drivers: stepper: add common helper header for tmc5xxx functions

[jilaypandya]

This commit adds a common helper header for tmc5xxx driver in #80784

[bjarki-andreasen]

I would update the signature to:

static inline uint32_t tmc5xxx_calculate_velocity_from_hz_to_fclk(uint64_t velocity_hz,
						                  uint32_t clock_frequency)
{
        __ASSERT_NO_MSG(clock_frequency);
        return (velocity_hz << TMC5XXX_CLOCK_FREQ_SHIFT) / clock_frequency;
}

• static inline to allow the compiler to either inline it or make it a static function, whichever is most efficient.
• since the function "can't fail", just return the result. Otherwise we would return a status, and copy out the result.
• add an assert to prevent divide by zero error
• make velocity_hz 64 bit removing the need to promote it and make it mutable (no const) to allow the compiler to work directly on the variable which is already on the stack
• since we are copying the arguments entirely, const really does not help us :) const is useful when referencing a structure outside of the function, so we don't modify something we don't own, inside the function its not really helpful :)

[bjarki-andreasen]

the indentation here can be helped by keeping the definition of uint32_t velocity_fclk in the header (where it should be according to MISRA), and even defining tmc5041_config->clock_frequency:

uint32_t clock_frequency = tmc5041_config->clock_frequency;
uint32_t velocity_fclk;
...
velocity_fclk = tmc5xxx_calculate_velocity_from_hz_to_fclk(velocity, clock_frequency);

the compiler will optimize the clock_frequency away :)

[bjarki-andreasen]

Nice :)

[faxe1008]

Doesn't this rely on the correct include order? The autoconf is not pulled in otherwise.