This causes num of packets command handling to dead lock since it
is done on RX buffer destroy if used on TX it may not be destroyed
on time causing.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
The specification states that the AdvA in a Scan Request packet should
be identical to the one sent in the original advertising packet, so
check this when processing a Scan Request.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
When the initiator is using an RPA and we match it using the filtering,
we should return early and allow the device packet through instead of
going through the whole resolving list.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The specification states that filter policies shall be ignored for
directed advertising, so reflect this behaviour in the code.
Additionally when the local device is using RPAs but the peer uses an
identity address, the resolving list index needs to be updated when
traversing the RL to reflect that there has indeed been a device match
even though the IRK match did not happen.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The AdvA in the CONN_IND packet must match the AdvA in the advertising
packet that triggered it regardless of the advertising type.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Added BUILD_ASSERT check for Tx and Rx thread priorities.
The Tx thread priority shall be higher than Rx thread
priority in order to correctly detect transaction violations
in ATT and SMP protocols. The Number of Completed Packets
for a connection shall be processed before any new data is
received and processed for that connection.
The Controller's priority receive thread priority shall be
higher than the Host's Tx and the Controller's Rx thread
priority.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
When the ARP message is received when the device is starting up,
the network interface might not yet have IPv4 address setup
correctly. In this case, the IP address pointer could be NULL
and we must not use it for anything.
Fixes#752
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Coverity reported false positives, add comment about these in
the code.
Jira: ZEP-2344
Jira: ZEP-2345
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This patch fixes a regression which the original patch introducing
this code (improving concurrent connection handling) had on
*sequential* connection handling. Without this patch, with the
default CONFIG_NET_TCP_BACKLOG_SIZE of 1, after each connection
request, there was 1s (ACK timeout) "dead time" during which new
connection wasn't ptocessed.
This is because k_delayed_work_remaining_get() was checked the
wrong way. But there's no need to use k_delayed_work_remaining_get()
at all, instead just call k_delayed_work_cancel() and dispatch on
its return code.
Note that there's still a problem of synchronizing access to
the global array tcp_backlog, as worker (which modifies it) may
preempt packet handling code (which also modifies it).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The sequence number validator was checking the seq numbers
incorrectly. This caused some valid RST packets to be dropped
and the TCP stream to hang.
Added also a TCP test case that tests the seq validator.
Jira: ZEP-2289
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Usually it is not enough to have just one IPv6 multicast
address defined for the network interface. So allocate three
IPv6 multicast addresses for the network interface as IPv6
by default uses multicast a lot. This hopefully will avoid
some mysterious errors if the addresses run out.
Note that this will increase memory usage a bit so you might
need to lower the count in your conf file if memory is low.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The commit 210c30805b ("net: context: Close connection fast
if TIME_WAIT support is off") was not a proper way of closing
the connection. So if Zephyr closes the connection (active close),
then send FIN and install a timer that makes sure that if the peer
FIN + ACK is lost, we close the connection properly after a timeout.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The commit 00ac0487b0 ("net: context: Remove tcp struct SYN-ACK
timer handling") removed also the passive close ACK timer.
Adding that ACK timer back so that we can close the connection
properly even if the last ACK from peer is lost.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This adds a second vendor service testing if service changed indications
works with more than one change in a row and as a bonus it implements
echo attribute which notifies any data that is written to it:
> ACL Data TX: Handle 3585 flags 0x00 dlen 9
ATT: Write Command (0x52) len 4
Handle: 0x0013
Data: 0000
< ACL Data RX: Handle 3585 flags 0x02 dlen 9
ATT: Handle Value Notification (0x1b) len 4
Handle: 0x0013
Data: 0000
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
With introduction of bt_gatt_service_unregister it is now possible to
unregister service at runtime.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
There could be situations where many services are changed in a row which
would cause k_sem_take to block on the second change, but if the calling
thread is actually the RX thread then this will deadlock since the RX
thread is the one processing the confirmations of indications and it is
blocked k_sem_give is never called.
To solve this the services changes are now offloaded to the system wq
and the code will attempt to consolidate the range being changed so only
one indication is send. If for some reason another changes is caused
while confirmation is pending we just reschedule it to run later to
avoid blocking the system wq in the same way.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
For consistency with the Resolving List, rename the whitelist filter to
match its type and the privacy-enabled version to align it with the
short name used in the RL.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
To correctly handle the allocation and deallocation of resolving list
items, the ll_rl_find and wl_pees_find functions have been modified to
return an invalid index instead of a negative value in case of error.
This is to avoid the ambiguity when mixing indices and negative numbers,
which does not play well with the first index 0.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Upcoming memory protection features will be placing some additional
constraints on kernel objects:
- They need to reside in memory owned by the kernel and not the
application
- Certain kernel object validation schemes will require some run-time
initialization of all kernel objects before they can be used.
Per Ben these initializer macros were never intended to be public. It is
not forbidden to use them, but doing so requires care: the memory being
initialized must reside in kernel space, and extra runtime
initialization steps may need to be peformed before they are fully
usable as kernel objects. In particular, kernel subsystems or drivers
whose objects are already in kernel memory may still need to use these
macros if they define kernel objects as members of a larger data
structure.
It is intended that application developers instead use the
K_<object>_DEFINE macros, which will automatically put the object in the
right memory and add them to a section which can be iterated over at
boot to complete initiailization.
There was no K_WORK_DEFINE() macro for creating struct k_work objects,
this is now added.
k_poll_event and k_poll_signal are intended to be instatiated from
application memory and have not been changed.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This fixes coverity CID 171565 which may be valid in case of the
connection is not properly setup, or its memory is corrupted, it
may cause use of invalid addresses to be set using
net_if_set_link_addr.
JIRA: ZEP-2344
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
If the TCP data packet needs to be re-sent after the packet is lost,
then the acknowledgment number will be changed. This then means that
the TCP checksum needs to be recalculated too.
Signed-off-by: june li <junelizh@foxmail.com>
When a node receives consistent DIO messages with same data from
Border Router just ignore those messages. Need not to proceed
further.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Link metric is part IPv6 neighbour data struct. But RPL code is
trying to access it from RPL parent table where link metric doesn't
exist. So provided an api to get IPv6 neighbour data from RPL parent
data.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
DAG rank will be properly written with net_pkt_write_be16() in
network packet. API will take care of endianness. So need
not to convert it using htons().
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Sender rank (16 bit uint) was properly read with net_frag_read_be16()
api and need not to convert it again using ntohs().
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
We had various asserts when checking network packet length but
printed also error when there was none. Fix this by checking
do we really have a too short message.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Remove NET_TCP_HDR() macro as we cannot safely access TCP header
via it if the network packet header spans over multiple net_buf
fragments.
Fixed also the TCP unit tests so that they pass correctly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Remove NET_UDP_HDR() macro as we cannot safely access UDP header
via it if the network packet header spans over multiple net_buf
fragments.
Fixed also the UDP unit tests so that they pass correctly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Remove NET_ICMP_HDR() macro as we cannot safely access ICMP header
via it if the network packet header spans over multiple net_buf
fragments.
Jira: ZEP-2306
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The IPv6 HBH option PAD1 ext header was not parsed properly
as the code read one extra byte from the ext header. The
PAD1 length is only 1 byte.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add utility function that helps to figure out if the
protocol headers can be directly accessed when they fit one
net_buf fragment, or if they need to accessed using various
net_pkt helpers that know about reading data from two
different net_buf's.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of directly setting pointer to where to start to calculate
the various IP related checksums, use the net_frag_skip() to first
find out what is the fragment where the calculation should start.
This needs to be like this so that if the IP header + possible
extension are so long that they do not fit the first fragment,
we need to be prepared to look into second fragment.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The following error and warning is received when connecting
to certain central devices:
[bt] [ERR] hci_num_completed_packets: packets count mismatch
[bt] [WRN] bt_att_recv: Ignoring unexpected request
This could happen if Tx-ed packet is not added to pending
list before a num of completed packets event and/or new Rx
packet is received.
This is fixed by reducing the Rx thread priorities in the
Host and the Controller.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
net_context_unref() is protocol agnostic, after being un-referenced,
the same context might end up being used then for UDP,
if context->tcp is not reset to NULL, calling net_context_unref()
after this UDP usage will again try to release this TCP pointer
which might lead to random error.
Signed-off-by: june li <junelizh@foxmail.com>
The code in send_frag() depends on being able to allocate a new
pending tx contexts with 100% certainty. We must therefore notify
*all* connections instead of the current one in order to ensure that
there are free contexts available in free_tx.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In some cases the lladdr might not be set, currently this is
seen with RPL unit tests, in which case we must not access
the lladdr.
Jira: ZEP-2330
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is no route to the neighbor, then do not try to delete
it because the route pointer is NULL.
Jira: ZEP-2329
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Move a way from misc/ and put in its own subsystem to allow enhancements
in the future and make it a core part of Zephyr, not just something
misc.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The num_handles parameter of the Number Of Completed Packets event is
8-bits and not 16-bits, so no helper variable or byte order conversion
is needed.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In case both conn->tx_notify and conn->tx_queue have data in them we
should first process the notify queue and only then new outgoing
packets.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Due to the missing remainder support in the ticker_start
function for first interval, fix implementation so as to
round the first interval/offset during master connection
setup and connection update to within +/-16us.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The SYN-ACK timer is now handled by the TCP backlog functionality,
while the remaining ACKs for established connections use the tcp
struct ack timer. With this, code setting tcp struct SYN-ACK state
timers can now be removed.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Add timer for sent TCP SYN-ACKs. If the timer is already
scheduled to run before canceling it is attempted, set
the cancelled flag and let the callback remove the
delayed work.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Add an array of configurable size that holds TCP backlog entries.
The array is shared between all incoming TCP connections in order
to make it possible to get away with less memory consumed than
with a connection based approach.
The backlog entries are created when a SYN is received for a
listening TCP socket and removed once the corresponding ACK is
seen. With an incoming RST the corresponding backlog entry is
cleared, if any.
The size of the global backlog array is defined with the
CONFIG_NET_TCP_BACKLOG_SIZE Kconfig variable.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
