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>
Make the IP address parameter const because we are not
modifying the IP address in net_if_ipv6_maddr_add() or
net_if_ipv6_maddr_rm()
Change-Id: I98c19de132e58c386f661e8a76a349d562a82c71
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
As the function does not change the original data, make
the corresponding parameter const.
Change-Id: I1125a2f9205dc73de2f0aac0c30110591baace1e
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Implement RFC2131 4.1.1 requirement that at dhcp implementation wait
for a random delay of 1 to 10s before sending the initial DISCOVER.
Implement RF2131 4.1 requirement that at dhcp implementation set the
initial retransmit timeout at 4 seconds and exponentially backoff on
each retransmit.
Change-Id: Id7029f3ed16a5f886dbd555fed87320aeffe31aa
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
Rather than reuse XID's on retransmissions, always use a unique XID.
Either behaviour is permitted by rfc2131. Debugging the dhcp client
in the presence of multiple dhcp servers with significant packet loss
in the network is much easier if we don't unnecessarily reuse request
identifiers.
Change-Id: I5c82cbdbf3dfc0ef88cae036aa863946844c144f
Signed-off-by: Marcus Shawcroft <marcus.shawcroft@arm.com>
finalize_segment() will call net_ipv4_finalize() or
net_ipv6_finalize(). Both the functions perform net_nbuf_compact().
But after finalize_segment(), net_nbuf_compact() called again, which
is unnecessary.
Change-Id: I9fab63bcc44eec87061a4b55edd5053cf6556a75
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Protocol family is checked in prepare_segment() and in same function
it's again verified by finalize_segment(). So remove the double checking
in finalize_segment().
Change-Id: I17123ab8741d017d7e3ff1ef3fb07371b0d4aa66
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Using net_buf_ref() technically works but debugging the network buffer
allocations is more difficult if done like that.
Change-Id: Iac81bd3ab95547741d49f32763baaa54e97b4877
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
If a net context is not connected or listening, we can go ahead and
call net_context_unref() to free it up instead of waiting for
FIN_ACK which will never happen.
Change-Id: Ice06f572df64f2edb5918c10c92087ce0b7b254a
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Several error messages are currently being logged as NET_DBG which
requires the user to have CONFIG_SYS_LOG_NET_LEVEL=4.
Let's show these as errors so they are more visible.
Change-Id: I28c9a1aedb78787ef098a9bf565472a437373933
Signed-off-by: Michael Scott <michael.scott@linaro.org>
Current API description of net_nbuf_compact() is not very clear.
The first parameter needs to be the first net_buf in the chain.
The changes to this API are needed in order to clarify following
use cases:
1) User provides fragment that is not first of the chain and compact is
successfully done. In this case there is no free space in fragment list
after the input fragment. But there might be empty space in previous
fragments. So fragment chain is not completely compacted.
2) What if input fragment has been deleted and api returns the same
buf?
So this commit simplifies the API behavior. Now net_nbuf_compact()
expects the first parameter to be either TX or RX net_buf and then it
compacts it. It fails only if the input fragment is a data fragment.
Change-Id: I9e02dfcb6f3f2e2998826522a25ec207850a8056
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
The net_nbuf_push() API is not used by anyone. Semantics are not
clear and following patch requires changes to push api, so removing
this API for now. If needed this can be re-introduced later.
Change-Id: I1d669c861590aa9bc80cc1ccb08144bd6020dac5
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
The ethernet header location was incorrectly calculated
and the result pointer had some random value.
Change-Id: I6b2deee787a78444f3ee3be805d4b82ebb6c3664
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Currently fragmentation introduces empty fragment in the
beginning of the list, and adds fragmentation header, and moves frags
data to previous fragments. This is done based on condition that max
data should be multiples of 8 bytes and offset is based on before
compression.
It will fail at scenario when current fragment has not enough space
to move from next fragment and next fragment has more than allowable
max bytes. This will cause memory overflow which is typically seen
as double-free error to the user.
This is solved more simple way by this commit. First detach frags list,
prepare new fragment and attach that to buffer. Then move data from
detached frag list and free the fragments in the old one.
Change-Id: I5e3693d47828ff3b92db4ba5f6c00c0b751daadc
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
If we are in the end of the fragment chain, then we can just
bail out as there is nothing more to do. There will be a
double free if we continue as the last entry is already
removed at this point.
Change-Id: I0f9782b408244d283dc7e3e087359dd00bada7a9
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit changes the net_buf getter functions in nbuf.h
by adding a timeout parameter. These function prototypes
are changed to accept a timeout parameter.
net_nbuf_get_rx()
net_nbuf_get_tx()
net_nbuf_get_data()
net_nbuf_get_reserve_rx()
net_nbuf_get_reserve_tx()
net_nbuf_get_reserve_data()
net_nbuf_copy()
net_nbuf_copy_all()
net_nbuf_push()
net_nbuf_append()
net_nbuf_write()
net_nbuf_insert()
Following convinience functions have not been changed
net_nbuf_append_u8
net_nbuf_append_be16
net_nbuf_append_be32
net_nbuf_insert_u8
net_nbuf_insert_be16
net_nbuf_insert_be32
net_nbuf_write_u8
net_nbuf_write_be16
net_nbuf_write_be32
so they call the base function using K_FOREVER. Use the
base function if you want to have a timeout when net_buf
is allocated.
Change-Id: I20bb602ffb73069e5a02668fce60575141586c0f
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Prior to commit df201a0e4b ("net: context: Assign a random port
number when context is created"), TCP clients were assigned a random
port number when the incoming sockaddr parameter's port value was 0.
After the above commit, this is now broken as it will bind to port 0.
If left this way, every TCP client would need to add a line of code
copying the context->local sockaddr_ptr's port value into the
src_sockaddr port prior to calling net_context_bin().
Instead, we can reinstate this behavior in net_context_bind(), by
making sure we only overwrite the randomly assigned port in the
context->local sockaddr if the incoming sockaddr parameter has a
port which is != 0.
Change-Id: I0f27f031f743d50c351ecf9ab55b5282a20ff292
Signed-off-by: Michael Scott <michael.scott@linaro.org>
If we receive a neighbor advertisement, we need to free its
net_buf because we are returning NET_OK to the caller. This
return code means that we consumed the net_buf but we did
not call net_nbuf_unref() in this case.
Change-Id: Ia6d8f1b440be87eff5d2b14a23336a37be6d6a04
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The pending buf needs to be freed after it has been sent
to the network. We took the ref when pending buf was saved,
now we need to unref it when it is about to be sent.
Change-Id: I1e429969895700000a8aa124bd645db2d52d036c
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Doing it only in net_context, prevented to do it once NS succesfully
finished. This generated an error in 15.4, where pending data had wrong
ll reserve size.
Change-Id: I0f917fb76171457e5dff2c29e44edb8f00662150
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
We must use the nexthop instead of destination address
when sending the packet. The current code mixed destination
and nexthop addresses and ignored the nexthop when sending
neighbor solicitation message.
Change-Id: I53887c16ef6fcf8365f1f47ab5792cb208dd273e
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Keep track of amount of bytes that are sent or received from
all network interfaces.
Change-Id: I706481aab1a7e0cf2bc78d032f2ef4ebbabe3184
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Current behaviour has an issue when UDP context is created with local
port number 0, net_conn_input() happens to treat zero port as
a wildcard ("receive packets for all ports"). net_context_bind()
for a UDP context doesn't affect its existing connection in any way.
Proposed solution is, context should be created with a random free
port assigned and bind() updates connection information from context.
Jira: ZEP-1644
Change-Id: Idb3592b58c831d986763312077b0dcdd95850bc9
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
The port numbers in the range from 0 to 1023 are the well-known ports
or system ports. Do not allocate them for users.
Change-Id: I4d7b4e1314759e4d8b260669946b9880282642c0
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Commit-ID 83ed3a29be changed the way
fragments are managed after being sent, and this brought a bug in the
active scan logic. Up to active scan to keep the buffer referenced
relevantly now, which is simpler than it used to be.
Reported-by: Johann Fischer <j.fischer@phytec.de>
Change-Id: I23db4bab878328b1ca3cb313e737de819177c281
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This adds a NET_L2_BLUETOOTH_ZEP1656 which sets NET_IF_POINTOPOINT to
avoid using nbr cache with Linux peers as they send wrong link
addresses.
Jira: ZEP-1656
Change-Id: I842f4cbb99ae9d9c004494739c07189c191929fe
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
This flag can be used by driver to indicate pointopoint links which should
not require destination link address to be resolved.
Jira: ZEP-1656
Change-Id: I58dd3bf48485d6203e75373497e00668317b9825
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
The net_nbuf_get*() functions were calling net_buf_alloc()
with K_FOREVER. This can cause issue if called from isr
context. The fix is to check if we are in isr and then try
to alloc net_buf with K_NO_WAIT.
Change-Id: I809170f2cd059480d436763e19a35386d9bdf048
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Channel, PAN-ID, short and extended address
Change-Id: Icdf94dff6f59cd155a072a609290197b37bd084c
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Channel, pan-id, short and extended addresses.
Change-Id: Ib63dadac37d649df3efc8fdd67f5312d3a7c8e20
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Fix setting short address through net mgmt API. It's not about
coordinator's short address, but local one.
Change-Id: I320143e40d336a1085bf12b17a88a2c35da83504
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
ORFD or RFD does not matter: the extended address should be set
according to device's MAC address.
Change-Id: I39d09c3a953283eeaa30b908ea159638604bd72b
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This makes sure ll addresses are properly set before sending reducing the
duplicated code on L2 drivers.
Change-Id: I5330c1d00a344e77555c6f31033ae42af20214bf
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Fix comparisons in net_if_get_by_index and net_if_get_by_iface:
__net_if_end is not a pointer to a net_if structure.
Change-Id: Ie8e3a457c3f0fa97c080b38b5b7d2b420c50252b
Signed-off-by: Julien Chevrier <julien.chevrier@intel.com>
There is dead code that is never executed so removing it.
Coverity-CID: 157585
Change-Id: I6926289b5735b78fcb99ad493d1b05198b9e36cc
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Previous commit: 6e6281af96
"net: tcp: Only return -ETIMEDOUT if timeout>0 in connect"
missed that K_FOREVER needs a semaphore taken, but has a
value of -1.
Change the logic here to timeout!=0 to handle this case.
Change-Id: Iae6a784443810130a7de267226483426fbd4f0d4
Signed-off-by: Michael Scott <michael.scott@linaro.org>
The RFC requires we honor the 2MSL TIME_WAIT timeout, support for
which was just removed with the FIN cleanup. Add it back, but make it
optional (proper sequence number and ephemeral port randomization
makes true collisions a birthday problem in a ~80 bit space!).
Change-Id: I176c6250f43bba0c914da1ee7f0136dcb1008046
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The context may live for a while after the user closes it (c.f. TCP),
and of course the documentation specifies that the user must not use
it after calling net_context_put(). Don't confuse them by invoking
their callbacks on the "closed" connection; it's likely that the user
has destroyed her own tracking data and the user_data pointers would
be garbage.
Change-Id: Iba9cc7025c6ea4a94cc4796903966f8d1b831996
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The connection close paths were a little tangle. The use of separate
callbacks for "active" and "passive" close obscured the fundamentally
symmetric operation of those modes and made it hard to check sequence
numbers for validation (they didn't). Similarly the use of the
official TCP states missed some details we need, like the distinction
between having "queued" a FIN packet for transmission and the state
reached when it's actually transmitted.
Remove the state-specific callbacks (which actually had very little to
do) and just rely on the existing packet queuing and generic sequence
number handling in tcp_established(). A few new state bits in the
net_tcp struct help us track current state in a way that doesn't fall
over the asymmetry of the TCP state diagram. We can also junk the
FIN-specific timer and just use the same retransmit timer we do for
data packets (though long term we should investigate choosing
different timeouts by state).
Change-Id: I09951b848c63fefefce33962ee6cff6a09b4ca50
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
A FIN packet generated by a net_context_put() must go into the normal
transmit queue and not be sent synchronously. Previously sent data is
expected to be delivered and acknowledged before the connection is
terminated.
An advantage we get with this change is unified timeout and retry
handling for FIN packets.
Note that there remains a misfeature here where the queing of the FIN
results in a synchronous switching of the connection callback to
tcp_active_close(), which will prevent any further data received from
being provided to the user.
Change-Id: I2d41316549da9fee383b4f32af5e8b3adf4cb122
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The death of a network context was sort of a mess. There was one
function, net_context_put(), which was used both by the user as a way
to "close" the connection and by the internals to delete it and to
"clean up" a TCP connection at the end of its life.
This has led to repeated gotchas where contexts die before you are
ready for them (one example: when a user callback decides the
transation is complete and calls net_context_put() underneath the
receive callback for the EOF, which then returns and tries to inspect
the now-freed memory inside the TCP internals). I've now stepped into
this mess four times now, and it's time to fix the architecture:
Swap the solitary put() call for a more conventional reference
counting implementation. The put() call now is a pure user API (and
maybe should be renamed "close" or "shutdown"). For compatibility,
it still calls unref() where appropriate (i.e. when the context can be
synchronously deleted) and the FIN processing will still do an unref()
when the FIN packets have been both transmitted and acked. The
context will start with a refcount of 1, and all TCP callbacks made on
it will increment the refcount around the callback to prevent
premature deletion.
Note that this gives the user a "destroy" mechanism for an in-progress
connection that doesn't require a network round trip. That might be
useful in some circumstances.
Change-Id: I44cb355e42941605913b2f84eb14d4eb3c134570
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The old default 2 is too low especially if TCP is enabled.
So in order not to confuse the application developer,
increase the default number of network contexts to 6.
Change-Id: I263bb4b6f31354a11d921d94aa97214abd85ae24
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
net_addr_ntop() will convert IPv4|6 address to string form.
Renamed existing net_sprint_ip_addr_buf() to net_addr_ntop()
and adjusted parameters as per API.
Jira: ZEP-1638
Change-Id: Ia497be6bf876ca63b120529acbadcfd9162a96e3
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
This make sure the link-local address of the destination is added to the
nbr cache as that is accessed when calling net_ipv6_prepare_for_send,
this is needed when following RFC 7668 since link-local addresses are
never registered using nbr discovery.
Change-Id: I2bc578d33d1061726d0cbf46e4464df74d79e992
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
