19057: fuzzing: Add uri_parser setup r=benpicco a=Teufelchen1
Hello!
### Contribution description
This PR is a replacement for PR #18802
In this contribution:
* The variable `AFL_FLAGS` is renamed to `FLAGS_FOR_AFL` because AFL is always complaining that `AFL_FLAGS` is not a valid env var for it. While this is not a bug nor an issue, I found it to be annoying.
* A generic input reader is added to simplify building a test harness
* The usage of this reader is demonstrated by adding a harness for fuzzing the uri_parser
(needs squashing after review)
### Testing procedure
Go to `fuzzing/uri_parser` and run `make all-asan` and `make fuzz` to get some action going.
Also mildly interesting: `./dist/tools/compile_test/compile_like_murdock.py -b native -a fuzzing/uri_parser`
### Issues/PRs references
The original PR #18802 is replaced because the generic input reader is present in both PRs but this PoC harness is much simpler.
19151: examples/gcoap: Fix shell parameter validation r=benpicco a=maribu
### Contribution description
Executing the shell command with an URI-Path that doesn't start with a slash results in an assertion error while composing the client side message. This is suboptimal user experience, so add an explicit check for a valid URI-Path and a dedicated error message.
### Testing procedure
#### In `master`
```
$ make BOARD=microbit-v2 -C examples/gcoap flash term
[...]
2023-01-15 22:23:32,512 # coap get [::1] /.well-known/core
2023-01-15 22:23:32,516 # gcoap_cli: sending msg ID 52272, 23 bytes
2023-01-15 22:23:32,520 # gcoap: response Success, code 2.05, 46 bytes
2023-01-15 22:23:32,524 # </cli/stats>;ct=0;rt="count";obs,</riot/board>
> coap get [::1] foo
2023-01-15 22:23:34,763 # coap get [::1] foo
2023-01-15 22:23:34,763 # 2329
2023-01-15 22:23:34,765 # *** RIOT kernel panic:
2023-01-15 22:23:34,767 # FAILED ASSERTION.
2023-01-15 22:23:34,767 #
2023-01-15 22:23:34,775 # pid | name | state Q | pri | stack ( used) ( free) | base addr | current
2023-01-15 22:23:34,784 # - | isr_stack | - - | - | 512 ( 200) ( 312) | 0x20000000 | 0x200001c8
2023-01-15 22:23:34,793 # 1 | main | running Q | 7 | 1536 ( 1072) ( 464) | 0x200006c0 | 0x2000095c
2023-01-15 22:23:34,802 # 2 | 6lo | bl rx _ | 3 | 1024 ( 328) ( 696) | 0x200036c0 | 0x200039c4
2023-01-15 22:23:34,810 # 3 | ipv6 | bl rx _ | 4 | 1024 ( 460) ( 564) | 0x20001294 | 0x20001574
2023-01-15 22:23:34,819 # 4 | udp | bl rx _ | 5 | 512 ( 300) ( 212) | 0x20003e98 | 0x20003f9c
2023-01-15 22:23:34,828 # 5 | coap | bl anyfl _ | 6 | 1112 ( 704) ( 408) | 0x20000e38 | 0x200011c4
2023-01-15 22:23:34,837 # 6 | nrf802154 | bl anyfl _ | 2 | 896 ( 288) ( 608) | 0x20001a90 | 0x20001d54
2023-01-15 22:23:34,843 # | SUM | | | 6616 ( 3352) ( 3264)
2023-01-15 22:23:34,843 #
2023-01-15 22:23:34,844 # *** halted.
2023-01-15 22:23:34,844 #
```
#### This PR
```
$ make BOARD=microbit-v2 -C examples/gcoap flash term
[...]
make: Entering directory '/home/maribu/Repos/software/RIOT/examples/gcoap'
/home/maribu/Repos/software/RIOT/dist/tools/pyterm/pyterm -p "/dev/ttyACM0" -b "115200"
2023-01-15 22:22:27,842 # Connect to serial port /dev/ttyACM0
Welcome to pyterm!
Type '/exit' to exit.
coap get [::1] /.well-known/core
2023-01-15 22:22:40,042 # coap get [::1] /.well-known/core
2023-01-15 22:22:40,046 # gcoap_cli: sending msg ID 25182, 23 bytes
2023-01-15 22:22:40,050 # gcoap: response Success, code 2.05, 46 bytes
2023-01-15 22:22:40,054 # </cli/stats>;ct=0;rt="count";obs,</riot/board>
> coap get [::1] foo
2023-01-15 22:22:43,858 # coap get [::1] foo
2023-01-15 22:22:43,862 # ERROR: URI-Path must start with a "/"
2023-01-15 22:22:43,866 # usage: coap <get|post|put|ping|proxy|info>
```
### Issues/PRs references
None
Co-authored-by: Teufelchen1 <bennet.blischke@haw-hamburg.de>
Co-authored-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
19010: bootloaders/riotboot: add tinyUSB DFU support r=benpicco a=gschorcht
### Contribution description
This PR provides
- the tinyUSB DFU and DFU Runtime support and
- the `riotboot_tinyusb_dfu` bootloader that uses the tinyUSB DFU mode to flash new application images.
~This PR includes PR #18983 for now to be compilable.~
### Testing procedure
1. Use any board that supports the `riotboot´ and `tinyusb_device` features and flash the bootloader first, for example
```
BOARD=nucleo-f767zi make -C bootloaders/riotboot_tinyusb_dfu flash
```
and check that the `riotboot_tinyusb_dfu` bootloader is in DFU mode:
```
dfu-util --list
```
3. Flash a first application using the following command:
```
FEATURES_REQUIRED=riotboot USEMODULE=tinyusb_dfu BOARD=nucleo-f767zi \
make -C tests/saul PROGRAMMER=dfu-util riotboot/flash-slot0
```
and check that the application starts and is seen as upgradable:
```
dfu-util --list
```
4. Restart the node in bootloader DFU mode by:
```
dfu-util -e
```
Flash a second application, for example
```
FEATURES_REQUIRED=riotboot USEMODULE=tinyusb_dfu BOARD=nucleo-f767zi \
make -C tests/shell PROGRAMMER=dfu-util riotboot/flash-slot1
```
and check that the second application starts and is seen as upgradable:
```
dfu-util --list
```
### Issues/PRs references
~Depends on PR #18983~
19149: SECURITY: Describe that declassification is an option r=benpicco a=chrysn
### Contribution description
Our security policy does not contain provisions for the case when what is reported is not what we consider an actual security issue. As it is described now, everything reported through security@ would go through the full treatment, including a point release.
I'm not sure it belongs into the text itself (as it's more about how security reporters interact with the project than internals), but declassification should IMO be backed at least by 3 maintainers, and no strong NACK.
### Issues/PRs references
#19141 followed that procedure after some chat on it on the maintainers channel. (In the discussion, I proposed declassification, with 2.5 people supporting it and one "I was about to, but can we be sure nobody is using it?" voice).
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
Co-authored-by: chrysn <chrysn@fsfe.org>
17066: sys/irq: Add C++ wrapper using RAII r=maribu a=jenswet
### Contribution description
This adds a C++ wrapper around the `irq.h` API. The wrapper uses RAII to accomplish a convenient and bug resistent use.
A little background: I'm currently writing my master thesis on using C++ for embedded development, at the working group that `@maribu` is part of. For that I will try to add better C++ support to several parts of RIOT and then do some benchmarking and metrics to compare it with the C implementation. For example, I also plan to add a wrapper around i2c, a std::cout drop-in replacement and probably some more about networks or threads.
### Testing procedure
I've added a unit test to verify that the IRQ wrapper calls the original `irq` functions as expected. As C++ and wrapper testing isn't done much so far in this project, I've added two additional headers to ease testing:
1. #17076 - fake functions framework, already merged
2. As there is no framework for C++ unit tests yet, I've added something for this too. Unfortunately the existing frameworks like GoogleTest, CppUTest or CppUnit don't easily compile for embedded or are difficult to integrate in to the RIOT build process. That's why I wrote some (simple) helper functions and macros inspired by the above frameworks. That allows to create C++ tests based on a fixture class with set up and tear down methods. It also allows some simple assertions and is easily extendable for other use cases. It wraps some of the fff functionality too.
Both of this is obviously not required for the initial reason of this PR. But I'd like to provide unit tests for the features that I suggest to introduce where possible. So I'd appreciate some feedback on that too. If you'd prefer a PR without or different tests please let me know.
You can run the test `irq_cpp` locally or on the CI to test the implementation.
Please feel free to give feedback or suggest improvements!
Co-authored-by: Jens Wetterich <jens@wetterich-net.de>
18763: sys/tiny_strerror: add missing error codes r=benpicco a=maribu
### Contribution description
When double-checking the error codes provided by newlib by default (without magic defines, such as `__LINUX_ERRNO_EXTENSIONS__` or `__CYGWIN__`), some where still missing in `tiny_strerror()`. This adds the missing ones.
This in turn showed that three errno codes were missing in the avr-libc compat `errno.h`, which are added as well.
### Testing procedure
Murdock should double check that the added errno codes indeed are defined by default.
### Issues/PRs references
None
Co-authored-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
19141: shell/rtc: Fix out of bounds access; document error behavior r=maribu a=chrysn
### Contribution description
### Testing procedure
Should be trivial enough, especially as the difference is hard to spot interactively.
On native, run the default example (and wait for the traffic to settle).
Then, run
```
> rtc poweron
> rtc settime 2022-01-01 00:00:00
> rtc settime 2022-99-01 00:00:00
```
Both still work, but the latter doesn't access unassigned memory any more
### Issues/PRs references
This fixes an issue that was submitted anonymously.
Co-authored-by: chrysn <chrysn@fsfe.org>
When double-checking the error codes provided by newlib by default
(without magic defines, such as `__LINUX_ERRNO_EXTENSIONS__` or
`__CYGWIN__`), some where still missing in `tiny_strerror()`. This
adds the missing ones.
17810: drivers/slipdev: implement sleep states r=benpicco a=benpicco
18348: sys/net/gnrc/pktbuf_static: make use of alignas() r=maribu a=maribu
### Contribution description
Since we are now using C11, we can make use of `alignas()` provided by `<stdalign.h>` to make the alignment code easier to read.
### Testing procedure
I didn't expect this to change binaries, but is safes 4 bytes. `elf_diff` shows that the compiler (at least GCC 11.3.0) was not able to detect that `gnrc_pktbuf_static_buf` was just an alias for `_pktbuf_buf`. That makes sense since it would be hard without LTO to rule out external writes to `gnrc_pktbuf_static_buf`, unless one would have added a `const` (to the pointer, not to the data the pointer points to).
The [output of `elf_diff`](https://mari-bu.de/pr_18348_gnrc_pktbuf_static_elf_diff.html) looks otherwise quite unscary.
Also:
```
$ make BOARD=nucleo-f767zi -C tests/unittests/ tests-pktbuf flash test
make: Entering directory '/home/maribu/Repos/software/RIOT/tests/unittests'
Building application "tests_unittests" for "nucleo-f767zi" with MCU "stm32".
[...]
Welcome to pyterm!
Type '/exit' to exit.
READY
s
START
.............................................
OK (45 tests)
make: Leaving directory '/home/maribu/Repos/software/RIOT/tests/unittests'
```
### Issues/PRs references
None
19120: CI: seperate check-labels and check-commits workflows r=maribu a=kaspar030
Co-authored-by: Benjamin Valentin <benjamin.valentin@ml-pa.com>
Co-authored-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
Co-authored-by: Kaspar Schleiser <kaspar@schleiser.de>
19106: core/lib: Add macros/utils.h header r=aabadie a=maribu
### Contribution description
The macros CONCAT(), MIN(), and MAX() are defined over and over again in RIOT's code base. This de-duplicates the code by moving the macros to a common place.
### Testing procedure
Generated binaries don't change, as this only a de-duplication of macros that doesn't change their definition.
### Issues/PRs references
None
Co-authored-by: Marian Buschsieweke <marian.buschsieweke@ovgu.de>
The macros CONCAT(), MIN(), and MAX() are defined over and over again in
RIOT's code base. This de-duplicates the code by moving the macros to a
common place.
18752: nanocoap_sock: deprecate nanocoap_get() r=benpicco a=benpicco
19100: cpu/esp_common: allow configuration of UART0 r=benpicco a=gschorcht
### Contribution description
This PR
- fixes the issue for ESP32 SoCs that UART0 signals can't be routed to arbitrary GPIOs and
- allows the configuration of the UART device used by the bootloader.
The UART interface and its configuration used by the STDIO are defined in RIOT using the define `STDIO_UART_DEV` and the configuration of the corresponding UART device in `periph_conf.h`.
However, the bootloader compiled directly in ESP-IDF uses its own definitions `CONFIG_ESP_CONSOLE_UART_*` for the UART configuration. To be able to use a consistent UART configuration in RIOT and the bootloader, e.g. to see the output of the 2nd stage bootloader, these `CONFIG_ESP_CONSOLE_UART_*` can be defined via a set of KConfig variables in RIOT (not yet implemented in Kconfig):
- `CONSOLE_CONFIG_UART_NUM` defines the UART device to be used by the bootloader and by `STDIO_UART_DEV`
- `CONSOLE_CONFIG_UART_RX` and `CONSOLE_CONFIG_UART_TX` define the GPIOs to be used by the bootloader and should be the GPIOs as defined in `periph_conf.h` for the corresponding UART device.
### Testing procedure
Any ESP32 node should still work with `stdio_uart` and the default configuration. To test an alternative configuration, use
```
CFLAGS='-DUART1_TXD=5 -DUART1_RXD=4 -DCONFIG_CONSOLE_UART_NUM=1 -DCONFIG_CONSOLE_UART_TX=5 -DCONFIG_CONSOLE_UART_RX=4' USEMODULE=esp_log_startup BOARD=esp32-wroom-32 make -C tests/shell flash
```
The bootloader output and the STDIO should be routed to UART1 at GPIO4 and GPIO5.
### Issues/PRs references
Prerequisite for PR ##18863
19104: tests/periph_uart: only exclude STDIO_UART_DEV if stdio_uart is used r=benpicco a=benpicco
Co-authored-by: Benjamin Valentin <benjamin.valentin@ml-pa.com>
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
Co-authored-by: Benjamin Valentin <benjamin.valentin@bht-berlin.de>
18773: nanocoap_sock: fix handling empty ACKs with separate response r=maribu a=benpicco
Co-authored-by: Benjamin Valentin <benjamin.valentin@ml-pa.com>
Co-authored-by: Benjamin Valentin <benjamin.valentin@bht-berlin.de>
19037: sys/usb, pkg/tinyusb: move USB board reset from highlevel STDIO to CDC ACM r=dylad a=gschorcht
### Contribution description
The USB board reset function `usb_board_reset_coding_cb` can be used on any CDC-ACM interface, even if the CDC ACM interface is not used as high-level STDIO. Therefore, this PR provides the following changes:
- The call of the board reset function `usb_board_reset_coding_cb` from USBUS stack has been moved from the STDIO CDC ACM implementation to the CDC ACM implementation and is thus a feature of any USBUS CDC ACM interface which does not necessarily have to be used as highlevel STDIO.
- The call of the board reset function `usb_board_reset_coding_cb` from tinyUSB stack been moved from module `tinyusb_stdio_cdc_acm` to module `tinyusb_contrib` and is compiled in if the `tinyusb_class_cdc` module is used together the `tinyusb_device` module. Thus, it is now a feature of the tinyUSB CDC ACM interface, which does not necessarily have to be used as highlevel STDIO.
- The `usb_board_reset` module defines the `usb_board_reset_in_bootloader` function as a weak symbol to be used when reset in bootloader if no real implementation of this function is compiled in and the `riotboot_reset` module is not used. It only prints an error message that the reset in bootloader is not supported. This is necessary if the module `usb_board_reset` is used to be able to restart the board with an application via a USB CDC ACM interface, but the board's bootloader does not support the reset in bootloader feature.
- A test application has been added that either uses the highlevel STDIO `stdio_acm_cdc` or creates a CDC-ACM interface to enable board resets via USB. If the `usbus_dfu` module is used, it also initializes the DFU interface to be able to work together with the `riotboot_dfu` bootloader.
### Testing procedure
1. Use a board with a bootloader that supports the reset in bootloader via USB, but don't use the highlevel STDIO to check that it works with `usbus_cdc_acm`, for example:
```python
USEMODULE=stdio_uart BOARD=arduino-mkr1000 make -C tests/usb_board_reset flash
```
After reset in application with command
```python
stty -F /dev/ttyACM0 raw ispeed 600 ospeed 600 cs8 -cstopb ignpar eol 255 eof 255
```
command `dmesg` should give an output like the following with RIOT's test VID/PID:
```python
dmesg
[1745182.057403] usb 1-4.1.2: new full-speed USB device number 69 using xhci_hcd
[1745182.160386] usb 1-4.1.2: New USB device found, idVendor=1209, idProduct=7d01, bcdDevice= 1.00
[1745182.160390] usb 1-4.1.2: New USB device strings: Mfr=3, Product=2, SerialNumber=4
[1745182.160392] usb 1-4.1.2: Product: arduino-mkr1000
[1745182.160393] usb 1-4.1.2: Manufacturer: RIOT-os.org
[1745182.160395] usb 1-4.1.2: SerialNumber: 6B6C2CA5229020D8
[1745182.170982] cdc_acm 1-4.1.2:1.0: ttyACM0: USB ACM device
```
After reset in bootloader with command
```python
stty -F /dev/ttyACM0 raw ispeed 1200 ospeed 1200 cs8 -cstopb ignpar eol 255 eof 255
```
command `dmesg` should give an output like the following with vendor VID/PID:
```python
[1746220.443792] usb 1-4.1.2: new full-speed USB device number 70 using xhci_hcd
[1746220.544705] usb 1-4.1.2: New USB device found, idVendor=2341, idProduct=024e, bcdDevice= 2.00
[1746220.544708] usb 1-4.1.2: New USB device strings: Mfr=0, Product=0, SerialNumber=0
[1746220.553471] cdc_acm 1-4.1.2:1.0: ttyACM0: USB ACM device
```
2. Test the same as in 1., but this time use the highlevel STDIO to check that there is no regression and it still works with `stdio_cdc_acm`, for example:
```python
BOARD=arduino-mkr1000 make -C tests/usb_board_reset flash
```
3. Use a board that supports `riotboot_dfu` but doesn't use the highlevel STDIO and flash the `riotboot_dfu` bootloader, for example:
```python
BOARD=stm32f429i-disc1 make -C bootloaders/riotboot_dfu flash term
```
Once the bootloader is flashed, command `dfu-util --list` should give something like the following:
```python
Found DFU: [1209:7d02] ver=0100, devnum=14, cfg=1, intf=0, path="1-2", alt=1, name="RIOT-OS Slot 1", serial="6591620BCB270283"
Found DFU: [1209:7d02] ver=0100, devnum=14, cfg=1, intf=0, path="1-2", alt=0, name="RIOT-OS Slot 0", serial="6591620BCB270283"
```
If the output gives only
```python
Found Runtime: [1209:7d00] ver=0100, devnum=123, cfg=1, intf=0, path="1-2", alt=0, name="RIOT-OS bootloader", serial="6591620BCB270283"
```
an application is already running in DFU Runtime mode. Use `dfu-util -e` to restart it in bootloader DFU mode.
Then flash the test application, for example:
```python
FEATURES_REQUIRED=riotboot USEMODULE='usbus_dfu riotboot_reset' \
BOARD=stm32f429i-disc1 make -C tests/usbus_board_reset PROGRAMMER=dfu-util riotboot/flash-slot0
```
Once the test application is flashed, command `dfu-util --list` should give:
```python
Found Runtime: [1209:7d00] ver=0100, devnum=123, cfg=1, intf=0, path="1-2", alt=0, name="RIOT-OS bootloader", serial="6591620BCB270283"
```
Now, use command
```python
stty -F /dev/ttyACM1 raw ispeed 600 ospeed 600 cs8 -cstopb ignpar eol 255 eof 255
```
to restart the board in application. Command `dfu-util --list` should give again the following:
```python
Found Runtime: [1209:7d00] ver=0100, devnum=123, cfg=1, intf=0, path="1-2", alt=0, name="RIOT-OS bootloader", serial="6591620BCB270283"
```
That is, the application is running in DFU Runtime mode. Then use command
```python
stty -F /dev/ttyACM1 raw ispeed 1200 ospeed 1200 cs8 -cstopb ignpar eol 255 eof 255
```
to restart the board in bootloader DFU mode. Command `dfu-util --list` should now give the following:
```python
Found DFU: [1209:7d02] ver=0100, devnum=50, cfg=1, intf=0, path="1-2", alt=1, name="RIOT-OS Slot 1", serial="7D156425A950A8EB"
Found DFU: [1209:7d02] ver=0100, devnum=50, cfg=1, intf=0, path="1-2", alt=0, name="RIOT-OS Slot 0", serial="7D156425A950A8EB"
```
That is, the bootloader is in DFU mode and another application can be flash.
4. After a hard reset of the board under 3., try the commands `reboot` and `bootloader`.
5. To check the same for tinyUSB, use the existing tinyUSB application with a CDC ACM interface and add module `usb_board_reset`, for example:
```python
USEMODULE=usb_board_reset BOARD=stm32f429i-disc1 make -C tests/pkg_tinyusb_cdc_msc flash term
```
After flashing, it should be possible to restart the application with command:
```python
stty -F /dev/ttyACM1 raw ispeed 600 ospeed 600 cs8 -cstopb ignpar eol 255 eof 255
```
When using command
```python
stty -F /dev/ttyACM1 raw ispeed 1200 ospeed 1200 cs8 -cstopb ignpar eol 255 eof 255
```
the following error message should be shown in terminal
```python
[cdc-acm] reset in bootloader is not supported
```
### Issues/PRs references
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
The board reset function can be used on any CDC ACM interface. It is not necessary that the tinyUSB CDC ACM STDIO is used. Therefore, the board reset function is now a feature of the CDC ACM interface that don't require any other functionality.
Definition of a function as weak symbol for reset in bootloader which prints an error message if no real implementation is compiled in and
the module `riotboot_reset` is not used. This is required if the module `usb_board_reset` is used to restart the board with an application via an USB CDC ACM interface, but the board's bootloader does not support a reset in the bootloader.
The board reset function `usb_board_reset_coding_cb` can be used on any CDC ACM interface. It is not necessary that the highlevel STDIO is used. Therefore, the call of the board reset function `usb_board_reset_coding_cb` has been moved from the USBUS STDIO CDC ACM implementation to USBUS CDC ACM implementation and is thus a feature of the USBUS CDC ACM interface that don't require any other functionality.
In `dac_util_map(...)` the expression `((value - min) * UINT16_MAX)` was
cast to a 16 bit unsigned, then divided by `(max - min)`. This means
that anytime `(value - min) != 0` the numerator was truncated prior to
being divided and then returned.
This patch modifies the expression so that the downcast to 16 bits is
performed as the last operation.
STM32F2/4/7 MCUs use sectors instead of pages, where the minimum sector size is defined by FLASHPAGE_MIN_SECTOR_SIZE, which is 16KB or 32KB (the first sector) depending on the CPU_MODEL. In this case SLOT0_OFFSET must be a multiple of the minimum sector size to cover a whole sector.
The start/stop overhead that might by introduced by ztimer_acquire() and ztimer_release() called during ztimer_set() resp. ztimer_handler() should not be mesured here. It has its own adjustment field.
Furthermore, the overhead mesaurement uses ztimer_now(). It is allowed to called it only after the clock has been acquired.
Having to cast a password provided as `const char *` to
`const uint8_t *` is a needless pain in the ass when using the API.
Hence, fix it by accepting passwords and salts as `const void *`
instead.
Since fmt no longer has a significant advantage in stack consumption,
we need to bump the `MIN_SIZE` guard that prevents causing stack
overflows due to the printing of the stack consumption.
Hooking into the existing wrappers for `malloc()`, `calloc()`,
`realloc()`, and `free()`, the new (pseudo) module `malloc_tracing`
prints out the calls to the given functions, the program counter of
the caller, as well as the return result.
The intent is to aid debugging double-frees, invalid frees, or memory
leaks.
The previous implementation used creative construct for impedance
mismatching between the core list API (which returns a ptr to the
removed element if found) and the GNRC pkt list API (which returns a
ptr to the new list head) that creates a temporary list head on the
stack.
I'm not entirely sure if the previous implementation is containing
undefined behavior that is used against us with GCC >= 12.x, or if this
is a compiler bug. In either case, not reusing the core list API here
and just having a textbook linked list delete function here is not much
less readable and fixes the issue for our users.
If a NON confirmable message is sent with a callback function,
not receiving a response in time would lead to a retransmission.
This is of course an error, as only CON messages are to be retransmitted.
This allows including C headers from C++. It sadly reduced the
diagnostics on C++ code as well, were there warning may make sense as
unintended side effect. We may be able to drop that later on, when more
C APIs are properly wrapped in native C++ APIs, so that C headers do no
longer need to be compatible with C++ compilers.
Nightlies are failing due to kconfig mismatch.
It would seem this is a result of bringing in the USB stuff.
I assume that this uses ztimer periph_timer as a backend as periph_timer is already selected.
However, kconfig only resolves one and not recursively making it hard to match.
For not a hack is added to override for these boards.
In situations with high packet loss, if all sync packets are lost, conflicting
subnets will be configured.
If the network is static, this can be prevented by always using the highest number
of subnets that has been observed.
This assumes no nodes are physically added / removed from the network.
The `gnrc_udp` shell command uses the function `netutils_get_ipv6()` but
does not include the corresponding module `netutils`. The only reason
most applications that use `shell_cmd_gnrc_udp` link is because they
also include the `shell_cmd_gnrc_icmpv6_echo` module (e.g. implicit via
`gnrc_ipcmpv6_echo`), which includes this dependency.
This eliminates race conditions around unregistering netdevs.
gnrc_netreg_lookup-style functions perform a DEVELHELP check on whether
that lock is held.
