Re: [PATCH] kobject: avoid unnecessary assignment of kobj->parent

[Wei Yang <weiyang@linux.vnet.ibm.com>]

On Sat, Jun 06, 2015 at 01:22:30PM +0900, Greg KH wrote:
>On Thu, Jun 04, 2015 at 10:13:50AM +0800, Wei Yang wrote:
>> kobj->parent is changed only when it is NULL originally.
>> 
>> This patch moves the assignment in the "if" to avoid unnecessary assignment.
>> 
>> Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
>> ---
>>  lib/kobject.c |    3 +--
>>  1 file changed, 1 insertion(+), 2 deletions(-)
>> 
>> diff --git a/lib/kobject.c b/lib/kobject.c
>> index 03d4ab3..37b10bb 100644
>> --- a/lib/kobject.c
>> +++ b/lib/kobject.c
>> @@ -216,9 +216,8 @@ static int kobject_add_internal(struct kobject *kobj)
>>  	/* join kset if set, use it as parent if we do not already have one */
>>  	if (kobj->kset) {
>>  		if (!parent)
>> -			parent = kobject_get(&kobj->kset->kobj);
>> +			parent = kobj->parent = kobject_get(&kobj->kset->kobj);
>>  		kobj_kset_join(kobj);
>> -		kobj->parent = parent;
>
>What is this "fixing"?  I hate multiple assignments on the same line,
>it makes it harder to read and understand.
>

Hi Greg

No. This one is not "fixing" anything. What it does is to not assign the
kobj->parent again when it is not changed.

No, this is not a "fixing". So it doesn't help on the functionality or
improve the performance. It just wants to make it clear that the kobj->parent
is changed when originally it is NULL, not for other cases. Hope this would
help a little.

I didn't noticed that you don't like this format, so I didn't want to offend
you intendedly. How about move "kobj->parent = parent" above and add a brace
to the if? For example like this:

 	/* join kset if set, use it as parent if we do not already have one */
 	if (kobj->kset) {
-		if (!parent)
+		if (!parent) {
			parent = kobject_get(&kobj->kset->kobj);
+			kobj->parent = parent;
+      		}
 		kobj_kset_join(kobj);
-		kobj->parent = parent;

>greg k-h

-- 
Richard Yang
Help you, Help me