Usage of (uint32_t) in device headers.

[Uwe Bonnes]

Posted on March 10, 2016 at 16:14

STM32 device headers use e.g.

#define FLASH_BASE            ((uint32_t)0x08000000U)

  /*!< FLASH(up to 1 MB) base address in the alias region  */

while e.g. Freescale uses in a similar situation

#define ADC0_BASE                                (0x4003B000u)

The usage of the (uint32_t) cast by ST denied to use FLASH_BASE in preprocessor expressions. Please consider dropping the sometimes harmfull (uint32_t) cast where not explicit needed!

[Walid FTITI_O]

Posted on March 11, 2016 at 11:09

Hi bonnes.uwe,

Thanks for your feedback, I will report this to the team.

-Hannibal-

[Walid FTITI_O]

Posted on March 31, 2016 at 13:20

Hi bonnes.uwe, 

Your feedback will be implimented in the new release. Thanks for the contribution.

-Hannibal-

[Radosław]

Posted on March 31, 2016 at 14:05

Using macro sufix and cast to uint32 is double solution. 

First st add cast in last revision add ''U'' for misra compilance.  This is chaotic. 

[Uwe Bonnes]

Posted on April 13, 2016 at 12:31

I can confirm the confusion.

[Uwe Bonnes]

Posted on November 30, 2016 at 13:45

A construct like  ((uint32_t)0x08000000U is still used in some places in the most recent Cube library update, e.g. in the L4 library in all xxx_BASE definitions and

for F4 in the F423 AES definitions.

[AvaTar]

Posted on November 30, 2016 at 13:56

> A construct like  ((uint32_t)0x08000000U is still used in some places in the most recent Cube library update, ...

 

I'd guess as a kowtow towards (potential) MISRA-obsessed automotive customers ...