intel_pstate: Correct rounding in busy calculation
authorDoug Smythies <doug.smythies@gmail.com>
Tue, 17 Jun 2014 20:36:10 +0000 (13:36 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 9 Jul 2014 18:18:26 +0000 (11:18 -0700)
commitf4e368a53c65f8acd462b2b038c18a21d30eaeb8
tree258a519724d8802f31b67c5ffb523036cc809f16
parent15a446b98f735f427927a251e82fed0a5a59c9ee
intel_pstate: Correct rounding in busy calculation

commit 51d211e9c334b9eca3505f4052afa660c3e0606b upstream.

There was a mistake in the actual rounding portion this previous patch:
f0fe3cd7e12d (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.

Fixes: f0fe3cd7e12d (intel_pstate: Correct rounding in busy calculation)
Signed-off-by: Doug Smythies <dsmythies@telus.net>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/cpufreq/intel_pstate.c