Implement the remaining functions from the POSIX_FD_MGMT Option
Group that are part of POSIX, and add the
CONFIG_REQUIRES_FULL_LIBC dependency to CONFIG_POSIX_FD_MGMT, to
pull in the remaining C89 functions.
The POSIX_FD_MGMT Option Group is required for PSE52, PSE53, and
PSE54 Subprofiling Option Groups.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Mark the POSIX_SIGNALS Option Group as supported to the extent
possible under Zephyr, as Zephyr does not yet support processes.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Since Zephyr itself does not currently support processes, but
conformant applications should still be able to link, add stubs
for the remaining POSIX functions in the POSIX_SIGNALS Option
Group.
The POSIX_SIGNALS Option Group is required for PSE51, PSE52,
PSE53, PSE54, and likely many other POSIX Subprofiles.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Newlib requires an alias for the getpid() function. There was
a Kconfig already present for doing so, but the actual alias
was missing.
So this is technically a bugfix.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
After RC1 (feature freeze), new tests for existing features are allowed.
Afterwards, only bug fixes and documentation to existing tests are
allowed.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This update includes bugfixes related to esp_timer,
responsible for handling Wi-Fi thread events.
This also removes all constructor attributes from non-used
hal functions.
Fixes#74368Fixes#74277Fixes#75063
Signed-off-by: Sylvio Alves <sylvio.alves@espressif.com>
The macro was changed to return a const pointer, but the
BTP use the returned pointer in a non-const way.
Modify the usage so that the macro is used to initialize
the adv_param instead, and then provide a non-const pointer
to tester_gap_create_adv_instance.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Check for existence of APPLICATION_CONFIG_DIR in sysbuild.
sysbuild.conf is an optional file, meaning that sysbuild will not fail
if APPLICATION_CONFIG_DIR points to a non-existing folder.
A normal Zephyr build already fails when passing an invalid
APPLICATION_CONFIG_DIR but only because it cannot find a required
prj.conf in the folder.
Checking the existence of APPLICATION_CONFIG_DIR in sysbuild and fail
the build when a non-existing folder is passed to sysbuild ensures
sysbuild behaves similar to a normal Zephyr build regarding a
non-existing folder, while still allow sysbuild.conf itself to be
optional.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Some controllers, like the SoftDevice Controller can miss the first few
advertisements after the host tells it to start scanning. The DUT would
assumes the first RPA advertisements report it gets is the first that
was sent, but it was actually not the first. This would skew the
tester's judgement about the timing of RPA rotations.
To remedy this, the tester will scan continuously and switch between
expecting the identity address and the RPA without stopping the scanner.
The tester will tell the DUT to switch to RPA after it has received the
identity address advertisement. This ensures that the tester will not
miss the first RPA advertisement.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
Several help prompts for non-user-selectable Kconfig options
included the phrase "select 'y' here", which does not make
any sense in this situation.
Adjust the help sections to use more appropriate language.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Mark the POSIX_DEVICE_IO Option Group as commplete. This Option
Group is required for PSE51, PSE52, PSE53, PSE54, and most of
the interesting applications involving file descriptors and
sockets.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Move the implementation of zsock_select() to zvfs_select(). This
allows other types of file descriptors to also make use of
select() functionality even when the network subsystem is not
enabled.
Additionally, it partially removes a dependency cycle between
posix and networking by moving functionality into a mutual
dependency.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Move the implementation of zsock_poll to zvfs_poll. This allows
other types of file descriptors to also make use of poll()
functionality even when the network subsystem is not enabled.
Additionally, it partially removes a dependency cycle between
posix and networking by moving functionality into a mutual
dependency.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Add pread() and pwrite() implementations, which are nearly
identical to read() and write() but differ in that they do not
update the file-descriptor offset and instead read from a
specific file offset.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Create a compatibility Kconfig option for libc's that provide
their own stdin, stdout, stderr variables.
These are POSIX variables, so eventually we may want to
consider defaulting this option to 'y' and maybe providing
some zvfs abstraction for them.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
The POSIX_DEVICE_IO Option Group requires a number of c89
functions mainly from stdio.h .
Namely,
clearerr(), fclose(), feof(), ferror(), fflush(), fetc(),
fgets(), fprintf(), fputc(), fputs(), fread(), freopen(),
fscanf(), fwrite(), getc(), getchar(), gets(), perror(),
printf(), putc(), putchar(), puts(), scanf(), setbuf(),
setvbuf(), ungetc(), vfprintf(), vfscanf(), vfprintf(), and
vscanf().
These should be provided by any conformant C library
(not by the POSIX API).
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Only invoke vtable methods read, write, and close if they are
non-NULL.
The close() vtable method is optional, so that should not return
an error if zvfs_close() is called and that method is
unimplemented.
Otherwise, if zvfs_read() or zvfs_write() are called and the
corresponding vtable method is unimplemented, fail setting
errno to EIO.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
fixes an incorrect dma configuration. When lpc dma driver was extended
with gather/scatter support the spi dma driver stopped working.
Signed-off-by: Johan Carlsson <johan.carlsson@teenage.engineering>
Remove address-of operator ('&') when assigning `clock_control_xxx_init`
function pointer in `DEVICE_DT_INST_DEFINE` macro.
This change aims to maintain consistency among the drivers in
`drivers/clock_control`, ensuring that all function pointer assignments
follow the same pattern.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
This nucleo board inherit from the root `h5` dts some special nodes:
- `vbat` on adc1 channel 2
- `vref` on adc1 channel 17
- `die_temp` on adc1 channel 16
The issue is discussed here:
https://github.com/zephyrproject-rtos/zephyr/issues/72914
Signed-off-by: Alex Fabre <alex.fabre@rtone.fr>
1. Changing default ADC sens line for ch0 on PA0
- `adc1` was documented to be sensing ch14 on PB1.
- This seems impossible because of a conflict between ch14
and UART2.
- `adc1` node is not defined in the board's DTS.
2. Set SPI1 NSS pin to PA4
- hardware NSS signal was documented to output on PC9.
- PC9 cannot be assigned as HW NSS for SPI1.
- HW NSS on SPI1 is by default on PA4, and remapeable to PA15.
3. Update number of UART lines available
- There are a total of 7 U(S)ARTs available on this board.
Signed-off-by: Alex Fabre <alex.fabre@rtone.fr>
The two functions did not have unit tests, and at least one
issue with input validation have been indentified and fixed
prior to these.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The function did not check if the provided string had a zero
length before starting to truncate, which meant that last_byte_p
could possible have pointed to the value before the string.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Sockets service thread need to be cooperative for best performance.
Enable CACHE_MANAGEMENT as some boards don't do it by default, it has
devastating impact on the performance as well.
Finally, increase FDS count, so that uploads still work when both TCP
and UDP receiver are enabled.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The nvs_calc_free_space() function does not return 0 when the NVS
is considered full, because some special ATEs are not taken into account.
This commit takes into account the ATE that is reserved for deletion in
each sector, in addition of the 'GC done' ATE when present.
Signed-off-by: Adrien Ricciardi <aricciardi@baylibre.com>
