Re: [PATCH] intel_ips: quieten "power or thermal limit exceeded" messages

[Joe Perches <joe@perches.com>]

On Thu, 2010-08-26 at 21:11 -0300, Cesar Eduardo Barros wrote:
> Em 26-08-2010 20:33, Joe Perches escreveu:
> > On Thu, 26 Aug 2010, Cesar Eduardo Barros wrote:
> >> On my Dell Inspiron N4010, one of these messages is printed every five
> >> seconds. Change both to dev_dbg to quieten them even more.
> > I think you should instead fix your hardware or maybe change
> > your thermal throttling settings.

I was probably a bit hasty in writing that.

> Is there a way to know if all this is just an oddness of this model, or 
> if there is something which is not working quite right?
> 
> (All the output above is from 2.6.35.3; I am not running 2.3.36-rc2+ 
> right now because it hangs on resume, and I have not yet had the time to 
> look at it.)

Perhaps you might try this patch and get a bit more information.

It seems a sensible patch and perhaps should be applied anyway.

Signed-off-by: Joe Perches <joe@perches.com>
---
 drivers/platform/x86/intel_ips.c |   58 +++++++++++++++++++++++++++++--------
 1 files changed, 45 insertions(+), 13 deletions(-)

diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c
index 9024480..73f9ad1 100644
--- a/drivers/platform/x86/intel_ips.c
+++ b/drivers/platform/x86/intel_ips.c
@@ -598,17 +598,36 @@ static bool mcp_exceeded(struct ips_driver *ips)
 {
 	unsigned long flags;
 	bool ret = false;
+	u16 mcp_avg_temp;
+	u16 mcp_temp_limit;
+	u16 mcp_power_limit;
+	u32 cpu_avg_power;
+	u32 mch_avg_power;
 
 	spin_lock_irqsave(&ips->turbo_status_lock, flags);
-	if (ips->mcp_avg_temp > (ips->mcp_temp_limit * 100))
-		ret = true;
-	if (ips->cpu_avg_power + ips->mch_avg_power > ips->mcp_power_limit)
-		ret = true;
+
+	mcp_avg_temp = ips->mcp_avg_temp;
+	mcp_temp_limit = ips->mcp_temp_limit;
+	mcp_power_limit = ips->mcp_power_limit;
+	cpu_avg_power = ips->cpu_avg_power;
+	mch_avg_power = ips->mch_avg_power;
+
 	spin_unlock_irqrestore(&ips->turbo_status_lock, flags);
 
-	if (ret)
+	if ((cpu_avg_power + mch_avg_power) > mcp_power_limit) {
 		dev_info(&ips->dev->dev,
-			 "MCP power or thermal limit exceeded\n");
+			 "MCP power limit %d exceeded: %d\n",
+			 mcp_power_limit,
+			 cpu_avg_power + mch_avg_power);
+		ret = true;
+	}
+	if (mcp_avg_temp > (mcp_temp_limit * 100)) {
+		dev_info(&ips->dev->dev,
+			 "MCP thermal limit %d exceeded: %d\n",
+			 mcp_temp_limit * 100,
+			 mcp_avg_temp);
+		ret = true;
+	}
 
 	return ret;
 }
@@ -623,20 +642,33 @@ static bool mcp_exceeded(struct ips_driver *ips)
 static bool cpu_exceeded(struct ips_driver *ips, int cpu)
 {
 	unsigned long flags;
-	int avg;
 	bool ret = false;
+	int avg;
+	int core_temp_limit;
+	u16 core_power_limit;
+	u32 cpu_avg_power;
 
 	spin_lock_irqsave(&ips->turbo_status_lock, flags);
+
 	avg = cpu ? ips->ctv2_avg_temp : ips->ctv1_avg_temp;
-	if (avg > (ips->limits->core_temp_limit * 100))
-		ret = true;
-	if (ips->cpu_avg_power > ips->core_power_limit * 100)
-		ret = true;
+	core_temp_limit = ips->limits->core_temp_limit;
+	core_power_limit = ips->core_power_limit;
+	cpu_avg_power = ips->cpu_avg_power;
+
 	spin_unlock_irqrestore(&ips->turbo_status_lock, flags);
 
-	if (ret)
+	if (cpu_avg_power > (core_power_limit * 100)) {
+		dev_info(&ips->dev->dev,
+			 "CPU power limit %d exceeded: %d\n",
+			 cpu_avg_power, core_power_limit * 100);
+		ret = true;
+	}
+	if (avg > (core_temp_limit * 100)) {
 		dev_info(&ips->dev->dev,
-			 "CPU power or thermal limit exceeded\n");
+			 "CPU thermal limit %d exceeded: %d\n",
+			 core_temp_limit * 100, avg);
+		ret = true;
+	}
 
 	return ret;
 }