Unfortunately, Zephyr SDK 0.10.0 ships with outdate Newlib 2.0.0
(from 2015 or earlier) which lacks sys/_timeval.h header, requiring
ugly workaround of defining struct timeval inline (the whole idea
was to standardize on sys/_timeval.h header for different libc's).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
According to POSIX, that's the header which defines this function.
Similarly, nothing in POSIX indicates that <time.h> should have
access to struct timeval, so it's removed (it's made accessible
to <sys/time.h> via <sys/_timeval.h> introduced earlier).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
This is implementation-level header which defines struct timeval, and
intended to be included by headers which need this structure. This
implementation scheme is compatible with Newlib, and thus provides a
step to use minlibc vs Newlib interchangeably.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Move PROPERTY_LINKER_SCRIPT_DEFINES to toolchain_ld_base.
No functional change expected.
This is motivated by the wish to abstract Zephyr's usage of toolchains,
permitting non-intrusive porting to other (commercial) toolchains.
Signed-off-by: Mark Ruvald Pedersen <mped@oticon.com>
gpio-map is a property of "nexus node", defined in dts v0.3.
It allows to describe a pin connector so it can be referenced
through phandles and hence used in expansion device nodes like a
shield header (typically implemented through overlays).
This change implements gpio controller resolution through these maps.
Few assumptions were taken in order to simplify the implementation.
These assumptions bring some limitations to the use of gpio-map
but my understanding is that this should still allow to cover most
use cases.
Assumptions:
-gpio-size is the same for all gpio-controllers referenced in a map
-optional properties gpio-map-mask and gpio-map-pass-thru are
supposed to be omitted
The understanding of this last assumption is that flags provided in
the expansion device node will overwrite the connector flags.
In a latter stage, when need happen, these limitations can be
revisited to unlock fully fledged gpio-map usage.
Fixes#15637
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Some applications using DMA, such as UART RX, could
need to complete the current DMA transaction earlier than
predefined, based on other termination conditions,
like UART's IDLE interrupts.
In that case, the client needs to know how many data are
still left in DMA transfer buffer so that it can figure
out how many data has been transfered. However, the current
DMA API doesn't provide any information for the client
to learn the transfer buffer information.
And some other information, like whether DMA transfer is busy
or not, transfer direction, etc, could interest a client.
So, added a dma API function to retrieve the current DMA
runtime status.
And implemented the API for STM32F4's DMA while keeping
others unimplemented.
Signed-off-by: Jun Li <jun.r.li@intel.com>
reload function is not implemented by every DMA driver.
So, add api's NULL check to make sure it is protected if not
implemented.
Signed-off-by: Jun Li <jun.r.li@intel.com>
It is convenient to have a blocking version of
`mqtt_read_publish_payload` function, for cases when it is called from
the event handler. Therefore, extend the 'mqtt_read_publish_payload'
argument list with information whether the call should block or not.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The ARM Cortex-M Exception Stack Frame (ESF) may consist of
several stack frame contexts (basic state context, additional
state context, FP context, etc.). To reflect these structural
properties, this commit re-factors the ESF, splitting out the
basic stack frame, holding the state context, into its own
struct container. The commit does not introduce behavioral
changes.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
If the local node keeps getting bombarded with messages, it's possible
that the storage timer gets rescheduled over and over again and never
expires. Add the necessary code to only reschedule the timer if the
new deadline is earlier than an existing one.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The default values for the timeouts, as well as non-defaults in most
Mesh samples, use a higher value for the RPL than then generic mesh
storage timeout. This hasn't had any effect in practice since the code
only uses the RPL timeout if it is *smaller* than the generic one.
The original intention of the code was to use the RPL timeout,
regardless of what the generic one is, whenever the RPL is the only
thing that needs updating. Add some helper macros to track the various
groups of pending flags, and perform the appropriate checks to apply
the RPL timeout whenever it's smaller than the generic timeout, or if
there are no other items to store besides the RPL.
Fixes#15904
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When run as "west -v build", make sure that the underlying build tool
is run in verbose mode as well (if the generator is known to support
it, which is the case for Unix Makefiles and Ninja based generators).
The per-generator hacks here are needed to support CMake 3.13. If we
move to CMake 3.14 or later, we can just run "cmake --build BUILD -v"
and be done with it.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Fixed format error when compiling with gcc and newlib.
Used standard formats (%u) instead of inttypes formats (PRIxxx)
since Zephyr redefines the standard formats, that way it should always
be in line with the Zephyr types.
Compiled with and without newlib using gcc.
Signed-off-by: Dennis Wildmark <dennis.wildmark@assaabloy.com>
This adds interrupt support to the SAM0 GPIO driver. This is heavily
inspired by @nzmichaelh work in #5715. The primary difference
from that implementation is that here the External Interrupt
Controller (EIC) is separated out into an interrupt controller driver
that is less tightly coupled to the GPIO API. Instead it implements
more of a conversion from the EIC's own odd multiplexing to a more
traditional port and pin mask IRQ-like callback. Unfortunately,
through the EIC on the SAMD2x are relatively well behaved
in terms of pin to EIC line mappings, other chips that share the
peripheral interface are not. So the EIC driver implements a
per-line lookup to the pin and port pair using definitions extracted
from the ASF headers.
The EIC driver still makes some assumptions about how it will be used:
mostly it assumes exactly one callback per port. This should be fine
as the only intended user is the GPIO driver itself.
This has been tested with some simple programs and with
tests/drivers/gpio/gpio_basic_api on a SAMD21 breakout and an
adafruit_trinket_m0 board.
Signed-off-by: Derek Hageman <hageman@inthat.cloud>
This adds a SERCOM I2C driver for SAM0 series chips.
Tested with a SAMD21 chip on a SSD1306 display and a MLX90393
sensor. Only compile tested for SAMD20 and SAMR21.
Signed-off-by: Derek Hageman <hageman@inthat.cloud>
This adds generic support for any board using the SAM0 ADC driver.
The test selects the internal input from the scaled I/O voltage,
which is always available on the ADC.
Signed-off-by: Derek Hageman <hageman@inthat.cloud>
It is possible that the device driver API pointer is null.
For example if the device driver returns an error, the device
code will make the API pointer NULL so that the API would not
be used. This can cause errors in networking code where we
typically do not check the NULL value.
Fixes#15003
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
- renaming functions to better names
- reordering functions place (register, then unregister for instance)
- centralizing logs to relevant place
Fixes#8722
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Also, there is no need for unspecified address bit. If specified address
bit is not set, then it will be obvious address is unspecified. Reducing
the amount of bits from 6 to 4.
This permits to reduce net_conn structure of 4 bytes. Its size is as
before indroducing node attribute.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
This will optimize path when unused an connection is required or when
looking up a used one.
That said, at this stage, it bloats up the net_conn structure with 4
added bytes. More optimization will overcome this drawback.
Fixes#8722
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Most of present #ifdef can be removed via using IS_ENABLED() macro.
Only small part of cache related logic still require #ifdef.
Fixes#8722
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
There were various flaws in it that motivated its removal:
- No hash collision handling mechanism. In case that would happen, the
behavior of the network connection would be unknown. This is the main
drawback
- The lookup is not that much more efficient than the default one. The
only difference of gain is in connection comparison (a u32t comparison
vs a full connection compare). But the list handling is the same. It's
made worse by the presence of a negatives match array which can be
easily filled in and becomes then fully usless, appart from consuming
CPU. As well as adding a new connection: it requires the whole cache
to be cleared which is unefficient.
- Not memory efficient, even compared to a proper hash table.
Two arrays instead of one etc...
All of this could be fixed by using a proper hash table, though it
remains to be seen if such object could fit in Zephyr core.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
During net_pkt/net_context API changes, some ip handling blocks were
ordered ipv4 first, ipv6 second. While it is the contrary everywhere
else. So reordering to get things consistent.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
- Not all #ifdef can be removed: those which have a dedicated attribute
in struct net_context.
- For CONFIG_NET_CONTEXT_CHECK: switching the NET_ASSERT_INFO to
NET_DBG (simpler to read and anyway an error code is returned)
Fixes#8725
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
It will help to use IS_ENABLED in place of #ifdef in relevant place.
Only struct net_if uses this structure.
In case only IPv4 is used, it will bloat up this struct by 12 bytes.
There are few reasons why this is "ok" in this case:
- On limited rom/ram system it will be unlikely to find a lot of
network interfaces so it should not harm much to raise the size of
struct net_addr.
- If IPv4 is the only enabled IP version, it gains a good amount of
rom/ram to discard IPv6 support so it is fine to steal a bit of this
gain to bloat up a bit struct net_addr.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Thit will help removing usage of #if defined(CONFIG_NET_OFFLOAD) in
relevant places.
Note that static inlines are used instead of #define foo(...), to keep
the parameter check at build time.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Partially revert commit ea177e785c
("usb: dfu: set bwPollTimeout dynamically")
Introduced fix does not work proper because there is no way to be
sure that a control stage had success before start erase process.
Instead IMG_ERASE_PROGRESSIVELY configuration should be used
if the erase of the flash takes longer time.
resolves: #15497
Signed-off-by: Johann Fischer <j.fischer@phytec.de>
The Holyiot YJ-16019 board is a small, coin cell driven board based on
the Nordic Semiconductors nRF52832. It provides one LED and one push
button.
Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Zephyr implementation of OpenThreads utilsFlashErasePage platform
function did not disable flash protection before calling `flash_erase`
function. This resulted in an error instead of actual flash erase on
platforms that properly implement flash write protection.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
This reverts commit bd24b31139.
While the test case failure described in #14186 is associated with the
cycle-based busy-wait implementation, that test is fragile, and fails
less frequently once the incongruence between ticks-per-second and the
32 KiHz RTC clock are resolved. It also assumes that the system clock
is more stable than the infrastructure underlying the the busy-wait
implementation, which is not necessarily true.
The gross inaccuracies in the standard busy-wait on Nordic described in
issue #11626 justify restoring the custom solution.
As this applies to all Nordic devices, move the setting to the top-level
Kconfig.defconfig.
See: https://github.com/zephyrproject-rtos/zephyr/issues/11626#issuecomment-487243369
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
The default system clock on all Nordic devices is based on a 32 KiHz
(2^15 Hz) timer. Scheduling ticks requires that deadlines be specified
with a timer counter that aligns to a system clock. With the Zephyr
default 100 clocks-per-sec configuration this results in 100 ticks every
32700 ticks of the cycle timer. This reveals two problems:
* The uptime clock misrepresents elapsed time because it runs 0.208%
(68/32768) faster than the best available clock;
* Calculation of timer counter compare values often requires an integer
division and multiply operation to produce a value that's a multiple
of clock-ticks-per-second.
Integer division on the Cortex-M1 nRF51 is done in software with a
(value-dependent) algorithm with a non-constant runtime that can be
significant. This can produce missed Bluetooth deadlines as discussed
in upstream #14577 and others.
By changing the default divisor to one that evenly divides the 2^15
clock rate the time interrupts are disabled to manage timers is
significantly reduced, as is the error between uptime and real time. Do
this at the top level, moving SYS_CLOCK_HW_CYCLES_PER_SEC there as well
since the two parameters are related.
Note that the central_hr configuration described in upstream #13610 does
not distinguish latency due to timer management from other
irq_block/spinlock regions, and the maximum observed latency will still
exceed the nominal 10 us allowed maximum. However this does occur
much less frequently than changing the timer deadline which can happen
multiple times per tick.
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Inside function test_net_pkt_basics_of_rw result of function call
net_pkt_read_be16 is not checked which might result performance
of the program and cause errors.
Coverity-CID 198003
Fixes#15776
Signed-off-by: Maksim Masalski <maxxliferobot@gmail.com>
Making a clean slate for a pylint test in CI.
'_' is a common name for non-problematic unused variables in Python.
pylint knows not to flag it.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Szymon is no longer actively looking at Bluetooth code, whereas Joakim
from Nordic has been assigned for Bluetooth host support.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This is a follow up to commit 0eaa5e53a3.
`HAL_SW_SWITCH_RADIO_ENABLE_S2_PPI_BASE` is a base number for two PPI
channels, so two bits need to be marked in the used channel bit mask.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
This make use of NET_IF_NO_AUTO_START flag so Bluetooth interfaces are
not automatically enabled after initialized.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Fix calling bt_gatt_foreach_attr with start handle parameter set
to last static attribute handle.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
The approved trademark name is Wi-Fi so update references to WiFi and
other spellings to Wi-Fi in documentation and Kconfig help strings.
(Note that use of spelling variatios of "wifi" in module names, CONFIG
names, link names and such are untouched.)
https://www.wi-fi.org/
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
Within the past few days, an update to the Ubuntu 18.04 toolchain has
begun emitting code sections during link that are messing with our
stub generation. They are appearing in the 32 bit stub link despite
not being defined in the single object file, and (worse) being
included in the output segment (i.e. at the start of the bootloader
entry point!) despite not being specifically included by the linker
script. I don't understand this behavior at all, and it appears to be
directly contrary to the way the linker is documented.
Marc Herbert discovered this was down to gcc being called with
--enable-default-pie, so -no-pie works to suppress this behavior and
restore the default. And it's correct: we aren't actually generating
a position independent executable, even if we don't understand why the
linker script is being disregarded (to include sections we don't
include). See discussion in the linked github issue.
Fixes#15877
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
k_stack_alloc_init() was creating a buffer that was 4 times
too small to support the requested number of entries, since
each entry in a k_stack is a u32_t.
Fixes: #15911
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Paul Sokolovsky