Fix stm32 periph headers #20531
Fix stm32 periph headers #20531
Conversation
This patch moves the SPI related defines and declarations out of the shared stm32 periph header into the SPI specific one.
This patch moves the USB related defines and declarations out of the shared stm32 periph header into the USB specific one.
github-actions
bot
added
Platform: ARM
| @@ -39,6 +39,65 @@ extern "C" { | |||
| */ | |||
| #define USBDEV_CPU_DMA_REQUIREMENTS __attribute__((aligned(USBDEV_CPU_DMA_ALIGNMENT))) | |||
|
|
|||
| #if !DOXYGEN /* hide implementation details */ | |||
There was a problem hiding this comment.
Should I have moved this to cpu/stm32/include/usbdev_stm32.h instead?
| #endif | ||
|
|
||
| /** | ||
| * @name Override the SPI mode values |
There was a problem hiding this comment.
I am assuming if I do the following it will make CI/CD happy (based on the HAVE_SPI_CLK_T below). Is this the right thing to do? Or should I document the struct and macro separately?
| * @name Override the SPI mode values | |
| * @brief Override the SPI mode values |
There was a problem hiding this comment.
you can test if the CI is happy with a change by running the dist/tools/doccheck locally
reading
https://www.doxygen.nl/manual/grouping.html#memgroup
https://www.doxygen.nl/manual/commands.html#cmdname
i think it would be better to have it like i will write in the suggesting comment
There was a problem hiding this comment.
There was a problem hiding this comment.
Thanks for the links. I setup to build the docs locally and am now playing around with your suggestion. I am seeing some weird behavior. According to those links, member groups may not be nested. The entire file is in a groups. This is a common idiom used through-out the RIOT source tree. This means that anywhere the member group mechanism is used, we are nesting them inside the one that wraps the entire file. Maybe this explains the weird things I am seeing?
There was a problem hiding this comment.
After doing more research, I'm 99% certain that is what is going on here. The file wide member group is being canceled as soon as Dozygen sees a nested group inside it. Then at the end of the nested group, Doxgen is placing all artifacts outside of the file-wide one. Essentially, Doxygen does not seem to maintain a stack for managing groups. It just enters the group it last sees and when it hits the end of a group it pops out to non-group namespace.
There was a problem hiding this comment.
I created #20537 to track this issue separately.
benpicco
added
the
CI: ready for build
| /** | ||
| * @name Override the SPI mode values | ||
| * | ||
| * As the mode is set in bit 3 and 2 of the configuration register, we put the | ||
| * correct configuration there | ||
| * @{ | ||
| */ | ||
| #define HAVE_SPI_MODE_T | ||
| typedef enum { | ||
| SPI_MODE_0 = 0, /**< CPOL=0, CPHA=0 */ | ||
| SPI_MODE_1 = STM32_SPI_CPHA_Msk, /**< CPOL=0, CPHA=1 */ | ||
| SPI_MODE_2 = STM32_SPI_CPOL_Msk, /**< CPOL=1, CPHA=0 */ | ||
| SPI_MODE_3 = STM32_SPI_CPOL_Msk | STM32_SPI_CPHA_Msk, /**< CPOL=1, CPHA=0 */ | ||
| } spi_mode_t; | ||
| /** @} */ | ||
|
|
There was a problem hiding this comment.
| /** | |
| * @name Override the SPI mode values | |
| * | |
| * As the mode is set in bit 3 and 2 of the configuration register, we put the | |
| * correct configuration there | |
| * @{ | |
| */ | |
| #define HAVE_SPI_MODE_T | |
| typedef enum { | |
| SPI_MODE_0 = 0, /**< CPOL=0, CPHA=0 */ | |
| SPI_MODE_1 = STM32_SPI_CPHA_Msk, /**< CPOL=0, CPHA=1 */ | |
| SPI_MODE_2 = STM32_SPI_CPOL_Msk, /**< CPOL=1, CPHA=0 */ | |
| SPI_MODE_3 = STM32_SPI_CPOL_Msk | STM32_SPI_CPHA_Msk, /**< CPOL=1, CPHA=0 */ | |
| } spi_mode_t; | |
| /** @} */ | |
| /** | |
| * @{ | |
| * @brief Override the SPI mode values | |
| * | |
| * As the mode is set in bit 3 and 2 of the configuration register, we put the | |
| * correct configuration there | |
| * | |
| */ | |
| #define HAVE_SPI_MODE_T | |
| typedef enum { | |
| SPI_MODE_0 = 0, /**< CPOL=0, CPHA=0 */ | |
| SPI_MODE_1 = STM32_SPI_CPHA_Msk, /**< CPOL=0, CPHA=1 */ | |
| SPI_MODE_2 = STM32_SPI_CPOL_Msk, /**< CPOL=1, CPHA=0 */ | |
| SPI_MODE_3 = STM32_SPI_CPOL_Msk | STM32_SPI_CPHA_Msk, /**< CPOL=1, CPHA=0 */ | |
| } spi_mode_t; | |
| /** @} */ |
the @brief is not needed (the first sentence is the brief-desciption if no @brief exists
Contribution description
This PR should not change what is built. It just moves some things out of shared headers into peripheral specific headers that already existed.
Testing procedure
I checked that it still builds by running
make -C ../../tests/periph/uart BOARD=nucleo-f767zi. I could not find a similar test for usbdev.Issues/PRs references
none known