[patch] drm: sysfs files error handling

[patch] drm: sysfs files error handling

[Dan Carpenter]

In the original code we used "j" as an iterator but we used "i" as an
index.

-               for (j = 0; j < i; j++)
-                       device_remove_file(&connector->kdev,
-                                          &connector_attrs[i]);

Smatch complained about that because "i" was potentially passed the end
of the array.  Which makes sense if we should be using "j" there.

I also thought that we should remove the files for &connector_attrs_opt1
but to do that I had to add separate iterators for &connector_attrs and
&connector_attrs_opt1.

Signed-off-by: Dan Carpenter <error27@gmail.com>
---
Found by static analysis and compile tested only.

diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 014ce24..f144487 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -353,7 +353,10 @@ static struct bin_attribute edid_attr = {
 int drm_sysfs_connector_add(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	int ret = 0, i, j;
+	int attr_cnt = 0;
+	int opt_cnt = 0;
+	int i;
+	int ret = 0;
 
 	/* We shouldn't get called more than once for the same connector */
 	BUG_ON(device_is_registered(&connector->kdev));
@@ -376,8 +379,8 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 
 	/* Standard attributes */
 
-	for (i = 0; i < ARRAY_SIZE(connector_attrs); i++) {
-		ret = device_create_file(&connector->kdev, &connector_attrs[i]);
+	for (attr_cnt = 0; attr_cnt < ARRAY_SIZE(connector_attrs); attr_cnt++) {
+		ret = device_create_file(&connector->kdev, &connector_attrs[attr_cnt]);
 		if (ret)
 			goto err_out_files;
 	}
@@ -393,8 +396,8 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 		case DRM_MODE_CONNECTOR_SVIDEO:
 		case DRM_MODE_CONNECTOR_Component:
 		case DRM_MODE_CONNECTOR_TV:
-			for (i = 0; i < ARRAY_SIZE(connector_attrs_opt1); i++) {
-				ret = device_create_file(&connector->kdev, &connector_attrs_opt1[i]);
+			for (opt_cnt = 0; opt_cnt < ARRAY_SIZE(connector_attrs_opt1); opt_cnt++) {
+				ret = device_create_file(&connector->kdev, &connector_attrs_opt1[opt_cnt]);
 				if (ret)
 					goto err_out_files;
 			}
@@ -413,10 +416,10 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 	return 0;
 
 err_out_files:
-	if (i > 0)
-		for (j = 0; j < i; j++)
-			device_remove_file(&connector->kdev,
-					   &connector_attrs[i]);
+	for (i = 0; i < opt_cnt; i++)
+		device_remove_file(&connector->kdev, &connector_attrs_opt1[i]);
+	for (i = 0; i < attr_cnt; i++)
+		device_remove_file(&connector->kdev, &connector_attrs[i]);
 	device_unregister(&connector->kdev);
 
 out:

[patch] drm: sysfs files error handling

[Dan Carpenter]

In the original code we used "j" as an iterator but we used "i" as an
index.

-               for (j = 0; j < i; j++)
-                       device_remove_file(&connector->kdev,
-                                          &connector_attrs[i]);

Smatch complained about that because "i" was potentially passed the end
of the array.  Which makes sense if we should be using "j" there.

I also thought that we should remove the files for &connector_attrs_opt1
but to do that I had to add separate iterators for &connector_attrs and
&connector_attrs_opt1.

Signed-off-by: Dan Carpenter <error27@gmail.com>
---
Found by static analysis and compile tested only.

diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 014ce24..f144487 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -353,7 +353,10 @@ static struct bin_attribute edid_attr = {
 int drm_sysfs_connector_add(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
-	int ret = 0, i, j;
+	int attr_cnt = 0;
+	int opt_cnt = 0;
+	int i;
+	int ret = 0;
 
 	/* We shouldn't get called more than once for the same connector */
 	BUG_ON(device_is_registered(&connector->kdev));
@@ -376,8 +379,8 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 
 	/* Standard attributes */
 
-	for (i = 0; i < ARRAY_SIZE(connector_attrs); i++) {
-		ret = device_create_file(&connector->kdev, &connector_attrs[i]);
+	for (attr_cnt = 0; attr_cnt < ARRAY_SIZE(connector_attrs); attr_cnt++) {
+		ret = device_create_file(&connector->kdev, &connector_attrs[attr_cnt]);
 		if (ret)
 			goto err_out_files;
 	}
@@ -393,8 +396,8 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 		case DRM_MODE_CONNECTOR_SVIDEO:
 		case DRM_MODE_CONNECTOR_Component:
 		case DRM_MODE_CONNECTOR_TV:
-			for (i = 0; i < ARRAY_SIZE(connector_attrs_opt1); i++) {
-				ret = device_create_file(&connector->kdev, &connector_attrs_opt1[i]);
+			for (opt_cnt = 0; opt_cnt < ARRAY_SIZE(connector_attrs_opt1); opt_cnt++) {
+				ret = device_create_file(&connector->kdev, &connector_attrs_opt1[opt_cnt]);
 				if (ret)
 					goto err_out_files;
 			}
@@ -413,10 +416,10 @@ int drm_sysfs_connector_add(struct drm_connector *connector)
 	return 0;
 
 err_out_files:
-	if (i > 0)
-		for (j = 0; j < i; j++)
-			device_remove_file(&connector->kdev,
-					   &connector_attrs[i]);
+	for (i = 0; i < opt_cnt; i++)
+		device_remove_file(&connector->kdev, &connector_attrs_opt1[i]);
+	for (i = 0; i < attr_cnt; i++)
+		device_remove_file(&connector->kdev, &connector_attrs[i]);
 	device_unregister(&connector->kdev);
 
 out:

Re: [patch] drm: sysfs files error handling

[Andi Kleen]

, Dan Carpenter wrote:
>
> In the original code we used "j" as an iterator but we used "i" as an
> index.
>
> -               for (j = 0; j<  i; j++)
> -                       device_remove_file(&connector->kdev,
> -&connector_attrs[i]);

I guess this really should be a attribute group anyways?

Typically when there's such a open coded loop it means the wrong
interfaces are being used.

-Andi

Re: [patch] drm: sysfs files error handling

[Andi Kleen]

, Dan Carpenter wrote:
>
> In the original code we used "j" as an iterator but we used "i" as an
> index.
>
> -               for (j = 0; j<  i; j++)
> -                       device_remove_file(&connector->kdev,
> -&connector_attrs[i]);

I guess this really should be a attribute group anyways?

Typically when there's such a open coded loop it means the wrong
interfaces are being used.

-Andi

Re: [patch] drm: sysfs files error handling

[Andi Kleen]

, Dan Carpenter wrote:
>
> In the original code we used "j" as an iterator but we used "i" as an
> index.
>
> -               for (j = 0; j<  i; j++)
> -                       device_remove_file(&connector->kdev,
> -&connector_attrs[i]);

I guess this really should be a attribute group anyways?

Typically when there's such a open coded loop it means the wrong
interfaces are being used.

-Andi


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [patch] drm: sysfs files error handling

[Dan Carpenter]

On Sun, Mar 28, 2010 at 10:55:14PM +0200, Andi Kleen wrote:
> , Dan Carpenter wrote:
>>
>> In the original code we used "j" as an iterator but we used "i" as an
>> index.
>>
>> -               for (j = 0; j<  i; j++)
>> -                       device_remove_file(&connector->kdev,
>> -&connector_attrs[i]);
>
> I guess this really should be a attribute group anyways?
>
> Typically when there's such a open coded loop it means the wrong
> interfaces are being used.

My graphics card is crap and doesn't use this code at all.  I'd feel 
uncomfortable those changes without being able to test it.

So while, you are probably right, someone else should probably do that.

regards,
dan carpenter

> -Andi

Re: [patch] drm: sysfs files error handling

[Dan Carpenter]

On Sun, Mar 28, 2010 at 10:55:14PM +0200, Andi Kleen wrote:
> , Dan Carpenter wrote:
>>
>> In the original code we used "j" as an iterator but we used "i" as an
>> index.
>>
>> -               for (j = 0; j<  i; j++)
>> -                       device_remove_file(&connector->kdev,
>> -&connector_attrs[i]);
>
> I guess this really should be a attribute group anyways?
>
> Typically when there's such a open coded loop it means the wrong
> interfaces are being used.

My graphics card is crap and doesn't use this code at all.  I'd feel 
uncomfortable those changes without being able to test it.

So while, you are probably right, someone else should probably do that.

regards,
dan carpenter

> -Andi