Add OpenThread to Zephyrs net stack as data link layer.
OpenThread requires to call process function when an event occurs.
This process function is called from cooperative thread.
Packet conversion and dispaching is implemented in openthread.c
as well as addresses forwarding.
Signed-off-by: Kamil Sroka <kamil.sroka@nordicsemi.no>
OpenThread requires platform definition with standarized API
so we have to add wrappers to make it compatible with Zephyr.
OpenThread is based on autoconf, this requires
more specific CMakeLists.txt which allows to clone specific
commit or point to local copy of openthread.
Signed-off-by: Kamil Sroka <kamil.sroka@nordicsemi.no>
The meaning of this address type is the same as NET_ADDR_MANUAL,
but with a provision that DHCP can override such an address.
It's intended for the usecase when there's a default static
configuration for when DHCP is not available, but DHCP should
override it.
Before going to add another address type, there was an attempt
to repurpose TENTATIVE address state, but it doesn't work as
expected, as indeed, all existing address types/states already
have clearly semantics, and it makes sense to just another
address type to avoid confusion and unexpected behavior.
Fixes: #5696
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The idea is that static config is used unless/until DHCP values
arrive. This allows to have the same network configuration values
for both a case of direct board - workstation connection (where
DHCP is usually not available), and a case where both a board and
workstation connect to a router (which serves DHCP).
The changes in this commit however take care of netmask and gateway
settings, but not about IP address itself. This is addressed in the
next patch.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Actually set Zephyr's default DNS server based on the corresponding
DHCP option received. This makes DHCP-based setup Zephyr complete:
now it's possible to connect Zephyr DHCP-enabled system to a typical
router, and it will fully auto-configure to access Internet.
This initial implementation uses just first DNS server address as
returned in DHCP message, it may need to be extended in the future
based on the need.
Signed-off-by: John Andersen <john.s.andersen@intel.com>
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
In some cases, we need to initialize DNS servers from a binary
addresses, e.g. in case of DHCP processing. With existing API,
such addresses would need to be converted to strings, just to
be converted back to struct sockaddr in dns_resolve_init().
This is not efficient, and with a number of addresses quite
cumbersome. So instead, allow to pass DNS server either as
strings, or as struct sockaddr's (or both).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The offset of the IP header in a received packet depends on the L2
header size. For Ethernet this is 14 bytes which puts the u32 IPv4
addresses on a non-u32 byte boundary. This causes chips that don't
support unaligned access (like the Cortex-M0) to fault.
The fixes in this patch are enough to ping the board and run the
http_server sample.
Signed-off-by: Michael Hope <mlhx@google.com>
For some reason, ieee802154_verify_channel had a device parameter
instead of an iface like all the others.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
It will be up to the user to configure a valid channel, through
net_mgmt, and call net_if_up() in order to get the device up.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Depending on device's band, the upper channel limit can vary a lot in
Sub-Ghz. Thus verifying it directly in L2 before requesting it to the
device.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Sub-Ghz bands have different limit of channels. 10, or more than a
thousand is actually possible. Thus the device needs to expose such
limit to the L2 which is unaware of frequency band logic. L2 will
then allow user to select a proper channel.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This field is set and maintained, but not actually used for anything.
The only purpose for it would be to validate ACK numbers from peer,
but such a validation is now implemented by using send_seq field
directly.
Fixes: #4653
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Case #1: If ACK received and our retransmit (i.e. unacked) queue is
empty, it's error. It's incorrect because TCP requires ACK to set for
every packet of established connection. For example, if we didn't
send anything to peer, but it sends us new data, it will reuse the
older ack number. It doesn't acknowledge anything new on our side,
but it's not an error in any way.
Case #2: If retransmit queue is only partially acknowledged, it's an
error. Consider that we have 2 packets in the queue, with sequence
numbers (inclusive) 100-199 and 200-399. There's nothing wrong if
we receive ACK with number 200 - it just acknowledges first packet,
we can remove and finish processing. Second packet remains in the
queue to be acknowledged later.
Fixes: #5504
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Per RFC 793:
A new acknowledgment (called an "acceptable ack"), is one for which
the inequality below holds:
SND.UNA < SEG.ACK =< SND.NXT
If acknowledgement is received for sequence number which wasn't yet
sent, log an error and ignore it.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
As title, we should update the existing observe_node when new attributes
are written from server side. Add the implementation to handle that.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
1. Read notification attributes set by server to setup the
minimum/maximum notification period of a observation request.
2. Reordering to check observe_node duplication first
(bailout earlier)
3. Simplify remove observe_node condition checking
NOTE: attributes are inheritable, priority: res > obj_inst > obj
Reference: LwM2M spec V1_0_1-20170704-A, section 5.1
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
Since we've added storing notification attributes written by server.
We can now append these attributes as part of link-format for discover
op.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
Implement write-attribute on obj/obj_inst/res according to LwM2M spec
20170704-A, sec 5.1.2. Support pmin/pmax/st/gt/lt parameters on WRITE
operation.
The basic idea is to add sys_slist_t to obj/obj_inst/res structure.
And attach struct lwm2m_attr to the list when attributes are written
from server side (implement lwm2m_write_attr_handler accordingly)
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
The ICMPv4 handler needs to unref the received echo reply packet
because we are returning NET_OK to caller. Similar change was done
for IPv6 earlier.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of accessing the environment variable ZEPHYR_BASE every time we
require accessing the source code root, use an intermediate variable
that has OS path separators correctly set to '/' to avoid issues on
Windows.
Note: This removes the ZEPHYR_SOURCE_DIR CMake variable. External
applications using that will need to change to use the new ZEPHYR_BASE
variable.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
When testing ping6 with net shell, it was noticed that after some
sucessive calls the applications stopped to handle rx packets.
Analyzing other icmpv6 register callbacks it was verified that is
necessary to unref packets before returning NET_OK.
Signed-off-by: Pedro Martucci <pedropaulomartucci@gmail.com>
commit 2a7546fb5a ("net: lwm2m: add support for coap2coap proxy")
erroneously changed the COAP_OPTION_* used to specify the coap2coap
or coap2http proxy resource used from COAP_OPTION_URI_PATH to
COAP_OPTION_PROXY_SCHEME.
Changing it back to COAP_OPTION_URI_PATH requires us to re-order how
the coap options are appended to the packet as the coap options must
be added in the order specified by the numbers in:
https://tools.ietf.org/html/rfc7252#section-12.2
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
CoAP packet w/ confirmation flag set is required to be retransmitted
before it got the ACK message from the peer.
However, the packet is usally unreference once it's sent to the network.
Although we set the timeout as no wait when calling function
net_app_send_pkt(), it's still possible that the packet is unreferenced
before we got a chance to increase the packet reference by calling
coap_pending_cycle().
Usually, the IP stack will generate an ARP packet first and then send
out the packet. However, this is not the case when the remote is a
loopback address.
As issue #5101 described, when asking client to perform a firmware pull
on URL "coap://127.0.0.1:7783/large". The packet will be unreferenced
immediately after calling net_app_send_pkt(). Which then result in
client hang.
The solution to the issue is to increase the reference count on the
sending packet and decrease it after the process is finished.
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
We shouldn't limit the amount of data appended to RX packets based on
the max send size of TX packets. Skip this check for packets in the
RX slab.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
This reverts commit 3599d793c2.
Setting a packet's data_len here doesn't fix the fact that we shouldn't
be using it at all on RX packets. Fix belongs in net_pkt_append().
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
When testing TLS connections on production web server, it was noticed
that the TLS mainloop was getting "hung" after a connection was made
via HTTPS and then closed by the server. The TLS mainloop was never
being notified that the connection was closed and was stuck waiting
for more data.
The next time that connection was used, TLS would fail to start.
Let's force trigger a closure of the TLS process when the net-app layer
is notified of the connection closure. This allows the connection to
be successfully reused later.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
The net code doesn't use libc stdio stdout in any way, so there's no
need tweak those options.
Fixes: #5565
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
This fixes a bug when calling net_pkt_append[_all] which uses
pkt->data_len as part of the maximum packet length calculation
when the net_context is set.
Without this change the maximum packet length is calculated as 0
(the value of pkt->data_len) and an ENOMEM error is returned.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Commit 753daa6 ("net: pkt: Compute TX payload data length")
removed the default packet setup on incoming packets when they
belong to the rx_pkt pool.
Let's restore this behavior, as MBEDTLS processing in net_app library
needs to use packet family to determine IP header length on
incoming packets.
NOTE: A future cleanup patch could set the IP header length based
on the context IP family. However, there are many places in the code
where this is being set, so care should be taken.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Do not reset net_context information in net_app_close. In case of
TCP behind the curtains connection might wait for some timers to
expire and send some messages (e.g. ACK). If we set source port to
'0', unexpected behaviour might happen with peer connection.
Only reset net_app context related information on net_app_close.
Let net_context_put will take care of proper connection closure.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Right now in FIN_WAIT1 state, if we receive FIN+ACK message, then
tcp state changed to FIN_WAIT2 on ACK flag and immediately on FIN
flag state changed to TIME_WAIT. Then final ACK is prepared and sent
(in queue at-least) to peer. Again immediately state changed to
TCP_CLOSED, where context is freed. net_context_put frees context
and releases tcp connection. Final ACK packet which is in queue
is dropped.
As a side effect of freed ACK packet, peer device keep on sending
FIN+ACK messages (that's why we see a lot of "TCP spurious
retransimission" messages in wireshark). As a result
of context free (respective connection handler also removed), we see
lot of packets dropped at connection input handler and replying with
ICMP error messages (destination unreachable).
To fix this issue, timewait timer support is required. When tcp
connection state changed to TIMEWAIT state, it should wait until
TIMEWAIT_TIMETOUT before changing state to TCP_CLOSED. It's
appropriate to close the tcp connection after timewait timer expiry.
Note: Right now timeout value is constant (250ms). But it should
be 2 * MSL (Maximum segment lifetime).
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
fin_timer will be started after sending FIN to peer. After successful
reception of FIN+ACK message in FIN_WAIT1 state, fin_timer should be
cancelled.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Of these, only struct net_ipv6_nbr_data::send_ns is a descriptive
change:
send_ns is used for timing Neighbor Solicitations in general, not
just for DAD.
The rest are typo/grammar fixes.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Consolidate and standardize error handling throughout
lwm2m_obj_firmware_pull.c. As well as handle previously
unhandled errors returned from transfer_request().
NOTE: in general, unhandled errors will now result in
RESULT_UPDATE_FAILED. Previously, unhandled errors in
transfer_request() would result in RESULT_CONNECTION_LOST
which might or might not be over-written with another
result later.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
The CoAP samples use an MBEDTLS config "config-coap.h" which could be
re-used by the LwM2M sample, except that most servers use a larger
maximum content length setting of 1500 bytes.
Let's add a CONFIG to set this for users of the CoAP lib and set the
CONFIG value for the samples to the 256 size currently used.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Currently, LwM2M firmware download only supports coap2http proxy.
Let's add support for coap2coap proxy as well.
This was tested running Californium demo app cf-proxy on the host
machine with the following setting changed in Californum.properties:
MAX_RESOURCE_BODY_SIZE=524288
Add the following to the samples/net/lwm2m_client/prj.conf:
CONFIG_LWM2M_FIRMWARE_UPDATE_PULL_COAP_PROXY_SUPPORT=y
CONFIG_LWM2M_FIRMWARE_UPDATE_PULL_COAP_PROXY_ADDR="coap://[2001:db8::2]:5682"
Build the sample for qemu_x86 as you would normally, but now
you can use a real world coap address to pull firmware using the 5/0/1
resource. The host machine running cf-proxy will pull the remote
resource and then deliver it to the running qemu sample.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
- Add needed settings for DTLS support to the lwm2m_ctx structure.
- Add initialization of MBEDTLS to the LwM2M lib based on the
user application settings in lwm2m_ctx.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
The default net_context remote address is scrambled when using a
connection via DTLS. Instead let's use the dtls context remote.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
While looping through possible lwm2m_ctx matches, we're referencing
remote before checking that the context itself is valid.
Also, reduce indentation issues.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Instead of building under the "app" context, let's build the
LwM2M library as a separate static library. This will be helpful
later when adding support for DTLS as w/o this configuration,
the build breaks on MBEDTLS config includes.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
This will avoid exposing IEEE 802.15.4 Zephyr's L2 private context data
to unrelevant places.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
