The current instrumentation point for CONFIG_TRACING added in
PR #28512 had two problems:
- If userspace and KPTI are enabled, the tracing point is simply
never run if we are resuming a user thread as the
z_x86_trampoline_to_user function is jumped to and calls
'iret' from there
- Only %rdi is being saved. However, at that location, *all*
caller-saved registers are in use as they contain the
resumed thread's context
Simplest solution is to move this up near where we update page
tables. The #ifdefs are used to make sure we don't push/pop
%rdi more than once. At that point in the code only %rdi
is in use among the volatile registers.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Instead of hardcoding the priority1 and priority2 values used
in BMCA, let the user tweak the values via Kconfig.
Fixes#28151
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit fixes the zephyr_get_system_include_directories_for_lang()
function in the event that no system include directories are specified
for Zephyr interface.
This fixes an issue where the value returned by this function is
`-isystem` with no directory.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
From 2.3 to 2.4 the net samples pair echo-server/client increase the
SRAM requirements. Since CI doesn't build all combinations, (even for
a release like 2.4.0-rc1) the problem was detected only now. Decrease
buffers for atsamr21_xpro on both echo_server/client sample pair.
Fixes#28341.
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
When added bluetooth over USB shell overlay, the
file was by mistake added with wrong extension.
Signed-off-by: Emil Obalski <emil.obalski@nordicsemi.no>
We only need to offset the start of the non-secure image
by 0x400, if TFM is built with BL2 support. In this case
we use the ROM_START_OFFSET Kconfig switch and set to
0x400. This instructs the linker to offset the beginning
of the ROM section by 0x400. In other words, we do not need
to statically move the start of the image by 0x400. This
fixes an issue that prevents from running Zephyr + TFM
without BL2 on Musca B1.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
We only need to offset the start of the non-secure image
by 0x400, if TFM is built with BL2 support. In this case
we use the ROM_START_OFFSET Kconfig switch and set to
0x400. This instructs the linker to offset the beginning
of the ROM section by 0x400. In other words, we do not need
to statically move the start of the image by 0x400. This
fixes an issue that prevents from running Zephyr + TFM
without BL2 on MPS2 AN521.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When we build Zephyr for a Non-Secure image, using TFM
as the Secure counterpart, we may or may not use BL2 (e.g.
MCUboot in TFM). If we build with BL2 we need to account
for a ROM offset before the start of the non-secure image
where TFM BL2 expects the BL2 header. This offset is not
needed when TFM is built without BL2. We use the existing
Kconfig options to determine whether we need a ROM offset
or not. We use the value of 0x400 for the BL2 header,
which is the default value used in TFM, but the option is
non-hidden, allowing the user to overwrite the default
value if needed.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Fix bt_le_oob_get_local and bt_le_ext_adv_oob_get_local stuck forever
waiting for the sc_local_pkey_ready semaphore when SC HCI commands
are not supported in the controller.
By using the le_sc_supported helper function the runtime check of HCI
commands and the feature check of CONFIG_BT_SMP_OOB_LEGACY_PAIR_ONLY
is combined to be handled int the same way.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Move the le_sc_supported helper function up in the source file so that
it can be re-used in other places without a forward declaration.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
The sam4l pinctrl contains two entries with wrong values. Fix the
typo to create valid pinctrl map.
Fixes#28538.
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
Since conn->send_data_total is of time size_t we need to use %zu or
we'll get build errors in sanitycheck on 64-bit platforms
Fixes#28605
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
For boards with (relatively) large memory, the test which dumps
page tables takes a long time to finish. The default timeout of
sanitycheck is not enough for those boards. UP Squared board is
such a board. So limits to pagetable dumping to boards with
less than 32MB.
Fixes#28548
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
If the network driver for some reason did not set the data in
the network packet properly, then just drop it as we cannot do
anything with just plain net_pkt.
Fixes#28131
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The function will check that the net_pkt will contain some
data and return true in that case, and false if the net_pkt
is empty.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Algorithm for waiting for clock stabilization was failing in case when
it was waiting for clock availablity and clock was already available
before function was called. That is because nrfx_clock_is_running
was returning false because XTAL was already started but not yet
running.
Added a check for current LF source, if XTAL is picked that indicates
that RC is already running because of two stage startup procedure.
Added documentation of lfclk_spinwait with explanation of two stage
approach.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
If hannel don't have alloc_buf and the RX MTU is configured to require
segmentation this warn the user and truncate the RX MTU.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Channel that requires segmentation must set alloc_buf in order for the
stack to be able to reassemble segments of an SDU.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
In case where ull_adv::disable() is disallowed, disable_mark is
erroneously re-mark instead of un-marked
Signed-off-by: Erik Brockhoff <erbr@oticon.com>
Initialize device pointer in driver's context.
It was not done quite right in commit a1708cf2f2
("drivers: ethernet: Fix device instance const qualifier loss"),
and then completely removed in commit 113d9274ea
("drivers: ethernet: remove stray expression")'
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
To improve Zephyr security, I decided to create a test
to check upon exit of a system call back to the calling thread,
the kernel scrubs CPU registers for sensitive data.
Signed-off-by: Maksim Masalski <maksim.masalski@intel.com>
If there is no handler for IPv4 broadcast packet, then ignore it
instead of trying to send an ARP message to resolve the senders
address.
Fixes#21016
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit fixes the following issue:
When using slip without TAP, i.e. CONFIG_SLIP_TAP=n (default), while
another ethernet interface is enabled and sets CONFIG_NET_L2_ETHERNET=y.
This causes ethernet_init() to be wrongly called with struct dummy_api
instead of struct ethernet_api.
ethernet_init() expects struct ethernet_api (by cast), so we end up
with the get_capabilities field pointing to garbage!
Actually, as we are using the dummy api, we don't need to call
ethernet_init() at all.
Sole dependency on CONFIG_NET_L2_ETHERNET is wrong because it
can be enabled by another interface.
Signed-off-by: David Komel <a8961713@gmail.com>
When we wake-up of deep sleep power state, we want to disable it.
Otherwise, when the cpu will go next to idle mode during a
SYS_POWER_STATE_ACTIVE, it will go into deep sleep mode
instead of a sleep mode.
fixes: #26896
Signed-off-by: Julien D'Ascenzio <julien.dascenzio@paratronic.fr>
Accourding to the spec the CCC descriptor may occur in any position
within the characteristic definition after the Characteristic Value.
Fixes#28324
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
BT_UUID_ES_TRIGGER_SETTING was incorrectly using
BT_UUID_ES_MEASUREMENT_VAL as value instead of
BT_UUID_ES_TRIGGER_SETTING_VAL.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Fix for handling REJECT_IND PDU received for PHY Update,
Connection Parameter Request and Data Length Update control
procedures.
If a link layer control procedure collision occurs, example
with local initiated PHY Update Procedure, and peer sends a
REJECT_IND PDU, then the PHY Update Procedure is stalled.
Fixes#28282.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The on-board LED of the blue and black variants of these boards have LED
logic inverted. This was never observed as most of the time, the LED was
used as a blinky. Fix this by setting its DT bindings to active low.
Signed-off-by: Siddharth Chandrasekaran <siddharth@embedjournal.com>
Named choice is needed in order to be able to extend it
in other modules.
Closes: #28559
Signed-off-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@nordicsemi.no>
According Mesh Profile Spec 5.4.2.4 Authentication, if device
use Input OOB Authentication method, should send input complete
pub after local public key has been acked.
`bt_mesh_input_string` or `bt_mesh_input_number` directly send
`input_complete`, however does not check whether the pub key has
been sent.
Mesh Provisioning timeout set to `60` seconds, so even this
probability is extremely low, it does not mean that there is no
such probability.
Signed-off-by: Lingao Meng <mengabc1086@gmail.com>
CPU Cortex-M implies Mainline Cortex-M, therfore, the dependency
on ARMV6_M_ARMV8_M_BASELINE is redundant and can be removed. The
change in this commit is a no-op.
We also add the ARMV6_M_ARMV8_M_BASELINE dependency on option
CPU_CORTEX_M0_HAS_VECTOR_TABLE_REMAP to make sure it cannot be
selected for non Cortex-M Baseline SoCs (at least, not without
a warning).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This STM32 serie redefines function relocate_vector_table()
It should take into account features:
SW_VECTOR_RELAY and SW_VECTOR_RELAY_CLIENT
fixes#28289
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
Tracing switched in threads in C code does not work, it needs to happen
in the arch_switch code. See also Xtensa and ARC.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
