HAL GPIO assert fails for GPIO_AF3_TIM10 on STM32F4 devices

[OleksandrMovchan]

Description
Hello,

I am using STM32F411RET and STM32F407ZGT microcontrollers.
Using STM32CubeMX, I configure a PWM output on GPIOB Pin 8 using TIM10 Channel 1.
After code generation, the following GPIO initialization code is produced:

GPIO_InitStruct.Pin = GPIO_PIN_8;
GPIO_InitStruct.Mode = GPIO_MODE_AF_PP;
GPIO_InitStruct.Pull = GPIO_NOPULL;
GPIO_InitStruct.Speed = GPIO_SPEED_FREQ_LOW;
GPIO_InitStruct.Alternate = GPIO_AF3_TIM10;
HAL_GPIO_Init(GPIOB, &GPIO_InitStruct);

The PWM output works correctly on hardware.
However, when full assertions are enabled by defining USE_FULL_ASSERT, an assertion failure occurs inside HAL_GPIO_Init() at the following check:

assert_param(IS_GPIO_AF(GPIO_Init->Alternate));

After inspecting the IS_GPIO_AF() macro for STM32F4 devices, I noticed that GPIO_AF3_TIM10 is not included in the list of valid alternate functions, even though TIM10 uses AF3 on STM32F4 devices and is correctly generated by STM32CubeMX.
This causes a false assertion failure when using a valid and supported peripheral configuration.

Expected behavior
GPIO_AF3_TIM10 should be included in the IS_GPIO_AF() validation macro for STM32F4 devices (e.g. STM32F407 / STM32F411), or the assertion logic should be updated to avoid rejecting valid alternate functions generated by CubeMX.

Additional notes
The issue occurs only when USE_FULL_ASSERT is enabled.
The generated code and hardware configuration are otherwise correct.
Disabling full assertions or manually extending IS_GPIO_AF() resolves the issue.

[mƎALLEm]

Hello,

Thank you for raising that issue. I'm escalating it internally for fix.

Indeed:

It's available with STM32F401x device (in stm32f4xx_hal_gpio_ex.h):

#if defined(STM32F401xC) || defined(STM32F401xE)
#define IS_GPIO_AF(AF) (((AF) == GPIO_AF0_RTC_50Hz) || ((AF) == GPIO_AF12_SDIO) || \
 ((AF) == GPIO_AF0_MCO) || ((AF) == GPIO_AF0_TAMPER) || \
 ((AF) == GPIO_AF0_SWJ) || ((AF) == GPIO_AF0_TRACE) || \
 ((AF) == GPIO_AF1_TIM1) || ((AF) == GPIO_AF1_TIM2) || \
 ((AF) == GPIO_AF2_TIM3) || ((AF) == GPIO_AF2_TIM4) || \
 ((AF) == GPIO_AF2_TIM5) || ((AF) == GPIO_AF3_TIM9) || \
 ((AF) == GPIO_AF3_TIM10) || ((AF) == GPIO_AF3_TIM11) || \
 ((AF) == GPIO_AF4_I2C1) || ((AF) == GPIO_AF4_I2C2) || \
 ((AF) == GPIO_AF4_I2C3) || ((AF) == GPIO_AF5_SPI1) || \
 ((AF) == GPIO_AF5_SPI2) || ((AF) == GPIO_AF5_SPI4) || \
 ((AF) == GPIO_AF6_SPI3) || ((AF) == GPIO_AF7_USART1) || \
 ((AF) == GPIO_AF7_USART2) || ((AF) == GPIO_AF8_USART6) || \
 ((AF) == GPIO_AF9_I2C2) || ((AF) == GPIO_AF9_I2C3) || \
 ((AF) == GPIO_AF10_OTG_FS) || ((AF) == GPIO_AF15_EVENTOUT))
#endif /* STM32F401xC || STM32F401xE */

While it's missing with STM32F411xE device (in stm32f4xx_hal_gpio_ex.h)::

#if defined(STM32F411xE) 
#define IS_GPIO_AF(AF) (((AF) == GPIO_AF0_RTC_50Hz) || ((AF) == GPIO_AF9_TIM14) || \
 ((AF) == GPIO_AF0_MCO) || ((AF) == GPIO_AF0_TAMPER) || \
 ((AF) == GPIO_AF0_SWJ) || ((AF) == GPIO_AF0_TRACE) || \
 ((AF) == GPIO_AF1_TIM1) || ((AF) == GPIO_AF1_TIM2) || \
 ((AF) == GPIO_AF2_TIM3) || ((AF) == GPIO_AF2_TIM4) || \
 ((AF) == GPIO_AF2_TIM5) || ((AF) == GPIO_AF4_I2C1) || \
 ((AF) == GPIO_AF4_I2C2) || ((AF) == GPIO_AF4_I2C3) || \
 ((AF) == GPIO_AF5_SPI1) || ((AF) == GPIO_AF5_SPI2) || \
 ((AF) == GPIO_AF5_SPI3) || ((AF) == GPIO_AF6_SPI4) || \
 ((AF) == GPIO_AF6_SPI3) || ((AF) == GPIO_AF5_SPI4) || \
 ((AF) == GPIO_AF6_SPI5) || ((AF) == GPIO_AF7_SPI3) || \
 ((AF) == GPIO_AF7_USART1) || ((AF) == GPIO_AF7_USART2) || \
 ((AF) == GPIO_AF8_USART6) || ((AF) == GPIO_AF10_OTG_FS) || \
 ((AF) == GPIO_AF9_I2C2) || ((AF) == GPIO_AF9_I2C3) || \
 ((AF) == GPIO_AF12_SDIO) || ((AF) == GPIO_AF15_EVENTOUT))

#endif /* STM32F411xE */

Internal ticket 227747 for fix.

Thank you for your your contribution.

[OleksandrMovchan]

It also related to the STM32F407xx

[mƎALLEm]

Ok thanks. I requested to do a check for all devices for AF3.

[mƎALLEm]

Issue confirmed and it seems IS_GPIO_AF() macros currently omitted all possible peripheral instances (TIM9, TIM10, and TIM11) for GPIO_AF3 selection.

That will be fixed in the next version of the HAL.