After the Publish Retransmit state was introduced the Publish Period
measurement would begin once the previous Publish message has finished
transmitting. This will however cause inaccurate periods, which is
particularly an issue with the PTS that expects accuracy of less than
0.5 seconds (apparently).
Since the publication timer is also used for the retransmissions we
can't simultaneously use if for the period as well. Therefore, we
introduce a new variable called period_start which makes a note of
when the period was supposed to start, and then once all
retransmissoins are done initializes the timer with the send duration
taken into account.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The only generally available model supporting publication that's
convenient to be used for testing is the Health Server Model.
Unfortunately since this model supports period publication, the
non-periodic side got less attention and had some bugs.
The first thing that needs to be done is to verify that the period
returned by bt_mesh_model_pub_period_get() is positive. If it's zero
then no periodic publication should take place.
Another thing that this patch cleans up is the naming of the callback
used for periodic publishing. There's no need do require the callback
to call bt_mesh_model_publish() since this must happen no matter what,
so instead rename the callback from 'func' to 'update' and have the
access layer call bt_mesh_model_publish() if the callback was
successful.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When a small amount of supported connections is configured (especially
the default number of 1), connectable advertising may fail. This is
perfectly fine since as soon as a disconnection event happens the
advertising will be successfully restarted. To avoid causing
unnecessary user worries, downgrade the resulting errors to warnings.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There were some things that were working only when receiving a Node
Reset message from someone else, but not when the app called
bt_mesh_reset() directly. There was also some state cleanup missing
for the transport layer. This patch addresses all of these issues.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In 90b471fe4, there was a change to make k_poll() return EINTR error
if it was cancelled with k_fifo_cancel_wait(). Handle this change, or
otherwise sockets EOF handling was broken.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Previously, the connection will be reset easily due to a forged TCP
reset with a random sequence number.
As described in RFC793 p.69, we should check if the sequence number
falls into the receiver window at first.
Signed-off-by: Aska Wu <aska.wu@linaro.org>
The Configuration Client is such a generally useful feature for the
shell that it makes sense to have it as a mandatory dependency (the
shell wasn't anyway compiling at the moment without it).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Optimised the parameter passing order of memq interface such
that the compiled code uses less space and execution time.
Having a parameter that gets returned as the first parameter
passed to a function avoids instructions required to have
the result in the return register.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The Del and Overwrite operations have the exact same parameters and
expected status response as the Add operation, so we can reuse most of
the code.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Even though we have LPN enabled, we might still receive messages
through other network interfaces than the advertising one (e.g. the
local network interface).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Using the start callback, especially with multi-segment messages, may
not be super useful for applications, but we should support if if they
do provide it. One application could e.g. be to calculate the duration
it takes for a multi-segment message to be completely received by the
remote end.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Model publication was broken in a couple of ways:
- The Publish Retransmit State was not taken into account at all
- Health Server used a single publish state for all elements
To implement Publish Retransmit properly, one has to use a callback to
track when the message has been sent. The problem with the transport
layer sending APIs was that giving a callback would cause the
transport layer to assume that segmentation (with acks) is desired,
which is not the case for Model Publication (unless the message itself
is too large, of course). Because of this, the message sending context
receives a new send_rel ("Send Reliable") boolean member that an app
can use to force reliable sending.
Another challenge with the Publish Retransmit state is that a buffer
is needed for storing the AppKey-encrypted SDU once it has been sent
out for the first time.To solve this, a new new net_buf_simple member
is added to the model publication context. The separate 'msg' input
parameter of the bt_mesh_model_publish() API is removed, since the
application is now expected to pre-fill pub->msg instead.
To help with the publishing API change, the Health Server model gets a
new helper macro for initializing the publishing context with a
right-sized publishing message.
The API for creating Health Server instances is also redesigned since
it was so far using a single model publishing state, which would
result in erratic behavior in case of multiple elements with the
Health Server Model. Now, the application needs to provide a unique
publishing context for each Health Server instance.
The changes are heavily intertwined, so it's not easily possible to
split them into multiple patches, hence the large(ish) patch.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Model Publish Retransmit Interval is in units of 50ms and not 10ms
like the other transmit/retransmit states. Create dedicated macros for
the Publish Retransmit State and use them where appropriate.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There's no need for callback exposed in the public API to be something
different than what's used internally. In fact this would just
complicate things. This patch exposes the internal callback under a
bt_mesh_adv_cb name and uses it throughout the mesh stack.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In case an outgoing message gets only delivered to the proxy interface
we should not forget to notify the send callbacks of this.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Some places of the code are interested in when the first advertising
event goes out. Others, on the other hand, are interested when the
last advertising event goes out. Some are even interested in both of
these. Instead of providing a single callback, provide a struct with
two possible callbacks for sending advertising PDUs.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This simplifies the callback implementations since they no-longer need
to do their own look-ups of the needed context.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
We've so far been trying to keep the per-buffer mesh meta-data to
a maximum of 8 bytes in anticipation of upcoming net_buf refactoring
that'll have all net_bufs in the system with the same sized user data.
It's however slowly becoming unfeasible to do this - e.g. simply
adding user data to the sending callback would already fill up the
full 8 bytes.
To solve this issue, treat the net_buf user data as just a pointer to
the actual user data, and keep the actual user data in a separate
array. We still keep taking advantage of net_buf_id() however, so
buffers can cheaply be mapped to their meta-data.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add in6addr_any and in6addr_loopback which are defined in RFC2553 Basic
Socket Interface Extensions for IPv6.
Signed-off-by: Aska Wu <aska.wu@linaro.org>
Previously, if passive close is peformed, the net context is released
after FIN is received and FIN,ACK is sent. The following last ack from
the peer will be treated as an improper packet, RST is sent to the peer.
This patch refines tcp_established() by centralizing the tcp state
transition and releases the net context only if NET_TCP_CLOSED is
reached.
Besides, the logic that releases the net pkt without appdata (i.e. ACK
or FIN) is moved from packet_received() to tcp_established(). This makes
packet_received() less dependent on the protocol and make the usage of
net pkt more clear in tcp_established().
Fixes: #4901
Signed-off-by: Aska Wu <aska.wu@linaro.org>
When testing memq implementation used by controller, a
missing check on NULL pointer return could lead to NULL
pointer deferencing.
Current implementation of controller and mayfly do not
by design lead to NULL pointer dereferencing, this fix
is only for correct-ness and complete-ness.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When resetting there is no point trying to start clearing the
Friendship since there will not be any security material left to send
out PDUs.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The common interpretation (among other implementations) seems to be
that Model Publication for a given AppKey Index implies a binding for
that AppKey. This isn't currently explicitly stated in the spec, but
in order to improve interoperability go with this interpretation as
well.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The variable for tracking that a buffer from the Friend Queue hasn't
been sent yet was not being properly set.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If net_context_recv() returns a error, net pkt will not be released. For
example, net_context_recv() returns -EBADF because the TCP connection is
closed by the peer.
Handle the return value instead of using SET_ERRNO().
Signed-off-by: Aska Wu <aska.wu@linaro.org>
If we cannot send a DHCP message, then unref the net_pkt
in order to avoid a buffer leak. Earlier we tried to
unref NULL net_pkt which is not correct.
Signed-off-by: june li <junelizh@foxmail.com>
This simplifies the API since there is no-longer a need to pass a huge
number of function arguments around.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This simplifies the API since there is no-longer a need to pass a huge
number of function arguments around.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add a command for getting and setting the model publication. We also
have to adjust the app's configuration, since both the Model
Publication Set and Status messages are segmented messages, meaning we
need two TX and RX segment contexts.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The value 0xffff is reserved, so we can use that to indicate whether
there's a vendor model in question or not. Simplifies things over the
previously used separate boolean value.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Token is missing when we jump to the error and token is not yet setup.
To correct it, we grab the token from the input packet at the beginning
of the handle_request()
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
There were some needed changes missing after the update to have manual
control of PB-ADV & PB-GATT provisioning bearers. E.g. the test for
CONFIG_BT_MESH_LOW_POWER in net.c was no-longer indicative of scanning
being on or not.
Do the scan enabling selection within the LPN module if LPN is
supported, or otherwise enable scanning unconditionally in
bt_mesh_provision().
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Semaphore state_sem initial count is set to 0 but k_sem_give()
is invoked first in sys_rand32_get() which will block the caller
forever.
Fix the issue by setting the initail count to count_limit.
Signed-off-by: Ramakrishna Pallala <ramakrishna.pallala@intel.com>
It's much more intuitive to read "if (module)" instead of
"if (module != -1)" when checking for a valid module. Update the code
to use struct shell_module pointers instead of integers for tracking
modules.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Make the internal commands (exit, select & help) as any other
commands, so that e.g. "help help" works as expected. Also redesign
the way commands are looked up to avoid duplicate lookups.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The print_cmd_unknown() is just a two-liner and only called from a
single place, so just do the printk calls inline.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The code was completely broken wrt command lookups when the command is
given in the format "<module> <cmd> <args...>". It would only work if
the default module is not set, which is almost never the case (as most
apps set it explicitly).
Refactor the command handling by moving more logic up to shell_exec(),
so that get_cb() does a lookup for a single module
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The way get_command_and_module() is used it's impossible for it to be
given an argv where argv[0] is NULL.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The way the command line parsing works, argv[0] is always guaranteed
to be non-NULL as well as a non-empty string, so doing checks for this
in get_cb() is redundant.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
All code paths in shell_exec() are guaranteed to set err before
returning it, so doing this (rather curious) initialization is
completely unnecessary.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Copy/paste error was checking minimum measurements where it
should have been checking maximum measurements.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Initial values for the min/max measurements were 0 and this caused
issues with sensors maximums that weren't above 0 and minimums that
went below 0. Let's update those to MAX_INT so the first sensor
value update will set those to correct values.
When resetting the measured values, let's use the current sensor
value not 0.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
10 seconds is quite long for configuration messages, and way too much
currently since we only talk through the local networking interface.
Set the default timeout to 2 seconds, and provide APIs through which
the timeout may be changed at run-time (mainly useful for the shell).
Note: The timeout_set() API is normally assumed to be called just once
for an application, based on the expected size of the network (hops &
latency). Trying to change it e.g. in a multi-threaded environment for
every message may not yield the expected results.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Convert the hb-ub-set command to a more generic hb-sub that can be
used both for getting and setting the Heartrate Subscription State.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Convert the hb-pub-set command to a more generic hb-pub that can be
used both for getting and setting the Heartrate Publication State.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Avoid having to go do "init" for the bt module before issuing "init"
for the mesh module. Instead perform Bluetooth init implicitly. The
bt_enable() API will cleanly fail with -EALREADY if it was previously
called.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If the app does direct provisioning, it may still want to do common
handling through its provisioning complete callback (if it has one
registered). This also means that we always require a non-NULL
provisioning context provided to bt_enable(), and that it needs to
fail if NULL was given.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
It may be useful for the app to know what the initial NetKeyIndex that
it was given during provisioning is.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This is in anticipation of soon adding health client support, which
could then cause confusion due to the ambiguous API names.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Now that there's support for configuration client as well, rename cfg
to cfg_srv to avoid any confusion.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add the ability to track the provisioning bearer through an extra
parameter to link_open/close. Also introduce new public functions to
enable/disable specific provisioning bearers. This also means that one
now needs to explicitly enable provisioning bearers after calling
bt_mesh_init().
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The original fragment chain of incoming packet will be lost and leaked
in case of early error, add frag back to packet and
let the caller do unref.
Fixes#4323
Signed-off-by: june li <junelizh@foxmail.com>
No need to do any IPv6 neighbor checks if the packet is routed back
to us by loopback driver.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is loopback interface, then let it handle all local
traffic. Loopback interface is only needed for test applications.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The command name and a shortened form of valid parameters is not
necessarily enough to understand its usage. Add the option of
providing a more lengthy description of the command usage.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If the packet is too short, the TCP header pointer might be
NULL. In this case we just need to bail out.
Coverity-CID: 178787
Fixes#4787
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Most commands provide a short string to describe the parameters it
takes. Provide this help text as part of the list of supported
commands for each module.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add commands for 1-byte states such as Default TTL, Friend and GATT
Proxy, as well as the 2-byte Relay state.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add support to the Configuration Client Model for getting and setting
1-byte states (which can be nicely generalized in code) as well as the
2-byte Relay state.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This patch adds read permission for client characteristic configuration
descriptor. This is required by MESH/NODE/MPS/BV-06-C,
MESH/NODE/MPS/BV-07-C PTS tests.
Signed-off-by: Grzegorz Kolodziejczyk <grzegorz.kolodziejczyk@codecoup.pl>
802.15.4, as other radio tech, works in little endian on network level.
To keeps things simple, the inner context per-interface, stores the
extended address that way. But it can be confusing in shell then, so
let's work handle these addreses through EUI-64 format there.
Fixes#4936
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Add a basic shell skeleton for Mesh, containing basic command for
initialization, provisioning and reset.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Logic for sending chunks of data is incompatible with adding
Content-Length: header.
Per https://tools.ietf.org/html/rfc7230#section-3.3.1:
"A sender MUST NOT send a Content-Length header field in any
message that contains a Transfer-Encoding header field."
Going a bit further in my mind: also don't send Transfer-Encoded
chunked data either when the Content-Length header is present.
In general, there will be problems if the http client library
makes payload changes without the user code knowing about it.
This patch removes the use of http_send_chunk() from the new
HTTP client code and instead sends the payload directly to
http_prepare_and_send()
This fixes an issue where every available buffer would be allocated
with repeating payload data because the for loop in http_request()
wasn't ending until we ran out of memory.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
We should not use the user suppied timeout setting in
http_client_send_req() for the connection timeout. In the
previous API the call to tcp_connect() used
CONFIG_HTTP_CLIENT_NETWORK_TIMEOUT as the timeout setting.
Let's do that here too.
This fixes -ETIMEDOUT error generation when using K_NO_WAIT
for http_client_send_req().
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Add status error string when sending a error message from
HTTP server to client as described in RFC 2616 ch 6.1.
Previously only error code was sent except for 400 (Bad Request).
This also fixes uninitialized memory access in error message.
Coverity-CID: 178792
Fixes#4782
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is a build setup problem where a device driver has not been
setup for the entropy driver then the call to device_get_binding()
will return a NULL value and the code will continue to use this NULL
value. The result is a hard fault later in code execution.
Note that CONFIG_ASSERT is by default off so one has to turn this
configuration on to catch this problem.
Signed-off-by: David Leach <david.leach@nxp.com>
The restoring of the buffer parsing state was only correct for the
friend queue (which needs the app-layer data). Relaying on the other
hand requires the network layer data, i.e. it needs a different state
to be restored.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
We never access the local network interface queue in a blocking
fashing, so it's unnecessary to have all the infrastructure that
k_fifo provides. Use the simpler sys_slist_t instead.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
It may be useful for the app to know when the provisioning link is
active and when it has been closed. This can be used e.g. to signal
the user the state of the device. Some PTS tests also require
verifying the link state.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The code was missing the sending of the Input Complete PDU, and was
also setting the link.auth value too late.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The ping command was not checking if the user gave target
host as a parameter. This would lead to NULL pointer access.
Fixes#4827
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The parameter structs for advertising and scanning contain many
members that may not get explicitly set when enabling these states. Do
a memset to zero on them to make sure we don't operate on
uninitialized memory.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
A value of 20ms means it's possible the LPN will end up doing
simultaneous advertising & scanning, which increases the risk of lost
packets. Use a default of 100ms to keep these separate.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Handling Friend Clear messages and sending the Friend Clear
Confirmation responses wasn't so far properly implemented. One of the
requirements is to keep sending the reponses even though we no-longer
have a friendship. This means that we need to keep the net_idx, frnd
and lpn_counter values valid, which in turn requires the introduction
of a separate "valid" boolean value.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The PTS tests for the Friend node expect a minimum of 16 queued
messages and the ability to have two LPNs. Set these as defaults.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When sending PDUs from Friend to LPN we should adhere to the
configured network transmit count & interval to get better
reliability for the PDUs to be received by the LPN.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When a Friend Node receives a Friend Request with a unicast
PreviousAddress that's not an element on the Friend Node, it needs to
start the Friend Clear procedure. This procedure involves sending
periodic Friend Clear messages to the old Friend of the LPN.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The only messages that should be encrypted using the friendship
credentials are those coming through the Friend Queue on the Friend
node, most request-response pairs between LPN & Friend (exceptions are
Friend Request - Friend Offer, and Friend Clear - Friend Clear
Confirm), as well as Model Publication messages when the Friendship
Credentials Flag has been enabled in the model publication.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When the node is in LPN node, the LPN-Friend messages has its own
retries, so doing this on the advertising level (i.e. following the
network transmit state) is sub-optimal.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This makes it possible (in a subsequent patch) to fine-tune some
special cases, like the LPN poll messages.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Update the advertising callback to include the exact duration that we
will be sending out the packet. This is useful since sometimes we want
to use the end point of the advertising as the reference time to count
when some other action should take place.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Lower the attempts and make sure we track the old Friend address for
subsequent Friend Requests, in case we never receive a Clear
Confirmation.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Add some automated policies for starting LPN establishment and make it
possible to perform the establishment in a "low power" way, i.e.
switching to low duty-cycle already when starting to send Friend
Requests.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Even if endpoint addresses are configurable by each platform,
it would be better to make the default configuration compatible
with a larger board range.
e.g. STM32 OTG FS device has only four endpoints (0x84 is out).
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
If we were asked to add 10KB to a packet, adding it won't help -
such packet won't be even sent by hardware on our side, and if
it is, it will be dropped by receiving side. So, make sure we
never add more data than MTU as set for the owning interface.
This actually gets a bit tricky, because we need also to account
for protocol header space. Typically, when net_pkt_append() is
called, protocol headers aren't even added to packet yet (they
are added in net_context_send() currently), so we have little
choice than to assume the standard header length, without any
extensions.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
According to the Mesh Spec value 0x00 of ReceiveWindow parameter is
prohibited. This is needed to pass MESH/NODE/FRND/LPN/BI-03-C.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The intention of the code was to access client->conn and not
clients->conn (which would always access the first client struct).
Fixes#4738
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The function return type was declared incorrectly. Also, the magic
checking for i after exiting the loop could cause some false positive
compiler warnings. Instead, return directly from the loop (with the
downside of duplicating a few lines of code).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
It's in practice impossible for the time between two beacons to be
more than 50 hours (the approximate wrap-around time for a 32-bit
millisecond timer), so we can use a 32-bit timestamp instead of a
64-bit one for the beacon tracking.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The LE scanning and advertising implementations were allocating and
holding buffers (the scan & advertising parameters respectively) while
at the same time potentially sending other commands (such as setting
the local private address). If these APIs would end up being called
simultaneously from different contexts, this could lead to a deadlock
in trying to allocate HCI command buffers, especially considering that
the default HCI command buffer count is 2.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When sending a packet with AR flag set, the ACK frame that should be
replied to it must holp the same sequence number, so let's verify this
properly.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
There will be place where validating only this part of the frame will be
necessary. This will avoid to run the little bit heavier
ieee802154_validate_frame().
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Introducing CMake is an important step in a larger effort to make
Zephyr easy to use for application developers working on different
platforms with different development environment needs.
Simplified, this change retains Kconfig as-is, and replaces all
Makefiles with CMakeLists.txt. The DSL-like Make language that KBuild
offers is replaced by a set of CMake extentions. These extentions have
either provided simple one-to-one translations of KBuild features or
introduced new concepts that replace KBuild concepts.
This is a breaking change for existing test infrastructure and build
scripts that are maintained out-of-tree. But for FW itself, no porting
should be necessary.
For users that just want to continue their work with minimal
disruption the following should suffice:
Install CMake 3.8.2+
Port any out-of-tree Makefiles to CMake.
Learn the absolute minimum about the new command line interface:
$ cd samples/hello_world
$ mkdir build && cd build
$ cmake -DBOARD=nrf52_pca10040 ..
$ cd build
$ make
PR: zephyrproject-rtos#4692
docs: http://docs.zephyrproject.org/getting_started/getting_started.html
Signed-off-by: Sebastian Boe <sebastian.boe@nordicsemi.no>
Sometimes we need to select Endpoint addresses manually to get it
working with certain USB controllers having limit for endpoints. In
this case default values break endpoint limit check. The proper
solution would be automatic endpoint allocation.
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Add a generic function for TCP option parsing. So far we're
interested only in MSS option value, so that's what it handles.
Use it to parse MSS value in net_context incoming SYN packet
handler.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Calculates full TCP header length (with options). Macro introduced
for reuse, to avoid "magic formula". (E.g., it would be needed to
parse TCP options).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
MSS is Maximum Segment Size (data payload) of TCP. In SYN packets,
each side of the connection shares an MSS it wants to use (receive)
via the corresponding TCP option. If the option is not available,
the RFC mandates use of the value 536.
This patch handles storage of the send MSS (in the TCP structure,
in TCP backlog), with follow up patch handling actual parsing it
from the SYN TCP options.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Add support for selecting Endpoint numbers and move it under USB
Device Networking menu.
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
If we are a Friend node with multiple LPNs, we need to iterate through
all available Friendship credentials to find the right keys.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Mesh spec expects us to fall back to master credentials if
friendship ones are not available. Also remove an unnecessary branch
with the help of a new 'idx' variable.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The friend_cred and new_key information is common for all segments of
a segmented transaction, so it makes sense to store them as part of
struct seg_tx instead of each buffer's user data.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The friend_cred hint needs to be set already at the point of
segmenting, i.e. doing it in bt_mesh_net_send() is too late. Move the
setting to bt_mesh_trans_send() and bt_mesh_ctl_send().
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
MESH/NODE/FRND/LPN/BI-02-C requires us to ignore unknown Transport
OpCodes instead of treating them as valid responses to a Friend Poll.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Mesh Specification recommends retrying up to 6 times the Friend
Poll when establishing Friendship as LPN.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
We should not have any valid key material in key slot 0 if the KR flag
is set, since then the new key/old key information will be incorrect
when network PDUs get decrypted.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
According to the Mesh Profile Specification: "The acknowledgment timer
shall be set to a minimum of 150 + 50 * TTL milliseconds".
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Mesh Profile Specification states that replay protection must be
done for all control and access messages. Furthermore, the replay
protection list must be updated with the sequence from the last
segment of a segmented message (the code was only updating based on
SeqZero).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Now that Friend support is complete we can create a full
implementation of the LPN PollTimeout Get message.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The commit 971da9d0 ("net: pkt: adjust_offset: Simplify and optimize
code") changed the adjust_offset() function but left the error print
intact. This print is now invoked even if there is no error which
looks bad in debug prints.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Even with the privacy feature disabled, the stack has so far defaulted
to using an NRPA for active scanning, in order to protect privacy.
This is mainly because it is not always clear that scanning for other
devices may risk revealing the local identity.
There may however be use cases where such revealing is actively
desired, so introduce a new option for this (which defaults to
disabled).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
conf member represents the configuration state of the device.
Change its type from u8 to bool and clean related functions.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
netusb endpoint config is specific to the function (ECM, RNDIS, EEM..).
Move this config to the function interface.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Move class_handler and send_pkt to netusb function interface.
This makes netusb 'function' agnostic.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Today, we support/use only one USB configuration descriptor.
Moreover I assume multi-config support should be managed at
usb core level and not by each class driver.
Let's track one netusb function per netusb instance for now.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
The addrlen of accept() and recvfrom() is a value-result argument. It
should be updated to the actual size of the source address after
calling accept() and recvfrom().
Signed-off-by: Aska Wu <aska.wu@linaro.org>
Update the firmware update_result accordingly by checking return
value of the firmware data write callback registered by application.
Also, set response code according.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
The expire function can call net_context_unref() which tries to
get a semaphore with K_FOREVER. This is not allowed in interrupt
context. To overcome this, run the expire functionality from
system work queue instead.
Fixes#4683
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
An edge condition was handled in a special way, even though the main
condition covered it well. More code, more jumps == slower code,
bigger binaries.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Create http library that uses net-app instead of net_context
directly. The old HTTP API is deprecated.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Now that net_buf has "native" support for sys_slist_t in the form of
the sys_snode_t member, there's a danger people will forget to clear
out buf->frags when getting buffers from a list directly with
sys_slist_get(). This is analogous to the reason why we have
net_buf_get/put APIs instead of using k_fifo_get/put.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The code parsing received net pkt to get source or destination
sockaddr repeats multiple times in net_context.c.
Eliminate the duplication by net_pkt_get_src_addr() and
net_pkt_get_dst_addr() which can handle different internet protocol
(i.e. ipv4 or ipv6) and transport protocol (i.e. tcp or udp)
Fixes: #4421
Signed-off-by: Aska Wu <aska.wu@linaro.org>
Rename net_pkt_get_src_addr() to net_pkt_get_addr() and make it able to
handle source or destination address.
Signed-off-by: Aska Wu <aska.wu@linaro.org>
Add support for loading IRKs into the controller as well as the LE
Enhanced Connection Complete HCI event. To simplify things, the old LE
Connection Complete handler translates its event into the new enhanced
one which is then the single place of processing new connection
events.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
MESH/NODE/CFG/HBS/BV-02-C expects it to be possible to do a Set with
the existing src & dst addresses but with a zero period in order to
"cancel" the current subscription. In such a case the addresses should
remain set but the period be set to zero, similar to what would happen
if the period would expire.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The heartbeat subscription Count, MinHops & MaxHops should only be
reset when enabling heartbeat subscription. Any other actions should
keep it unchanged.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The heartbeat should only be sent in case the relay state actually
changes. This fixes MESH/NODE/CFG/HBP/BV-03-C.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
"4.1.2 Log field transformation
In order to compress two-octet values into one-octet fields, the
following logarithmic transformation is used: any two-octet value is
mapped onto a one-octet field value representing the largest integer
n, where 2^(n-1) is less than or equal to the two-octet value."
Log field transformation table:
Log Field Value 2-octet Value
0x01 0x0001
0x02 0x0002 through 0x0003
0x03 0x0004 through 0x0007
0x04 0x0008 through 0x000F
0x05 0x0010 through 0x001F
0x06 0x0020 through 0x003F
0x07 0x0040 through 0x007F
0x08 0x0080 through 0x00FF
0x09 0x0100 through 0x01FF
0x0A 0x0200 through 0x03FF
0x0B 0x0400 through 0x07FF
0x0C 0x0800 through 0x0FFF
0x0D 0x1000 through 0x1FFF
0x0E 0x2000 through 0x3FFF
0x0F 0x4000 through 0x7FFF
0x10 0x8000 through 0xFFFF
"4.2.17.2 Heartbeat Publication Count Log
The Heartbeat Publication Count Log value between 0x01 and 0x11 shall
represent that smallest integer n where 2^(n-1) is greater than or
equal to the Heartbeat Publication Count value. For example, if the
Heartbeat Publication Count value is 0x0579, then the Heartbeat
Publication Count Log value would be 0x0C."
According to this definition 2^(n-1) is an upper bound for n log
value.
Proposed Publication Count Log transformation table:
Pub Count Log Value 2-octet Value
0x01 0x0001
0x02 0x0002
0x03 0x0003 through 0x0004
0x04 0x0005 through 0x0008
0x05 0x0009 through 0x0010
0x06 0x0011 through 0x0020
0x07 0x0021 through 0x0040
0x08 0x0041 through 0x0080
0x09 0x0081 through 0x0100
0x0A 0x0101 through 0x0200
0x0B 0x0201 through 0x0400
0x0C 0x0401 through 0x0800
0x0D 0x0801 through 0x1000
0x0E 0x1001 through 0x2000
0x0F 0x2001 through 0x4000
0x10 0x4001 through 0x8000
0x11 0x8001 through 0x10000
According to Log field transformation table 0x0579 would be
transformed to 0x0B and should be to transformed to 0x0C.
This is required to pass MESH/NODE/CFG/HBP/BV-01-C.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Section 4.2.18.2 in the Mesh Profile Specification states:
"The Heartbeat Subscription Destination shall be the unassigned address,
the primary unicast address of the node, or a group address, all other
values are Prohibited."
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This patch fixes issue when receiving iv index greater than current
index + 42 in update mode. According to Specification when node is in
update state it should only accept iv index equal to the current iv
index. When node is in normal mode it should ignore index that is
greater than current index + 42.
This allows to pass MESH/NODE/IVU/BI-02-C.
Also this patch cleans up the iv update procedure, to make it easier
to read.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The lower transport layer is responsible e.g. for the Friend Queue, so
we need to have the buffer in its original parsing state there.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
A regression by commit 9728179757 ("Allow net_context re-connect").
The code did not create IPv4 listener if IPv6 listener was successfully
created.
Fixes#4697
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
A CPU fault occurs when sys_k_event_logger_get_wait()
api invoked with config KERNEL_EVENT_LOGGER_THREAD
enabled.
sys_k_event_logger_get_wait() supposed to pend on a
semaphore when all the events from kernel event logger
are read.But when sys_k_event_logger_get_wait() inovked
with config KERNEL_EVENT_LOGGER_THREAD(i.e captures thread
events) subsquent call inside this function will write to
kernel event log buffer to capture pend event.This will
release the semaphore on which sys_k_event_logger_get_wait()
was pending hence thread gets unpend before swap gets called.
Which in other words a thread which is invoking
sys_k_event_logger_get_wait()(i. e sem_count = 0) get pends
and unpends in single function flow when KERNEL_EVENT_LOGGER_THREAD
enabled.
This would cause overlapping of the stack address where
return address of "_pend_current_thread" stored with esp of
callee saved(i. e thread->callee_saved.esp). Thus return adrress
of "_pend_current_thread" would be overwitten with zero. Which
in turn causes CPU fault.
The thread invoking sys_k_event_logger_get_wait() supposed
to only read the events of the threads which logged to kernel event
logger buffer. But it should not write to kernel event logger
buffer. Otherwise it would cause the race condition explained above.
Signed-off-by: Savinay Dharmappa <savinay.dharmappa@intel.com>
Fail on Mesh initialization if provisioning is enabled and keys were
not generated. This make it simpler to debug misconfigured devices.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
On targets where non-Zephyr controllers are likely, such as qemu, it
may be harmful to try to issue any of the vendor HCI commands, since
non-Zephyr controllers may interpret them in completely different
ways.
Introduce a Kconfig option that, when enabled, uses some simple
heuristics (HCI version & lack of public address) to try to guess in
advance whether the Zephyr HCI vendor extensions are supported or not.
The new option is available for any host-only configuration and is
enabled by default for the qemu targets.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fixes the following conformance test regression failure
introduced in commit 7dd5fbee26 ("Bluetooth: controller:
Fix MIC error due to parallel Enc Proc")
TP/CON/MAS/BV-28-C [Initiating Connection Parameter Request
different procedure collision encryption]
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
We should call coap_update_from_block() which will determine the minimum
size of the BLOCK1 SIZE between server/client and update the current
offset and total size(if available) accordingly.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
Fix to disallow initiating LE Start Encryption while another
procedure is in progress. Similarly, disallow initiating
another procedure while Encryption procedure is in progress.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
This adds an implementation of Xoroshiro128+ PRNG, based on the
original implementation at [1]. This pseudorandom number generator
will use the entropy driver to obtain the seed.
While it uses only 128 bits of state, it's pretty robust for non-crypto
safe applications.
[1] http://vigna.di.unimi.it/xorshift/xoroshiro128plus.c
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Instead of every hardware number generator driver providing an
implementation of this function, use the random device API to
centralize the implementation of this function.
This is a very simplistic function that can be seen as a stepping stone
to refactor the random number generation in Zephyr.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This should clear up some of the confusion with random number
generators and drivers that obtain entropy from the hardware. Also,
many hardware number generators have limited bandwidth, so it's natural
for their output to be only used for seeding a random number generator.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Some "random" drivers are not drivers at all: they just implement the
function `sys_rand32_get()`. Move those to a random subsystem in
preparation for a reorganization.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Add a net_buf_id() API which translates a buffer into a zero-based
index, based on its placement in the buffer pool. This can be useful
if you want to associate an external array of meta-data contexts with
the buffers of a pool.
The added value of this API is slightly limited at the moment, since
the net_buf API allows custom user-data sizes for each pool (i.e. the
user data can be used instead of a separately allocated meta-data
array). However, there's some refactoring coming soon which will unify
all net_buf structs to have the same fixed (and typically small)
amount of user data. In such cases it may be desirable to have
external user data in order not to inflate all buffers in the system
because of a single pool needing the extra memory.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In a case we get packet without fragments return -ENODATA.
The bug was discovered by Coverity.
Fixes#4637
Coverity-CID: 178334
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Previous max range value for RTO was 2 seconds, increase to 60 seconds
as setting larger values can be useful when debugging retransmission
issues on slow networks.
Signed-off-by: Ricardo Salveti <ricardo@opensourcefoundries.com>
Due to parameters used, net_context_recv() call cannot fail (it just
installs a callback, no I/O performed).
Coverity-CID: 178247
Fixes: #4581
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Add option to set initial Retransmission Timeout value. The value is
different from NET_TCP_ACK_TIMEOUT since latter affects TCP states
timeout when waiting for ACK for example.
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
If we receive lot of data fragments, then yield after initial
processing so that TLS thread can start to work on these.
If we do not yield here, we pile up data buffers and might run
out of memory more easily.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It might happen in TCP client, that the TCP connection is terminated
in which case net_context is freed. Check this and mark corresponding
net_context inside net_app to NULL. This way there will be no issue
to access already freed net_context.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The client TLS code did not handle server issued close properly.
Now the connection is terminated properly and TLS thread is left up to
wait more requests from the user.
This commits adds new boolean field to net_app context. Because there
are already multiple boolean flags there, convert them all to bitfields
to save space.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
As the TLS handshake might take long time before connection is ready,
check this before trying to send user data.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Print information that we are sending plain data and receiving
encrypted data, the code claimed that we are sending encrypted
data which is not the case here.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If user closes the client connection, then make sure that
user can just call net_app_connect() instead of calling the
client init. The client initializes everything in net_app but
for simple re-connect that is not necessary.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Now that objects and samples have their return values fixed, let's
propagate them back up to the user if there's an error.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Previously, post_write and execute callbacks returned 1 when handled
and 0 for error condition. However, this wasn't detailed enough and
the engine can't propagate any sort of error back to users -- so it
doesn't even check the return values in many cases!
Let's adjust the resource callback functions of all objects and the
lwm2m_client sample to return 0 for success or a valid error code.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Now that we can access resource data in the lwm2m subsys, let's use
the user provided firmware push buffer (5/0/0) to also store the
firmware pull data.
This way the size of the firmware pull buffer is completely up to the
application.
NOTE: This patch adds a 64 byte firmware buffer to the lwm2m_client
sample for this purpose.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
With the change to support multi-fragement buffers in the LwM2M subsys,
the OPAQUE data type was direct write methods were broken.
Let's fix OPAQUE handling by using the newly introduced getter methods
which can use multiple user callbacks (depending on the size of the
user provided buffer). Let's also add public methods for users to set
/ get OPAQUE data in resources for future use with DTLS key data.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
The lwm2m_engine_get_resource() function needs to be made available to
other portions of the lwm2m subsys in order for firmware resource data
to be used in the future.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
During conversion from the ZoAP to CoAP APIs the use for this variable
was removed, but the variable itself was left in place.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
No need for 2 different defines to specify URI lengths in the source
for firmware pull method. Let's combine them.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Each content formatter should have a way of handling opaque data.
For instance TLV data will individually be able to specify a length
but plain text will take up the rest of the packet.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
The existing LwM2M framework expected contiguous buffers and this
was the reason for the 384 byte buffer sizes. This was previously
a limitation of the ZoAP API. The new CoAP API doesn't have this
limitation and the LwM2M library has already been migrated to use
it.
Let's finish the process by replacing any contiguous buffer handling
with the correct net_pkt APIs to parse across multiple fragments.
Signed-off-by: Michael Scott <michael.scott@linaro.org>
application/octet-stream is used to indicate opaque payload format.
Use plain text handler to handle the opaque format.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>