This application provides an example of Azure RTOS USBX stack usage on NUCLEO-H723ZG board, it shows how to develop USB Device Human Interface "HID" mouse based application.
The application is designed to emulate a USB HID mouse device, the code provides all required device descriptors framework and associated class descriptor report to build a compliant USB HID mouse device.
At the beginning ThreadX calls the entry function tx_application_define(), at this stage, all USBx resources are initialized, the HID class driver is registered and the application creates 2 threads with the same priorities :
- app_ux_device_thread_entry (Prio : 10; PreemptionPrio : 10) used to initialize USB_OTG HAL PCD driver and start the device.
- usbx_hid_thread_entry (Prio : 20; PreemptionPrio : 20) used to send HID reports to move automatically the PC host machine cursor.
To customize the HID application by sending the mouse position step by step every 10ms:
- For each 10ms, the application calls the GetPointerData() API to update the mouse position (x, y) and send the report buffer through the ux_device_class_hid_event_set() API.
When plugged to PC host, the NUCLEO-H723ZG must be properly enumerated as a USB HID mouse device. During the enumeration phase, device provides host with the requested descriptors (device, configuration, string). Those descriptors are used by host driver to identify the device capabilities. Once the NUCLEO-H723ZG USB device successfully completed the enumeration phase, the device sends an HID report after a user button press. Each report sent should move the mouse cursor by one step on host side.
Host PC shows that USB device does not operate as designed (mouse enumeration failed, the mouse pointer doesn't move).
User is familiar with USB 2.0 "Universal Serial BUS" specification and HID class specification.
The remote wakeup feature is not yet implemented (used to bring the USB suspended bus back to the active condition).
- Some code parts can be executed in the ITCM-RAM (64 KB up to 256kB) which decreases critical task execution time, compared to code execution from Flash memory. This feature can be activated using '#pragma location = ".itcmram"' to be placed above function declaration, or using the toolchain GUI (file options) to execute a whole source file in the ITCM-RAM.
- If the application is using the DTCM/ITCM memories (@0x20000000/ 0x0000000: not cacheable and only accessible by the Cortex M7 and the MDMA), no need for cache maintenance when the Cortex M7 and the MDMA access these RAMs. If the application needs to use DMA (or other masters) based access or requires more RAM, then the user has to:
- Use a non TCM SRAM. (example : D1 AXI-SRAM @ 0x24000000).
- Add a cache maintenance mechanism to ensure the cache coherence between CPU and other masters (DMAs,DMA2D,LTDC,MDMA).
- The addresses and the size of cacheable buffers (shared between CPU and other masters) must be properly defined to be aligned to L1-CACHE line size (32 bytes).
- It is recommended to enable the cache and maintain its coherence:
- Depending on the use case it is also possible to configure the cache attributes using the MPU.
- Please refer to the AN4838 "Managing memory protection unit (MPU) in STM32 MCUs".
- Please refer to the AN4839 "Level 1 cache on STM32F7 Series and STM32H7 Series"
-
ThreadX uses the Systick as time base, thus it is mandatory that the HAL uses a separate time base through the TIM IPs.
-
ThreadX is configured with 100 ticks/sec by default, this should be taken into account when using delays or timeouts at application. It is always possible to reconfigure it, by updating the "TX_TIMER_TICKS_PER_SECOND" define in the "tx_user.h" file. The update should be reflected in "tx_initialize_low_level.S" file too.
-
ThreadX is disabling all interrupts during kernel start-up to avoid any unexpected behavior, therefore all system related calls (HAL, BSP) should be done either at the beginning of the application or inside the thread entry functions.
-
ThreadX offers the "tx_application_define()" function, that is automatically called by the tx_kernel_enter() API. It is highly recommended to use it to create all applications ThreadX related resources (threads, semaphores, memory pools...) but it should not in any way contain a system API call (HAL or BSP).
-
Using dynamic memory allocation requires to apply some changes to the linker file. ThreadX needs to pass a pointer to the first free memory location in RAM to the tx_application_define() function, using the "first_unused_memory" argument. This requires changes in the linker files to expose this memory location.
- For EWARM add the following section into the .icf file:
place in RAM_region { last section FREE_MEM };
- For MDK-ARM:
either define the RW_IRAM1 region in the ".sct" file or modify the line below in "tx_initialize_low_level.S to match the memory region being used LDR r1, =|Image$$RW_IRAM1$$ZI$$Limit|
- For STM32CubeIDE add the following section into the .ld file:
._threadx_heap : { . = ALIGN(8); __RAM_segment_used_end__ = .; . = . + 64K; . = ALIGN(8); } >RAM_D1 AT> RAM_D1
The simplest way to provide memory for ThreadX is to define a new section, see ._threadx_heap above. In the example above the ThreadX heap size is set to 64KBytes. The ._threadx_heap must be located between the .bss and the ._user_heap_stack sections in the linker script. Caution: Make sure that ThreadX does not need more than the provided heap memory (64KBytes in this example). Read more in STM32CubeIDE User Guide, chapter: "Linker script".
- The "tx_initialize_low_level.S" should be also modified to enable the "USE_DYNAMIC_MEMORY_ALLOCATION" flag.
- The DTCM (0x20000000) memory region should not be used by application in case USB DMA is enabled
- Should make sure to configure the USB pool memory region with attribute "Non-Cacheable" to ensure coherency between CPU and USB DMA.
RTOS, ThreadX, USBXDevice, USB_OTG, Full Speed, HID, Mouse.
- This example runs on STM32H723xx devices.
- This example has been tested with STMicroelectronics NUCLEO-H723ZG boards revision MB1364-H723ZG-E01 and can be easily tailored to any other supported device and development board.
In order to make the program work, you must do the following :
- Open your preferred toolchain
- Rebuild all files and load your image into target memory
- Run the application