the tx_sem
but that is memset to 0, furthermore there exists a flag
ATT_DISCONNECTED to indicate the context is no longer valid so instead
move memset to bt_att_accept so it is cleared when it is about to be
reused.
Fixes#8083
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
The adv_send() function was incorrectly decoding the 5-bit value (it
was using it directly as milliseconds), which effectively lead to the
code always picking the controller's minimum supported interval.
Fix this issue, but do it by simplifying the (re)transmission state
tracking so that the state is always stored in the original "packed"
8-bit value, where 5 bits are reserved for the interval, and 3 for the
count.
Fixes#7972
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In the function zephyr_smp_write_at the offset was checked
for negative values while it type is unsigned -
which cause static analyses issue.
Fixes#7737
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
We should check for valid lengths, not just because flash may have
become corrupted, but also because this fixes coverity errors, such as
CID 186030.
Fixes#7739
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
DISK_ERASE_BLOCK_SIZE, DISK_VOLUME_SIZE was described in misleading
manner. This patch changes descriptions for both prompts to
appropriate.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
IPv4 header's checksum was set when 'setup_ipv4_header', but in that
times IPv4 header's source ip and dest ip address haven't setted. So
it may be set an incorrect checksum.
Fixes: #7989
Signed-off-by: qianfan Zhao <qianfanguijin@163.com>
We init the net_tc tx/rx work queues during net_init() with a
call to init_rx_queues(). The L2/L3 and networking drivers have been
setup at this point. If we yield the current thread, we risk
a call to net_recv_data() which calls net_queue_rx() which calls
net_tc_submit_to_rx_queue() on an RX work queue which hasn't been
setup yet.
This manifests as a boot hang under seemingly random circumstances.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Do not set protocol in interface descriptor. Although this field has
been ignored by major OSs during the last decades, this creates a big
deal of problems with ModemManager sending unexpected AT commands when
plugging in a dev board.
Fixes: #6646
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Security layer was tested only against CC2520, which does not mandate to
begin the session with a valid key. However, the crypto API tells it
must do so. And indeed, starting a session on mtls shim crypto device
will fail due to missing key. Thus moving the relevant code where it
should.
Reported-by: Johann Fischer <j.fischer@phytec.de>
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This one converts "raw" timeout value to use K_MSEC() macro
in order to make clear how long the timeout is.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Use of K_SECONDS() macro is more intuitive so use that instead of
plain MSEC_PER_SEC define.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Convert couple of MSEC() calls to K_MSEC() as the timeouts
when using MSEC() are just too long.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This patch prevents eem_read_cb from trying to allocate a lot of memory.
It may happen that EEM payload size is zero, the eem_read_cb then tries
to allocate a buffer which is 0xfffc bytes large and luckily blocks.
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
_nvs_sector_is_used() never uses the offset argument. As a consequence,
it only check the first sector of the flash. Fix that.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The two functions that compute the crc16 when writing (nvs_append_close)
and when reading (nvs_check_crc) currently assume that the flash is
also mapped in read mode at address 0. This is not true on all SoCs, and
even less on an SPI flash.
Fix this by adding a new nvs_compute_crc() function which compute the
CRC16 of an entry using the flash using nvs_flash_read, in blocks of
write_block_size. This might not be the optimal size, but it keeps the
stack usage small.
Use this function in both nvs_append_close() and nvs_check_crc() instead
of accessing the flash from address 0.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
GCC complains that last_entry.len and last_entry.data_addr might be
uninitialized in _nvs_gc. Fix that.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that the flash writes are padded up to the write block size, there
is no need to have explicit padding fields in the _nvs_sector_hdr and
_nvs_data_slt structure. This allow to save space when the write block
size equals to 1 or 2
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Writing more than the source buffer means that some random data,
possibly coming from the stack, ends-up in the flash. This could be
a security issue.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reading more than the destination buffer means that data is overwritten
possibly on the stack. This causes unpredictable behaviours like
crashes.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Some SoCs do not allow shorter writes than the write block size, usually
when they have ECC memory. This patch first write all data in multiple
of the write block size and then do a last write with the data padded.
It uses 0xff as the padding byte to avoid wearing-out the flash.
nvs_append_close() is slightly modified to compute the crc16 that will
be put in the slot over the size defined in the header, to match the way
it is checked on read.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
During normal use, a CoAP packet can be resent due to network congestion
and other causes. During block transfer the LwM2M client checks to make
sure the block received is the one we expect and if not generates a
"duplicate" warning. When this happened, we were releasing the reply
handler and when the correct block was received the client would
generate a "No handler" error.
To avoid releasing the reply handler too early, let's set the coap_reply
"user_data" field to an error condition (1). Then, once the reply
processing is complete we can check the user_data field to be sure that
it's ok to release the reply handler.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
In certain cases the response to a command can come in the form of a
non-priority event. This is the case of LE Create Connection Cancel,
which generates a Command Complete and then an LE (Enh) Connection
Complete. Take this case (and other future ones) into account by calling
the correct Host recv function.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The Kconfig symbol CONFIG_USB_DEVICE_HID_BOOTP isn't defined anywhere,
so remove dead code associated with it.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This patch fixes possible null pointer dereference in
net_stats_update_rpl_resets(...).
net_rpl_set_root_with_version(...) does not initialize instance->iface
and calls net_rpl_reset_dio_timer(...), which then calls
net_stats_update_rpl_resets(instance->iface).
fixes: #7862
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
Some versions of gcc do not seem to compile out the inaccessible code
in this case and instead give the following error:
subsys/bluetooth/host/mesh/transport.c:419: undefined reference to
`bt_mesh_lpn_poll'
This happens at least when building samples/bluetooth/mesh for
native_posix on Fedora 28.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Implement the new entropy_get_entropy_isr() function to allow the kernel
to collect entropy before the scheduler and kernel data structures are
ready. Switch to an nrf-specific version for high-performance
requirements in the BLE Link Layer.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
This keeps biting us; sanitycheck turns on assertions by
default for tests, but standalone builds, or builds done
for other test infrastructure do not. Put all builds in
the same state.
Specific tests (such as benchmarks) can override this with
CONFIG_FORCE_NO_ASSERT.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Rename CONFIG_RNDIS_TX_BUF_* to CFG_RNDIS_TX_BUF_* and rename
CONFIG_RNDIS_CMD_BUF_* to CFG_RNDIS_CMD_BUF_*. The CONFIG options where
not exposed in Kconfig so limit use of CONFIG_ to Kconfig only symbols.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Remove non-existent Kconfig symbol references. An additional (but
related) change is the removal of all persistent storage symbols from
the Arduino 101 Bluetooth shell app, since BT_STORAGE no longer
exists.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Not only removes a branch during normal operation, but also ensures
that, by initializing at the PRE_KERNEL_2 stage, and granting
privileges to all threads to the semaphore, this code will work in
early boot situations and in user mode.
This assumes that entropy drivers will all initialize during
PRE_KERNEL_1 stage. All in-tree drivers do that.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
The code was using MSEC() macro in few places instead of more
proper K_MSEC(). The MSEC() takes seconds as a parameter and
K_MSEC() takes milliseconds.
Fixes#7657
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
