The number of arguments for a format string is approximated by the
number of conversion specifications. This count may exceed the maximum
supported argument count. Limit the extraction to the available space.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Fix a race which seems to have been presenting itself
very sporadically on loaded systems.
The race seems to have caused tests/kernel/sched/schedule_api
to fail at random on native_posix.
The case is a bit convoluted:
When the kernel calls z_new_thread(), the POSIX arch saves
the new thread entry call in that new Zephyr thread stack
together with a bit of extra info for the POSIX arch.
And spawns a new pthread (posix_thread_starter()) which
will eventually (after the Zephyr kernel swapped to it),
call that entry function.
(Note that in principle a thread spawned by pthreads may
be arbitrarily delayed)
The POSIX arch does not try to synchronize to that new
pthread (because why should it) until the first time the
Zephyr kernel tries to swap to that thread.
But, the kernel may never try to swap to it.
And therefore that thread's posix_thread_starter() may never
have got to run before the thread was aborted, and its
Zephyr stack reused for something else by the Zephyr app.
As posix_thread_starter() was relaying on looking into that
thread stack, it may now be looking into another thread stack
or anything else.
So, this commit fixes it by having posix_thread_starter()
get the input it always needs not from the Zephyr stack,
but from its own pthread_create() parameter pointing to a
structure kept by the POSIX arch.
Note that if the thread was aborted before reaching that point
posix_thread_starter() will NOT call the Zephyr thread entry
function, but just cleanup.
With this change all "asynchronous" parts of the POSIX arch
should relay only on the POSIX arch own structures.
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Fix issue unable to connect to bonded peer when host resolution is used
either because the controller does not support privacy, or the
controller resolving list was exceeded.
In this case we need to use the RPA from the advertising report
directly, there is a small chance of the peer cycling the RPA here, in
which case the connection might be unsuccessful.
Bug introduced here: 45da629b24Fixes: #18306
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
When the database changes and a client becomes change unware it should
also clear the out of sync flag as the following request should return
an error:
Core Spec 5.1 | Vol 3, Part G, 2.5.2.1 Robust Caching:
'The error response is sent only once after the client becomes
change-unaware, unless the client disconnects or the database changes
again before the client becomes change-aware in which case the error
response shall be sent again.'
Fixes: #18183
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
This fixes the following crash:
ASSERTION FAIL [0] @ ZEPHYR_BASE/subsys/logging/log_core.c:180
argument 2 in log message "%s: Stored CCCs for %s (%s)"
missing log_strdup().
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
The code for checking space in the Friend queue was faulty in the case
that we receive a message with more segments than the configure Friend
Queue size. This is not an issue for the default configuration but
still a possible one. Move the check for exceeding Friend Queue Size
to the per-LPN function, so that bt_mesh_friend_queue_has_space()
iterates all LPNs before delivering its verdict. This allows us to
return success in case no LPN matched (which is how the code was
intended to work).
Fixes#18522
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Qualification test case MESH/NODE/FRND/FN/BV-08-C requires that we do
not store more messages than the reported Friend Queue size. The
implementation was so far opportunistic and stored more if it could
(it would later discard if necessary to make sure all queues can store
the required amount). The spec also requires the queues to have new
messages overwrite old ones (in the style of a circular buffer), so we
have to keep track of which buffers are part of the same segmented
message (so we discard all buffers belonging to the same message).
To pass the test case, add APIs to check for space in the Friend
queue, and track the number of buffers for each incoming segmented
message.
Fixes#18090
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The bt_mesh_trans_resend() function had no users, and had in fact not
even a prototype in a header file. Just remove it.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When sent solely to the Friend Queue the send callbacks were not
getting called for unsegmented messages.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
According to Mesh Profile Specification 1.0.1 Figure 5.17, the
unprovisioned device should send its confirmation value after the
provisioner sends theirs. Previously, the confirmation value would be
sent immediately after OOB input complete. Now it first waits for the
input data, then from confirmation from the provisioner before sending
the confirmation.
Fixes: #18178.
Signed-off-by: Trond Einar Snekvik <Trond.Einar.Snekvik@nordicsemi.no>
The SDK version is of the form X.Y.Z. Change the cmake scripts to be
based on X.Y of the version. This allows us to easily support newer
toolchains without having to explicitly add cmake files for the version
as well as removes duplication between those files.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
These files got inadvertently added to the 1.14 branch with the
manual backport commit d7d9988d55.
Remove these since they're not supposed to be in the 1.14 branch and
there are not even any hooks to build them.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There's a couple of missing imports from when the
203021bedb backport was added to this
branch.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
If we receive IPv4 packet where source address is unspecified
(all zeros), then we need to drop it.
Fixes#17427
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Having a checksum of zeros in UDP means "missing checksum" and is a
valid case as per RFC 768:
"An all zero transmitted checksum value means that the transmitter
generated no checksum (for debugging or for higher level protocols
that don't care)."
Such support is made possible by adding a new Kconfig option named
CONFIG_NET_UDP_MISSING_CHECKSUM.
However, that is valid only for IPv4. For IPv6, see the RFC 2460
section 8.1:
"Unlike IPv4, when UDP packets are originated by an IPv6 node, the UDP
checksum is not optional."
So the UDP checksum will always be verified in IPv6.
Fixes#16375
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
From RFC 768, in "Fields":
"If the computed checksum is zero, it is transmitted as all ones"
Fixes#16379
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Before calling socket callback function, make sure the callback
function exists so that we do not get NULL pointer reference.
Fixes#18021
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Package "hub" has been renamed some time ago to "git-spindle".
Recently package with the same name "hub" has been released on PyPi
page, which causes installing pip requirements.txt to fail since
it is trying to install different package.
Signed-off-by: Marcin Sloniewski <marcin.sloniewski@gmail.com>
If the allocated net_pkt is larger than the network interface MTU,
then check if the IPv6 fragmentation is enabled and allow larger
net_pkt length as the IPv6 fragmentation will split the packet into
suitable parts.
Fixes#16354
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The net_pkt_pull() needs to move all the remaining data in the
net_buf instead of just the pull amount.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Previous ATE searching is accelerate by reading sector
close ate, which allow to skip reading whole storage.
ATE is already covered by crc8. That patch introduce
check on correctness offset read form close ate. This
increases meta-data integrity check level.
This also preserves against possible looping when read
incorrect or not a NVS storage image.
fixes#16899
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
When code relocation enabled, there will be serval regions holding
text. And then there will be function call between these .text
regions, when distance between caller and callee is too far, linker
will automatically generate and insert veneer functions. And these
veneer functions will be located right after the last instruction
in the .text region by the linker. So these code will be put in the
memory reserved for priv_stacks text and kobject text if they don't
consume all the reserved memory. Or the veneer functions will be put
before the reserved memory if there isn't code in the reserved
memory. And then in the user mode building process, there will be
different memory layout and it will cause usr mode not working.
And moving the memory reserved for priv_stacks text and kobject text
at the beginning of .text will avoid above problem. The detailed
analysis for this issue can be found on Github issue #17038.
Fixes: #17038.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
No need to write deletion entry for non-existing entry.
Furthermore such a write to filled up storage will make another
writes impossible as it occupied delete nvs_ate space, but it
does not remove anything.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
During initialization NVS analyzes open sector for
find write addresses. It was possible to infinite loop
in case when there was no space for any new ATE in the sector.
This patch introduce check for space available in the sector which
fix the issue.
fixes#17891
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
NVS always leave space for deletion ate in a full sector.
Even that it was not possible to write delete nvs_ate to the
full sector. Because of that it was possible to fill up NVS,
and be able to delete nothing.
This patch introduce recognition of case
the delete ate is writing, and allow to write such nvs_ate to
opened full sector.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
If we receive IPv6 packet where source address is unspecified
(all zeros), then we need to drop it.
Fixes#17450
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Noticed this error print when running in mps2_an385 platform.
<err> net_if: There is no network interface to work with!
Adding CONFIG_NET_LOOPBACK=y removes this error. After this
the CONFIG_NET_IPV6_MLD=n needed to be added in order to avoid
crashes when IPv6 MLD is done which is not needed for this test.
Removed also the whitelisting as that is not really needed and
we want to run these in as many platforms as possible.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
CONFIG_NET_IPV* checks are not needed in lwm2m_parse_peerinfo().
The functions used are always available. Worse, having these checks
forces the need to enable CONFIG_NET_IPV4 or IPV6 when it's not really
needed (LwM2M could be using an offloaded IP stack).
Fixes: https://github.com/zephyrproject-rtos/zephyr/issues/17401
Signed-off-by: Michael Scott <mike@foundries.io>
CONFIG_NET_IPV* checks are not needed in lwm2m_sprint_ip_addr(). The
functions used are always available. Worse, having these checks
forces the need to enable CONFIG_NET_IPV4 or IPV6 when it's not really
needed (LwM2M could be using an offloaded IP stack).
NOTE: Also fixes an issue where a NULL is returned when the IP address
is unknown. This usually ends up with a crash/abort in the logging
code.
Fixes: https://github.com/zephyrproject-rtos/zephyr/issues/17401
Signed-off-by: Michael Scott <mike@foundries.io>
Skip the TCP options before giving the data to application.
Without this, the TCP options would be passed to the application.
Fixes#17055
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If a sample wants to use the Zephyr implementation of mbedtls, it
enables CONFIG_MBEDTLS and sets any needed Zephyr-specific mbedtls
options.
Currently, the LwM2M subsystem selects MBEDTLS automatically when
LWM2M_DTLS_SUPPORT is enabled. Let's remove this and let the
LwM2M client sample enable mbedtls and it's options.
This mimics the behavior of several other network-related samples
and removes conflicts when selecting alternate implementations of
MBEDTLS.
Fixes: https://github.com/zephyrproject-rtos/zephyr/issues/17399
Signed-off-by: Michael Scott <mike@foundries.io>
That is, specifically fetch v1.14.1, which will be the release tag to
check out on the LTS branch the next time a point release is cut.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
West is moving on. While later versions of west will remain compatible
with Zephyr LTS as long as it's supported, let's still recommend the
best tested version.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
The commit 1c7b668804 tried to fix
resending segments to the GATT bearer, however it got the buffer
refernce counting wrong. The bt_mesh_net_resend() function doesn't
own a reference to the buffer, i.e. it's not responsible for unrefing
it. E.g. bt_mesh_adv_send() takes its own reference.
What was missing however was the handling of the callbacks. Use the
recently introduced send_cb_finalize() helper to make sure they're
called.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There are several places that require the send callbacks to be called
immediately. Reduce the code a bit by introducing a helper for this.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Mesh Profile Specification states in section 3.10.5 IV Update
procedure: "The IV Update procedure is initiated by any node that is a
member of a primary subnet", meaning if we're not on the primary
subnet we should not try to initiate the procedure.
Additionally this fixes initiating IV Update in all cases where the
sequence is updated, by putting the code into bt_mesh_next_seq().
Fixes#17977
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This branch is inherited from old code that used to start the timer as
soon as a message was queued for transmission rather than when it was
actually transmitted (the case today). It'll also cause a race in case
the publication goes over the GATT layer since the "sent" callback
happens synchronously in that case.
Fixes#17821
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The GATT bearer is a reliable one so there should theoretically never
be a need to resend segments. If however for some strange reason the
proxy client doesn't immediately ack all segments we should do the
resending on the GATT bearer, rather than sending them over
advertising.
Fixes#17907
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The code was incorrectly bailing out with "return 0" rather than
continue. Also, it was incorrectly making a reference to
tx->seg[seg_o] since when a PDU goes through the friend queue we don't
use the usual retransmission mechanism.
Fixes: #17932
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In case a queued buffer is canceled before sending we have to unref
it, since that's what adv_send() would do as well.
Fixes#17936Fixes#18013
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Mesh Profile 3.6.7.2 Publishing Heartbeat messages:
"
Triggered publishing of Heartbeat messages is enabled by
the Heartbeat Publication Features state (see Section 4.2.17.5):
...
- If the Low Power bit is set to 1, a Heartbeat message shall be
published when the node establishes or loses Friendship (see Section
3.6.6.1).
"
Fixes#18194
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When sending heartbeat messages triggered by feature changes the code
was trying to look up the configured publication subnet, in an
apparent attempt to figure out if publication is enabled or not. A
more appropriate way is to check for the heartbeat publication
destination address, and since we have a helper function this can be
done in a single place.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The heartbeat is a transport layer feature, so move it to transport.c.
This also opens the way to properly fix Friendship-established/lost
triggering for LPN role.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When in Low Power mode an LPN may receive messages sent by nodes other
than its friend during its brief receive window. These messages get
rejected by the transport layer. At some point in the future the LPN
will receive these messages from its friend, however they will have
already been added to the network message cache earlier.
When the transport layer rejects a message due to it being received from
a non-friend node while in Low Power mode it must be removed from the
network message cache so that it can be correctly received from its
friend in the future.
Fixes#17809
Signed-off-by: Rich Barlow <rich@bennellick.com>
The RPA timer should always be running, even with privacy supported by
the controller. In order to select an IRK to generate the private
the controller is instructed by provided an identity address.
If we want to advertise privately without providing an identity address
the host has to set the private address.
Fixes: #18150
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix missing interval min copy in bt_conn_set_param_le. Application is
unable to override BT_GAP_INIT_CONN_INT_MIN for interval_min.
Fixes: #17789
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix assert in the RPA timeout when no resolving entry exists in the
resolving list, and the advertiser was requested to use a local IRK
to generate it's own address.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
When removing a service using bt_gatt_service_unregister, its attributes
handles will keep their value which can cause a problem when adding the
service again to the GATT database.
When re-adding, the gatt_register routine is taking the last handle of
the GATT database to compare it with the handles of the service to be
added.
If a service has handles with a lower value than the last handle of
the database an error will occur.
If we add/remove/add the last service, no error will occur as its
handles are always greater than the last one of the database.
Instead of resetting the handles when unregistering a service and
reassign new ones we make sure that the handles of the services are
not in use in the GATT database in order to place the service back
where it used to be.
Signed-off-by: Arthur Crepin-Leblond <arthur.crepin@proglove.de>
The original creates a pointer to a compiler-generated temporary that is
destroyed when the scope is exited. The pointer is stored in a structure
defined in an enclosing scope and is invalid by the point it's used.
The fix holds the structure in a variable with the same lifetime as
the pointer.
Signed-off-by: Findlay Feng <i@fengch.me>
The call to bt_conn_security will send a SMP security request for
peripheral, and for central it will initiate LL encryption.
A call to bt_conn_security with no IO capabilities but authenticated
keys has been distributed, would succeed on central side, but fail on
peripheral side with error code -22.
The keys could have been either:
- Preprogrammed
- IO capabilities may have changed.
- OOB bonding may been used.
Fix so that Peripheral can send a security request if the bond
information is already established.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix an issue where the generated identity was not permanently stored.
This resulted in being unable to reconnect after bonding when using
privacy, since a new local IRK was generated on reboot.
When settings is enabled the application is responsible for loading
identities and possible creating its own identities.
When settings_load is called and no identities has been created or found
in persistent storage a new identity will be created.
Since bt init has not been finalized bt_id_create will not make a call
to bt_settings_save_id. So we need to make sure that this identity will
be stored.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
This patch updates the GAP ICS to 7.4.1.
It updates the Bluetooth Spec to 5.1 and added missing ICS tables.
Signed-off-by: Tedd Ho-Jeong An <tedd.an@intel.com>
Handle optional argument UUID in bt_gatt_discover with type
DISCOVER_DESCRIPTOR, bt_uuid_cmp doesn't check for NULL pointer.
On system with MMU (nrf52_bsim) this can result in segfault.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
This introduces a new flag (BT_GATT_SUBSCRIBE_WRITE_PENDING) which is
set when a write operation requires canceling before the parameters can
be reused.
Fixes#17534
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
This adds a dummy cancel request which is used instead of NULL when
clearing and outstanding request thus prevent new request to take
place before a response is received while allowing the original request
to be reused.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Set advertisers state flags so that previous advertise params does not
affect the current advertiser state.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix gatt indicate when conn is NULL and called with characteristic
declaration as the attribute argument. In this case the handle was not
advanced to the characteristic value. This is inconsistent with the rest
of the notify and indicate API
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix an issue in the gatt_register routine where removing a service and
adding it again would cause the database to have non-ascending
orderdered handles numbers and lead to an incomplete service discovery.
Fix: Go through the database and look for a place where to insert
the new service.
Signed-off-by: Arthur Crepin-Leblond <arthur.crepin@proglove.de>
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix an issue in the SMP identity distribution where the IRK for the
default identity was always distributed.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
The code was already rejecting the key, however that rejection
happened only after we had already sent our public key as response,
which got interpreted as acceptance by the tester (PTS).
This fixes issue that d4fd267086
supposed to fix. The problem is bt_dh_key_gen is async.
Local public key cannot be sent from the same context
the bt_dh_key_gen is called because we don't know yet
if remote key is valid.
Fixes MESH/NODE/PROV/BI-13-C.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Fix bug introduced by: 45da629b24
Mistakenly set the own address type to the destination address type.
Also this uses the RPA_OR_RANDOM in case local IRKs exists in the
controller.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix an issue with enhanced connection complete event raised by the
advertiser role. The advertiser reported that it has resolved the
the identity address of the peer, even when the peer is connecting
using it's identity address.
The host will not have the correct on-air address type i.e public.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
The init addr should contain the on-air address used to establish the
connection. The dst address contains either the current RPA of the
unknown peer, or the identity address after identity information has
been exchanged.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
The 'registered' flag was not cleared when 'gatt metrics off'
was called so that one was unable to register the service again.
Remove 'registered' variable that is actually not needed as there
is no tracking if service is already registered in similar
cmd_register_test_svc command. If the service is already registered,
the host will log an error.
Fixes: #17882
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
This will handle BTP Discover All Primary Services command.
Related TC: GATT/CL/GAD/BV-01-C
Fixes: #17861
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Core 5.1, Vol 3, Part F, 3.3.1
"Multi-octet fields within the GATT profile shall be sent least
significant octet first (little-endian) with the exception of the
Characteristic Value field. The Characteristic Value and any fields
within it shall be little-endian unless otherwise defined in the
specification which defines the characteristic."
Fixes: GATT/SR/GAS/BV-02-C
Fixes: #17857
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
This changes RPA timeout so that the address is changed every 60
seconds. This change is needed to speed up privacy related test cases.
Currently, the new RPA was initiated on request that was a bit hackish
and could be done via BT_RPA_TIMEOUT Kconfig option.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
This enables support for GAP writable device name in tester application
for testing purposes.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
This updates tester application to recent changes in BTP related
to Connection Parameters Update.
Connected event has been extended with connection parameters.
Connection Parameters Update command and event have beend added to
initiate and track current connection parameters.
Needed to automate qualification test GAP/CONN/CPUP/BV-06-C
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Having this option disabled, MITM flag state can be controlled by
bt_conn_security state. This option is enabled by default to not
change the current implementation behavior.
Related to SM/MAS/SCPK/BV-01-C.
Fixes#17463
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
When performing OTA using the LwM2M subsys, several logging errors
regarding log_strdup were noted. Let's fix these.
Signed-off-by: Michael Scott <mike@foundries.io>
zero slice_ticks when can't time slice so that next_timeout will
ignore slice_ticks of _current_cpu and system can stay low power
state longer time.
Fixes: #17368.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
OpenThread did not verify if the interface provided in the net_mgmt
handler is actually an OpenThread interface. In result, when multiple
network interfaces were used, different interfaces were processed by the
OpenThread handler, ending up in a crash.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Add self.require() checks before running commands. Increase test
coverage, including for this feature, while we are here.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
The runners/jlink.py script has a mechanism for erroring out if a host
tool is not installed. Abstract it into runners/core.py and handle it
from run_common.py. This will let it be used in more places.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
* if thread switchs in interrupt, the target sp must be in
thread's kernel stack, no need to do hardware sp switch
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Fix assert in the RPA timeout when no resolving entry exists in the
resolving list, and the advertiser was requested to use a local IRK
to generate it's own address.
Relates to #18091 and #17846.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix the nRF52840 Coded PHY radio timings based on testing
with conformance tester.
Fixes the following conformance tests:
LL/CON/MAS/BV-52-C [Master Receiving Data, LE Coded, CI Change]
LL/CON/MAS/BV-54-C [Slave Receiving Data, LE Coded, CI Change]
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix the regression in Coded PHY CI change implementation
introduced by reverting the commit 9d1ca9c390 ("Bluetooth:
controller: remove redundant PPI channel and TIMER CC").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Fix the conditional compile of Data Length Update event
generation on PHY Update Procedure when Data Length
Extensions are not supported.
The regression was introduced in
commit 70a89876d0 ("Bluetooth: controller: Fix missing
data length update event")
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix bug in Data Length Update procedure that caused the
connection to drop due to the implementation sending bigger
PDU before the peer has acknowledged the receipt of Length
Response PDU.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Enable the Fast Encryption design and connection RSSI
measurement when CONFIG_BT_HCI_RAW is selected.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix the controller implementation to perform connection
event length reservation based on the completed Data Length
Update and/or PHY Update Procedure.
This fix with avoid states/roles from stepping on each
others event length. Connection would have supervision timed
out or have stalled data transmissions due to insufficient
reserved air time.
Relates to #15171.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix missing generation of data length update HCI event when
effective tx and rx timings change due to PHY update
procedure.
Fixes BT LL TS 5.1.0 test:
LL/CON/MAS/BV-52-C [Master Receiving Data, LE Coded, CI
Change]
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
flash_erase cannot erase a page from the STM32xx SOC flash. It seems
that the erase wait time is not enough and against to what the SOC's
datasheet states. As a result this patch doubles the wait time.
Signed-off-by: Ioannis Konstantelias <ikonstadel@gmail.com>
Because NXP MPU's regions are dynamically enabled/disabled, ENET
device's access maybe restricted when switching out of a task.
Background DMA transfers from RAM to MAC/PHY may happen during MPU
region reconfiguration or core idling.
Enabled ENET (Kinetis MPU Master 3) to always have access to RAM address
space.
Signed-off-by: Andrei Gansari <andrei.gansari@nxp.com>
Stack canaries require that the z_libc_partition be added to
the memory domain, otherwise user thread access to the
stack canary value will result in an MPU/MMU fault.
These tests define their own domains to test specific userspace
features. Adding another partition to them would be invasive,
would potentially break some platforms with a limited number
of MPU regions, and these tests are not designed to validate
stack canaries anyway, we have other tests for that.
Fixes: #17595
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
In z_sys_mem_pool_block_alloc() the size of the first level block
allocation is rounded up to the next 4-bite boundary. This means one
or more of the trailing blocks could overlap the free block bitmap.
Let's consider this code from kernel.h:
#define K_MEM_POOL_DEFINE(name, minsz, maxsz, nmax, align) \
char __aligned(align) _mpool_buf_##name[_ALIGN4(maxsz * nmax) \
+ _MPOOL_BITS_SIZE(maxsz, minsz, nmax)]; \
The static pool allocation rounds up the product of maxsz and nmax not
size of individual blocks. If we have, say maxsz = 10 and nmax = 20,
the result of _ALIGN4(10 * 20) is 200. That's the offset at which the
free block bitmap will be located.
However, because z_sys_mem_pool_block_alloc() does this:
lsizes[0] = _ALIGN4(p->max_sz);
Individual level 0 blocks will have a size of 12 not 10. That means
the 17th block will extend up to offset 204, 18th block up to 216, 19th
block to 228, and 20th block to 240. So 4 out of the 20 blocks are
overflowing the static pool area and 3 of them are even located
completely outside of it.
In this example, we have only 20 blocks that can't be split so there is
no extra free block bitmap allocation beyond the bitmap embedded in the
sys_mem_pool_lvl structure. This means that memory corruption will
happen in whatever data is located alongside the _mpool_buf_##name
array. But even with, say, 40 blocks, or larger blocks, the extra bitmap
size would be small compared to the extent of the overflow, and it would
get corrupted too of course.
And the data corruption will happen even without allocating any memory
since z_sys_mem_pool_base_init() stores free_list pointer nodes into
those blocks, which in turn may get corrupted if that other data is
later modified instead.
Fixing this issue is simple: rounding on the static pool allocation is
"misparenthesized". Let's turn
_ALIGN4(maxsz * nmax)
into
_ALIGN4(maxsz) * nmax
But that's not sufficient.
In z_sys_mem_pool_base_init() we have:
size_t buflen = p->n_max * p->max_sz, sz = p->max_sz;
u32_t *bits = (u32_t *)((u8_t *)p->buf + buflen);
Considering the same parameters as above, here we're locating the extra
free block bitmap at offset `buflen` which is 20 * 10 = 200, again below
the reach of the last 4 memory blocks. If the number of blocks gets past
the size of the embedded bitmap, it will overlap memory blocks.
Also, the block_ptr() call used here to initialize the free block linked
list uses unrounded p->max_sz, meaning that it is initially not locating
dlist nodes within the same block boundaries as what is expected from
z_sys_mem_pool_block_alloc(). This opens the possibility for allocated
adjacent blocks to overwrite dlist nodes, leading to random crashes in
the future.
So a complete fix must round up p->max_sz here too.
Given that runtime usage of max_sz should always be rounded up, it is
then preferable to round it up once at compile time instead and avoid
further mistakes of that sort. The existing _ALIGN4() usage on p->max_sz
at run time are then redundant.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The block_fits() predicate was borked. It would check that a block
fits within the bounds of the whole heap. But that's not enough:
because of alignment changes between levels the sub-blocks may be
adjusted forward. It needs to fit inside the PARENT block that it was
split from.
What could happen at runtime is that the last subblocks of a
misaligned parent block would overlap memory from subsequent blocks,
or even run off the end of the heap. That's bad.
Change the API of block_fits() a little so it can extract the parent
region and do this properly.
Fixes#15279. Passes test introduced in #16728 to demonstrate what
seems like the same issue.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Logger had false assumption that once log_panic is called then
context switch will never occur and was not protecting against
reentrancy in panic mode. Added interrupt locking when accessing
unprotected part.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
z_arm_do_syscall is executing in privileged mode. This implies
that we shall not be allowed to use the thread's default
unprivileged stack, (i.e push to or pop from it), to avoid any
possible stack corruptions.
Note that since we execute in PRIV mode and no MPU guard or
PSPLIM register is guarding the end of the default stack, we
won't be able to detect any stack overflows.
This commit implement the above change, by forcing
z_arm_do_syscall() to FIRST switch to privileged
stack and then do all the preparations to execute
the system call.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
System call arguments are indexed from 1 to 6, so arg0
is corrected to arg1 in two occasions. In addition, the
ARM function for system calls is now called z_arm_do_syscall,
so we update the inline comment in __svc handler.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
We need to correct the inline comment in swap_helper.S,
which is suggesting that system call attempts with
invalid syscall IDs (i.e. above the limit) do not force
the CPU to elevate privileges. This is in fact not true,
since the execution flow moves into valid syscall ID
handling.
In other words, all we do for system calls with invalid
ID numbers is to treat them as valid syscalls with the
K_SYSCALL_BAD ID value.
We fix the inline documentation to reflect the actual
execution flow.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
device_get_binding() compares pointers first before doing strcmp().
However, enabling coverage forces -O0 to disable any compiler
optimizations. There would be multiple copies of the same string,
and the code pathing doing pointer comparsion would not be tested
at all. So add this flag to merge string constants such that
the pointer comparison would be exercised.
This also adds a bad driver which fails initialization. This is
to make sure that execution path is covered.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The special 'unittest' target has largely been supersesed by
native_posix, and converting this to a regular test will allow
us to see code coverage for the CRC functions in our coverage
reports.
Fixes: #16943
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This test uses ztest, there is no need to set this and it
was breaking builds with CONFIG_COVERAGE=y
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Adds coverage for the uncovered json_calc_encoded_len()
and covers a bunch more error cases.
This gets us up to 90.1% line coverage and 100% function
coverage.
Fixes: #16944
Partial fix for: #16011
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This mechanism had multiple problems:
- Missing parameter documentation strings.
- Multiple calls to k_thread_name_set() from user
mode would leak memory, since the copied string was never
freed
- k_thread_name_get() returns memory to user mode
with no guarantees on whether user mode can actually
read it; in the case where the string was in thread
resource pool memory (which happens when k_thread_name_set()
is called from user mode) it would never be readable.
- There was no test case coverage for these functions
from user mode.
To properly fix this, thread objects now have a buffer region
reserved specifically for the thread name. Setting the thread
name copies the string into the buffer. Getting the thread name
with k_thread_name_get() still returns a pointer, but the
system call has been removed. A new API k_thread_name_copy()
is introduced to copy the thread name into a destination buffer,
and a system call has been provided for that instead.
We now have full test case coverge for these APIs in both user
and supervisor mode.
Some of the code has been cleaned up to place system call
handler functions in proximity with their implementations.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If the offset within the thread struct to the
ARC arch-specific 'relinquish_cause' member is too
large, ld_s instructions referencing it will not
compile. This happens easily if CONFIG_THREAD_NAME
reserves a name buffer within the thread struct, since
all the arch-specific members come last.
Use the regular 'ld' instruction instead.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
sflists have a couple APIs related to sfnodes that aren't
present for slists. There were uncovered, write some tests
for them.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We were testing all the slist APIs, but not the sflist
variant. Make a copy of the slist tests for sflist,
with the names properly changed.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Some of the slist APIs were only being indirectly exercised;
add to the slist test case to cover everything explicitly.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Added a fix handling L2CAP start frame with payload length
of zero which otherwise sent zero length data start PDU on
air.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
This function doesn't do anything, and only exists so that
it can be overridden later, exclude from coverage reports.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The attribute handle used to read next attribute has to be incremented
to not loop reading the same attribute.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
When using the procedure Read By Type the response may contain multiple
instances so it needs to be parsed properly. When dealing with long
values only the beggining will be read, for the remaining bytes the
application should issue another bt_gatt_read with offset so Read Blob
procedure is used as recommended by the spec:
BLUETOOTH CORE SPECIFICATION Version 5.1 | Vol 3, Part F page 2312:
The Read Blob Request would be used to read the remaining octets of a
long attribute value.
Fixes#16107
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
This patch adds support for Read Using Characteristic UUID which is one
of the procedure to read the characteristic value especially when the
client only knows the characteristic UUID and does not know the handle
of the characteristic.
Signed-off-by: Tedd Ho-Jeong An <tedd.an@intel.com>
Log all send errors, but don't try to call e.g. prov_send_fail_msg()
since that'll almost certainly fail as well.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Mesh Profile Spec v1.0.1 | Section 5.4.2.3:
"The Provisioner and the device shall check whether the public key
provided by the peer device or obtained OOB is valid (see Section
5.4.3.1).
When the Provisioner receives an invalid public key, then provisioning
fails, and the Provisioner shall act as described in Section 5.4.4.
When the device receives an invalid public key, then provisioning
fails, and the device shall act as described in Section 5.4.4."
This is also in Erratum 10395 which is Mandatory for Mesh v1.0.
The code was already rejecting the key, however that rejection
happened only after we had already sent our public key as response,
which got interpreted as acceptance by the tester (PTS).
Fixes MESH/NODE/PROV/BI-13-C.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Mesh Profile Spec v1.0.1 Section 5.4.2.3: "If the public key was not
available using an OOB technology, then the public keys are exchanged
between the Provisioner and the unprovisioned device. For each
exchange, a new key pair shall be generated by the Provisioner and the
unprovisioned device."
This allows passing MESH/NODE/PROV/BV-12-C.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Clear the callback list once generation is complete and we've done
calling all callbacks. This lets us use bt_pub_key_gen() multiple
times, which before this patch could have resulted in a corrupt linked
list.
Also remove redundant callback dispatching from bt_pub_key_gen() since
the function checks for the PUB_KEY_BUSY flag in the beginning, i.e.
there cannot be other pending generation actions at this point in the
code.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Update code to handle other users of the public key generation APIs
by fetching the current public key at the beginning of each SMP
session. This is particularly important if someone creates the (rather
odd) combination of Mesh and SMP where Mesh will regenerate a new
key pair after provisioning.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Mismatch in Public Key type will cause device to send Invalid Format
error, and treat any further PDU's as unexpected.
This affects MESH/NODE/PROV/BI-03-C test case.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Provisioning errors shouldn't cause device to close link. Upon error,
device will send Provisioning Failed PDU, and any further PDU's will
be considered as unexpected as per Mesh Profile section 5.4.4.
Also a timer is started every time device sends or receives a PDU.
This affects MESH/NODE/PROV/BV-10-C test case.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Introduce a separate buffer pool for events which the HCI driver
considers discardable. Examples of such events could be e.g.
Advertising Reports. The benefit of having such a pool means that the
if there is a heavy inflow of such events it will not cause the
allocation for other critical events to block and may even eliminate
deadlocks in some cases.
Also update all mesh samples not to specify explicit RX buffer counts
anymore. Instead, create appropriate defaults in Kconfig so that we
only need to override this in the app for cases like the bbc:microbit
with limited memory.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This event is a priority one, so it's not safe to have it use the RX
buffer pool which may be depleted due to non-priority events (e.g.
advertising events). Since the event is consumed synchronously it's
safe to have a single-buffer pool for it. Also introduce a new
bt_buf_get_evt() API for HCI drivers to simplify the driver-side code,
this effectively also deprecates bt_buf_get_cmd_complete() which now
has no in-tree HCI driver users anymore.
Fixes#16864
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Test error cases and alternative implementation to bring code
coverage up to 100% file / 100% line.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Current socket connect call implementation always takes
K_FOREVER timeout value, which blocks TCP connections
in case failure. TCP connections waits until it receives
SYN ACK. If there is no SYC ACK means, connect call is
blocked forever.
Added a Kconfig option to define timeout value. Default
value is 3000 milliseconds. User can modify it.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Modify dns_unpack_answer() function to check if the answer is
compressed or not, and return correct values regardless.
Fixes#16594
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We had no system call coverage for k_thread_suspend
and k_thread_resume.
Some unnecessary cleanup tasks in the test case have
been removed.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We had plenty of coverage for k_cycle_get(), but not its
32-bit variant. Run a case in user mode so that the system
call handler gets covered.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If the net_mgmt event has some info, like IP address, that
could be sent, then send it the same time. This is very useful
for the receiver of the event in order to know that is happening
in the system.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The info parameter is difficult to use if the caller does not
get information how long the info struct is. So add info_length
parameter to net_mgmt_event_wait_on_iface() and
net_mgmt_event_wait() APIs.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We didn't have code coverage for this function anywhere
except indirectly through some network tests; exercise it
in the suite of userspace tests.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Address a coverage gap in kernel/userspace.
Unfortunately, in the process of fixing this, a bug was
discovered, see #17023.
This test is user mode specific, filter the testcase
on whether userspace is available instead of ifdefing
the code.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Return value from close() can be ignored in sntp_close()
as it is not returning value to caller anyway.
Coverity-CID: 198863
Fixes#16584
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This test case is so timing sensitive that gathering code
coverage data screws up the results.
Since this is an abnormal execution environment anyway,
just skip the assertions if CONFIG_COVERAGE=y.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We didn't have any coverage of the system call handlers for
k_wakeup() and k_is_preempt().
Increase RAM requirements due to stack alignment constraints
on MPU platforms when user mode is enabled.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Fix how the tstacks array was declared extern so this
actually compiles on all platforms with user mode enabled.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
When advertising with different identities we need to flag any
programmed RPA as invalid if it was generated using a different
identity.
Fixes#16893
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
- Erase operations must be aligned to the erase-size.
- Don't need to perform an alignment check on a full erase. The offset
is not used in this case.
- Don't need to perform alignment check on a sector sized erase, as
this alignment is checked on entrance to the function.
- Removed some whitespace.
Signed-off-by: Ryan C Johnson <ryan.johnson@flex.com>
Fix iterating past the response which causes an invalid memory to be
accessed and passed over to the callback as if there were more
attributes found.
Fixes#16602
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Update the Replay Protection List handling for segmented messages to
be more in line with Figure 3.43 in Mesh Profile Specification 1.0.
This means that the RPL check and update need to be split into two
independent steps rather than always doing these together.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This patch fixes following bug:
After first GC operation the 1st sector had become scratch
and the 2nd sector had became write sector. After that NVS
was initialize (via reboot) again - it recognized the 1st
sector as write sector and 2nd as undone GC destination sector,
therefore it cleared 2nd sector and re-run GC, which implied data loss.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
The syscall handler for k_poll() returns error values
instead of killing the caller for various bad arguments,
cover these cases.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If multithreading is disabled, thread_entry() never runs
since we cannot create threads; the non-multithreading case
was simply dead code.
Indicate to code coverage that CODE_UNREACHABLE should be
skipped.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
These were never getting called anywhere from user mode,
except for k_queue_alloc_append(), but only by virtue of
some workqueue tests.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Threads that are sleeping forever may be woken up with
k_wakeup(), this shouldn't fail assertions.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
LCOV/gcovr doesn't understand what CODE_UNREACHABLE means.
Adding LCOV_EXCL_LINE to the macro definition unfortunately
doesn't work.
Exclude a bit of code which spins endlessly when multi-
threading is disabled that runs after the coverage report
is dumped.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We don't get any coverage past when we dump the coverage data,
so exclude the end of the function and move setting the main
thread as nonessential to immediately before the coverage dump.
The comment was also amended.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
data copying and bss zero are called from arch code
before z_cstart(), and coverage data gathering doesn't
work properly at that point. Not all arches use this
code anyway, some do it in optimized assembly instead.
Weak main() is also excluded; it does nothing and every
test overrides it.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Addresses coverage gaps. Some changes were made so that exited
threads do not have k_thread_abort() called on them.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
It is possible that iface is NULL when selecting IPv4 destination
address for a sent packet.
Coverity-CID: 198877
Fixes#16570
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
When building without support for user mode (CONFIG_USERSPACE=n)
we need to correct the starting address of the MPU Guard, before
passing it to the function that evaluates whether a stack
corruption has occurred. The bug was introduced by commit
(60bae5de38) in
PR-13619, where the start address of the MPU guard was properly
corrected, but the guard start at the corresponding stack-fail
check was not adjusted accordingly.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When trying to establish friendship the Friend must respond to the
initial Friend Poll with a Friend Update. If this initial Friend Update
response is not received the Friendship establishment process must start
again.
When starting a second Friendship establishment processes the `sent_req`
field of the `lpn` struct was left set to `TRANS_CTL_OP_FRIEND_POLL`.
This prevented the initial Friend Poll being sent out on the second
attempt. Since the Friend Poll was not sent, no timeout is set and
nothing happens ever again. No more Friendship Requests are sent.
This commit clears `sent_req` back to zero when no Friend Update
response has been received after the initial Friend Poll.
Fixes#16678
Signed-off-by: Rich Barlow <rich@bennellick.com>
For the request:
DCID shall map to rx.cid:
'This field specifies the endpoint of the channel to be disconnected
on the device receiving this request.'
SCID shall map to tx.cid:
'This field specifies the endpoint of the channel to be disconnected
on the device sending this request.'
For the response when receiving the roles are inverted.
Fixes#16799
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Although unlikely it is possible that a remote may attempt to send just
1 byte as the write request allows to do that:
BLUETOOTH CORE SPECIFICATION Version 5.1 | Vol 3, Part F
page 2320:
'If the attribute value has a fixed length and the Attribute Value
parameter length is less than or equal to the length of the attribute
value, the octets of the attribute value parameter length shall be
written; all other octets in this attribute value shall be
unchanged.'
Fixes#16734
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Mesh Profile 1.0 Section 6.6:
"The timeout for the SAR transfer is 20 seconds. When the timeout
expires, the Proxy Server shall disconnect."
This will let qualification test case MESH/SR/PROX/BV-05-C pass
without requiring an explicit disconnect.
Fixes#16600
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
It is possible that UDP header cannot be accessed so we need
to check that we do not do null pointer dereference.
Coverity-CID: 198866
Fixes#16581
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
GATT long messages require buffers to store multiple packets. The actual
configuration for the bluetooth peripheral sample defines only two
buffers to store these messages. As the default ATT payload size is
18 bytes, the maximum payload size for a long ATT message is 36 bytes.
As the variable `vnd_long_value[]` has a defined size of 74 bytes, it
cannot be fully modified due to the 36 bytes limitation.
Set CONFIG_BT_ATT_PREPARE_COUNT=5 so all the maximum payload size goes
to 90 bytes and all the 74 bytes can be written to `vnd_long_value[]`.
Signed-off-by: Joao Cordeiro <jvcc@cesar.org.br>
When there is no need to divide the PWM clock (i.e. the requested
period cycles fit the 15-bit PWM counter), the prescaler value
should be 0, not 1.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Since there is no NFC antenna connector on the dongle and the pins
P0.09 and P0.10 that are dedicated to NFC functionality are in the
group of just a few ones available for external connections, it seems
more reasonable to configure these pins by default as regular GPIOs,
as users will most likely want to use them in this way.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
The _k_mutex linker section is used to gather instances of
struct k_mutex into a list so that init_mutex_module() could iterate
that list to perform runtime initialization tasks. In this case, we're
not defining a struct k_mutex but rather a struct pthread_mutex which is
a completely different structure. Not only those struct pthread_mutex
would be corrupted with unexpected data, but since they're not the
same size as struct k_mutex, the actual struct k_mutex instances that
follow in the list would be misaligned and get corrupted too.
There is nothing that requires runtime initialization in the static
definition of a struct pthread_mutex so let's remove the section
attribute.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
OpenThread uses CONFIG_OT_PLAT_FLASH_PAGES_COUNT to calculate the # of
pages at the end of flash to use for storing OpenThread settings.
This calculation has an off-by-one error which sets the offset for
the storage area as 1 page of flash too low.
For example, on nRF52840:
- default setting for CONFIG_OT_PLAT_FLASH_PAGES_COUNT is 4
- flash size is 1MB (0x100000)
- flash page size is 4096 (0x1000)
- expected offset is 0xfc000
Using the current logic we get an offset of: 0xfb000
Fixes: https://github.com/zephyrproject-rtos/zephyr/issues/16339
Signed-off-by: Michael Scott <mike@foundries.io>
Partially revert commit ea177e785c
("usb: dfu: set bwPollTimeout dynamically")
Introduced fix does not work proper because there is no way to be
sure that a control stage had success before start erase process.
Instead IMG_ERASE_PROGRESSIVELY configuration should be used
if the erase of the flash takes longer time.
resolves: #15497
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
Fix the encryption setup queueing implementation to avoid
overlapping with local initiated Length Update Procedure.
Fixes#15733.
Relates to #15335, and #15186.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
According to documentation Compare event will not be triggered
if CC=0 and CLEAR task is set. Added handling of that situation.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Add implementation to toggle GPIO Debug pins on HFCLK
request and release by the controller.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix XTAL advanced feature by adding the missing
implementation to calculate and, retain or release the XTAL
clock source after a Bluetooth state or role is stopped.
Fixes#15817.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix the missing reset of Encryption Procedure state when the
peripheral responded with error reason as pin or key missing
which otherwise caused connection disconnection on next
reception of data or control packet.
Relates to #15570.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Non-XIP system with FLASH_SIZE = 0 is no-flash system. And no-flash
system makes text, rodata, and data all in SRAM, so define the marco
ROM_ADDR to RAM_ADDR.
Fixes: #16027.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
This is a workaround for IOP issue, where peer rejects LLCP Slave
Connection Parameter Request with LMP Error Transaction Collision
error code even if previous request is complete at the instant.
Relates to #15366.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
on non-XIP system, SRAM is the default region, and relocated .data
section and .bss section of SRAM shouldn't be inserted between
_image_rom_start and _image_rom_end, because the memory region between
_image_rom_start and _image_rom_end will construct the mpu ro region.
Also for the newly added memory region on non-XIP system, the
relocated .text secition and .rodata section should also be mpu aligned.
Fixes: #16090.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
When some header are included into C++ source file, this kind of
compilations errors are generated:
error: invalid conversion from 'void*'
to 'u32_t*' {aka 'unsigned int*'} [-fpermissive]
Signed-off-by: Benoit Leforestier <benoit.leforestier@gmail.com>
Just like board's can be placed in out-of-tree BOARD_ROOT's, we now
support DeviceTree sources and bindings being placed in out-of-tree
DTS_ROOT's.
This required for out-of-tree drivers that use DeviceTree.
To implement this we get rid of various user-settable CMake variables
like DTS_APP_BINDINGS, DTS_APP_INCLUDE, and instead have ZEPHYR_BASE,
APPLICATION_SOURCE_DIR and out-of-tree directories conform to using
the same DTS_ROOT concept and directory structure.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
All DT include directories are system directories, except for
dts/common. There is no reason for dts/common to diverge, so we align
it with the other directories to be consistent.
Also, we move it closer to the other include directories.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Bits are actually shifted by filter number instead of bank number.
This results in wrong mode and filter_index calculation.
Fix shifting of mode bits by using bank_number instead of filter_number.
Signed-off-by: Alexander Wachter <alexander.wachter@student.tugraz.at>
Backport to 1.14 the move of the configuration options that control the
west build command to the page where the command itself is described.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
When making a build folder pristine until now we were running the
'pristine' build target. The issue with that is that ninja/make or
whatever build tool is being used might decide to re-run CMake itself if
some of the dependencies have changes. This might trigger an error that
is unfriendly and unnecessary, since the user is explicitly asking for
the build folder to be wiped before starting a fresh build.
To avoid this issue restor to running directly the CMake script that the
'pristine' build target itself uses, so as to make sure that the build
folder is wiped unconditionally regardless of changes made to the tree.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Add a new command-line and build config option, `pristine`, that the
user can pass to `west build` or set in its configuration file(s) in
order to automatically trigger a pristine build on every build or
whenever west considers it required.
The option can take the following values:
- never: Never run the target
- always: Always run the pristine target before building
- auto: Run the pristine target when required
With `auto`, the pristine target will be run when running
west with an existing build folder containing a build system and:
- Selecting a different board from the one currently in the build system
- Selecting a different application from the one currently in the build
system
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The -B option has always existed but was first officially documented in
CMake 3.13.0. In that same release the -S option was introduced,
replacing the old undocumented -H. Switch to using the officially
documented options.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Move includes to outside of the extern "C" statements to fix C linkage
issue.
| /include/misc/util.h:53:1: error: template with C linkage
| template < class T, size_t N >
| ^~~~~~~~
Signed-off-by: Greg S. Woods <gwoods@lexmark.com>
Enable pull-up on UART TX pin to reduce power consumption.
If the board is powered by battery the SoC consumes more
power than expected.
The consumption increases because TX pin is floating
(High-Impedance state of pin B from Dual-Supply Bus Transceiver).
Similar to commit b5b728495b
("boards: reel_board: enable pull-up on UART RX pin")
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
Libraries that use mbedTLS have been invoking
zephyr_link_interface(mbedTLS). It is not clear what the intent of
this code has been, but it is redundant with the mbedTLS build
scripts, so it can be safely removed.
In addition to being redundant, it causes problems as it introduces an
ordering dependency, with this code mbedTLS must be declared before
users of mbedTLS are declared. Since this code is redundant, this
ordering dependency is also unnecessary.
This code is believed to have been added early on by accident and
copied through cargo-cult programming since.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
'west --version' output has the path to the west manifest directory.
If there is a semantic version string in this path name, it will
raise a false version check failure.
Signed-off-by: Thomas Stilwell <Thomas.Stilwell@nordicsemi.no>
The upper limit of the timeout should not be 0.
tests/drivers/watchdog/wdt_basic_api checks for this and fails as the
driver currently only checks that the timout does not exceed the upper
bound.
This also makes it check the lower bound, so that the test passes.
Signed-off-by: Benjamin Valentin <benjamin.valentin@ml-pa.com>
Fix regression by the addition of CONFIG_BT_CONN conditional
compilation that disabled the advanced clock (crystal)
oscillator management when only observer and/or advertiser
states supported in the controller build.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Previous version calculated rt_clock_base incorrectly by subtracting
clock_gettime from the specified time. Effectively the following
formula was used.
rt_clock_base := new_time - clock_gettime()
This is clearly incorrect when we consider what should happen if we
call clock_settime with the result of clock_gettime. It ought to be
approximately a no-op, but instead we end up zeroing the clock.
rt_clock_base := clock_gettime() - clock_gettime() = 0
This patch fixes clock_settime by instead using k_uptime_get to
calculate rt_clock_base, like so:
rt_clock_base := new_time - k_uptime_get()
Trying the earlier thought experiment we get:
rt_clock_base := clock_gettime() - k_uptime_get()
Using the definition of clock_gettime this expands to:
rt_clock_base := (rt_clock_base + k_uptime_get()) - k_uptime_get()
The two k_uptime_get() terms cancel out, leaving:
rt_clock_base := rt_clock_base
I.e. the no-op that we expect when calling clock_settime with
the result of clock_gettime.
Note: The bug is only observable when rt_clock_base is non-zero.
So when clock_settime is called for the first time, it will appear
to work correctly since rt_clock_base is initialized to 0.
Signed-off-by: Alexander Mihajlovic <alexander.mihajlovic@endian.se>
The net_if_ipv4_select_src_addr() should return global address
in the interface if nothing else is being found.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we receive a multicast ICMPv4 packet, then send the reply back
with correct source address and not with multicast address.
Fixes#16257
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Let's filter out on a state parameter.
There is no impact as this function is not used anywhere yet.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Make sure we are able to receive UDP packets with broadcast
destination address. If CONFIG_NET_IPV4_ACCEPT_ZERO_BROADCAST
is set, then check here also non-standard broadcast address
that is described in RFC 1122 chapter 3.3.6.
Fixes#11617
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
With -O0 optimizion, gcc compiler doesn't inline "static inline"
marked function. So when function call return from function
set_and_switch_to_psp which is to switch sp from MSP to PSP, the
ending "mov sp, r7" instruction will overwrite the just updated
sp value(PSP) with the beginning stack pointer(should be MSP)
stored in r7 register, so the switch doesn't happen. And it causes
unpredictable problems in the initialization process, the backward
analysis for this problem can be found on Github issue #15794.
Fixes: #15794.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
k_stack_alloc_init() was creating a buffer that was 4 times
too small to support the requested number of entries, since
each entry in a k_stack is a u32_t.
Fixes: #15911
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The function net_pkt_get_info() is found in net_pkt.h and not
in buf.h. This caused
"implicit declaration of function ‘net_pkt_get_info’"
warning if CONFIG_NET_BUF_POOL_USAGE was enabled.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Send ZLP when host asks for a bigger length and the last chunk
is wMaxPacketSize long, to indicate the last packet.
Signed-off-by: Marcin Szymczyk <Marcin.Szymczyk@nordicsemi.no>
We are just at the knife edge with 512, with stack
overflows being observed with stack canaries enabled.
Given the special case for the idle thread stack size
on this arch, seems reasonable to increase it here
for that arch.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This test uses ztest anyway, the default should be fine
just like any other test running under ztest.
k_thread_create() uses a lot of stack, and the main
stack size is very small if ztest is enabled. Do it in
another ztest task instead.
We don't need to mess with the main thread's priority,
just have the alt thread run cooperatively.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The source address in unicast DHCPv4 Request packets was found out
to be all zeros address 0.0.0.0. This address is only acceptable if
the destination is a multicast one, where the host in question is
acquiring a DHCP address lease. This is true for the DHCP Discover
and the initial DHCP Request message from the client towards the
server. As subsequent DHCP Request renewal messages are sent as
unicast to the server, the server will drop such packets.
Fix this issue by explicitely specifying what source IP address is
to be used, if none is specified, the all zeros address 0.0.0.0 is
used in multicast addresses. The source address in the other
unicast cases is identical to the 'ciaddr' in the DHCP message.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Depending on optimization level, the TC_RESULT_STR[]
array could actually be placed in memory instead of
being expanded at compile time, resulting in memory
access errors from user mode.
Just replace TC_RESULT_TO_STR() with an inline function
containing a switch statement instead.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
There are limits to CMSIS stack sizes, we can't just
add a number to it. Use the configured maximum.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Set to same default as regular stacks. This doesn't use
any extra memory until CMSIS_V2_THREAD_DYNAMIC_MAX_COUNT
is set. 0 is not a valid default if that is set.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The various struct pipe_sequence were not located in memory
accessible to user mode. With optimization turned on, they
weren't in memory at all, but with code coverage enabled
the arrays were actually being read, resulting in memory
access failures from user mode.
Fix them by placing in ROM, they never get modified.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Depending on configuration, this value could end up as
a variable and not an array symbol, causing a crash if
newlib decides to call _sbrk on behalf of a user thread,
which needs to perform arithmetic on it.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
After the rework in #12403, specifying a shield which has overlay
out of the tree causes unnecessary inclusions of overlays.
For every board root, overlays that have same index as
expected overlay are being included.
Fix this by removing already included overlays from SHIELD list.
Signed-off-by: Marcin Szymczyk <Marcin.Szymczyk@nordicsemi.no>
If the user attempts to send data before the USB connection is
established (see the HID sample for an example of such code), the
DataInCallback never gets called which leaves the write semaphore in a
taken state forever.
Signed-off-by: Josef Gajdusek <atx@atx.name>
The SYSCFG_CFGR1_PA11_PA12_RMP define is present even on packages where
the remap isn't strictly required. This commit makes the remap optional
based on a DT property.
Also fixes syntax error caused by a missing );.
Signed-off-by: Josef Gajdusek <atx@atx.name>
SAMD20 does not have the DMA or USB peripherals and as a result
the IRQs for all subsequent lines are shifted down from SAMD21.
This splits the interrupt assignment for the SERCOMs into the
SoC specific DTS file and moves the USB definition to SAMD21 only.
Signed-off-by: Derek Hageman <hageman@inthat.cloud>
The radio API expects the setting of the TX-power in dBm. The actual
TX-power is however set by a 6bit register, which mapping between
register value and power in dBm is a lookup-table in the datasheet.
This mapping for the kw41z was off, which not only lead to incorrect
output power, but also to a maximal output power of only -1.9 dBm
instead of the possible +3.5 dBm.
Signed-off-by: Tobias Aschenbrenner <tobias.aschenbrenner@blik.io>
Signed-off-by: Franco Saworski <f.saworski@posteo.de>
This implements three API functions that are required for
tests/subsys/usb/device to build:
- usb_dc_ep_disable()
- usb_dc_ep_halt()
- usb_dc_ep_flush()
While halt and disable are trivial, flush is just a stub for now.
Signed-off-by: Benjamin Valentin <benpicco@googlemail.com>
Make sure the parameters for the API functions are valid, return error
otherwise.
This is expected by the tests/subsys/usb/device test case.
Signed-off-by: Benjamin Valentin <benpicco@googlemail.com>
For example it should be possible to compile and use the SLIP module
with NET_L2_DUMMY. This required the following changes:
* Fix a typo in the initializer for struct dummy_api
* Only define eth_capabilities if CONFIG_NET_L2_ETHERNET is defined to
silence a -Wunused-function compiler warning
* Unconditionally include net/dummy.h
Signed-off-by: Sören Tempel <soeren+git@soeren-tempel.net>
Without this change compilation of the SLIP module would fail if LLDP
wasn't enabled. There is also no need to include net/lldp.h explicitly
as net/ethernet.h will include it conditionally if CONFIG_NET_LLDP is
defined.
Signed-off-by: Sören Tempel <soeren+git@soeren-tempel.net>
We need to re-initialize the trickle->timer delayed work to use
trickle_timeout() if there are multiple triggerings.
Fixes#15606
Signed-off-by: ling wei <lingwei@cisco.com>
Enable the clock for GPIO ports on the RV32M1 SoC before attempting to
access the port controller registers.
Fixes: #15339
Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Excerpt from the manual:
If ptr is NULL, then the call is equivalent to malloc(size) [...]
Without this commit, such calls end with a BUS FAULT.
Signed-off-by: Tomasz Gorochowik <tgorochowik@antmicro.com>
There was a stray close group comment @} that was causing a large chunk
of the watchdoc API documentation to be missing.
Fixes: #15678
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
The size of the ROM region is now rounded up to the
nearest power of two; we no longer assume that RAM
is in a different part of memory.
Fixes: #15558
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
In addition to checking that the ARP entry does not exist as the
implementation is done currently, also check if the ARP packet
is due to IPv4 link local address configuration. In both cases
use the provided IPv4 address instead of the one set for the
interface.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
IPv4 link local uses ARP to detect conflicting addresses. Properly
set the ethernet packet type to NET_ETH_PTYPE_ARP when probing
for address duplicates.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
If the local node keeps getting bombarded with messages, it's possible
that the storage timer gets rescheduled over and over again and never
expires. Add the necessary code to only reschedule the timer if the
new deadline is earlier than an existing one.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The default values for the timeouts, as well as non-defaults in most
Mesh samples, use a higher value for the RPL than then generic mesh
storage timeout. This hasn't had any effect in practice since the code
only uses the RPL timeout if it is *smaller* than the generic one.
The original intention of the code was to use the RPL timeout,
regardless of what the generic one is, whenever the RPL is the only
thing that needs updating. Add some helper macros to track the various
groups of pending flags, and perform the appropriate checks to apply
the RPL timeout whenever it's smaller than the generic timeout, or if
there are no other items to store besides the RPL.
Fixes#15904
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If promiscuous mode support is disabled in Kconfig and promiscuous.h is
included the build will fail. The returned errno should be ENOTSUP.
Signed-off-by: Brett Witherspoon <spoonb@cdspooner.com>
2019-04-17 08:35:18 -05:00
320 changed files with 6619 additions and 1881 deletions
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
