In the case of an unexpected PDU we need to send the right response.
This was already taken care of for PB-ADV, but not for PB-GATT.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The Provisioner might have missed our earlier Link Acknowledgement, so
if we receive another one with matching Link ID and link.expect state,
simply send another acknowledgement.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When sending a segmented message, the state could get stuck if the
advertising bearer fails in transmitting and we don't detect that it
happened. Add a send_start callback for all packets so we can always
know if sending fails.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When both PB-ADV and PB-GATT are enabled, the PB-ADV code (prov.c)
uses the bt_mesh_proxy_get_buf() API to get a net_buf_simple buffer.
Unfortunately this function also suffers from the same issue that was
fixed by commit 2b273444c1.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In net_app_init_client(), the remote port is set after an initial
set of checks based on the remote_addr's assigned family.
However, if the peer_addr_str is a host name which needs to be
resolved via DNS, the family won't be set yet, and the port is
left as 0.
To fix this behavior let's move the port assignment after the
DNS lookup section to be sure that the remote sa_family is set.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
Difference being that the data is not, then, allocated from the pool.
Only the net_buf is.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
The conversion to the new net_buf_simple API was done incorrectly
here. The buffer initialization should use net_buf_simple_init()
instead of net_buf_simple_reset(), so that buf->__buf gets properly
initialized (and not left pointing at NULL).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There's a risk of a deadlock if we use the same pool for ACL fragments
as we use for general ACL TX buffers: all TX buffers are queued up,
and we try to segment one of them, a segment buffer will never become
available. To work around this risk, introduce a dedicated fragment
pool.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Refactor ticker execution context dependency out into HAL
folder. This decouples ticker from mayfly, enabling porting
towards a more tasklet (if and when kernel gets the support)
style execution contexts type implementation support.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Convert the mesh code to use the new net_buf_siple APIs. This has the
benefit of saving 4 bytes off the stack due to the not needed pointer.
Also update the publication context helpers to map to the new
net_buf_simple API in an intuitive way.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If context is bound to IPv6 unspecified addresss and some port
number, then unspecified address is passed in TCP reset packet
message preparation. Eventually packet dropped at the peer.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Zephyr doesn't have luxury to create re-transmit timer per packet. So
it has different methods to handle packets in queue. But re-sending
packets on valid ack messages causing issues.
E.g. A TCP node sent two packets (packet-1, packet-2). Peer replied
two ACKs (ACK-1 and ACK-2), and these two ACK's are at rx_thread
queue. Now ACK-1 is handled and reference of packet-1 is freed
from sent list. Then if condiftion (valid ACK and connection
state is ESTABLISHED) notices that, sent list is not empty.
Restart the timer, modify sent flag and resend packets in a
list. Here packet-2 is sent again, even though ACK-2 is already
received. Situation is worse if there are more packets in the
list.
So only start the re-transmit timer in-case queue is not empty. It
allows rx_thread to handle all incoming packets (in this e.g ACKs).
When the re-trasmit timer expires, it sends the packets which
are left in queue.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
In case of failed to get source address from the net packet,
release the assgined tcp backlog entry. Otherwise it will
never be freed.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
k_delayed_work_cancel(&context->tcp->fin_timer) called twice
immediately one after other.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Application should normally declare a bt_uuid with proper type and then
use bt_uuid_cmp.
Fixes#5162
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
RFC 7252, sec 4.8 [https://tools.ietf.org/html/rfc7252#section-4.8]
The default MAX_RETRANSMIT is 4, current implementation only
retry 3 times. Update next_timeout() to return one more retry.
Also, add a TODO for random generated initial ACK timeout
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
The existing implementation assumed DNS resolv callback will be
called just once, but that's not always the case (apparently,
for multi-homes hosts or something). So, apply array bounds
checking (and do pointer arithmetic only after it, as the C
standard otherwise warns of "undefined behavior"). In such a case,
the port number wasn't set in each entry too, so rework how it's
done.
The issues discovered while resolving archive.ubuntu.com.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
This feature is X86 only and is not used or being tested. It is legacy
feature and no one can prove it actually works. Remove it until we have
proper documentation and samples and multi architecture support.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
New advertising started while scanning is already enabled
would delay the first advertisement event until the end of
the current overlapping scan window in the Zephyr native BLE
controller implementation. Hence, consider this scan window
duration when calculating the advertising stop.
Relates to: #6083
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix, add the missing code for the removal of any accumulated
soft latencies or negative drift ticks when scheduling next
interval expiry with added laziness.
Typically a first interval would accumulate soft latencies
and this has to be removed if the interval is rescheduled
with any added laziness (scheduled to the next soft real
time interval).
Example, scan windows block any new scheduling until the end
of the window, adding latencies to any soft real time ticker
expiry which should try to execute as early as possible after
the scan window.
Fixes: #6083
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Refactor BT_CTLR_ADV_INDICATION feature by moving the
implementation closer to the implementation that is closing
the advertisement event so that in the future new members
can be added as necessary (example, advertised channels).
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
No one was setting this to any other value than its default, which
happens to be the same as BT_BUF_USER_DATA_MIN.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Since all in-tree users are good with 4 bytes of user data, reduce the
default to 4. Also fix the default and rage values where apparently
'0' and '8' had been swapped.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The HCI command buffers are the only ones with more than 4 bytes of
required user data. Move the user data into a separate array and do
the mapping with the help of net_buf_id(). After this, it will be
possible to reduce the default net_buf user data size from 8 to 4.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fix incorrect radio timing for Tx chain delay for S2 and S8
coding. Regression introduced in commit 55d3ce111c
("Bluetooth: controller: Refactoring nRF5 radio driver").
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Having these in the beginning of a switch statement without any case
statement makes no sense.
Fixes#6135
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Introduce two new "standard" data allocators to net_buf. There are now
three in total:
NET_BUF_POOL_FIXED_DEFINE: This is the closes to the old
implementation, i.e. fixed size chunks. It's also what the old
NET_BUF_POOL_DEFINE macro maps to.
NET_BUF_POOL_HEAP_DEFINE: uses the OS heap
NET_BUF_POOL_VAR_DEFINE: defines a variable sized allocator using
k_mem_pool (this is all that there was in my first draft of this
feature)
Currently the variable length allocators (HEAP & VAR) support
reference counted data payloads, i.e. cheap cloning. The FIXED
allocator does not currentlty support this to allow for the simplest
possible implementation, but the support can be added later if
desired.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Redesign of the net_buf_simple and net_buf structs, where the data
payload portion is split to a separately allocated chunk of memory. In
practice this means that buf->__buf becomes a pointer from having just
been a marker (empty array) for where the payload begins right after
the meta-data.
Fixes#3283
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fix conditional compilation in nRF5 radio interface so that
when Coded PHY feature is not selected then it does not use
the PPIs required for supporting the feature.
This will allow pwm_nrf5_sw driver to use PPI channels 14 to
19 and support 3 PWM channels. Without the Coded PHY feature
disabled, only PPI 14, 15, 18 and 19 are available for PWM.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Added a new API to get net pkt based on dst address. Destination
address will be used to find correct net_context.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
When net_app_ctx has multiple net_contexts, selecting net_context
based on dst address has some glitches.
E.g. One net_app_ctx and two net_contexts (net_ctx1, net_ctx2).
Both net_contexts are mapped to net_app_ctx.
When a caller looking for net_ctx (e.g. net_ctx2) with
matching dst address. Loop goes through net_ctx1 and dst
doesn't match. But another if condition checks does this
net_context mapped to net_app_ctx, yes it matches.
So return net_ctx1, which is wrong.
So first go through all elements in array of net_contexts for
matching dst address, if it fails to find then go for matching
net_app_ctx.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Default network interface is the first interface if there are multiple
interfaces exist. But this creates an issue when different parts of
the network subsystem wants to choose particular default interface.
At-least with this Kconfig option user can select default network
interface.
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Prior to this commit, the old `FLASH_DRIVER_NAME` macro was being used
in the definition of the flash drivers array. This caused the array to
have a size of 0.
This commit changes the code to use the newer `FLASH_DEV_NAME` macro,
causing the configured flash device to be present in the table.
Signed-off-by: Christopher Collins <ccollins@apache.org>
The code in the LwM2M "Device" object was checking for this max
power source config value, except that it was never added as a
Kconfig option. Let's go ahead and add it so apps can set the
value appropriately.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
When attempting to read an unused multi-value resource such as
"Available Power Sources", lwm2m client return a successful coap packet
w/o any values. This looks like a timeout to any lwm2m servers which
make sure a valid response is returned.
Let's fix this by returning the correct NOT_FOUND error code instead.
NOTE: To test this I commented out the portion of the lwm2m client
sample which initializes the LWM2M_DEVICE_PWR_SRC_TYPE_BAT_INT and
LWM2M_DEVICE_PWR_SRC_TYPE_USB values. By default, the sample will
setup dummy values to be returned by the client.
Signed-off-by: Michael Scott <michael@opensourcefoundries.com>
As per review of PR #5893, this is a follow up patch to update
select_writer() API to match the behavior of select_reader() API.
Quote from OMA-TS-LightweightM2M-V1_0_1-20170704-A. 8.2.5
"An Object Instance or Resource is Read by sending a CoAP GET to the
corresponding path. The response includes the value in the
corresponding Plain Text, Opaque, TLV or JSON format according to
the specified Content-Format (see section 6.4).The request MAY
specify an Accept option containing the preferred Content-Format
to receive. When the specified Content-Format is not supported by
the LwM2M Client, the request MUST be rejected."
Therefore, we do not attempt to assign a content-format when the
requested one is not supported. Instead, we return an error code.
0 is returned when reader or writer has been successfully selected by
select_reader() or select_writer()
Signed-off-by: Robert Chou <robert.ch.chou@acer.com>
It is faster to operate directly on flash_area pointer instead
of fetch it all the time using the fcb flash area id.
Also as f_area_id was needed only for get appropriate flash_area
pointer, so it is better to pass it only while initialization
the fcb and not store it in fcb instance data at all.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Previously flash_area structure was the parameter for the callback
function for transferring sector location info.
Now flash_sector is used so relevant flash_area is missing for
the callback.
This patch introduces structure fcb_entry_ctx which incorporates
entries' location and relevant flash area. It is used to pass complete
information to the callback. Additional pointer to the flash_area fcb
speeds up operations.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Prior to this commit: `struct fcb` contained a pointer to an array of
`struct flash_area`. Each entry in the array represented a flash sector
that the fcb comprised.
After commit: `struct fcb` contains a pointer to an array of `struct
flash_sector`, not `struct flash_area`.
Rationale: The FCB needs to know which sectors it occupies for its
rotate operation. However, the `flash_area` type is meant to represent
a collection of sectors, not a single sector. `struct flash_area`
contains information that the fcb does not need (e.g., area ID and
device ID). Furthermore, the flash_map API provides a means of
converting a flash area ID to an array of sectors
(`flash_area_get_sectors()`), but no way to convert to an array of
areas.
Signed-off-by: Christopher Collins <ccollins@apache.org>
Fixed a missing ticks unit to microsecond unit conversion
potentially caused incorrect window offsets being used while
establishing connections.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
ticker timespace reservation can range up to 10.24 seconds,
needing 19-bits to represent in 32KHz clock units. Hence,
fix controller implementation to use u32_t to store ticks
slot values.
Without this fix, the controller is asserting in scan_adv
sample when using continuous scanning with 2 second interval
and window.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Johan Hedberg