Skip to contentSkip to conversation

OMAP2+: CPUfreq: Allow the CPU scaling when secondary CPUs are offline.

PATCH In review Santosh Shilimkar 2 participants 2 messages Jun 2, 2011 – Jun 3, 2011

Subsystems (3)
Arm (32-bit) Arm sub-architectures OMAP2+ support
lore
Share & tools
lore

Santosh Shilimkar

author
#0 lore raw

PATCH 1/1 OMAP2+: CPUfreq: Allow the CPU scaling when secondary CPUs are offline.

Missed Kevin in cc. :(
Sorry about that.

-------- Original Message --------
Subject: [PATCH] OMAP2+: CPUfreq: Allow the CPU scaling when secondary CPUs are offline.
Date: Thu, 2 Jun 2011 20:21:10 +0530
From: Santosh Shilimkar <[email protected]>
To: <[email protected]>
CC: Santosh Shilimkar <[email protected]>

Current OMAP2PLUS CPUfreq tagret() functions returns when all the CPU's are not online. This will break DVFS when secondary CPUs are offlined.

The intention of that check was just avoid CPU frequency change during the window when CPU becomes online but it's cpufreq_init is not done yet.

Fix the check accordingly.

Thanks for Nishant Menon <[email protected]> for reporting it.

Signed-off-by Santosh Shilimkar <[email protected]>
Reported-by Nishanth Menon <[email protected]>
Tested-by Vishwanath BS <[email protected]>

---
There were some question of making the variable atomic etc in an internal discussion. After some thinking, I realised there is no need of that since this is just a counter which maintains the count for online_cpus = cpufreq_init_cpus.

 arch/arm/mach-omap2/omap2plus-cpufreq.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)
arch/arm/mach-omap2/omap2plus-cpufreq.c +8 -1
index 33a91ec..909bfcb 100644
@@ -44,6 +44,7 @@ static struct cpufreq_frequency_table *freq_table;
 static struct clk *mpu_clk;
 static char *mpu_clk_name;
 static struct device *mpu_dev;
+static int cpus_initialized;
 
 static int omap_verify_speed(struct cpufreq_policy *policy)
 {

@@ -82,7 +83,7 @@ static int omap_target(struct cpufreq_policy *policy,
 	struct cpufreq_freqs freqs;
 
 	/* Changes not allowed until all CPUs are online */
-	if (is_smp() && (num_online_cpus() < NR_CPUS))
+	if (is_smp() && (cpus_initialized < num_online_cpus()))
 		return ret;
 
 	/* Ensure desired rate is within allowed range.  Some govenors

@@ -194,6 +195,8 @@ static int __cpuinit omap_cpu_init(struct cpufreq_policy *policy)
 		policy->shared_type = CPUFREQ_SHARED_TYPE_ANY;
 		cpumask_or(cpumask, cpumask_of(policy->cpu), cpumask);
 		cpumask_copy(policy->cpus, cpumask);
+		cpus_initialized++;
+		smp_wmb();
 	}
 
 	/* FIXME: what's the actual transition time? */

@@ -206,6 +209,10 @@ static int omap_cpu_exit(struct cpufreq_policy *policy)
 {
 	clk_exit_cpufreq_table(&freq_table);
 	clk_put(mpu_clk);
+	if (is_smp()) {
+		cpus_initialized--;
+		smp_wmb();
+	}
 	return 0;
 }
Signature · 1 line
1.6.0.4
#1 lore raw

Hi,

Quoted 49 lines · ext Santosh Shilimkar

Missed Kevin in cc. :(
Sorry about that.

-------- Original Message --------
Subject: [PATCH] OMAP2+: CPUfreq: Allow the CPU scaling when secondary CPUs are offline.
Date: Thu, 2 Jun 2011 20:21:10 +0530
From: Santosh Shilimkar <[email protected]>
To: <[email protected]>
CC: Santosh Shilimkar <[email protected]>

Current OMAP2PLUS CPUfreq tagret() functions returns when all the CPU's are not online. This will break DVFS when secondary CPUs are offlined.

The intention of that check was just avoid CPU frequency change during the window when CPU becomes online but it's cpufreq_init is not done yet.

Fix the check accordingly.

Thanks for Nishant Menon <[email protected]> for reporting it.

Signed-off-by Santosh Shilimkar <[email protected]>
Reported-by Nishanth Menon <[email protected]>
Tested-by Vishwanath BS <[email protected]>

---
There were some question of making the variable atomic etc in an internal discussion. After some thinking, I realised there is no need of that since this is just a counter which maintains the count for online_cpus = cpufreq_init_cpus.

  arch/arm/mach-omap2/omap2plus-cpufreq.c |    9 ++++++++-
  1 files changed, 8 insertions(+), 1 deletions(-)
arch/arm/mach-omap2/omap2plus-cpufreq.c +1 -0
index 33a91ec..909bfcb 100644
@@ -44,6 +44,7 @@ static struct cpufreq_frequency_table *freq_table;
  static struct clk *mpu_clk;
  static char *mpu_clk_name;
  static struct device *mpu_dev;
+static int cpus_initialized;
 
  static int omap_verify_speed(struct cpufreq_policy *policy)
  {

arch/arm/mach-omap2/omap2plus-cpufreq.c static int omap_target(struct cpufreq_policy *policy, · line 85

Quoted 3 lines
arch/arm/mach-omap2/omap2plus-cpufreq.c +0 -0
@@ -82,7 +83,7 @@ static int omap_target(struct cpufreq_policy *policy,
  	struct cpufreq_freqs freqs;
 
  	/* Changes not allowed until all CPUs are online */

This comment are no more relevant and should be changed also...

Quoted 16 lines
arch/arm/mach-omap2/omap2plus-cpufreq.c +3 -1
-	if (is_smp() && (num_online_cpus() < NR_CPUS))
+	if (is_smp() && (cpus_initialized < num_online_cpus()))
  		return ret;
 
  	/* Ensure desired rate is within allowed range.  Some govenors

@@ -194,6 +195,8 @@ static int __cpuinit omap_cpu_init(struct 
 cpufreq_policy *policy)
  		policy->shared_type = CPUFREQ_SHARED_TYPE_ANY;
  		cpumask_or(cpumask, cpumask_of(policy->cpu), cpumask);
  		cpumask_copy(policy->cpus, cpumask);
+		cpus_initialized++;
+		smp_wmb();
  	}

/* FIXME: what's the actual transition time? */

arch/arm/mach-omap2/omap2plus-cpufreq.c static int omap_cpu_exit(struct cpufreq_policy *policy) · line 218

Quoted 10 lines
arch/arm/mach-omap2/omap2plus-cpufreq.c +4 -0
@@ -206,6 +209,10 @@ static int omap_cpu_exit(struct cpufreq_policy *policy)
  {
  	clk_exit_cpufreq_table(&freq_table);
  	clk_put(mpu_clk);
+	if (is_smp()) {
+		cpus_initialized--;
+		smp_wmb();
+	}
  	return 0;
  }

Best Regards,
Igor Dmitriev