Do not run in CI all tests, as they have very similar configuration
and provide little extra coverage to justify the extra CI time.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
This test is quite heavy and long in simulation,
but its length is meant to find unlikely issues which
may be triggered only very rarely.
Let's provide a kconfig value to chose how long the
test is, and set it to a lower value when running in simulation
(in CI) to save time.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Increase the UART speed to 1Mbps for the simulated nrf52
to reduce the amount of time spent busy waiting, and
therefore the real time duration of the test.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
filtering by CONFIG_SERIAL always matches, we need a different filter
and be able to enable the driver if supported.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Add the nrf52_bsim to the platform_allow list
for tests which had such a filter.
Note that for this test to pass, you need
to connect the uart in loopback, passing the command
line option `-uart0_loopback` to zephyr.exe
Note this tests do not run right now by default in CI
as the fixture is preventing it.
They can be run with:
twister -T tests/drivers/uart/ -v -p nrf52_bsim \
--fixture gpio_loopback -- -uart0_loopback
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
These tests were filtering by CONFIG_UART_CONSOLE,
but that does not seem a strick requirement.
It seems the requirement would just be CONFIG_SERIAL,
but that is implicit if CONFIG_SERIAL_SUPPORT_ASYNC,
which is required for all these tests, so let's
just remove CONFIG_UART_CONSOLE.
(The simulated nrf5x boards do not set CONFIG_UART_CONSOLE
by default, as by default the console is routed to the
process stdout instead)
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
The option CONFIG_NATIVE_UART_0_ON_STDINOUT
is only usable for the native_sim/posix(_64) targets,
but this test is only allowed for the nrf52840dk_nrf52840.
Building with this option set causes a kconfig warning.
Let's just remove it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
On the async configurations, the first part of the test
(test_uart_pm_in_idle) can interfere with the second part
of the test, if the device has a fast enough CPU.
This is due to the first part of the test ending
as soon as it queues the last byte for transmission.
If the device is fast enough (and the simulated
nrf52 is), the 2nd part of the test will start executing
enabling the UART Rx, which can result in either
the Rx being enabled mid frame, which can result in a frame
error or even before the Tx HW started pushing the byte
in the line (resulting in that byte being received).
Neither of these cases are handled by the 2nd test,
which sees a spurious error or received byte
and fails.
Let's just add a small delay at the end of the first test
to allow the Tx of the last byte to be done.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
For uart_async_api, uart_mix_fifo_poll & uart_pm:
These test can be run fine in the simulated nrf52_bsim board,
just connecting the UART Tx and Rx in loopback,
but we need an overlay just like for the real boards.
Let's provide it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
For some reason the host gcc (11 & 12) does not believe
rx_buf_size is constant. Let's work around it by using the
sizeof() expression it is initialized to instead.
This fixes a build error when targetting native targets
which use the host gcc:
tests/drivers/uart/uart_async_api/src/test_uart_async.c:236:34:
error: expression in static assertion is not constant
236 | BUILD_ASSERT(rx_buf_size <=
sizeof(tdata.rx_first_buffer), "Invalid buf size");
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
The test was actually filtered out in the integration platforms
(native_sim) as SERIAL was not enabled in its prj.conf
and that is not enabled by default for this target.
Let's just enable it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Commit eb44414af9 modified
test_single_read for early rx buf release.
The commit assumed that tdata.last_rx_buf points to either &tdata.rx_buf[0]
if the driver releases the first buffer or at &rx_buf[5] if the first
buffer is not released. However, where tdata.last_rx_buf points to depends
on the timeout given to uart_rx_enable(), making the test flaky.
This commit modifies the test by keeping track how many bytes have been
received in the first and second buffers.
The function tdata_check_recv_buffer() validates that the sent data
matches the received bytes which may have been split between first and
second buffer.
This fixes the uart_async_api test on the xmc45_relax_kit.
Signed-off-by: Andriy Gelman <andriy.gelman@gmail.com>
On the STM32H735G-DK Discovery kit, when connecting a UART from
pin PF7 to PF6, the test uart_async_api passes now.
Signed-off-by: Benjamin Deuter <benjamin.deuter@gmail.com>
Add synchronization using semaphores for the UART ISR tx/rx
tests, instead of relying on a single k_yield() call to
ensure all UART ISR work can be completed before validating
the result.
Signed-off-by: Bjarki Arge Andreasen <bjarkix123@gmail.com>
Most tests define a buffer for UART data on the stack,
including the tests testing UART IRQ. Move buffers to
static memory, within the test fixture, which can then
be passed as the user_data to the UART IRQ callback.
Signed-off-by: Bjarki Arge Andreasen <bjarkix123@gmail.com>
rand32.h does not make much sense, since the random subsystem
provides more APIs than just getting a random 32 bits value.
Rename it to random.h and get consistently with other
subsystems.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Run this test on lpuart10 with internal loopback enabled.
Use DTCM region for this test because in some local
variables are used for dma buffer which requires
non-cacheable
Signed-off-by: Dat Nguyen Duy <dat.nguyenduy@nxp.com>
The test drivers.uart.basic_api.wide
can't work without keyboard
add harness: keyboard for this test
Signed-off-by: Marc Desvaux <marc.desvaux-ext@st.com>
change pins for the test uart_async_api loopback because the old pins
pb10/pb11 are now used for the I2C loopback test.
Signed-off-by: Marc Desvaux <marc.desvaux-ext@st.com>
This includes new overlays for boards supporting DMA, and one more test
to verify M0 boards both with and without DMA can be built when
requesting CONFIG_UART_ASYNC_API.
Signed-off-by: Diego Elio Pettenò <flameeyes@meta.com>
Signed-off-by: Gerson Fernando Budke <nandojve@gmail.com>
There are situations in which the async API for uart can be provided
by another driver (case in point, uart_rtt), and thus there is no valid
DMA controller for the uart_sam0 driver to talk with.
By separating the configuration, there's no need to exclude samd20-based
boards (that have no DMA peripheral) from the uart_async_api tests.
Signed-off-by: Diego Elio Pettenò <flameeyes@meta.com>
Co-authored-by: Gerson Fernando Budke <nandojve@gmail.com>
It is the arduino_i2c node on the board.
Also remove comment from arduino_r3_connector as the arduino_i2c
node can be used.
Signed-off-by: Andriy Gelman <andriy.gelman@gmail.com>
This board is a relatively inexpensive development kit for USB-PD
controllers, using an UPD301C controller.
This Zephyr config includes support for:
* UART (present on he debug header of the board), tested with the
hello_world sample application;
* the one standalone LED (CAP_MIS), tested with the blinky sample
application;
* the rotary encoder (PDO_SEL), via ADC, tested with the adc sample
application;
* the current sense amplifier (I_SENSE), currently untested;
* the SPI bus, connected internally in the UPD301C to the UPD350;
* the I2C bus, exposed on the debug header, currently untested.
Note that the drivers.uart.async_api.rtt has to be disabled, as it is
for other m0 boards with no dma or it fails to build.
Signed-off-by: Diego Elio Pettenò <flameeyes@meta.com>
Refactor the ESP32 target SOCs together with
all related boards. Most braking changes includes:
- changing the CONFIG_SOC_ESP32* to refer to
the actual soc line (esp32,esp32s2,esp32s3,esp32c3)
- replacing CONFIG_SOC with the CONFIG_SOC_SERIES
- creating CONFIG_SOC_FAMILY_ESP32 to embrace all
the ESP32 across all used architectures
- introducing CONFIG_SOC_PART_NUMBER_* to
provide a SOC model config
- introducing the 'common' folder to hide all
commonly used configs and files.
- updating west.yml to reflect previous changes in hal
Signed-off-by: Marek Matej <marek.matej@espressif.com>
Modify the test to not rely on assumption that RX buffer is released
when it is fully released. RX buffer maybe released after any RX_RDY
event within the buffer, e.g. after each timeout driver may be
switching to the new buffer.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
