STM32CubeMX (MCUs)

Dear @KDJEM.1 ,I have seen that in STM32CubeIDE Version: 1.16.0 this issue is solved if the "Generate peripheral inizialization as a pair of '.c/.h' files per peripheral" is NOT selected in the ".ioc". Otherwise, the problem keeps there and "HAL_PWREx_EnableVddUSB()" is not called.Here, you can see the code if the option is NOT selected:    void HAL_PCD_MspInit(PCD_HandleTypeDef* hpcd) { RCC_PeriphCLKInitTypeDef PeriphClkInitStruct = {0}; if(hpcd->Instance==USB_DRD_FS) { /* USER CODE BEGIN USB_DRD_FS_MspInit 0 */ /* USER CODE END USB_DRD_FS_MspInit 0 */ /** Initializes the peripherals clock */ PeriphClkInitStruct.PeriphClockSelection = RCC_PERIPHCLK_USB; PeriphClkInitStruct.UsbClockSelection = RCC_USBCLKSOURCE_HSI48; if (HAL_RCCEx_PeriphCLKConfig(&PeriphClkInitStruct) != HAL_OK) { Error_Handler(); } /* Enable VDDUSB */ HAL_PWREx_EnableVddUSB(); /* Peripheral clock enable */ __HAL_RCC_USB_CLK_ENABLE(); /* USB_DRD_FS interrupt Init */ HAL_

Using "Firmware Relative Path" and generate code with references f STM32CubeIDE will mix relative and absolute paths in generated project files STM32CubeMX 6.14.0 on Windows 11.1. Create a directory and unzip en.stm32cubeh5-v1-5-0.zip there 1. Create a minimalistic new project in STM32CubeMX for STM32H5.2. In tab Project Manager, tab Project:2.1 Set Project Location to the created directory and set the name of your project2.2 Set Toolchain / IDE to STM32CubeIDE2.3 Uncheck Use Default Firmware Location2.4 Use the browse button and select STM32Cube_FW_H5_V1.5.0 in the directory.3. In tab Project Manager, tab Code generator: choose option Add necessary library files as reference in the toolchain project configuration file4. Save the project and generate the code.Note that the path is correctly stored as a relative path in the .ioc-file with the key ProjectManager.CustomerFirmwarePackageBugsDespite that the actual relative path from Toolchain Folder Location i

I tried to use cmake with CubeMx but I cant make RTT works.Error:multiple definition of `_write_r'; /usr/lib/gcc/arm-none-eabi/13.2.0/../../../../arm-none-eabi/lib/thumb/v7-m/nofp/libc_nano.a(libc_a-writer.o):/build/arm-none-eabi-newlib/src/build-nano/arm-none-eabi/thumb/v7-m/nofp/newlib/../../../../../../newlib-4.3.0.20230120/newlib/libc/reent/writer.c:45: first defined hereRTT needs to define _write_r but seems already define.If I comment this line from the gcc tool chain it solves the problem:set(CMAKE_C_LINK_FLAGS "${CMAKE_C_LINK_FLAGS} -Wl,--start-group -lc -lm -Wl,--end-group")Before I used my custom cmake this way:```target_link_options(${EXECUTABLE} PRIVATE-T${MCU_LINKER_SCRIPT}${CPU_PARAMETERS}-Wl,-Map=${CMAKE_PROJECT_NAME}.map--specs=nosys.specs-Wl,--start-group-lc-lm-lstdc++-Wl,--end-group-Wl,--print-memory-usage)```and have no problem so I don't really understand why.Also tried -fcommon it does nothing.

Hi everyone,I noticed that CubeMX 6.14.0 ignores the octospi parameter settings. I created a new project for the NucleoU5A5 board, configured Octospi1 with a clock prescaler of 99. I notice the following behaviorGenerated code (CubeIDE) has a prescaler value of 1Next time I open the IOC file, the Octospi prescaler is back to 1I didn't check other settings. I have attached my ioc file Kind regards,Timon Zijnge 

MCU:STM32U585xxEnv:ThreadX + USBX + FileXI want to use USBX add my own class type, but I'm already have 3 classes, HID, MSC and CDC ACM, it means 4 classes after added my own.In file (ux_device_descriptors.h) have a define (USBD_MAX_SUPPORTED_CLASS  3U), this value determined the (tclasslist) array size, also the framework max class, can not be changed by user define or STM32CubeMX(no option find).If I change this define, it will be overwrite by STM32CubeMX when generated code. How can I do this correctly?

In many cases, especially for middleware, it may be necessary to extend the code generated by CubeMX.Since not all files have sections that are untouched by the code generator or some places inside a file are missing those user code sections, I'm often in the need of a feature that would allow to add my own sections by adding a comment such as /* USER CODE BEGIN */ /* USER CODE END */ to my source file. Unfortunately, those sections are overwritten as soon as the code generator gets triggered again.Would it be possible to add this feature?How does the code generator currently handle those sections anyway if not looking for the specific comments (since adding my own user code comments does not work)?

The GPIO clock is the only code that seemed to be generated under the initialization function. the other necessary code such as the following was not created. GPIO_InitTypeDef GPIO_InitStruct; GPIO_InitStruct.Pin = GPIO_PIN_8; GPIO_InitStruct.Mode = GPIO_MODE_OUTPUT_PP; GPIO_InitStruct.Pull = GPIO_NOPULL; GPIO_InitStruct.Speed = GPIO_SPEED_FREQ_LOW; HAL_GPIO_Init(GPIOC, &GPIO_InitStruct);In the setup GUI I had selected the pin PC8, and set it to a GPIO_Output yet the above code was not generated.

Hello everyone, I'm currently using the STM32H745BIT6 microcontroller and I'm having difficulty configuring the system so that the M7 core can use internal memory (Internal RAM/Flash) to store data from TouchGFX. I would like to avoid using any external memory (external RAM or Flash). I'm not sure how to properly configure this. I would greatly appreciate any support or guidance. Thank you ! 

Hi,A small but annoying bug with STM32CubeMX 6.14.0 with H7 package 1.12.1.  Every time I change anything in the LWIP general settings (e.g. DHCP enable/disable), the PHY setting changes by itself from USER_PHY to undefined. This is a nuisance as if you generate code without realizing, some files are deleted meaning you lose user defined code blocks.Example file attached.On a side note, ST please add a KSZ8081 PHY option.  Seems to be commonly asked for.  The code is 95% the same as the LAN8742 option.  I have attached a KSZ8081 example to show how similar it is.  Or at least abstract the PHY function call names so that just the part number specific function definitions need to be changed (as is done in the EthPHY.h example attached.The USER_DEFINED code template that is generated is not much use.  It does not allow returning error for init failure.  It also does not have access to the netif or PHY register read/write functions.Thanks

Hi,    I do not use PD8/PD9/PB0 pins in my design. Therefore I did not configure them in CubeMax.    Why NECLEO-H753ZI CUBEMX Pinout View PD8/PD9/PB0 pins are in RED? Thanks, Peter    

I am trying to get the USART2 working on my Nucleo-H7 board. I am using HAL_UART_Transmit for the TX side. This works fine. But when I use HAL_UART_Receive on the RX side, there is never any RX data coming into the device. I have stepped through the HAL lower layers and watched the SFRs for any activity, but there is never any activity on the RX side (the RXNE bit never toggles). I'm starting to wonder if there is a hardware issue with the Nucleo board.My project is an enhancement of the NX-TXP_Echo_Server example which receives 64 bytes received from the ethernet interface, and forwards the 64 byte command to the USART device to be used by another board. The remote board basically echoes the serial data received from the Nucleo board. Again the TX function of the Nucleo board works fine. But the RX side never sees any data received by the chip. I have a scope on the TX and RX lines and can verify that the remote board is indeed receiving the data sent to if from the Nucleo, and it ind

Hi,just made an update of CubeMX from 6.11.1 to 6.14.The problem I have is that our comments within user blocks were ISO-8859-1 encoded and this never was a problem until now and with 6.14, when we generate code, all Umlauts are replaced with the UTF-8 sequence EF BF BD.Is there an option to tell CubeMX to use ISO-8859-1 for reading and encoding?

I'm trying to configure SAI2 as a I2S master.I2S_CKIN has been enabledBut the configuration cannot be loaded when I changed SAI A Mode to master.Now I got a page like this.Is there anything I can do to fix it?

I've been working with USBX for a while now, and after updating to the latest X-CUBE-AZRTOS-H7 package (v3.2.0), I'm running into a couple of problems relating to the templates STM32CubeMX generates. Specifically, the files ux_device_descriptors.c and ux_device_descriptors.h1). In the function USBD_Get_String_Framework, the variable "count" is declared as uint8_t. This is a problem if the string framework is more than 256 bytes in length. This definition is outside of the user editable code, so I need to remember to edit it every time I regenerate my code. This should be set to uint16_t or unsigned.2). Also in USBD_Get_String_Framework, the USER CODE tags are missing BEGIN and END. As a result, any code I put in between them is deleted when regenerating code.3). In the header file,  the "USER CODE BEGIN/END Private_defines" tags appear twice, which results in any code in one of the sections begin duplicated when regenerating code.These are not a big deal, but they are a

Hi all,once again I stumble across CubeMX being so inflexible in its code generation.I use CubeMX to generate initialization code but I'm also using C++ and that's where CubeMX has major downsides.For example it is still not possible to tell CubeMX to generate the MX_XXX_Init function calls anywhere but in main.Further more an option to add a simple pointer to all HandleTypeDefs which is free to use would help a lot.From my experience STM HAL drivers are actually working pretty much ok but sometimes the are based on assumptions that a either flawed or don't apply in the project context.Hence you have to write your own driver.Just a simple pointer so I can pass additional information together with the handle struct would make life much easier especially when trying to implement a driver as C++ class.To any ST employee reading this:Are there any plans to make CubeMX more flexible?To all users of CubeMX:What would you really like ST to change in CubeMX?BR

I used to have the CubeMX generated code as a library in my main project: /BSP/CubeMX/ contains the .ioc and generated code.I could include the CubeMX 6.13 (and earlier I think) generated CmakeLists.txt and have all generated code compiled as a static library:project(BSP) add_library(${PROJECT_NAME} STATIC) add_subdirectory(CubeMX/cmake/stm32cubemx SYSTEM) target_link_libraries(${PROJECT_NAME} PRIVATE stm32cubemx) # Suppress compile warnings in 3rd-party C code target_compile_options(${PROJECT_NAME} PRIVATE -Wno-unused-parameter #-Wno-conversion) #expose header files from stm32cubemx to public target_include_directories(${PROJECT_NAME} SYSTEM PUBLIC "$<TARGET_PROPERTY:stm32cubemx,INTERFACE_INCLUDE_DIRECTORIES>")I needed some workarounds for getting everything to work, but I had a environment where I could re-generate CubeMX code and the whole project keeps on compiling without the need to fix something manually.Now the CubeMX generated CMakeLists.txt changed its content and I'm

Hi. Newby question about my STM32CubeMX (6.14.0) installation on Ubuntu. With the initial 'update', why do I get so many...:2025-03-31 14:17:40,313 [INFO] WebApp:194 - Connection lost2025-03-31 14:17:40,865 [INFO] WebApp:191 - Connection restablished...messages on the console? Is this normal? Its being going on for hours now !!!!(I have a good fiber connection to the internet)Thanks for your timeLeonBFollowed step described in the readme as close as possible:On Linux®:On STM32CubeMX-Lin line, click "Get software" to download the package.Extract (unzip) the downloaded package.Make sure you have administrator rights to access the target installation directory. You can run the installation as root (or sudo) to install STM32CubeMX in shared directories.Do chmod 777 SetupSTM32CubeMX-6.14.0 to change the properties, so that the file is executable.Double-click on the SetupSTM32CubeMX-6.14.0 file, or launch it from the console window.I did not do the last step but instead started the app

Originally a reply here, but seems to be a new question.Thank you, i worked in "Clock Configuration", but RTC is locked . It means for this MCU works only with internal clock for RTC?

I am working with the STM32F407ZGT6 microcontroller and have encountered a warning related to Ethernet and USB functionalities being assigned to the same pins. While I have attempted to resolve this by using alternative pin mappings, the conflict warning still persists. I would like to clarify the following: Understanding the Warning – Despite assigning Ethernet and USB to different available pins, the warning remains. Could you provide insights into why this occurs and whether it impacts functionality? Resolution Approach – Are there specific configurations, software settings, or modifications required to eliminate this conflict? Technical References – If there are any relevant technical documents or application notes addressing this issue, I would appreciate your guidance. Any insights or suggestions would be greatly appreciated! Thanks in advance! Best regards,Ishika Sapkal

Hello,Is it possible for ST to re-order the startup code in ST's implementation of AzRtos (CubeMX)?.  I think it would make everyone's life a lot easier if all the AzRtos services (ThreadX, FileX, NetX, USBX) start before the user apps/threads.Currently, in the file: app_azure_rtos.c, the process is:create TX memory poolStart User threads (app_threadx.c)create FX memory poolStart MX_FileX_Init()create NetX memory poolStart MX_NetXDuo_Init()create USBX memory poolStart MX_USBX_Host_Init()Start MX_USBX_Device Init()The problem is that user threads are started before the necessary system services are available.I think it was done this way to simplify how CubeMX generates the code (where the user threads are grouped with ThreadX).  In my opinion it makes more sense to move "step 2" after the USB services are up and running.  In my opinion it should be as follows:create TX memory poolcreate FX memory poolStart MX_FileX_Init()create NetX memory poolStart MX_NetXDuo_Init()creat

In MX can be choiced wrap or incr? No info . Primary i ask for usage with internal SRAM5 in U5x.Same for other optimal GPU2D setup. ST have many time to correct this.

I am working on X-CUBE-AZURE in B-U585I-IOT02A board. My IDE version is 1.18 and i am using Fedora 41. When i enable FDCAN peripheral in CubeMX it places the stm32u5xx_hal_fdcan.c and .h file into the X-CUBE-AZURE/Drivers/STM32U5xx_HAL_Driver/Src. Even though it places the correct .c and .h files the build fails when i enable FDCAN from CubeMX. Here is the build error message./home/erhangok/Documents/GitHub/IOT_Node_Azure_v01fw/Projects/B-U585I-IOT02A/Applications/NetXDuo/Nx_Azure_IoT/Core/Src/main.c:357:(.text.MX_FDCAN1_Init+0x72): undefined reference to `HAL_FDCAN_Init' collect2: error: ld returned 1 exit status make: *** [makefile:104: Nx_Azure_IoT.elf] Error 1 "make -j8 all" terminated with exit code 2. Build might be incomplete. 15:41:47 Build Failed. 2 errors, 478 warnings. (took 19s.801ms)Even though the file exists in Drivers/STM32U5xx_HAL_Driver it still cannot find it. When i look further into this i saw that subdir.mk file does not contain this line: "./Drivers/STM32U5xx_HA

Hi When I try to set DSI clock more than 500M, the STM32cubemx will display an error message "dsi clock source frequency must be =< 62MHz" as below. But it said STM32H747 DSI clcok can reach 1Ghz in datasheet.I use 25Mhz HSE and attach STM32cubemx project for reference. 

Hello!The code generated for lwIP and Ethernet (Internal MAC) has three threads for working of lwIP and the Ethernet driver.tcip_thread: main thread used to execute the lwIP stack all TCP/IP processing is done by this.EthIf: Process the RX interrupt on Ethernet peripheral, when new data packet is received passes it to lwip.EthLink: constanly monitors the Phy chip for any connection/disconnection and auto-negotiation events.I am using v 1.17.0 of the IDE (latest at the time of writting)The problem is that the default value of STACK size of EthIf (Rx Handler Thread) is very small: Only 350 bytes. which leads to stackoverflow. I have manually changed the value from code. But I cannot find any GUI to change this! Only we are allowed to configure the tcip_thread but their is NO option for last two threads.  Further there is only a single constant that is used for initialization of both the threads (EthIf and EthLink)#define INTERFACE_THREAD_STACK_SIZE ( 1024 )and this same constant is

We are using this the X-CUBE-SMBUS middleware for a simple SMBus like device, and noticed that the generated configuration file sets the wrong define for the "Condensed command table" setting.The generated config file sets the DENSE_CMD_TABLE define, but the library expects DENSE_CMD_TBLX-CUBE-SMBUS/2.1.0/STM32CubeMX/templates/stm32_config_stack_h.ftl:[#if SMB_densetable_value == "1"]#define DENSE_CMD_TABLE[/#ifshould read[#if SMB_densetable_value == "1"]#define DENSE_CMD_TBL[/#ifThe error is noticeable if the PMBUS version is set to 1.1 or 1.2. Setting PMBUS version to 1.3 (default) hides the error as  DENSE_CMD_TBL is a required feature for PMBUS 1.3 and is automatically enabled regardless of what the CubeMX config says.In our case we are using the library for a minimal SMBus like device, not using any PMBus features. So PMBUS feature level is set to 1.1 and most features are disabled.