| Message ID | 1403037370-3484-1-git-send-email-dsmythies@telus.net (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Headers | show |
On Tuesday, June 17, 2014 01:36:10 PM Doug Smythies wrote: > There was a mistake in the actual rounding portion this previous patch: > f0fe3cd7e12d8290c82284b5c8aee723cbd0371a intel_pstate: Correct rounding in busy calculation > such that the rounding was asymetric and incorrect. > > Severity: Not very serious, but can increase target pstate by one extra value. > For real world work flows the issue should self correct (but I have no proof). > It is the equivalent of different PID gains for positive and negative numbers. > > Examples: > -3.000000 used to round to -4, rounds to -3 with this patch. > -3.503906 used to round to -5, rounds to -4 with this patch. > > Signed-off-by: Doug Smythies <dsmythies@telus.net> Applied, thanks! > --- > drivers/cpufreq/intel_pstate.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c > index 4e7f492..924bb2d 100644 > --- a/drivers/cpufreq/intel_pstate.c > +++ b/drivers/cpufreq/intel_pstate.c > @@ -196,10 +196,7 @@ static signed int pid_calc(struct _pid *pid, int32_t busy) > pid->last_err = fp_error; > > result = pterm + mul_fp(pid->integral, pid->i_gain) + dterm; > - if (result >= 0) > - result = result + (1 << (FRAC_BITS-1)); > - else > - result = result - (1 << (FRAC_BITS-1)); > + result = result + (1 << (FRAC_BITS-1)); > return (signed int)fp_toint(result); > } > >
diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index 4e7f492..924bb2d 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c @@ -196,10 +196,7 @@ static signed int pid_calc(struct _pid *pid, int32_t busy) pid->last_err = fp_error; result = pterm + mul_fp(pid->integral, pid->i_gain) + dterm; - if (result >= 0) - result = result + (1 << (FRAC_BITS-1)); - else - result = result - (1 << (FRAC_BITS-1)); + result = result + (1 << (FRAC_BITS-1)); return (signed int)fp_toint(result); }
There was a mistake in the actual rounding portion this previous patch: f0fe3cd7e12d8290c82284b5c8aee723cbd0371a intel_pstate: Correct rounding in busy calculation such that the rounding was asymetric and incorrect. Severity: Not very serious, but can increase target pstate by one extra value. For real world work flows the issue should self correct (but I have no proof). It is the equivalent of different PID gains for positive and negative numbers. Examples: -3.000000 used to round to -4, rounds to -3 with this patch. -3.503906 used to round to -5, rounds to -4 with this patch. Signed-off-by: Doug Smythies <dsmythies@telus.net> --- drivers/cpufreq/intel_pstate.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)