Since callers of bt_recv() have so far anyway been required to know in
which context to call it (based on e.g. bt_hci_evt_is_prio) it's
cleaner to have two separate APIs: bt_recv and bt_recv_prio.
Change-Id: Icd0d9aed9c51ffd2def31432c4ffcc16a9f13ccd
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Use interrupt based transmission instead of polling.
Change-Id: Iebfd67372044d3e6de9bdbdd6f0c9c6d01d46cb9
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
So far the use of k_sem meant that there was no major benefit of
having more than 2 or 3 RX buffers since there was no queuing
mechanism. Instead of using k_sem, introduce a k_fifo and use that to
queue up incoming buffers. This way the RX buffer count can be
increased with measurable effects on throughput.
Change-Id: I8122b233aeee7c8e145de3fff5f10bcfe348efaa
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The H:5 HCI driver doesn't need the type-specific bt_buf allocation
helpers.
Change-Id: Iccb9803f32f3143f1cdf972d9229124816a99aa6
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Completely redesign the H:4 driver to utilize its own thread and
select the new RECV_IS_RX_THREAD Kconfig option.
Jira: ZEP-1483
Change-Id: I0ca0661b66d564e5edc1b8505706d6bb76632e79
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This option has not been used so far and will only contribute
unnecessary complexity in subsequent patches that introduce a concept
of combined RX buffers.
Change-Id: I53e0ce5155eebc352b84ba41b30ecb9d9958699f
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This helps simplify code that was previously combining net_buf_add()
with memcpy().
Change-Id: If44cf9cd651aba5e544e36567869f147468663eb
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Not all users are in an ISR context where we can't block, so give the
callers the freedom to choose if they want to block or not.
Jira: ZEP-1481
Change-Id: I19bd7e2df94c4eeb60886a17a78f872bd7bea887
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
There are static initializer macros available for most kernel objects
which we should use whenever possible.
Change-Id: I496f4d05d26801eddd21fae53bdd4fcdc3246fe3
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In order to keep the initialization process light-weight, remove
net_buf_pool_init() and instead perform the initialization of the pool
and buffers in a "lazy" manner. This means storing more information
in the pool, and removing any 'const' members from net_buf. Since
there are no more const members in net_buf the buffer array can be
declared with __noinit, which further reduces initialization overhead.
Change-Id: Ia126af101c2727c130651b697dcba99d159a1c76
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Until now it has been necessary to separately define a k_fifo and
an array of buffers when creating net_buf pools. This has been a bit
of an inconvenience as well as blurred the line of what exactly
constitutes the "pool".
This patch removes the NET_BUF_POOL() macro and replaces it with a
NET_BUF_POOL_DEFINE() macro that internally expands into the buffer
array and new net_buf_pool struct with a given name:
NET_BUF_POOL_DEFINE(pool_name, ...);
Having a dedicated context struct for the pool has the added benefit
that we can start moving there net_buf members that have the same
value for all buffers from the same pool. The first such member that
gets moved is the destroy callback, thus shrinking net_buf by four
bytes. Another potential candidate is the user_data_size, however
right not that's left out since it would just leave 2 bytes of padding
in net_buf (i.e. not influence its size). Another common value is
buf->size, however that one is also used by net_buf_simple and can
therefore not be moved.
This patch also splits getting buffers from a FIFO and allocating a
new buffer from a pool into two separate APIs: net_buf_get and
net_buf_alloc, thus simplifying the APIs and their usage. There is no
separate 'reserve_head' parameter anymore when allocating, rather the
user is expected to call net_buf_reserve() afterwards if something
else than 0 headroom is desired.
Change-Id: Id91b1e5c2be2deb1274dde47f5edebfe29af383a
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
We can locally reference the files we need, so don't add a -I we
don't need.
Change-Id: I764aea4177a8995489e0f15f71f7373427b43394
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
There's now snprintk available that's more light-weight on the stack
than snprintf.
Change-Id: I6b3e4409703ca92fe6b8f4146ff47c490ab826cb
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fix compilation issues that show up if SYS_LOG is mapped to printk
instead of printf. Unlike printf, printk is annotated so that the
compiler catches incorrect format specifiers passed to it.
Change-Id: I4d6f635a0ed61de698727028ea8767dc0ef28bb1
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Using the K_* macros makes it easier to read what exactly the various
timeouts are.
Change-Id: Ia405d3760b8e600af7e33a7221ef6ec717708973
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Switch left-over usage of TICKS_NONE and TICKS_UNLIMITED to the new
unified kernel counterparts K_NO_WAIT and K_FOREVER.
Change-Id: I2f2a16360e816f9f8791eb216deb3c70b8cc87df
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The NBLE driver is now deprecated and will be removed in a future
Zephyr version.
Change-Id: I50753f8cf8566ac1e4e73df480c96253c199e6a7
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Start using the k_work and k_delayed_work APIs.
Change-Id: Iac0525a444c5c4e0f28db08844d7b28e17e905fc
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Switch to using k_work and k_delayed_work APIs.
Change-Id: I3be18132cc417607adaec0ad711256b94e314c38
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Don't use the deprecated NANOKERNEL init level. The only requirement
for Bluetooth drivers is for them to be registered before the
application main() runs, so POST_KERNEL should be good enough.
Change-Id: I02a8609bf63e9d608b802576214a2e76211b3965
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Use the unified kernel API k_sleep() instead of the deprecated
fiber_sleep().
Change-Id: I587d72ca5b53aacc02647b32c3ebceb1d7fe067e
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Switch to using the unified kernel k_thread_spawn() API instead of
nano_fiber_start().
Change-Id: I325cf467ae2a52c6aec8fc166397c323929e3013
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Use the unified kernel API k_yield() instead of fiber_yield().
Change-Id: I8f52031f52f7ac8783033a51751dc22decdfa59a
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Use defines from k_fifo intead of legacy API.
Change-Id: Ib8cf0d88240ef145da550b8cf83d2580e7140521
Signed-off-by: Szymon Janc <ext.szymon.janc@tieto.com>
All boolean options default to 'n' without the need to explicitly
state this. It's only the cases where we want 'default y' where we
need to state this.
Change-Id: I47dbda62462ea437a2423b8508ea2cc640a22e41
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Restructure the Bluetooth options more logically.
- Both host and controller are now behind the same high level
CONFIG_BLUETOOTH.
- Selecting controller support disables other HCI driver selection, so
the controller isn't in the same list as HCI drivers any more.
- Under the top-level there's a "Custom stack" option, which when
enabled opens up the option of choosing CONFIG_NBLE.
There are various other cleanups and simplifications in this patch as
well, since splitting these up would have been fairly tricky while
making sure all test cases still build.
Change-Id: I5bb715cb9d20201cb8b72fbd149c8a09a4b2d7d2
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This suppresses warnings for platforms where SERIAL can't be provided.
All the platforms that these drivers are interesting for already
default to SERIAL=y.
Change-Id: Id692f99e018009b30903db8a2c046a6086be01c2
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Making a reference to the common work queue code should not necessarily
drag in the system workqueue, since it is possible to use a workqueue
that is not the system workqueue. This is done by moving the system
workqueue into its own code module.
Moving the system workqueue to its own code module allows removing the
NANO_WORKQUEUE and SYSTEM_WORKQUEUE kconfig options, and compiling the
common workqueue code and system workqueue all the time. They are only
linked in the final image if a reference to them exist, same as the
other kernel modules.
Change-Id: I6f48d2542bda24f4702e7c2e317818dd082b3c11
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Move controller code from drivers/controller to
subsys/bluetooth/controller.
Change-Id: I73f675188485aa3267507bad7647796e593a3da0
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Move the Bluetooth host stack from net/bluetooth to
subsys/bluetooth/host. This is preparation for having both host and
controller under the same root, i.e. subsys/bluetooth/.
Change-Id: I3bc796f7e331fca0c485f3890d62b9c03e027b96
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
In 3 of our UART driver implementations the ISR is exclusively
used and enabled for the RX path. The existing logic was
susceptible to a stall situation where a polled out transmission
would be interrupted by a reception and then the ISR code
would loop forever due to the TX event being signalled (although
the interrupt itself was disabled) causing the ISR to keep
looping for an RX interrupt.
Change-Id: Ic379e58b1c974aca3cee37d2d81f12c3726fb160
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
These options were only needed for a MyNewt-based nRF51 firmware on
these boards (the MyNewt BLE stack is called Nimble, hence the
prj_nimble.conf sample config files). With a Zephyr-based nRF51
firmware these options are no-longer needed, so it's not appropriate
to have them default to enabled. Instead, if they are needed, require
the app-specific configuration to enable them.
Change-Id: Iefbee4d97590af4e11bcedea05fe61f32a147b83
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fix issues with tab vs space as well as missing license/copyright
header in hci/Kconfig.
Change-Id: Id766308dc8eda2d598ea749e5e0d2166014db929
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The bt_driver API was created when Zephyr only had a Bluetooth host
stack, but no controller-side functionality. The only "driver" that
was needed for the host was the HCI driver, and hence "HCI" was
omitted from the name.
With support both for host and controller Zephyr will be getting more
Bluetooth driver types, in particular radio drivers. To prepare for
this, move all HCI drivers to drivers/bluetooth/hci/ and rename the
bt_driver API bt_hci_driver.
Change-Id: I82829da80aa61f26c2bb2005380f1e88d069ac7d
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Since RSSI events trigger interrupts and signal the RX path
semaphore with a resulting event lenght of 0. Due to this
fact the Controller HCI driver was leaking (i.e. not freeing)
those events, grinding the recv_fiber to a halt.
Issue identified by Szymon Janc.
Change-Id: I3e259b2823717b523ac331f8f787252414fb9290
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Report the HCI version supported by the Controller as 4.2,
since it was set as 0 (1.0b) and this confused certain Host
implementations, such as BlueZ.
Change-Id: I809721ee9c2b55e77e6a3ca63688c802a9ffa0ba
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Integrating the IRQ definitions for both nRF51 and nRF52, and defining
the set of IRQ values that are common to both.
Also switch the controller, uart and clock drivers to use the common
header definitions.
Change-Id: Id6816d7a97720896cfe4df83656fb8b1f8fb01fa
Signed-off-by: Ricardo Salveti <ricardo.salveti@linaro.org>
As part of an effort to closely integrate with Zephyr OS,
removed the custom implementation of clock control
interfaces used in controller code and use the driver model
in Zephyr OS.
Jira: ZEP-897
Change-id: I03efbff471e42b9bd9832818dd20dd5998a60227
Signed-off-by: Vinayak Chettimada <vinayak.kariappa.chettimada@nordicsemi.no>
This way the application can reuse the same callback for multiple CCC
since it can track what CCC is affect by checking the attribute pointer.
Change-Id: I608da643aea07de26b65d67e6db3268d717d0f53
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
nRF51x SoC is based on ARM Cortex-M0 and running at 16MHz clock.
Due to Radio ISR latency restriction, the CPU utilization within
BLE tIFS of 150 us should not be more than the time radio h/w
takes to get ready. In order to keep short CPU utilization
inside tIFS encryption procedure has been spread over several
connection intervals. In comparison to possible encryption setup
in 3 connection intervals, this commit introduces a encryption
setup being done in 5 connection intervals so that max. CPU time
is well within the permissible limits by nRF51 h/w and tIFS.
Change-Id: I8971d69675836b9e18bea59d95951e0dfe8a7d2d
Signed-off-by: Vinayak Chettimada <vinayak.kariappa.chettimada@nordicsemi.no>
The bt_storage_clear() function doesn't modify the data behind the
addr parameter, so the pointer should be decared const.
Change-Id: Icce676f9df80bac26ba4877bb63752559a43339f
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
As part of an effort to closely integrate with Zephyr OS,
removed the custom implementation of IRQ interfaces used
in controller code.
Jira: ZEP-841
Change-id: Ie427f45aeecad51053112371526cb7dc4817248f
Signed-off-by: Vinayak Chettimada <vinayak.kariappa.chettimada@nordicsemi.no>
Code under SCHED_ADVANCED macro in ctrl.c is used to
disable advanced radio time space scheduling calculations
used for placing non-overlapping observer and master roles;
to populate connection parameter request/response PDU
fields, and autonomous generation of slave initiated
connection parameter requests.
This macro shall always be enabled for optimal and correct
operation of the controller, It shall only be disabled to
debug failures in other part of the controller, to rule out
this code's influence, if any, on the failure being
debugged.
Change-id: I04e5f837ef0a5658361bd3668f583f1e13504570
Signed-off-by: Vinayak Chettimada <vinayak.kariappa.chettimada@nordicsemi.no>
Place __packed at the end of the structure definition in line
with the rest of the Zephyr codebase.
Jira: ZEP-732
Change-Id: I25aa731cbd188a6e23ca2035eb22fa919295bb25
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Johan Hedberg