STM32CubeMX errata 6.17.x

1. Main fixed issues for 6.17.0

ID	Summary
225210	After code generation, the DMA streams are linked to incorrect DMA handles. This issue occurs when using STM32CubeMX version 6.14.0 to set up a new NUCLEO-H723ZG project, configure the CORDIC with both write and read DMA streams enabled, and generate the associated code. STM32CubeMX version 6.17.0 provides a correction for this issue.
224930	When configuring an STM32CubeMX project with the STM32N645X0H3Q microcontroller or the NUCLEO-N657X0-Q board, it is not possible to assign any GPDMA1 or HPDMA1 channel to the I3C2_TX or I3C2_RX DMA requests. STM32CubeMX version 6.17.0 addresses this issue.
225670	When creating a new project for the STM32N6570‑DK board, enable the RAMCFG peripheral for the application context. Select AXISRAM from 3 to 6 in the mode section. Define a single continuous memory region for these SRAMs in the memory mapping tool (MMT). After this configuration, the generated .ioc file cannot be reopened in STM32CubeMX. When attempting to load the project, STM32CubeMX displays an internal error in a popup message. This issue prevents the .ioc file from being loaded after the previous RAM configuration. STM32CubeMX version 6.17.0 provides a correction for this issue.
222109	For STM32CubeMX projects that use STM32U575 products, when configuring multiple MDF filters—each assigned to its own DMA channel—two issues have been observed: The initialization sequence for all MDF filters executes twice. All MDF filters are linked to the same DMA channel in the generated code, resulting in incorrect DMA channel assignment. These issues cause runtime errors because the application attempts to start the same DMA channel more than once. STM32CubeMX version 6.17.0 resolves these issues.
227138	In STM32CubeMX version 6.16.0, projects generated by using Start from Board for the NUCLEO-H563 board produce the same compiler warning four times when built with IAR Embedded Workbench (EWARM). The warning concerns an incompatible redefinition of the macro USE_NUCLEO_144, which is defined both in the compiler command-line options and in the generated header file stm32h5xx_nucleo_conf.h. STM32CubeMX version 6.17.0 integrates a solution for this problem.

2. Known limitations

• The user manual embedded in STM32CubeMX 6.17.0 is provided for convenience only and is a work in progress. It does not reflect finalized documentation. Certain sections may change in the next revision.
  • The final validated revision of the user manual will be published soon on st.com and will be the unique reference for documentation.

• When integrating software packages, please note that the current configuration only allows the selection of one major component at a time. Attempting to select packages like X-CUBE-TOUCHGFX and X-CUBE-TCPP together will be prevented by the package selector.
• An issue can occur with MDK-ARM when the assembly file settings are modified from the "Option for Target" panel. More precisely, when a user adds include paths or defines from that panel, the settings are removed.
  • As a temporary workaround, users must manually add include paths or defines for assembly files by following these steps:

1. Right-click on the assembly file and select "Options for File'startup_stm32XXXxx.s"
2. Navigate to the Asm tab and add the necessary include paths and/or defines

• Generating an STM32 project with a minimum version of the IAR Embedded Workbench® toolchain creates a permanent association. The user cannot change this minimum version later or regenerate the project.
• For STM32 projects that use ThreadX with the starm-clang compiler, users may encounter build issues when using Expansion Packages that provide X-CUBE-AZRTOS for the following products: STM32G0, STM32G4, STM32H7, STM32H7RS, STM32L4, STM32L5, STM32WB, and STM32WL. To ensure ThreadX works correctly with the starm-clang compiler, proceed as follows:

1. Copy “Middlewares/ST/threadx/ports/cortex_m0/ac6” folder from the STM32Cube firmware to the project
2. Update the CMakeLists.txt file under “cmake/stm32cubemx/” by replacing “/gnu/” with “/ac6/”
3. Update the tx_initialize_low_level.S file under "Core/Src/"as illustrated below:

#if defined  (__clang__)                            to be replaced by ==> #if defined (__ARMCC_VERSION)

#if (defined(__GNUC__) && !defined(__clang__))      to be replaced by ==> #if (defined(__GNUC__) && !defined(__ARMCC_VERSION))

Only for STM32G0 users must change the following instruction in the tx_initialize_low_level.S

.text 32                  to be replaced by ==> .text

• For STM32 products that natively support the ThreadX feature, building projects with ThreadX enabled in low-power support mode by using Makefile or CMake toolchains results in warnings.
• The compare Projects tool cannot compare projects that have the direct memory access (DMA) feature enabled.
• Starting from Ubuntu version 23.10, when the user selects Tools > MCU commercial part number to configure Tools > MCU selector, the CAD viewer does not function.
• If you get a file access error during compilation in a partner IDE (for example, “cannot open source input file”), try moving the project to a shorter path on your disk. You can either:
  • Move the entire STM32 package manually, or
  • Move only the example by using the Example Selector in STM32CubeMX and placing it closer to the disk root (to avoid long path issues).
• If the “CAD resources” feature does not work correctly when “Use System Proxy parameters” is selected, try switching to “Manual Configuration of Proxy Server” or “No proxy” instead. You can change this in: [Updater Settings…] > [Connection Parameters].
• When unzipping a package on a macOS^® system, use an official macOS^® application such as the default graphical archive application, or the ditto tool in the command line:
  $ditto -x -k setupArchive.zip
  When a Linux^®-like application is used (such as unzip), the expanded files lose the signature and are not recognized as signed.
• For Keil® MDK users, especially those working with STM32CubeMX-generated projects, note that starting from Keil® MDK v5.37, Arm Compiler 6 (AC6) is now the default toolchain and Arm Compiler 5 (AC5) is deprecated. This change can cause compatibility issues for some STM32 series, including:
  • STM32F x
  • STM32L (0x, 1x, 4x)
  • STM32H7 (2x, 3x, 4x, 5x, Ax, Bx)
  • STM32G x
  • STM32WB (1x, 3x, 5x)
  • STM32MP15x With these products, you may see compilation errors when using AC6, including messages indicating that the required Arm Compiler is not available. Workaround:
  • To avoid issues with these STM32 projects, you can switch back and continue using Arm Compiler 5 by following the steps below:

1. Download Arm Compiler 5 manually from https://developer.arm.com/documentation/ka005198/latest/.
2. Follow the step-by-step instructions available in this guide [1] to add Arm Compiler 5 to Keil® MDK 5.37 or newer.
3. Once installed, select AC5 as the Arm Compiler of your project settings.

---

Note: a. Arm and TrustZone are registered trademarks of Arm Limited (or its subsidiaries) in the US and/or elsewhere.

2.1. When selecting microcontrollers in the STM32C0 series

• When generating code for the STM32C071K8U3 microcontroller using EWARM toolchain, the linker file incorrectly defines the __ICFEDIT_region_ROM_end__ symbol as 0x0801FFFF. According to the device's datasheet (DS14693 Rev 1:7),the STM32C071K8U3 has 64 KBytes of Flash memory,So that the user should set the ROM end address to 0x0800FFFF instead of 0x0801FFFF.

/*-Memory Regions-*/
define symbol __ICFEDIT_region_ROM_start__ = 0x08000000;
define symbol __ICFEDIT_region_ROM_end__   = 0x0800FFFF;

2.2. When selecting microcontrollers in the STM32G0 series

• Regarding the cohabitation between USB and USB Power Delivery, the VID and PID of USB Power Delivery and USB are not automatically aligned. VID/PID values must be aligned manually.
• When both the USB Power Delivery and FreeRTOS™ CMSIS_V1 middleware are enabled at the same time, the FreeRTOS™ eTaskGetState parameter must be enabled in the include parameters tab.

2.3. When selecting microcontrollers in the STM32G4 series

• When attempting to create a new STM32G411 project using STM32CubeMX with the firmware "STM32Cube_FW_G4_V1.6.0", the project creation fails. Instead of generating a correctly structured project, the process terminates unexpectedly.

2.4. When selecting microcontrollers in the STM32H5 series

• When creating a new STM32H5 project, enabling the JPEG peripheral and generating the code, the Utilities/JPEG folder containing the JPEG source files is not generated.
  • To avoid this issue the user should add the missing folder manually (copy that folder from the firmware under this path Users\<User_name>\STM32Cube\Repository\STM32Cube_FW_H5_Vx.y.z\Utilities) each code generation.
• The following configuration will be lost after migrating to STM32CubeMX version 6.15.0: An STM32H573ZITx project created with STM32CubeMX version 6.14.1, where the OctoSPI1 peripheral is enabled, the boot path is configured, and an MMT user application region is defined within the OctoSPI1-controlled memory range (0x90000000 to 0xA0000000). Additionally, the Application Regions Settings are applied to both the peripheral configuration and linker files.
  • Upon migration, a popup notification will inform the user about the MMT user application region lose. The Application Regions Settings will be disabled for both Peripherals and linker files Furthermore, the hardware memory range managed by OctoSPI1 will be split into two sub-regions:
    • The first region, labeled “External memory (OCTOSPI1)”, with a size of 2 bytes.
    • The second region, labeled “OCTOSPI1”, with a size of less than 256 Mbytes.

To avoid the issue the user is recommended to follow these steps:

1. Increase the Memory Size parameter under Octospi1 Configuration tab to 2 Gbits (256 M Byte).
2. Recreate manually the MMT user application region that was allocated to the memory range controlled by Octospi1
3. Turn on the “Apply Application Regions Settings to Peripherals” and “Apply Application Regions Settings to Linker Files” toggle buttons in the MMT panel.

• For STM32CubeMX 6.14.0 OEMiROT projects, the execution of provisionning.bat script might fail due to wrong file path settings.
  • To fix this failure:  the  user should  either keep the default boot path configuration in stm32CubeMX, or if it's necessary, edit env.bat and set oemirot_boot_path_project to %~dp0..\.
• When migrating projects with OEM-iROT boot path from STM32CubeMX V6.8.x to STM32CubeMX V6.10.x or later releases, the user might need to reconfigure ROT_Provisioning.
• The makefile generation no longer works as soon as the project structure is changed from "Secure Project only" or "Non Secure Project only" to "Secure Project and Non Secure Project enabled (dual project)".
• In Microsoft^® Azure^® NetX Duo, some configuration flags, which are not used in the STM32Cube firmware package examples, might be nonfunctional when combined.

2.5. When selecting microcontrollers in the STM32H7 series

• The function HAL_HSEM_IRQHandler() is wrongly called in the HSEMx_IRQHandler() scope after performing the following scenario on STM32CubeMX 6.15.0 with the STM32H745ZIT3 product:

1. Enable the HSEM1 global interrupt on the context Cortex‑M7 and the HSEM2 global interrupt on the context Cortex‑M4
2. Select LL instead of HAL in Project Manager → Advanced Settings and choose STM32CubeIDE as toolchain
3. Generate the code.

To prevent the HAL_HSEM_IRQHandler() function from being generated the user should, navigate to the NVIC IP, select the Code Generation tab, and disable the Call HAL handler option.

• When creating a new STM32H7RS project, enabling the JPEG peripheral and generating the code, the Utilities/JPEG folder containing the JPEG source files is not generated.
  • To avoid this issue the user should add the missing folder manually (copy that folder from the firmware under this path Users\<User_name>\STM32Cube\Repository\STM32Cube_FW_H7RS_Vx.y.z\Utilities) each code generation.
• A linker File mismatch after project regeneration: When regenerating an STM32CubeIDE project for the STM32H7S78-DK board using the X-CUBE-AZRTOS-H7RS pack, the system incorrectly selects STM32H7S7L8HXH_default.ld as the linker file. The expected linker file is STM32H7S7L8HXH_ROMxspi2.ld. This issue occurs after any STM32CubeIDE project has been removed and subsequently regenerated.
• For STM32H7RS OEMiROT projects in STM32CubeMX 6.14.0, the execution of provisionning.bat script might fail due to wrong file path settings.
  • To fix this failure:  the user should either keep the default boot path configuration in stm32CubeMX, or if it's necessary, edit env.bat and set oemirot_boot_path_project to %~dp0..\.
• For the projects using STM32H7B product lines, when activating the SAI feature the parameters are not displayed in the Configuration tab.
• During the migration from an older STM32H7 project (created with STM32CubeMX version coming earlier than V6.13.0) to the STM32CubeMX V6.13.0, The following files aren't automatically updated:

system_stm32h7xx.c and startup_stm32h7nnxx.s which are needed for the correct initialization of the system power supply.

• This limitation might appear: An early RAM access prior to the configuration of the system power supply which is not allowed.
  • For new projects created by STM32CubeMX v6.13.0 and later, those files are correctly updated.
  • The system power supply configuration is also kept in SystemClock_Config() to avoid runtime issue with existing projects.
  • The workaround :To successfully migrate the project:The user should Copy the new files from the firmware side and ensure that they are correctly included in the project. In addition he should verify the compiler compatibility: he should use the appropriate startup_stm32h7nnxx.s file for the chosen compiler.
• For the STM32H7Sx microcontrollers in the STM32H7 series, when using the NUCLEO-H7S3L8 board, the boot path is not supported. It is only supported for STM32H7S78-DK board or custom boards with the same flash and same pin mapping.
• For the STM32H7Rx microcontrollers in the STM32H7 series, the boot path is not supported and Debug authentication is disabled.
• MDMA and HSEM are not supported in LL. Only HAL is supported.
• When activating any mode under SPDIFRX:
  • The alternate function configuration code is not generated in the HAL_SPDIFRX_MspInit function in the stm32h7xx_hal_msp.c file.
  • Add the workaround code in the user section after “GPIO_InitStruct.Speed =GPIO_SPEED_FREQ_LOW;” as described in Table 2 depending on the channel and selected pin.

Table 2. Workaround for alternate function configuration code

Channel	Selected pin	Workaround code
IN0	PD7	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF9_SPDIF; /* USER CODE END alternate function configuration */
	PG11	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF8_SPDIF; /* USER CODE END alternate function configuration */
IN1	PD8	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF9_SPDIF; /* USER CODE END alternate function configuration */
	PG12	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF8_SPDIF; /* USER CODE END alternate function configuration */
IN2	PC4	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF9_SPDIF; /* USER CODE END alternate function configuration */
	PG8	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF8_SPDIF; /* USER CODE END alternate function configuration */
IN3	PC5	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF9_SPDIF; /* USER CODE END alternate function configuration */
	PG9	/* USER CODE BEGIN alternate function configuration */ GPIO_InitStruct.Alternate = GPIO_AF8_SPDIF; /* USER CODE END alternate function configuration */

2.6. When selecting dual-core microcontrollers in the STM32H7 series

• Only Boot0 is supported: both cores boot at once.
• Import from and to dual-core STM32H7 is not supported.
• For memory-to-memory DMA or BDMA or MDMA configuration, the initialization code is generated for both cores.
• OpenAMP issue when compiling under EWARM and MDK-ARM when OpenAMP is activated under STM32CubeMX:
  • To avoid link errors in OpenAMP when compiling in EWARM IDE, add the next four lines of code in the linker files (.icf):

/* Create region for OPENAMP */
define symbol __OPENAMP_region_start__ = 0x38000400;
define symbol __OPENAMP_region_size__ = 0xFC00;
export symbol __OPENAMP_region_start__;
export symbol __OPENAMP_region_size__;

• To avoid link errors in OpenAMP when compiling in MDK-ARM IDE, add the next four lines of code in the linker files (.sct):

; ***** Create region for OPENAMP *****
.resource_table +0 ALIGN 4 { ; resource table
*(.resource_table)
}
; Shared Memory area used by OpenAMP
__OpenAMP_SHMEM__ 0x38000400 EMPTY 0xFC00 {}

• OpenAMP under STM32CubeIDE needs linker file update:
  • The following lines must be added under the .ld file:

/* Specify the memory areas */
MEMORY
{
...
m_ipc_shm (RW) : ORIGIN = 0x38000400, LENGTH = 63K
}
/* Symbols needed for OpenAMP to enable rpmsg */
__OPENAMP_region_start__=ORIGIN(m_ipc_shm);
__OPENAMP_region_end__=ORIGIN(m_ipc_shm)+LENGTH(m_ipc_shm);

2.7. When selecting microcontrollers in the STM32L5 series

• Arm^® TrustZone^® support:
  • Fixed default SAU is configured in the application partition_stm32l552xx.h and partition_stm32l562xx.h files.
  • Import does not work for TrustZone^® projects (TZEN=1).
• To facilitate the peripheral privilege configuration, from STM32CubeMX V6.2.0, the GTZC peripheral is split into two domains: secure and nonsecure. The side effect is that, in the case of project migration from a release version before V6.2.0, if GTZC is configured in both the secure and nonsecure domains with previously configured privilege peripherals on a nonsecure peripheral, this configuration is lost during the migration. The GTZC_NS privileges must be configured again.
• Regarding the cohabitation between USB and USB Power Delivery, the VID and PID of USB Power Delivery and USB are not automatically aligned. VID/PID values must be aligned manually.

2.8. When selecting microcontrollers in the STM32N6 series

• For STM32N6 projects using the VENC feature, video can be compressed before sending it over UVC either by using the 'VENC (Video Encoder)' option or the 'standalone JPEG peripheral' option. Since the VENC middleware stack (VideoEncoder/common, EWL, H264, JPEG under Middlewares/Third_Party/VideoEncoder/...) is not supported by STM32CubeMX and therefore not generated, customers who want to use VENC (H264/JPEG) must manually copy this middleware from the VENC from the firmware to their projects.
• When creating a new STM32N6 project, enabling the JPEG peripheral and generating the code, the Utilities/JPEG folder containing the JPEG source files is not generated.
  • To avoid this issue the user should add the missing folder manually (copy that folder from the firmware under this path Users\<User_name>\STM32Cube\Repository\STM32Cube_FW_N6_V1.3.0\Utilities) each code generation.
• During migration from an older STM32N6 project (created with STM32CubeMX versions earlier than V6.16.0) to STM32CubeMX V6.16.0, the system_stm32n6xx_fsbl.c file will not update automatically.
  • To update it, the user needs to replace the system_stm32n6xx_fsbl.c file in the project with the latest version found in the STM32CubeN6 firmware package under \Drivers\CMSIS\Device\ST\STM32N6xx\Source\Templates
• An issue may occur when building an STM32N6 project in STM32CubeIDE where the AppS context is added incrementally. Specifically, if the project is initially created with components enabled only in the EXTMEM and FSBL contexts—but not in AppS—and later these components are activated in the AppS context by modifying the *.ioc file and regenerating the code, the AppS project fails to build due to missing files in the STM32CubeIDE\Appli\Application\User\Core directory.
• When using USBPD with STM32N6 projects,and generating code with MDK-ARM toolchain, compilation error occurs due to inconsistent compiler flags in the settings, For example the library USBPDCORE_NOPD_CM55_Keil is compiled with specific compiler flags that set the wchar_t size or enum packing,while other libraries and files in the project are compiled with different flags.
  • To fix this error the compiler option Short enums/wcharof the MDK-ARM project under Options for Target-> C/C++ (AC6) should be unchecked.
• For the full configuration of the FMC feature and whatever the chosen mode with the STM32N6xx products, the stm32n6xx_ll_fmc.c file is not automatically included in the project. To avoid facing compilation errors the manual inclusion of this file is required.
• To avoid linking issues with Keil® MDK-ARM when a USBX Core System is enabled, the following modes should be enabled :
  • "UX Host HS1" or "UX Device HS1" if USB1_OTG_HS IP is used.
  • "UX Host HS2" or "UX Device HS2" if USB2_OTG_HS IP is used.
• Warning messages appear after building a project that configure the ThreadX with 'Low Power Support' mode, regardless of the toolchain.
• In the full LL mode, the function that activates the VddIOx power domain LL_PWR_EnableVddIOx() is missing from STM32CubeMX generated code. Consequently, there's no signal in some pins. (Note that for LL_PWR_EnableVddIOx() and VddIOx the x could be 2,3,4,5 depends on the activated power domain)
  • To fix the issue the user should add manually the LL_PWR_EnableVddIOx() in the main() after calling SystemClock_Config() when working in full LL mode.

2.9. When selecting microcontrollers in the STM32U3 series

• For STM32U3(only 2 MEGA microcontrollers) projects using the HSP_ENGINE middleware a build issue may occurs, indicating a type incompatibility, after the following scenario:
  • Activating the HSP_ENGINE middleware in mode sequencer
  • Adding a new processing list under the Processing List tab
  • Creating a new buffer under the Buffers tab
  • Selecting that buffer as an input, output ,coef ... buffer
  • Adding or updating a processing function,
  • Going back to the Buffers tab and changing the C Type of the used buffer
  • Generating the code

2.10. When selecting microcontrollers in the STM32U5 series

When creating a new STM32U5 project, enabling the JPEG peripheral and generating the code, the Utilities/JPEG folder containing the JPEG source files is not generated. To avoid this issue the user should add the missing folder manually (copy that folder from the firmware under this path Users\<User_name>\STM32Cube\Repository\STM32Cube_FW_U5_Vx.y.z\Utilities) each code generation.

• An issue may be encountered, after upgrading the X-CUBE-FREERTOS package from version 1.3.0 to 1.3.1, after the upgrade all of the CMSIS-RTOS2 Tasks and Queues are no longer visible in the "Tasks and Queues" tab within the project's configuration.
• A preprocessor error occurred in the compilation of STM32U5 projects because the #include statements in main.h were not properly delimited by newlines (more precisely including "stdio.h" in main.h with no line ending).
• Regarding the cohabitation between Azure^® RTOS USBX and USB Power Delivery, the VID and PID of USB Power Delivery and USBX are not automatically aligned. VID/PID values must be aligned manually.
• To avoid linking issues with Keil^® MDK-ARM when a USBX Core System is enabled, "Device CoreStack FS" or "Host CoreStack FS" must be enabled according to the USB_OTG_FS selected mode.
• To avoid compilation issues when one DMA channel request is activated for one TIMx peripheral, replace htim_xx by htim_base. For example, replace:
  __HAL_LINKDMA(htim_oc, hdma[TIM_DMA_ID_CC1], handle_GPDMA1_Channel11);
  by
  __HAL_LINKDMA(htim_base, hdma[TIM_DMA_ID_CC1], handle_GPDMA1_Channel11);
• In Microsoft^® Azure^® NetX Duo, some configuration flags, which are not used in the STM32Cube firmware package examples, might be nonfunctional when combined.

2.11. When selecting microcontrollers in the STM32WB series

• STM32CubeMX incorrectly removes the STM32_WPAN middleware settings from CMake and Makefile configuration files during subsequent code generations, which leads to compilation errors.
• The value of the CFG_BLE_MAX_CONN_EVENT_LENGTH parameter in the file app_conf.h is wrongly formatted: uint16_t is used instead of uint32_t. This impacts the Data Throughput of the application since the max value is 0xFFFF, while it should be 0xFFFF FFFF. The defined value must be changed manually in the generated code.

2.11.1. When selecting microcontrollers in the STM32WB0 product lines

• The following expansions packages for STM32Cube are not supported by microcontrollers from the STM32WB0 product lines:
  • X-CUBE-NFC10
  • X-CUBE-NFC9
  • X-CUBE-NFC4
  • X-CUBE-NFC10
  • X-CUBE-NFC6
  • X-CUBE-WB05N
  • X-CUBE-IPS
  • X-CUBE-TCPP
  • X-CUBE-SAFEA1
  • X-CUBE-GNSS1
  • X-CUBE-SUBG2
  • X-CUBE-EEPRMA1
  • X-CUBE-TOF1
  • X-CUBE-ALS
  • X-CUBE-BLEMGR
  • X-CUBE-ISPU
  • X-CUBE-BLE1
  • X-CUBE-BLE2
  • X-CUBE-MEMS1
  • X-CUBE-SFXS2LP1

2.12. When selecting microcontrollers in the STM32WBA series

• Creating a new project with an STM32WBA microcontroller,enabling BLE in the STM32_WPAN middleware and generating the build files using CMake or Makefile may result in a compilation error.It's caused by the reversed linking order of the libraries stm32wba_ble_stack_full.a and WBA5_LinkLayer_BLE_Full_lib.a .
  • To solve the issue the user, needs to reverse default order of the two libraries:
    • Here the update listed respectively in Makefile (for the Makefile toolchain) and in the top level CMakeLists.txt file (for CMake toolchain):

 #libraries
 LIBS = -lc -lm -lnosys \
 -l:stm32wba_ble_stack_full.a \
 -l:WBA5_LinkLayer_BLE_Full_lib.a 

# Add linked libraries
target_link_libraries(${CMAKE_PROJECT_NAME}
  stm32cubemx
   :stm32wba_ble_stack_full.a
   :WBA5_LinkLayer_BLE_Full_lib.a
  
   # Add user defined libraries
)

• The execution in EWARM or KEIL IDEs of an STM32WBA projects enabling the TrustZone and configuring the GTZC,the SAU features is failing due to an issue at the jump from secure to non-secure domain, indeed the RAM addresses in the linker files are not aligned with the GTZC and SAU memory configurations in the STM32CubeMX side (a linker files mismatch)
  • To resolve this issue, a manually update of the RAM address definitions in the linker files is required from the user:

-The user should edit:

-For WBA6/WBA6M Devices:

the Secure Linker File (`stm32wba6xx_flash_s.icf/stm32wba6mxx_flash_s.icf`) and change the RAM start and end addresses as follows:

   define symbol __ICFEDIT_region_RAM_start__ = 0x30000000;  
   define symbol __ICFEDIT_region_RAM_end__   = 0x3003FFFF;  
   

the Non-Secure Linker File (`stm32wba6xx_flash_ns.icf/stm32wba6mxx_flash_ns.icf`) and change the RAM start and end addresses as follows:

   define symbol __ICFEDIT_region_RAM_start__ = 0x20040000;  
   define symbol __ICFEDIT_region_RAM_end__   = 0x2007FFFF;  
   

- For WBA5/WBA5M Devices:

the Secure Linker File (`stm32wba5xx_flash_s.icf/stm32wba5mxx_flash_s.icf`)and change the RAM start and end addresses as follows:

   define symbol __ICFEDIT_region_RAM_start__ = 0x30000000;  
   define symbol __ICFEDIT_region_RAM_end__   = 0x3000FFFF;  
     

the Non-Secure Linker File (`stm32wba5xx_flash_ns.icf/stm32wba5mxx_flash_ns.icf`)and change the RAM start and end addresses as follows:

   define symbol __ICFEDIT_region_RAM_start__ = 0x20010000;  
   define symbol __ICFEDIT_region_RAM_end__   = 0x2001FFFF;     
   

• Some unexpected behaviors might appear during the creation of STM32WBA63 project: a non-updated STM32_WPAN Advertising elements menu, erroneous warnings at the peripheral configuration, and failures of IOC file reopening that lead to STM32CubeMX's home screen.
• When the selected middleware is STM32WPAN for Bluetooth^® Low Energy with the parameter CFG_BLE_NUM_LINK set above 8, use the Bluetooth^® Low Energy link layer library LinkLayer_BLE_Basic_20_links_lib.a with the header file from the path \link_layer\ll_cmd_lib\config\ble_basic_20_links.
• When building the BLE_Audio_TMAP_Central example for the STM32WBA55G-DK1 in STM32CubeIDE after being processed by STM32CubeMX, some library files might be missing or incorrectly referenced, leading to build errors.
  • To ensure proper library inclusion after installing the example, choose one of the following methods:

1. Enable Library Copying: In the Project Manager tab, under "Code Generator > STM32CUBE MCU packages and embedded software packs," check the option "Copy all used libraries into the project folder."
2. Manually Update Paths: In the .extSettings file, replace all instances of paths starting with ../Middlewares with ../../../Repository/STM32Cube_FW_WBA_Vx.x.x/Middlewares.
   • Example 1: Change ../Middlewares/ST/STM32_WPAN/ble/audio/Inc to ../../../Repository/STM32Cube_FW_WBA_Vx.x.x/Middlewares/ST/STM32_WPAN/ble/audio/Inc.
   • Example 2: Change Middlewares/STM32_WPAN=../Middlewares/ST/STM32_WPAN/ble/audio/lib/ble_audio.a to Middlewares/STM32_WPAN=../../../Repository/STM32Cube_FW_WBA_Vx.x.x/Middlewares/ST/STM32_WPAN/ble/audio/lib/ble_audio.a.
   • Note: The second workaround (Option 2) assumes the following default installation paths:
     • Example Path: C:\Users\USER_NAME\STM32Cube\Example
     • Firmware Path: C:\Users\USER_NAME\STM32Cube\Repository\STM32Cube_FW_WBA_Vx.x.x

2.13. When selecting microcontrollers in the STM32WL series

• For STM32WL3 projects that operates with FreeRTOS ,when the timebase is set to SysTick in conjunction with a toolchain that utilizes the GCC compiler, build issues may be encountered.
  • To mitigate these issues and ensure stable behavior when the compiler is GCC, it is recommended to restrict the use of Timebase source to the other available timers (TIMx).
• If LoRaWAN^®, SubGHZ_Phy or Sigfox^TM middleware is used with the option Generate peripheral initialization as a pair of '.c/.h' files per peripheral disabled, it is advised to untick the visibility of the SUBGHZ and IPCC peripherals in [Project Manager]>[Advanced Settings] Generated Function Calls panel to avoid an issue when compiling.
• Import from and to dual-core is not working for the STM32WL series.
• For dual-core products, IPCC LL + RF middleware (LoRaWAN^®, Sigfox^TM, and SubGHZ_Phy) are not supported.
• If a Sigfox project is generated with MDK-ARM, the following option must be set in the project Options C/C++(AC6) panel: Misc Controls: -fshort-enums
• The SUBGHZ peripheral is forced on the Cortex^®-M0+.
• In the case of a project migration from a release version before V6.2.0, to facilitate the GTZC configuration, some previously default configuration ("enabled"/"disabled") has been moved to "default" configuration. This new state provides the default configuration of the MCU without any code generation. Users specifically needing this code generation must modify this state.
• When compiling under IAR Embedded Workbench^®, in case of dual-core products, the user must activate the share mode in both CM4 and CM0PLUS ST-LINK project options.
• The "Time base: System tick timer" value selected in the NVIC configuration panel is not used for the code generation. This value must be updated by the end user in the generated code.
• KMS and STM32Cube Azure^® RTOS ThreadX cannot be enabled in the same core.

2.13.1. When selecting microcontrollers in the STM32WL33 product lines

• The following expansions packages for STM32Cube are not supported by microcontrollers from the STM32WL33 product lines:
  • X-CUBE-NFC10
  • X-CUBE-NFC9
  • X-CUBE-NFC4
  • X-CUBE-NFC10
  • X-CUBE-NFC6
  • X-CUBE-WB05N
  • X-CUBE-IPS
  • X-CUBE-TCPP
  • X-CUBE-SAFEA1
  • X-CUBE-GNSS1
  • X-CUBE-SUBG2
  • X-CUBE-EEPRMA1
  • X-CUBE-TOF1
  • X-CUBE-ALS
  • X-CUBE-BLEMGR
  • X-CUBE-ISPU
  • X-CUBE-BLE1
  • X-CUBE-BLE2
  • X-CUBE-MEMS1
  • X-CUBE-SFXS2LP1

2.14. When selecting microprocessors in the STM32MP1 series

• DMA nodes are generated in the device tree, but the DMA properties in the IP clients nodes are not generated.
• Import from and to MPU projects does not work properly.
• Dual-core project structure compatibility break : the action to import (migrate or continue) projects created with versions earlier than STM32CubeMX V5.3.0 is supported but requires the manual copy of USER SECTIONS for the device tree and Cortex^®-M4 firmware from the former folder structure (DeviceTree/Inc/Src) to the new one (CA7/DeviceTree - CM4/Src - CM4/Inc).
• Additional software is not supported.
• RCC generation in Production mode is not supported on the coprocessor side. Only the Engineering mode is supported.
  By default, the STM32CubeMX-generated code is compliant with the Engineering mode. Therefore, the call of the following clock functions must be removed in the Production mode since these clocks are then managed by Linux^®: SystemClock_Config(), PeriphCommonClock_Config(), and HAL_RCCEx_PeriphCLKConfig():
  • The system parts (SystemClock_Config() and PeriphCommonClock_Config()) can be removed in STM32CubeMX by selecting [Not Generate function call] for RCC in the Project Manager, then Advanced Settings tabs.
  • HAL_RCCEx_PeriphCLKConfig() must be removed manually from file stm32mp1xx_hal_msp.c (STM32CubeMX-generated code).
  • To make the user code compatible with both the engineering and production modes, the above RCC functions can be put under dynamic condition
    if(IS_ENGINEERING_BOOT_MODE()).
• On macOS^®, installation issues may result from the fact that the install is not signed. Use the following procedure:
  • Download STM32CubeMX on a Window^® OS personal computer.
  • Copy the downloaded install into the macOS^® personal computer.
  • Launch the install.
• OpenAMP warning when compiling under MDK-ARM if OpenAMP is not activated under STM32CubeMX:
  • Remove the next four lines of code from the linker files (.sct) to avoid warning “No section matches pattern *(.resource_table)”:

; ***** Comment these 4 lines if OPENAMP is not used *****
.resource_table +0 ALIGN 4 { ; resource table
*(.resource_table)
}
; Shared Memory area used by OpenAMP
__OpenAMP_SHMEM__ 0x10040000 EMPTY 0x8000 {}

• Device tree generation: the "pwr" user section name is changed to "pwr_regulators".
  • Existing code in the "pwr" user section must be backed up and reported manually into the new "pwr_regulators" user section after code generation.
• About the HRTIM peripheral: for the "Set Source Selection" & "Reset Source Selection" parameters, the "No source is selected" must be manually set for any unused source before decreasing the number of sources.
• When using the STM32CubeMP13 bare-metal firmware, the secure mode of the peripheral RTC is not available for the STM32MP13xx microprocessors. It is only available when the STM32CubeMP1 MCU package is used.
• When using the STM32CubeMP13 bare-metal firmware and activating Azure^® RTOS ThreadX, the STM32CubeIDE linker file must contain the following section to avoid compilation issues:

.stack:
{
 _stack_bottom = ABSOLUTE(.) ;
 /* Allocate room for stack. This must be large enough for the
 IRQ, FIQ, and SYS stack if nested interrupts are
 enabled.*/
 . = ALIGN(8) ;
 . += 32768 ;
 _sp = . - 16 ;
 _stack_top = ABSOLUTE(.) ;
 } >RAM
 _end = .;

2.15. When using the Memory Management Tool

• For projects using STM32H573I-DK board with TrustZone feature enabled, the activation of the following Memory Management parameter 'Apply Application Regions Settings to peripherals' after initial code generation results in build failures.
• Linker file issues may result from configuring MMT after the initial code generation (more precisely activating the MMT in the second code generation).
• When using the Memory Management Tool in a project based on CMake or Makefile build generators, the following configurations are not supported:
  • The TrustZone activation.
  • The use of the STM32H7Rx/7Sx microcontrollers from the STM32H7 series.
• For the projects based on STM32WB series and using the MMT, the generated linker file in not taking into account the activation of the option "Boot Info".
• When setting the mode of the Extmem Manager in the boot runtime context, the Memory Management Tool generates only an 'Execute' memory region and omits the 'LOAD_REGION' one. The 'LOAD_REGION' is displayed as reserved in the UI. This limitation applies to all toolchains and IDEs.
• The linker script is not automatically updated to incorporate the 'isrvercor' configuration.

2.16. When using the Example Selector

• When generating STM32CubeMX (version 6.13.0) projects for RF applications of the STM32WB0x and STM32WBA5x series, with the STM32CubeIDE selected as a tool chain, users may encounter issues with the generated project structure. More precisely essential header files might be missing, resulting in compilation errors.
  • Workaround: To address this issue, it is recommended to regenerate the project within STM32CubeMX.

• When starting an example from the Example Selector and regenerating the code, some files can be duplicated. It can lead to potential compilation issues. In this case, remove the duplicated files.

2.17. When using additional software packs

• If a software component refers to a specific peripheral, this peripheral must be initialized with HAL APIs (not with LL APIs).
• When using dual-core devices, if a software component refers to a specific peripheral, this peripheral must be assigned only to the same core as the software component.
• When a pack is disabled, the generated files are not removed from the project. Workaround: remove these files manually.
• Functionalities that are not supported:
  • maxInstances, isDefaultVariant, generator attributes of the component element
  • Dtz, Ddsp, Dsecure attributes of the condition element
  • repository element
  • tag and url attributes of a release element
  • dominate element
  • public attributes for the file element
  • preIncludeLocal and preIncludeGlobal file category attributes
  • Pre_Include_Local_h element
  • Pre_Include_Global_h element
  • Pack components where attribute values come with a "." character
• Compilation issues when using both X-CUBE-AZRTOS-H7 v1.1.0 and OpenAMP middleware in the same project with minimum version MDK-ARM v5.32:
  X-CUBE-AZRTOS-H7 v1.1.0 requires Arm C compiler version 6 while OpenAMP is not compatible with this version (only Arm C compiler version 5 is supported). The workaround consists in using X-CUBE-AZRTOS-H7 v1.0.0 instead. This one is compatible with Arm C compiler version 5.
• In the case of an error message when downloading SEGGER I-CUBE-embOS:
  • Download SEGGER.I-CUBE-embOS manually from the SEGGER download site at https://www.segger.com/downloads/embos.
  • Open the STM32CubeMX Embedded Software Packages Manager.
  • Use the option Load from Local and select the pack.

2.18. STM32CubeIDE toolchain

• It is functional only with a Java 8 64-bit virtual machine with versions earlier than STM32CubeMX V5.6.1.

2.19. DDR test suite

• STM32CubeProgrammer (STM32CubeProg) is not supported from v2.10.0 onwards.

2.20. User manual

• Some chapters in the user manual STM32CubeMX for STM32 configuration and initialization C code generation (UM1718) are not fully up to date.

2.21. Symbolic links on Ubuntu

• When using Ubuntu 20.04 if you use a symbolic link to refer to the "Firmware Relative Path" in the STM32CubMX Project Manager tab, it will be translated to the absolute path before saving down.

3. Recommendations

3.1. Command line authentication required

Since introduction of authentication at the command level, we recommend our costumers to review their scripts. They should make sure they are authenticating before trying to download a package or before trying to generate code on the command line. Failure to do so will result on errors displayed on the command line.

3.2. DDR binaries for the STM32MP1 boards

The DDR binaries needed to connect to the STM32MP1 boards are available from the GitHub website at https://github.com/ with the keyword STMicroelectronics/STM32DDRFWUTIL.

3.3. List pinout-compatible microcontrollers

The loading of microcontrollers can take a long time if all series and all packages are selected. If the loading takes too long, stop it, refine the microcontroller filters on a dedicated series or package, and restart the loading.

3.4. How to improve the wake-up from Stop sequences without relying on the global SystemClock_Config()

The SystemClock_Config provided in the HAL examples is intended to be used to configure the system clock at startup. It is not recommended to use it to restore the clock settings after a Stop mode.
To keep the wake-up as quick as possible, it is highly recommended to restore only what has been changed by the hardware to enter the Low-power mode. This is typically the case for the oscillators and system clock sources. The RCC configuration registers are unchanged, and there is no need to restore them.
Typically, after a Stop mode, it is recommended to call a separate SystemClockConfig_STOP after the wake-up that enables the LSE and PLL, and selects PLL as system clock source (LSE and PLL are disabled in Stop mode). Find below an example of SystemClockConfig_STOP (provided in the HID_Standalone_LPM application):

static void SystemClockConfig_STOP(void)
{
#if defined (USB_USE_LSE_MSI_CLOCK)
/* Configures system clock after wake-up from STOP: enable LSE,
PLL and select PLL as system clock source (LSE and PLL are disabled in Stop mode) */
__HAL_RCC_LSE_CONFIG(RCC_LSE_ON);
/* Wait till HSE is ready */
while(__HAL_RCC_GET_FLAG(RCC_FLAG_LSERDY) == RESET)
{}
/* Enable the main PLL. */
__HAL_RCC_PLL_ENABLE();
/* Wait till PLL is ready */
while(__HAL_RCC_GET_FLAG(RCC_FLAG_PLLRDY) == RESET)
{}
/* Select PLL as SYSCLK */
MODIFY_REG(RCC->CFGR, RCC_CFGR_SW, RCC_SYSCLKSOURCE_PLLCLK);
/* Wait till system clock switch to PLL */
while (__HAL_RCC_GET_SYSCLK_SOURCE() != RCC_CFGR_SWS_PLL)
{}
#elif defined (USB_USE_HSE_CLOCK)
/* Configures system clock after wake-up from STOP: enable HSE,
PLL and select PLL as system clock source (HSE and PLL are disabled in Stop mode) */
__HAL_RCC_HSE_CONFIG(RCC_HSE_ON);
/* Wait till HSE is ready */
while(__HAL_RCC_GET_FLAG(RCC_FLAG_HSERDY) == RESET)
{}
/* Enable the main PLL. */
__HAL_RCC_PLL_ENABLE();
/* Wait till PLL is ready */
while(__HAL_RCC_GET_FLAG(RCC_FLAG_PLLRDY) == RESET)
{}
/* Select PLL as SYSCLK */
MODIFY_REG(RCC->CFGR, RCC_CFGR_SW, RCC_SYSCLKSOURCE_PLLCLK);
/* Wait till system clock switch to PLL */
while (__HAL_RCC_GET_SYSCLK_SOURCE() != RCC_CFGR_SWS_PLL)
{}
#endif /* USB_USE_LSE_MSI_CLOCK */
}