To be able to check during compilation that the number of endpoints provided by a USB peripheral is not exceeded, each interface class has to define the number of IN and OUT endpoints it requires.
19380: driver/usbdev_synopsys_dwc2: use correct number of EPs r=gschorcht a=gschorcht
### Contribution description
This PR fixes the problem that the driver uses the wrong number of EPs when using the USB OTG HS core.
The constant `DWC2_USB_OTG_FS_NUM_EP` was used in several places independent on whether the USB OTG FS core or the USB OTG HS core is used even though there is a function `_max_endpoints` which takes into account the configuration used. For most MCUs this was not a problem, because they have only a USB OTG FS core anyway. But for MCUs like the STM32, which have both a USB OTG FS core and a USB OTG HS core, it matters.
### Testing procedure
Use either
```
USEMODULE='stdio_cdc_acm' BOARD=stm32f429i-disc1 make -j8 -C tests/usbus_cdc_ecm flash
```
or
```
BOARD=stm32f429i-disco make -j8 -C tests/usbus_cdc_ecm flash
```
Without this PR it just stucks due to an `assert` here: 19da279ba5/sys/usb/usbus/cdc/ecm/cdc_ecm.c (L217-L221) The reason is that the USB OTG FS core has only 4 EPs but the application requires 5 IN EPS.
With the PR it works as expected.
### Issues/PRs references
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
This commit fixes the problem that the driver uses the wrong number of EPs when using the USB OTG HS core. The reason is that the constant `DWC2_USB_OTG_FS_NUM_EP` is used in several places even though there is a function `_max_endpoints` which takes into account the configuration used. For most MCUs this is not a problem, because they have only a USB OTG FS core anyway. But for MCUs like the STM32, which has both a USB OTG FS core and a USB OTG HS core, it matters.
19374: makefiles/boards/stm32: fix DFU_USB_ID handling r=benpicco a=gschorcht
### Contribution description
This PR fixes the still existing problem that an application can't be flashed to a STM32 board if it uses `riotboot_dfu` with default VID/PID (1209:7d02).
In PR #18964 item 1, the problem was already described that an application can't be flashed on a board that is using `riotboot_dfu`. Using for example
```python
FEATURES_REQUIRED+=riotboot USEMODULE+=usbus_dfu make -C examples/saul BOARD=nucleo-f767zi \
PROGRAMMER=dfu-util all riotboot/flash-slot0
```
always leads to error
```python
/home/gs/src/RIOT-Xtensa-ESP.esp-idf-4.4/makefiles/boards/stm32.inc.mk:28: *** DFU_USB_ID is not set. Stop.
/home/gs/src/RIOT-Xtensa-ESP.esp-idf-4.4/makefiles/boot/riotboot.mk:87: recipe for target 'riotboot/bootloader/binfile' failed
```
even if `DFU_USB_ID` variable is set as described in documentation.
```python
FEATURES_REQUIRED+=riotboot USEMODULE+=usbus_dfu make -C examples/saul BOARD=nucleo-f767zi \
PROGRAMMER=dfu-util DFU_USB_ID=1209:7d02 all riotboot/flash-slot0
```
The reason is that the variable `DFU_USB_ID` isn't exported and the check 8dc8bf3567/makefiles/boards/stm32.inc.mk (L27-L29) sees an empty `DFU_USB_ID` variable here. It prevents to use `dfu-util` event though the following `dfu-util.mk` will generate a default value for this variable from the `USB_VID` and `USB_PID` variables if necessary.
Commit 6a76b94c6e0ae97bc743cc90d0811e691a35869e of PR #18964 tried to fix this problem but wasn't merged for any reason.
To fix this problem, the check is completely removed. If a board such as `weact-f4x1cx` uses a DFU boorloader and requires a certain VID/PID combination, board's makefile is responsible to set `DFU_USB_ID` variable.
### Testing procedure
It is not necessary to use a real boad, checking the compilation process is sufficient.
1. Using default VID/PID as described in documentation:
```python
FEATURES_REQUIRED+=riotboot USEMODULE+=usbus_dfu make -C examples/saul BOARD=nucleo-f767zi \
PROGRAMMER=dfu-util all riotboot/flash-slot0
```
can't be compiled without this PR but calls `dfu-util` correctly with this PR using the default VID/PID:
```python
dfu-util --device 1209:7d02 --alt 0 --download examples/saul/bin/nucleo-f767zi/riotboot_files/slot0.1678440536.bin
```
2. Using a VID/PID as described in documentation:
```python
FEATURES_REQUIRED+=riotboot USEMODULE+=usbus_dfu make -C examples/saul BOARD=nucleo-f767zi \
DFU_USB_ID=1209:affe PROGRAMMER=dfu-util all riotboot/flash-slot0
```
can't be compiled without this PR but calls `dfu-util` correctly with this PR using the default VID/PID:
```python
dfu-util --device 1209:affe --alt 0 --download examples/saul/bin/nucleo-f767zi/riotboot_files/slot0.1678440536.bin
```
3. Compiling a board with DFU bootloader
```python
make -C examples/saul flash BOARD=weact-f411ce
```
should still call dfu-util with correct VID/PID:
```python
dfu-util --device 0483:df11 --alt 0 --download /home/gs/src/RIOT-Xtensa-ESP.esp-idf-4.4/examples/saul/bin/weact-f411ce/saul_example.bin --dfuse-address 0x8000000:leave
```
### Issues/PRs references
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
19375: tools/renode: add support for target reset r=benpicco a=aabadie
19376: boards/stm32f4discovery: use default port to access stdio via cdc acm r=benpicco a=aabadie
19377: pkg/tinyusb: fix default VID/PID configuration r=benpicco a=gschorcht
### Contribution description
This PR fixes the default VID/PID configuration if tinyUSB board reset feature is used.
While reviewing PR #19086 I was wondering why `esp32s2-wemos-mini` requires to set `USB_VID`/`USB_PID` explicitly to `USB_VID_TESTING`/`USB_PID_TESTING`. The reason was that tinyUSB board reset feature wasn't declared as RIOT internal.
### Testing procedure
Flashing `esp32s2-wemos-mini` should still work.
```
BOARD=esp32s2-wemos-mini make -C tests/shell flash
```
The VID/PID should be `1209:7d00` and not `1209:7d01`.
### Issues/PRs references
Co-authored-by: Alexandre Abadie <alexandre.abadie@inria.fr>
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
Defining the feature as RIOT internal avoids that the board has to set USB_VID and USB_PID explicit to USB_VID_TESTInG/USB_PID_TESTING if tinyUSB board reset feature is used.
The variable `DFU_USB_ID` is used by `dfu-util` to specify the device to be flashed. The STM32 make system prevents `dfu-util` from being used as programmer if this variable is not set, although `dfu-util.mk` generates a default value for this variable from the `USB_VID` and `USB_PID` variables if necessary. Therefore, checking the `DFU_USB_ID` variable is removed here. If a board requires a specific combination of VID/PID for `dfu_util`, it is responsible for setting the `DFU_USB_ID` variable.
19362: sys/usbus: handle exceeding number of endpoints r=dylad a=gschorcht
### Contribution description
This PR fixes issue #19359 for STM32 USB OTG cores partially:
1. It must not be silently ignored if the number of endpoints is not sufficient for an application. Instead of producing a non-working application, the application stops now with `kernel_panic` if the number of EPs is exhausted. This fixes the problem described in issue #19359 for USB cores with CID version 1.x, e.g. for STM32F439ZI FS interface (CID 1200) since they only have 4 IN and 4 OUT endpoints including the control endpoint EP0.
2. [Update: this part was fixed by PR #17086] ~If a feature is not supported, the device has to signal a STALL on the endpoint that should be used for the data phase in a control transaction. This means that for control read transactions the IN endpoint must signal a STALL and for control write transactions the OUT endpoint must signal a STALL. In former implementation, only the IN endpoint signaled a STALL independent on whether it was a control read or control write transaction. The change also fixes the problem that the enumeration stopped for about 5 seconds if module `usb_reset_board` isn't used. The reason is that the host sends a `SET LINE CODING` request to the CDC ACM interface and the device must signal a STALL on the OUT endpoint if it is not supported.~
### Testing procedure
1. Use a STM32 board with USB OTG version 1.x, for example a `nucleo-f439zi`:
```
USEMODULE='stdio_cdc_acm' BOARD=nucleo-f439zi make -j8 -C tests/usbus_cdc_ecm flash
```
Without this PR, the application seems to run but the CDC ECM interface is not working. The `ping` command can't be executed. With this PR, the application stops with `kernel_panic`. Because `stdio_cdc_acm` is used which doesn't work in this case, the `kernel_panic` has to be observed in debugger.
```
USEMODULE='stdio_cdc_acm' BOARD=nucleo-f439zi make -j8 -C tests/usbus_cdc_ecm debug
```
2. [Update: this part was fixed by PR #17086] ~Use a STM32 board with USB OTG version 2.x and USB FS connector, for example a `nucleo-f767zi`:
```
USEMODULE='stdio_cdc_acm' BOARD=nucleo-f767zi make -j8 -C tests/usbus_cdc_ecm flash
```
Without this PR a delay of 5 seconds in enumeration of the CDC ACM interface can be observed before the CDC ECM interface is enumerated. With this PR there is no delay anymore.~
### Issues/PRs references
Fixes issue #19359 patially.
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
If the number of endpoints is not sufficient for an application, it should not be silently ignored and cause a non-working application. Rather, should cause an assertion as it is a configuration issue.
17086: usbdev: Add dedicated stall functions r=benpicco a=bergzand
### Contribution description
This PR adds dedicated stall functions for usbdev peripherals. Two
functions are added. The first function (usbdev_ep_stall) to enable and
disable the stall condition on generic endpoints. The second function is
a dedicated function to set the stall condition on endpoint zero in both
directions. This status can only be set and should automatically be
cleared by the usbdev implementation (or hardware) after a new setup
request is received from the host.
### Testing procedure
- examples/usbus_minimal should still enumerate correctly on the host side.
- #17085 can be used to demonstrate the ep0_stall function with the `tests/usbus_cdc_acm_stdio/` test
### Issues/PRs references
None
Co-authored-by: Koen Zandberg <koen@bergzand.net>
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
19351: tools/compile_like_murdock: Add features r=MrKevinWeiss a=MrKevinWeiss
### Contribution description
This adds a few nice-to-have features:
- The `-b all` allows all supported boards to build
- The `-v, -vv, -vvv` allows different levels of verbosity, the most useful might be `-v` which filters only the kconfig diffs
- All warning or STDERR based output from `info-*` are hidden
- The `-m` that only runs a diff of modules and packages, skipping any compilation
### Testing procedure
In induce a mismatch and check the output with `-v`, `-vv` options, clean it up and see the full output with `-vvv`:
```
./dist/tools/compile_test/compile_like_murdock.py -a APP -d -b BOARD -v
```
Check all supported boards are listed with
```
./dist/tools/compile_test/compile_like_murdock.py -a examples/hello-world/ -d -b all
```
### Issues/PRs references
Co-authored-by: MrKevinWeiss <weiss.kevin604@gmail.com>
19365: bootloaders: fix bootloader button logic r=benpicco a=dylad
### Contribution description
In lastest master, the `BTN_BOOTLOADER_INVERTED` logic doesn't work as expected.
This PR fixes the underlying logic by replacing the `BTN_BOOTLOADER_INVERTED` macro definition by a runtime function.
In fact the current code:
```
#ifndef BTN_BOOTLOADER_INVERTED
#if (BTN0_MODE == GPIO_IN_PD)
#define BTN_BOOTLOADER_INVERTED false
#else
#define BTN_BOOTLOADER_INVERTED true
#endif
#endif
```
cannot work because both `BTN0_MODE` and `GPIO_IN_PD` are not known by the precompiler as they are enum values defined at cpu level.
Thus, replaces it by a runtime function in our bootloader applications.
I've also add `GPIO_OD_PU` along side `GPIO_IN_PU` and add a new define (which can be override at board level or app level) in case an external pullup is used.
### Testing procedure
Flash the riotboot_dfu bootloader:
`make BOARD=saml21-xpro -C bootloaders/riotboot_dfu flash`
Then, flash any test app:
`PROGRAMMER=dfu-util USEMODULE=usbus_dfu make BOARD=saml21-xpro -C tests/shell riotboot/flash-slot0`
With master, the application will not start.
With this PR, the application will start after flashing.
### Issues/PRs references
Fixes#19364
19366: nanocoap_sock: don't include token in empty ACK response r=benpicco a=benpicco
19367: cord: bump reference from draft to rfc r=benpicco a=bergzand
### Contribution description
The draft is an RFC, this bumps the "see also" in the docs to the rfc.
### Testing procedure
Check that the correct RFC is linked in the docs.
### Issues/PRs references
None
Co-authored-by: Dylan Laduranty <dylan.laduranty@mesotic.com>
Co-authored-by: Benjamin Valentin <benjamin.valentin@ml-pa.com>
Co-authored-by: Koen Zandberg <koen@bergzand.net>
19343: ztimer: add ztimer_stopwatch convenience functions r=benpicco a=benpicco
19349: cpu/native: Switch to ztimer for gettimeofday r=benpicco a=MrKevinWeiss
### Contribution description
A xtimer is somewhat taken over by ztimer this explicitly uses ztimer instead of relying on the compatibility layer.
### Testing procedure
`make all test -C tests/cpp11_mutex/`
and green murdock I guess.
### Issues/PRs references
19353: doc: add quicklink to boards in navbar r=benpicco a=OlegHahm
### Contribution description
Finding a list of supported boards and how to use them is an essential information. Currently this list is somewhat hidden under "Modules" which is not very intuitive. Hence, I propose to (at least) put a link in the side menu to this overview page.
### Testing procedure
1. Call `make doc`
2. Check the sidebar `${RIOT_BASE}/doc/doxygen/html/index.html` for an entry "Supported Boards"
19361: nanocoap_sock: ensure response address is the same as request address r=benpicco a=benpicco
19363: Fix stm32 timer periodic r=benpicco a=Enoch247
### Contribution description
From the commit msg:
> cpu/stm32/periph/timer: remove unneeded header
>
> I see no reason this header should be included. It does not exist in
> RIOT's source tree. This patch removes the include.
and
> cpu/stm32/periph/timer: fix execution flow
>
> The implmentation of `timer_set_absolute()` has The following problems.
> First, it attempts to restore the auto reload register (ARR) to it's
> default if the ARR was previosly set by `timer_set_periodic()` by
> comparing it to the channel's capture compare (CC) register _after_ it
> has already set the CC register. Secondly, it clears spurious IRQs
> _after_ the CC register has been set. If the value being set is equal to
> the timer's current count (or the two become equal before the supurios
> IRQ clearing happens), this could cause a legitimate IRQ to be cleared.
>
> The implmentation of `timer_set()` has the same error in handling the
> ARR as described above.
>
> This patch reorders the operations of both functions to do:
>
> 1. handle ARR
> 2. clear spurious IRQs
> 3. set channel's CC
> 4. enable IRQ
>
> Additionally, the calulation of `value` in `timer_set()` is moved
> earlier in the function's exec path as a pedantic measure.
### Testing procedure
I tested by doing the following:
1. `make -C tests/periph_timer BOARD=nucleo-f767zi all flash term`
2. press s
3. press [ENTER]
4. observe test passes
5. `make -C tests/periph_timer_periodic BOARD=nucleo-f767zi all flash term`
6. press s
7. press [ENTER]
8. observe test passes
9. `make -C tests/periph_timer_short_relative_set BOARD=nucleo-f767zi all flash term`
10. press s
11. press [ENTER]
12. observe test passes
### Issues/PRs references
- none known
Co-authored-by: Benjamin Valentin <benpicco@beuth-hochschule.de>
Co-authored-by: MrKevinWeiss <weiss.kevin604@gmail.com>
Co-authored-by: Oleg Hahm <oleg@hobbykeller.org>
Co-authored-by: Benjamin Valentin <benjamin.valentin@bht-berlin.de>
Co-authored-by: Joshua DeWeese <jdeweese@primecontrols.com>
19234: Makefile.dep: pull in FEATURES_REQUIRED for periph r=maribu a=OlegHahm
### Contribution description
Whenever the module of a peripheral driver, i.e., periph_* should be used, the corresponding entry in the FEATURES_REQUIRED should be added. Conflicts between these modules are only checked when this entry is present.
### Testing procedure
Use, for example, the `default` example on a board with `periph_rtt` and `periph_rtc` available (e.g., the iotlab-m3).
Change the Makefile from `FEATURES_OPTIONAL += periph_rtc` to `USEMODULE += periph_rtc`. Build and flash the application and call the `rtc gettime` shell command a couple of times.
#### On current master:
You should see incorrect timestamps because the `periph_rtc` conflicts with `periph_rtt` (which is pulled in by the `ztimer_periph_rtt` module).
#### With this PR:
You should see a warning like
```
The following features may conflict: periph_rtc periph_rtt
Rationale: On the STM32F1, the RTC and RTT map to the same hardware peripheral. Only one standard C library can be used. Only one GPIO IRQ implementation can be used. Package tinyUSB is not yet compatible with periph_usbdev.
EXPECT undesired behaviour!
```
Co-authored-by: Oleg Hahm <oleg@hobbykeller.org>
19358: sys/usbus/cdc/ecm: fix High-Speed mode r=dylad a=gschorcht
### Contribution description
This PR provides some changes to fix the USBUS CDC ECM interface in High-Speed mode.
In High-Speed mode, the EP data size has to be at least 512 bytes instead of 64 Byte in Full-Speed mode. To be able to define configurations like EP data sizes depending on whether Full-Speed or High-Speed USB device peripherals are used, the feature `periph_usbdev_hs`/`HAD_PERIPH_USBDEV_HS` is introduced.
### Testing procedure
Use `tests/usbus_cdc_ecm` and any board with USB HS connector, for example:
```
USEMODULE=periph_usbdev_hs_utmi BOARD=stm32f723e-disco make -j8 -C tests/usbus_cdc_ecm flash
```
`ping` command works with this PR but doesn't work without this PR.
### Issues/PRs references
Co-authored-by: Gunar Schorcht <gunar@schorcht.net>
Whenever the module of a peripheral driver, i.e., periph_* should be used, the corresponding entry in the
FEATURES_REQUIRED should be added. Conflicts between these modules are
only checked when this entry is present.
The implmentation of `timer_set_absolute()` has The following problems.
First, it attempts to restore the auto reload register (ARR) to it's
default if the ARR was previosly set by `timer_set_periodic()` by
comparing it to the channel's capture compare (CC) register _after_ it
has already set the CC register. Secondly, it clears spurious IRQs
_after_ the CC register has been set. If the value being set is equal to
the timer's current count (or the two become equal before the supurios
IRQ clearing happens), this could cause a legitimate IRQ to be cleared.
The implmentation of `timer_set()` has the same error in handling the
ARR as described above.
This patch reorders the operations of both functions to do:
1. handle ARR
2. clear spurious IRQs
3. set channel's CC
4. enable IRQ
Additionally, the calulation of `value` in `timer_set()` is moved
earlier in the function's exec path as a pedantic measure.
To be able to define configurations like EP data sizes depending on whether full-speed or high-speed USB device peripherals are used, the feature `HAS_PERIPH_USBDEV_HS` is introduced.
To be able to define configurations like EP data sizes depending on whether full-speed or high-speed USB device peripherals are used, the feature `periph_usbdev_hs` is introduced.
