Correct the logic that determines when CIADDR is set and when it is
left as all zeros. RFC2131 requires that CIADDR is only set when the
client is in specific states.
Change-Id: I93a45e355fbef8d11dba5ac194570b87c594656e
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
RFC2131 places explict requirements on a client w.r.t which messages
are broadcast and which are unicast directly to a server. Notable
rules as they apply to the current dhcpv4 implementation are that
DISCOVER and a REQUEST in response to an OFFER are broadcast. A
REQUEST in state RENEWAL is unicast. There are further rules relevant
to the REBINDING state which is not yet implementated.
Adjust the current implementation that always uses broadcast to use
unicast as required by RFC2131
Change-Id: I6edef4241bcd74623a804a73415888cd679888d0
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
RFC2131 requries that a client MUST NOT include server identifier in
DHCPREQUEST from state RENEWING.
Change-Id: I0fba703f3a3b218af1ac03f4f1f0daa290bd428f
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
The DHCPv4 client obtains resources from the server on a lease. Once
a lease expires those resources must be relinquished.
Update the DHCPv4 client to explicitly remove leased resources from
the network stack once a lease has expired.
Change-Id: I5d3b7a02e463041cfdee1d104f5962498bdd6a30
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
The existing dhcpv4 initialization code places all initialization in
the per iface dhcpv4_start() code. The net_udp_register() setup needs
to happen once for the dhcpv4 instance rather than once per iface.
Subsequent patches in this series also need a place to perform one
time initialization, independent of iface.
Factor the one time setup code out of the existing net_dhcpv4_start()
into a dhcpv4_init() function. Rather than use SYS_INIT() to run
dhcpv_init() we hook directly into net_init().
The prototype of dhcpv4_init() is shared within the net subsystem with
a new private header file deliberately to avoid exposing the function
via the public API in include/net/dhcpv4.h
Change-Id: I3502a53cc3bfe4db4e4cd22c02ae133c266fdf10
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Factor out common code associated with entering the REQUESTING state.
Change-Id: I0fc038a49309625c24b874dfda0355794cd41da9
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Factor out common code associated with entering the SELECTING state.
Change-Id: If34929e5f13dff7650388a2b491a01cbbf5dc100
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Refactor send_discover() to provide a clean separation between state
machine and action.
Change-Id: I96510cc6a7ab16216821fe65ea4289a78f6443aa
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Refactor send_request() to provide a clean separation between state
machine and action.
Change-Id: If4c61c5789c919ab3ce6c1f914afc243176760d8
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Split handle_reply into distinct state machine and actions.
Change-Id: Ib177c7a4007662f941f156acfa5791e1d27e3d73
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
The dhcpv4 state machine implementation names differ from the names
used by the relevant rfc. There is no good reason for them to be
different, so rename them.
Change-Id: I837f1bc4788c0ed4d2949b12eb6f5bfeef9a0be8
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
There are two implementations of dhcpv4 state name numbers to human
readable strings, unify the two.
Change-Id: I1d654918bb919108a0d8c5514b309b193c9c3f96
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
The dhcpv4 state machine has a dedicated context within each net_if
structure. For reasons unknown the timers used by the dhcpv4 state
machine have been placed in net_if outside of the dhcpv4 context area.
Relocate them into the dhcpv4 context.
Change-Id: I0531f493610dffda9ca9208993597a5665bde997
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Change the message type representation from integer to enum, this will
help prevent re-occurrence of recent issues where states and events
are interchanged by making it more likely the compiler will see a type
mismatch.
Change-Id: Ia235afda428a9e5dfbd933e02beeae468b4c84a2
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Refactor and simplify packet construction. This resulting code is
simplier, easier to read and compiles a lot smaller.
Change-Id: I43c67d79fbb77f85af54355eca41f516054cbba3
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
This function is used only within net_context.c, make the symbol static
then.
Change-Id: Ib2c00a21c25e8c6a1404d6345d4b8ae05b779525
Signed-off-by: Jesus Sanchez-Palencia <jesus.sanchez-palencia@intel.com>
'query->value' is a 'const uint8_t *' so it should be casted to a
char (signed) array before it is used in places where a 'char *'
is expected, strncmp() is an example.
Jira: ZEP-1810
Change-Id: I94cf780a40ad5fed29607d2302dc7a10387bb86f
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Since we try to match with POSIX behavior as much as possible, let's
not bother the user unless they need to be bothered. The recv(2) POSIX
syscall won't return 0 on stream sockets unless the connection has been
closed by the peer; however, that was happening with the callback set
by net_context_recv().
Change it so that the callback is never called if operating on a stream
socket and appdatalen is set to 0.
This is similar to a previous patch sent by Michael Scott in [1], but
not relying on the actual TCP flags: only on the appdatalen and the
socket type.
[1] https://gerrit.zephyrproject.org/r/#/c/9949/
Jira: ZEP-1632
Change-Id: Ib0c214fc9269d305a03e8d85eb606f106c45b038
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
If the DAD timeouts, then the pending pointer will be null
when we remove the neighbor. Fortunately this only prints
this error message and does not cause any issues in the code.
[net/nbuf] [DBG] net_nbuf_unref_debug: (0x00118350): *** ERROR *** \
buf 0x00000000 (nbr_clear_ns_pending():175)
Change-Id: I3e11d4aa1d90f205df591b5d5cdcf2ee7bde6c01
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It is possible that the buffer waiting for retransmission is modified
after it is sent, for example, it can be compacted by 6lo, and our
assumption of where is the message ID is located in the buffer is no
longer valid.
As the message ID is the only information that is necessary for
keeping track of retransmissions, we keep a copy of it in the pending
struct, as well as the destination address of the retransmission.
Change-Id: Id33d54353404628673541225a1a05e27ee08765f
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
When we do DAD (Duplicate Address Detection), the local IPv6
address gets added to the neighbor cache. This is useless so
remove it after DAD has finished.
Change-Id: I9625d367e96d8108a7d3d1d8b2e95f3c4ea11c45
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add net_ipv6_nbr_rm() utility function that can be used to
remove an IPv6 neighbor from the cache.
Change-Id: I9794856a4f65c5e943656970648e5c5762b0338c
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add utility function to go through all the stored neighbors
in the IPv6 neighbor cache.
Change-Id: I42fe0ec48c000215403aef63629d0763189ebdbb
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The sequence and identifier fields are 16-bit instead of 32-bit
long. This did not cause any issue in Echo-Reply but those two
fields should be set properly.
Change-Id: I5e4878f53d6bb37660d46d173159d27bbe0e94dc
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
TCP header was not sent back to originator when ICMPv6
error message was prepared to be sent.
Change-Id: I171bd724c4260b83d7d1c37e0894f9ed8cddd2c9
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
In order to see who is freeing the fragment, add function
and line information to net_buf_frag_del() when net_buf
debugging is activated.
Change-Id: I732f579fab2390cb16804cb35b83f46e65fca342
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
TCP maintains 'sent_list' for retransmission if it doesn't get ACK for it.
Same list is not freed on net_tcp_release() call. This causes memory leak.
Change-Id: I2b2def1ea19487cc48ea4fbb6343ef0c773f288f
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Due to commit fece856959 ("net: tcp: Clean up FIN handling") the
tcp_established() callback now handles TCP connections which are
in various ending/closing states other than TCP_ESTABLISHED.
Currently, these states are generating the following error and not
being processed:
Context 0x123456778 in wrong state 6.
(Shown when TCP is in LAST_ACK state).
This commit also fixes a memory leak issue discribed in
Jira: ZEP-1658
Analysis of the memory leak issue is here:
When TCP connection is established, tcp context is in
NET_TCP_ESTABLISHED state. Once it receives FIN message from client
it goes to NET_TCP_CLOSE_WAIT and then it turns to NET_TCP_LAST_ACK
after connection closing request from server. Now server gets final
ack from client, but tcp_established() will reject it because current
state is not in NET_TCP_ESTABLISHED. Even if server receives proper
ack, it is not handled by server. Hence 'sent_list' is not freed.
Change-Id: I41c8af2e6851809f87a02c271a4290cf3d823ebb
Signed-off-by: Michael Scott <michael.scott@linaro.org>
NET_ASSERT(net_nbuf_iface(buf)) should be called before setting
it on context [net_context_set_iface(context, net_nbuf_iface(buf))].
Change-Id: I9a1da1214857e96e03784bc98a9aae5cf59ef0fc
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Using char or uint8_t relevantly.
Jira: ZEP-1723
Change-Id: I512cb6ff4800cd23f6539e7a47c7f3c72dc94183
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
The interface L2 address type is set at the same time as the
L2 address is set to the network interface. This is most
convinient place to set the address type.
Change-Id: I712d7357d075959eb79df3463141cfbc6d163a74
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Currently, for the following MQTT msg fields:
- client_id
- will_topic
- user_name
- topic
their length is computed inside the routine that receives the MQTT msg.
Although this simplifies development, also imposes one restriction:
data must be null-terminated. Sometimes, data is received from other
sources and not generated by the application, so the null-terminated
constraint may be considered problematic for the user.
This patch removes the assumption that string fields are null-terminated.
Current data structures are already prepared to handle this case, so no
API change is required.
Change-Id: I5a147a5b21e0da49541cbe62baac363c8737cd3e
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
This patch updates the Remaining Length field from uint16_t to
uint32_t. The MQTT std specifies that this field must be
unsigned 4 bytes length.
Change-Id: I319d0745c673faece4bbd4db29b1bafad78ac199
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
The dhcpv4 client code builds ip and udp packets from scratch rather
than using the network stack to do the heavy lifting (why ?).
When it computes the udp checksum of each packet it builds it neglects
to clear any preexisting detritus from the checksum field. The result
of this is that some packets will be built with correct checksums and
some will be built with incorrect checksums.
This is the underlying reason that the dhcp client often taken many
retransmissions and elapsed time before in order to acquire an IP
address.
Change-Id: Iebd1ed34e06f7f2e53d45f6d1555e22f48490287
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Fix long standing issue where a dhcpv4 message type is compared
against a dhcpv4 state machine state name rather than a message type.
The issue probably arizes due to the similarity in names between
messages and states. By accident, the relevant message types and
states happen to share the same numbers, hence the implementation
works, but is ill defined.
Change-Id: I5c028de4336ff42f6696e28b3492c932c58b5a05
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Makes it cleared that zoap_update_from_block() doesn't modify the
packet.
Change-Id: I35429b153370c50eb5ae9c914b47a3144faf2f04
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This fixes the case that a request for block number (NUM) 0, using a
16 byte block was considered invalid.
This was because it is encoded as the value 0 (zero), which can be
expressed as the BLOCK1 option present but without any value
associated. The old code considered this the same as the option not
existing.
Change-Id: I0f3912803a88865e9f544a6d0078ed4231775a88
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Change the handling of iface parameter in net_if_ipv6_maddr_lookup()
function:
* If the *iface is set to NULL, then return the found
interface to the caller.
* If the *iface is not NULL, then use that interface
when doing the lookup.
Change-Id: Ia1f0365170ea9f3e615d189231160614a80d241a
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Because of the change in next commit "net: if: Change
the iface param in net_if_ipv6_maddr_lookup",
we must initialize the network interface to NULL.
Without this the multicast address lookup will fail.
Change-Id: I113b44ce23c5f2ecbbf1698972078f102995e891
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
