Skip to contentSkip to conversation

hwmon: add required idr locking

PATCH New Mark M. Hoffman 2 participants 4 messages Mar 5, 2006 Hardware monitoring

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hwmon: add required idr locking

Add required locking around idr_ routines, retry the idr_pre_get/idr_get_new pair properly, and sprinkle in some likely/unlikely for good measure.

(Lack of idr locking didn't hurt when all callers were I2C clients, as the i2c-core serialized for us anyway. Now that we have non I2C hwmon drivers, this is truly necessary.)

This patch should be considered for 2.6.16.

Signed-off-by Mark M. Hoffman <mhoffman at lightlink.com>
drivers/hwmon/hwmon.c +19 -5
@@ -17,6 +17,7 @@
 #include <linux/idr.h>
 #include <linux/hwmon.h>
 #include <linux/gfp.h>
+#include <linux/spinlock.h>
 
 #define HWMON_ID_PREFIX "hwmon"
 #define HWMON_ID_FORMAT HWMON_ID_PREFIX "%d"
@@ -24,6 +25,7 @@
 static struct class *hwmon_class;
 
 static DEFINE_IDR(hwmon_idr);
+static DEFINE_SPINLOCK(idr_lock);
 
 /**
  * hwmon_device_register - register w/ hwmon sysfs class
@@ -37,20 +39,30 @@ static DEFINE_IDR(hwmon_idr);
 struct class_device *hwmon_device_register(struct device *dev)
 {
 	struct class_device *cdev;
-	int id;
+	int id, err;
 
-	if (idr_pre_get(&hwmon_idr, GFP_KERNEL) = 0)
+again:
+	if (unlikely(idr_pre_get(&hwmon_idr, GFP_KERNEL) = 0))
 		return ERR_PTR(-ENOMEM);
 
-	if (idr_get_new(&hwmon_idr, NULL, &id) < 0)
+	spin_lock(&idr_lock);
+	err = idr_get_new(&hwmon_idr, NULL, &id);
+	spin_unlock(&idr_lock);
+
+	if (unlikely(err = -EAGAIN))
+		goto again;
+	else if (unlikely(err))
 		return ERR_PTR(-ENOMEM);
 
 	id = id & MAX_ID_MASK;
 	cdev = class_device_create(hwmon_class, NULL, MKDEV(0,0), dev,
 					HWMON_ID_FORMAT, id);
 
-	if (IS_ERR(cdev))
+	if (unlikely(IS_ERR(cdev))) {
+		spin_lock(&idr_lock);
 		idr_remove(&hwmon_idr, id);
+		spin_unlock(&idr_lock);
+	}
 
 	return cdev;
 }
@@ -64,9 +76,11 @@ void hwmon_device_unregister(struct clas
 {
 	int id;
 
-	if (sscanf(cdev->class_id, HWMON_ID_FORMAT, &id) = 1) {
+	if (likely(sscanf(cdev->class_id, HWMON_ID_FORMAT, &id) = 1)) {
 		class_device_unregister(cdev);
+		spin_lock(&idr_lock);
 		idr_remove(&hwmon_idr, id);
+		spin_unlock(&idr_lock);
 	} else
 		dev_dbg(cdev->dev,
 			"hwmon_device_unregister() failed: bad class ID!\n");
Signature · 2 lines
Mark M. Hoffman
mhoffman at lightlink.com
#1 lore raw

PATCH 1/1 hwmon: add required idr locking

Hi Mark,

Quoted 8 lines

Add required locking around idr_ routines, retry the idr_pre_get/idr_get_new pair properly, and sprinkle in some likely/unlikely for good measure.

(Lack of idr locking didn't hurt when all callers were I2C clients, as the i2c-core serialized for us anyway. Now that we have non I2C hwmon drivers, this is truly necessary.)

This patch should be considered for 2.6.16.

Thanks for reporting the problem and proposing a fix. It's all correct as far as I can tell, but I'd still have a few comments:

drivers/hwmon/hwmon.c static DEFINE_IDR(hwmon_idr); · line 55

Quoted 9 lines
drivers/hwmon/hwmon.c +7 -1
-	if (idr_get_new(&hwmon_idr, NULL, &id) < 0)
+	spin_lock(&idr_lock);
+	err = idr_get_new(&hwmon_idr, NULL, &id);
+	spin_unlock(&idr_lock);
+
+	if (unlikely(err = -EAGAIN))
+		goto again;
+	else if (unlikely(err))
 		return ERR_PTR(-ENOMEM);

It has just occured to me that the error value we are returning is not correct. Why -ENOMEM? As far as I know, idr_get_new does not allocate memory, by design, so it's probably the more inaccurate error value we could return.

Why don't we just return ERR_PTR(err) instead?

I realize that this isn't strictly related with your patch, the old code returned just the same, but while we're here changing that code...

drivers/hwmon/hwmon.c static DEFINE_IDR(hwmon_idr); · line 61

Quoted 2 lines
drivers/hwmon/hwmon.c +1 -1
-	if (IS_ERR(cdev))
+	if (unlikely(IS_ERR(cdev))) {

IS_ERR() already includes an unlikely(), so this is redundant.

And you're right, let's try to get that patch in 2.6.16.

Thanks,

Signature · 1 line
Jean Delvare
#2 lore raw

PATCH 1/1 hwmon: add required idr locking

In reply to Mark M. Hoffman

Hi Jean:

Quoted 27 lines · Jean Delvare
Quoted 8 lines

Add required locking around idr_ routines, retry the idr_pre_get/idr_get_new pair properly, and sprinkle in some likely/unlikely for good measure.

(Lack of idr locking didn't hurt when all callers were I2C clients, as the i2c-core serialized for us anyway. Now that we have non I2C hwmon drivers, this is truly necessary.)

This patch should be considered for 2.6.16.

Thanks for reporting the problem and proposing a fix. It's all correct as far as I can tell, but I'd still have a few comments:

drivers/hwmon/hwmon.c static DEFINE_IDR(hwmon_idr); · line 55

Quoted 9 lines
drivers/hwmon/hwmon.c +7 -1
-	if (idr_get_new(&hwmon_idr, NULL, &id) < 0)
+	spin_lock(&idr_lock);
+	err = idr_get_new(&hwmon_idr, NULL, &id);
+	spin_unlock(&idr_lock);
+
+	if (unlikely(err = -EAGAIN))
+		goto again;
+	else if (unlikely(err))
 		return ERR_PTR(-ENOMEM);

It has just occured to me that the error value we are returning is not correct. Why -ENOMEM? As far as I know, idr_get_new does not allocate memory, by design, so it's probably the more inaccurate error value we could return.

Quoted 1 line

Why don't we just return ERR_PTR(err) instead?

OK.

Quoted 7 lines

I realize that this isn't strictly related with your patch, the old code returned just the same, but while we're here changing that code...

drivers/hwmon/hwmon.c static DEFINE_IDR(hwmon_idr); · line 61

Quoted 2 lines
drivers/hwmon/hwmon.c +1 -1
-	if (IS_ERR(cdev))
+	if (unlikely(IS_ERR(cdev))) {

IS_ERR() already includes an unlikely(), so this is redundant.

So it does - struck.

New boot-tested (on x86) patch to follow...

Regards,

Signature · 2 lines
Mark M. Hoffman
mhoffman at lightlink.com
#3 lore raw

PATCH 1/1 hwmon: add required idr locking

In reply to Mark M. Hoffman

Add required locking around idr_ routines, retry the idr_pre_get/idr_get_new pair properly, and sprinkle in some likely/unlikely for good measure.

(Lack of idr locking didn't hurt when all callers were I2C clients, as the i2c-core serialized for us anyway. Now that we have non I2C hwmon drivers, this is truly necessary.)

This patch should be considered for 2.6.16.

Signed-off-by Mark M. Hoffman <mhoffman at lightlink.com>
drivers/hwmon/hwmon.c +20 -6
@@ -17,6 +17,7 @@
 #include <linux/idr.h>
 #include <linux/hwmon.h>
 #include <linux/gfp.h>
+#include <linux/spinlock.h>
 
 #define HWMON_ID_PREFIX "hwmon"
 #define HWMON_ID_FORMAT HWMON_ID_PREFIX "%d"
@@ -24,6 +25,7 @@
 static struct class *hwmon_class;
 
 static DEFINE_IDR(hwmon_idr);
+static DEFINE_SPINLOCK(idr_lock);
 
 /**
  * hwmon_device_register - register w/ hwmon sysfs class
@@ -37,20 +39,30 @@ static DEFINE_IDR(hwmon_idr);
 struct class_device *hwmon_device_register(struct device *dev)
 {
 	struct class_device *cdev;
-	int id;
+	int id, err;
 
-	if (idr_pre_get(&hwmon_idr, GFP_KERNEL) = 0)
+again:
+	if (unlikely(idr_pre_get(&hwmon_idr, GFP_KERNEL) = 0))
 		return ERR_PTR(-ENOMEM);
 
-	if (idr_get_new(&hwmon_idr, NULL, &id) < 0)
-		return ERR_PTR(-ENOMEM);
+	spin_lock(&idr_lock);
+	err = idr_get_new(&hwmon_idr, NULL, &id);
+	spin_unlock(&idr_lock);
+
+	if (unlikely(err = -EAGAIN))
+		goto again;
+	else if (unlikely(err))
+		return ERR_PTR(err);
 
 	id = id & MAX_ID_MASK;
 	cdev = class_device_create(hwmon_class, NULL, MKDEV(0,0), dev,
 					HWMON_ID_FORMAT, id);
 
-	if (IS_ERR(cdev))
+	if (IS_ERR(cdev)) {
+		spin_lock(&idr_lock);
 		idr_remove(&hwmon_idr, id);
+		spin_unlock(&idr_lock);
+	}
 
 	return cdev;
 }
@@ -64,9 +76,11 @@ void hwmon_device_unregister(struct clas
 {
 	int id;
 
-	if (sscanf(cdev->class_id, HWMON_ID_FORMAT, &id) = 1) {
+	if (likely(sscanf(cdev->class_id, HWMON_ID_FORMAT, &id) = 1)) {
 		class_device_unregister(cdev);
+		spin_lock(&idr_lock);
 		idr_remove(&hwmon_idr, id);
+		spin_unlock(&idr_lock);
 	} else
 		dev_dbg(cdev->dev,
 			"hwmon_device_unregister() failed: bad class ID!\n");
Signature · 2 lines
Mark M. Hoffman
mhoffman at lightlink.com