Re: [PATCH V5 1/4] Remove extra conversions in pid calculation

[Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>]

On Sun, 2016-03-06 at 08:34 +0100, Philippe Longepe wrote:
> pid->setpoint and pid->deadband can be initialize in float so we
> can remove the int_tofp in pid_calc.

Rafael had some comments on this patch:
"s/initialize/initialized/

> can remove the int_tofp in pid_calc.

This is not "float", but "fixed point".

Also "avoid" rather than "remove".
"
Not sure if Rafael is OK without change.

> Signed-off-by: Philippe Longepe <philippe.longepe@linux.intel.com>
> ---
>  drivers/cpufreq/intel_pstate.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/cpufreq/intel_pstate.c
> b/drivers/cpufreq/intel_pstate.c
> index eb0aef0..114e4e0 100644
> --- a/drivers/cpufreq/intel_pstate.c
> +++ b/drivers/cpufreq/intel_pstate.c
> @@ -198,8 +198,8 @@ static struct perf_limits *limits =
> &powersave_limits;
>  
>  static inline void pid_reset(struct _pid *pid, int setpoint, int
> busy,
>  			     int deadband, int integral) {
> -	pid->setpoint = setpoint;
> -	pid->deadband  = deadband;
> +	pid->setpoint = int_tofp(setpoint);
> +	pid->deadband  = int_tofp(deadband);
>  	pid->integral  = int_tofp(integral);
>  	pid->last_err  = int_tofp(setpoint) - int_tofp(busy);
>  }
> @@ -225,9 +225,9 @@ static signed int pid_calc(struct _pid *pid,
> int32_t busy)
>  	int32_t pterm, dterm, fp_error;
>  	int32_t integral_limit;
>  
> -	fp_error = int_tofp(pid->setpoint) - busy;
> +	fp_error = pid->setpoint - busy;
>  
> -	if (abs(fp_error) <= int_tofp(pid->deadband))
> +	if (abs(fp_error) <= pid->deadband)
>  		return 0;
>  
>  	pterm = mul_fp(pid->p_gain, fp_error);