[PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[dirk.brandewie]

From: Dirk Brandewie <dirk.j.brandewie@intel.com>

Change to using max P-state instead of max turbo P-state.  This change
resolves two issues.

On a quiet system intel_pstate can fail to respond to a load change.
   https://bugzilla.kernel.org/show_bug.cgi?id=59481

On CPU SKUs that have a limited number of P-states and no turbo range
intel_pstate fails to select the highest available P-state.

This change is suitable for stable v3.9+

Reported-by: arjan@linux.intel.com
Reported-by: dsmythies@telus.net
Tested-by: arjan@linux.intel.com
Tested-by: dsmythies@telus.net

CC: stable@vger.kernel.org
CC: dsmythies@telus.net
Signed-off-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
---
 drivers/cpufreq/intel_pstate.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c
index 07f2840..6d6a0b4 100644
--- a/drivers/cpufreq/intel_pstate.c
+++ b/drivers/cpufreq/intel_pstate.c
@@ -103,10 +103,10 @@ struct pstate_adjust_policy {
 static struct pstate_adjust_policy default_policy = {
 	.sample_rate_ms = 10,
 	.deadband = 0,
-	.setpoint = 109,
-	.p_gain_pct = 17,
+	.setpoint = 97,
+	.p_gain_pct = 20,
 	.d_gain_pct = 0,
-	.i_gain_pct = 4,
+	.i_gain_pct = 0,
 };
 
 struct perf_limits {
@@ -468,12 +468,12 @@ static inline void intel_pstate_set_sample_time(struct cpudata *cpu)
 static inline int intel_pstate_get_scaled_busy(struct cpudata *cpu)
 {
 	int32_t busy_scaled;
-	int32_t core_busy, turbo_pstate, current_pstate;
+	int32_t core_busy, max_pstate, current_pstate;
 
 	core_busy = int_tofp(cpu->samples[cpu->sample_ptr].core_pct_busy);
-	turbo_pstate = int_tofp(cpu->pstate.turbo_pstate);
+	max_pstate = int_tofp(cpu->pstate.max_pstate);
 	current_pstate = int_tofp(cpu->pstate.current_pstate);
-	busy_scaled = mul_fp(core_busy, div_fp(turbo_pstate, current_pstate));
+	busy_scaled = mul_fp(core_busy, div_fp(max_pstate, current_pstate));
 
 	return fp_toint(busy_scaled);
 }
-- 
1.8.3.1

Re: [PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[Viresh Kumar]

On Thu, Jul 18, 2013 at 9:18 PM,  <dirk.brandewie@gmail.com> wrote:
> From: Dirk Brandewie <dirk.j.brandewie@intel.com>
>
> Change to using max P-state instead of max turbo P-state.  This change
> resolves two issues.
>
> On a quiet system intel_pstate can fail to respond to a load change.
>    https://bugzilla.kernel.org/show_bug.cgi?id=59481
>
> On CPU SKUs that have a limited number of P-states and no turbo range
> intel_pstate fails to select the highest available P-state.
>
> This change is suitable for stable v3.9+
>
> Reported-by: arjan@linux.intel.com
> Reported-by: dsmythies@telus.net
> Tested-by: arjan@linux.intel.com
> Tested-by: dsmythies@telus.net
>
> CC: stable@vger.kernel.org
> CC: dsmythies@telus.net
> Signed-off-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
> ---
>  drivers/cpufreq/intel_pstate.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

Re: [PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[Guilherme Salazar]

 <dirk.brandewie <at> gmail.com> writes:

> 
> From: Dirk Brandewie <dirk.j.brandewie <at> intel.com>
> 
> Change to using max P-state instead of max turbo P-state.  This change
> resolves two issues.
> 
> On a quiet system intel_pstate can fail to respond to a load change.
>    https://bugzilla.kernel.org/show_bug.cgi?id=59481
> 
> On CPU SKUs that have a limited number of P-states and no turbo range
> intel_pstate fails to select the highest available P-state.
> 
> This change is suitable for stable v3.9+
> 
> Reported-by: arjan <at> linux.intel.com
> Reported-by: dsmythies <at> telus.net
> Tested-by: arjan <at> linux.intel.com
> Tested-by: dsmythies <at> telus.net
> 
> CC: stable <at> vger.kernel.org
> CC: dsmythies <at> telus.net
> Signed-off-by: Dirk Brandewie <dirk.j.brandewie <at> intel.com>

Hi,

I've just tested the patch and, even though it works (the frequency now sits 
around 1.3-1.5 GHz), I noticed that if I suspend the laptop and then resume 
it, the abnormal behavior (frequency sitting above 2 GHz) returns.

Before suspending, the value in 
/sys/devices/system/cpu/intel_pstate/min_perf_pct was 37; afterwards, 100.

cpuinfo: http://pastebin.com/C6zR90HC

I'd be glad to give any further details.

Best,
Guilherme

Re: [PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[Dirk Brandewie]

On 08/01/2013 10:56 AM, Guilherme Salazar wrote:
>   <dirk.brandewie <at> gmail.com> writes:
>
>>
>> From: Dirk Brandewie <dirk.j.brandewie <at> intel.com>
>>
>> Change to using max P-state instead of max turbo P-state.  This change
>> resolves two issues.
>>
>> On a quiet system intel_pstate can fail to respond to a load change.
>>     https://bugzilla.kernel.org/show_bug.cgi?id=59481
>>
>> On CPU SKUs that have a limited number of P-states and no turbo range
>> intel_pstate fails to select the highest available P-state.
>>
>> This change is suitable for stable v3.9+
>>
>> Reported-by: arjan <at> linux.intel.com
>> Reported-by: dsmythies <at> telus.net
>> Tested-by: arjan <at> linux.intel.com
>> Tested-by: dsmythies <at> telus.net
>>
>> CC: stable <at> vger.kernel.org
>> CC: dsmythies <at> telus.net
>> Signed-off-by: Dirk Brandewie <dirk.j.brandewie <at> intel.com>
>
> Hi,
>
> I've just tested the patch and, even though it works (the frequency now sits
> around 1.3-1.5 GHz), I noticed that if I suspend the laptop and then resume
> it, the abnormal behavior (frequency sitting above 2 GHz) returns.
>
There is a separate bug against the suspend resume behaviour. It is coming
from a bug in the cpufreq core.

https://bugzilla.kernel.org/show_bug.cgi?id=59781

The fix is disscussed in https://lkml.org/lkml/2013/7/29/692 and is queue for 3.12

> Before suspending, the value in
> /sys/devices/system/cpu/intel_pstate/min_perf_pct was 37; afterwards, 100.
>
> cpuinfo: http://pastebin.com/C6zR90HC
>
> I'd be glad to give any further details.
>
> Best,
> Guilherme
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

[PATCH] cpufreq/intel_pstate: Change to scale off of max P-state

[dirk.brandewie]

From: Dirk Brandewie <dirk.j.brandewie@intel.com>

Change to using max P-state instead of max turbo P-state.  This change
resolves two issues.

On a quiet system intel_pstate can fail to respond to a load change.
   https://bugzilla.kernel.org/show_bug.cgi?id=59481

On CPU SKUs that have a limited number of P-states and no turbo range
intel_pstate fails to select the highest available P-state.

This change is suitable for stable v3.9+

Reported-by: arjan@linux.intel.com
Reported-by: dsmythies@telus.net
Tested-by: arjan@linux.intel.com
Tested-by: dsmythies@telus.net

CC: stable@vger.kernel.org
CC: dsmythies@telus.net
Signed-off-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
---
 drivers/cpufreq/intel_pstate.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c
index 07f2840..6d6a0b4 100644
--- a/drivers/cpufreq/intel_pstate.c
+++ b/drivers/cpufreq/intel_pstate.c
@@ -103,10 +103,10 @@ struct pstate_adjust_policy {
 static struct pstate_adjust_policy default_policy = {
 	.sample_rate_ms = 10,
 	.deadband = 0,
-	.setpoint = 109,
-	.p_gain_pct = 17,
+	.setpoint = 97,
+	.p_gain_pct = 20,
 	.d_gain_pct = 0,
-	.i_gain_pct = 4,
+	.i_gain_pct = 0,
 };
 
 struct perf_limits {
@@ -468,12 +468,12 @@ static inline void intel_pstate_set_sample_time(struct cpudata *cpu)
 static inline int intel_pstate_get_scaled_busy(struct cpudata *cpu)
 {
 	int32_t busy_scaled;
-	int32_t core_busy, turbo_pstate, current_pstate;
+	int32_t core_busy, max_pstate, current_pstate;
 
 	core_busy = int_tofp(cpu->samples[cpu->sample_ptr].core_pct_busy);
-	turbo_pstate = int_tofp(cpu->pstate.turbo_pstate);
+	max_pstate = int_tofp(cpu->pstate.max_pstate);
 	current_pstate = int_tofp(cpu->pstate.current_pstate);
-	busy_scaled = mul_fp(core_busy, div_fp(turbo_pstate, current_pstate));
+	busy_scaled = mul_fp(core_busy, div_fp(max_pstate, current_pstate));
 
 	return fp_toint(busy_scaled);
 }
-- 
1.8.3.1