Adds a simple retry mechanism on top of the existing queueing
framework.
* On the first packet transmitted with an empty queue, the timer is
started (to an initial timeout of 200ms).
* On a received, valid ACK, the timer is restarted.
* The timer is stopped whenever the last packet is ACKed and the queue
becomes empty.
* On timer expiration, the timeout is doubled and the first
unacknowledged packet is retransmitted.
* When receiving an ACK when we are known to be retransmitting, we
mark all sent packets as needing to be sent again.
Change-Id: Iee3544ed43d0a2d50c29fa82f958a34cc10f7cda
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Create a variable in nbuf metadata area for storing TCP
buf_sent information.
Change-Id: I21a85c58183f1e4997d03d4a93546ddd9b32e977
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Retain transmitted packets in the queue, and remove them only when a
matching ACK arrives.
Change-Id: I76d6586879dc0499371c52979128bd47e1204132
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
TCP ACK values aren't known until packet transmission time. Set the
ACK value then, logging the last one sent for later comparison.
Note that this allows us to effect a neat trick: if your
net_context_send() callback produces a synchronous reply on the
stream, Zephyr can include the ACK in the sent data and not have to
transmit a separate packet. Traditional OSes with internal queing and
a syscall interface can't do this.
Change-Id: If8b8546dac6163d763b8f8dd75865192f2eb302a
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
The ACK field of a packet depends on the received sequence number from
the other side of a connection, which can change long after a packet
is prepared/queued. The IP finalize step (which computes things like
the TCP checksum) needs to be invoked just before a packet is sent,
not from prepare.
Split it out to a separate function.
Change-Id: Ibdae0c9b6833bdfb130cc381f823150dceab6637
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Hook up TCP transmission through the net_context_send() API. Queues
packets via a list in the net_buf header, but as of right now simply
transmits the queue synchronously, ignoring the need for retransmit
and the limits of the receive window on the other side.
Requires that the ACK transmission be moved ahead of the net_context
callback invokation. This to work around a glitch in the way ACKs
work with queueing (they depend on current state, but packets are
assembled just once) that will be fixed in a coming patch.
Change-Id: I7490333e4b314e7734fcc03f2a63d76ae89d698a
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Per spec, this is required to be zero when URG is unset, but we were
never initializing it. It seems benign AFAICT, but produces warnings
in wireshark.
Change-Id: Ie4e28070e32767aabc60c63bb85287c1215ebb0d
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Transmitted data coming in via net_context_send() must already have an
allocated TX nbuf. Let prepare_segment() work with one if available,
adding any trailing data to the end of the generated buf chain.
Change-Id: I4f084ce51e6713773a48b8c9c59450656bc87d8e
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Having this unset hits an error check in net_context_send(). Should
consider making this implicit in the destination address being valid
instead...
Change-Id: I52c273d018c3b780837f40678643255da34b6418
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This is dead code in the tree as it stands, and the upcoming rework is
so changed as to make trying to evolve it in place impossible. Also
removes some temporary tracking fields in the net_tcp and tcp_segment
structs which had been intended for this implementation.
Change-Id: I59fe00cd2d5c3e48bc5d38d3d096d9ff72fcdccf
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Port the zperf network performance tool to native IP stack.
TCP support is missing from this version.
Change-Id: I2648e7d81014bb0cae8cf50854b0ae5f89a6567c
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We must not set destination link layer address when generating
ICMPv6 Echo-Reply message. The reason is that the address must
come from neighbor cache right before the reply is sent to the
network.
Jira: ZEP-1360
Change-Id: I32a23f5689107b6827728cce67a5d27fa964ebd2
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we cannot send a network packet, do not quit as that is
really not necessary here.
Change-Id: I7613d46bdbc849b75aaf0fc71c58beba810fd658
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
As the default board is qemu_x86, make sure there is similarly
named config file that can be used by default.
Change-Id: Ibc480d3b4eecc2601e3670c8da19be9c01833c36
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Use by default a board specific config files for both
echo-server and echo-client sample applications.
This makes more sense than using the network technology specific
config file that was used earlier.
Change-Id: I155fc138f4b0ed1166f4e25267513afc3fb6fd76
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Remove legacy stack artifacts from the echo-server and echo-client
config files.
Change-Id: Icc591b9540495c4ee181517369b0df9bb91972bb
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Removed already done items related to IPv6, 6LoWPAN, user space
protocols and documentation from the TODO file.
Change-Id: I0d864675c544ae69ad3ef351c0d42680a6a0b4e6
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Check that we change the ll address in neighbor cache
properly when receiving NA.
Change-Id: Ib0373751abebdfe1c5e83fd38ec91b9587797e32
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Peer address is used when sending NS so that it is added
properly to neighbor cache.
Change-Id: I51c2e236a0b137f7ab086a782f3baced13aa3ee3
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Also disable prefix lifetime overflow test as it does not work
properly.
Change-Id: I54c8716f6a6d4f5bec76b0a61fe0e18e983924f4
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The value is useful debugging information.
Change-Id: I8b0b378299644edb9d887fa3248b3cd821360526
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This is needed so that we can call it from module test
application.
Change-Id: I1e92d1a62af1db823e4f89f30783c94fd7edf324
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Interface pointer was not given in parameter list which
caused garbage to be printed.
Change-Id: If6c4487bb21e0639d5ada283b9e478e5417f18a0
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we are suppose to send an IPv4 buffer that does not have
ethernet header, then add it here so that the packet can
be sent properly.
Change-Id: If873bd299d7f2d6762a6e593ef69043a9190e976
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The cc2520 driver RX thread did not like the stack information
collection macros defined in net_core.h. Both the cc2520 TX and
RX got the same variable name which did not compile.
This is now changed so that the first parameter is added to
the variable name to create a unique variable.
Change-Id: Ia41d01a71afd73af2ef31aa5f7a890a3cf0385aa
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we receive extension 59 (No Next Header), then drop the
packet as we do not support forwarding atm.
Change-Id: I8b1a2a3748f2e88dbf47eb24b732f5edfb9c006c
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If the fragment size is different than the value in IP
header, then drop the packet as there is clearly something
wrong in L2 layer. The fragment size must match the value in
protocol headers.
Change-Id: Ibfdb0ab37b3cd4e81e6e03c36beb1e0abfa8c00d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Same IPv4 address can only be added once to the network interface.
Change-Id: I5fb545dd07c11f0e4ed358759ea39afd1334606f
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Same IPv6 prefix can only be added once to the network interface.
Change-Id: Ib972bb76cf2589d0c873023d1dc2ad49f9b5cf25
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Same IPv6 address can only be added once to the network interface.
Change-Id: Iabc02ffb4026bdccc46e5b2e7ffee82b3c0a59ea
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It is expected that the caller of the net_send_data() function
will release the net_buf or do whatever it wants with it if
there is an error while sending the net_buf.
Change-Id: Id9ddeed3da8b5091c0a8b9d72377f9d12fa76b59
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This enables application debugging using qemu target:
export QEMU_EXTRA_FLAGS="-s -S"
make qemu
Was tested with echo_server
Change-Id: I1906dea2fbc5afe03b11729d2edb8f3629a629f8
Signed-off-by: Ruslan Mstoi <ruslan.mstoi@intel.com>
As the native IP stack is now the default, there is no need
for corresponding Kconfig option.
Change-Id: I08e4992f540f928a2b7378e8803e634e38725348
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit removes the old MQTT and NATS publisher and subscriber
applications that were based on the uIP stack.
Change-Id: I7ba8e155e32ba0aa53c8dec1b6c0cc5170d70943
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
This commit removes the legacy Contiki based uIP stack.
The new native IP stack must be used after this commit.
The commit also removes following things:
- legacy cc2520 driver
- legacy ethernet drivers
- legacy IP stack samples
and changes these things:
- disabled tests that only work for legacy IP stack
- select new IP stack by default
- enable random number generator by default as it is needed
by the new IP stack
Change-Id: I1229f9960a4c6654e9ccc6dac14a7efb9394e45d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The default kernel init priority is too low. Make this
configurable and set the default priority so that the
ethernet driver is started just before the network stack.
This commit adds generic ethernet priority and changes
currently available ethernet drivers to use it.
Change-Id: If695e52b6dd9ea227f10ba306bb145d72d2312b0
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We tried to use the received Echo-Request buffer to send
the Echo-Reply. This did not work that well as then the
source link layer address was set to be the same as
destination address by the lower level L2 driver.
Jira: ZEP-1350
Change-Id: I0e7c06866d5b863f091e1d5c2a79dfcc2f23df20
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Refactor code to exit earlier in case of error.
Change-Id: I9ff893f923ff877d69b0e985bb6e4159504e67cd
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
The "net conn" net-shell command will print information
about currently active TCP connections.
Change-Id: I6ef67887997553f4bfad14bcfd4fcde0e062a77b
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This is to be used in net-shell to view currently active
TCP connections.
Change-Id: I9c2a69e3ab6013835a42dfde47d580623998b3fa
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The "net conn" net-shell command will print out currently
active network contexts (sockets).
Change-Id: Iaf3ced1a4b7ab10291c0b87eecd9e9bedb486e06
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The address family of local endpoint address was not
set in net_context. This caused wierd issues as the
address family was always AF_UNSPEC.
Change-Id: Ibd05f9af86a8bdea3399ad20e3b84bd7d43a6415
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This is to be used in net-shell to view currently
available network connections.
Change-Id: Iadb6555ed6db4e8fb0639e6d0dcf1eccd970acca
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If IP stack offloading is enabled, then call corresponding
L2 driver when corresponding net_context API function is
called. The L2 driver can then do its magic to send/recv
IP packet or do what ever with it.
Change-Id: I26695cee6feb80a41923db0165f22d94477333be
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
For dual IPv6 and IPv4 stack, the limit of network connections needs
to be increased as the default value of 4 is too small in this case.
Change-Id: If031ae9ee1dd095d20ee6a677ef8c678055ecfb0
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Return -ESHUTDOWN when this flag is set.
Change-Id: Ia2f5d8890b70652c1c348f25d4879248540ac66e
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Andy Ross