Bug: CubeMX code generation with LL HAL for timer

[rleigh-electradx]

Hi,

CubeMX version 6.15.0 with STM32U5 HAL 1.8.0.

<project>/XP2/Core/Src/tim.c: In function 'MX_TIM5_Init':
<project>/XP2/Core/Src/tim.c:228:30: error: conversion from 'long unsigned int' to 'uint16_t' {aka 'short unsigned int'} changes value from '4294967196' to '65436' [-Werror=overflow]
228 | TIM_InitStruct.Prescaler = 156-LL_TIM_ETR_FILTER_FDIV1_N2;
| ^~~
cc1.exe: all warnings being treated as errors

 

The prescaler setting of "156-1" should be 155 when set, but the code is actually this:

 /* USER CODE END TIM5_Init 1 */
 TIM_InitStruct.Prescaler = 156-LL_TIM_ETR_FILTER_FDIV1_N2;
 TIM_InitStruct.CounterMode = LL_TIM_COUNTERMODE_UP;
 TIM_InitStruct.Autoreload = 4294967295;
 TIM_InitStruct.ClockDivision = LL_TIM_CLOCKDIVISION_DIV1;
 LL_TIM_Init(TIM5, &TIM_InitStruct);
 LL_TIM_DisableARRPreload(TIM5);
 LL_TIM_SetClockSource(TIM5, LL_TIM_CLOCKSOURCE_INTERNAL);
 LL_TIM_SetTriggerOutput(TIM5, LL_TIM_TRGO_RESET);
 LL_TIM_DisableMasterSlaveMode(TIM5);
 /* USER CODE BEGIN TIM5_Init 2 */

Where did the LL_TIM_ETR_FILTER_FDIV1_N2 come from?  This has replaced the calculation I entered in the prescaler configuration, above.

 

Kind regards,

Roger

[Souhaib MAZHOUD]

Hello @rleigh-electradx 

Thank you for your contribution!

I am currently checking this issue, and I will get back to you ASAP.

KR, Souhaib

[rleigh-electradx]

In case it matters, this was with an STM32U575ZIT part.

[Souhaib MAZHOUD]

Hello @rleigh-electradx 

This issue has been raised to the development team for analysis and revision of upcoming releases (Ticket 207335 This is an internal tracking number and is not accessible or usable by customers).)

KR, Souhaib