Re: [PATCH] fs:remove a no needed mutex_unlock in devtmpfs

[Wanlong Gao <gaowanlong@cn.fujitsu.com>]

On 07/29/2011 05:03 PM, Li Zefan wrote:
> Wanlong Gao wrote:
>> According ed75e95de574c99575e5f3e1d9ca59ea8c12a9cb, Al killed the
>> lookup_create(), this function will return with an mutex locked,
>> and it called by dev_mkdir() previously. then unlock this mutex
>> in dev_mkdir(), but when lookup_create() is replaced by kern_path_create(),
>> no locked mutex returned, so no need to unlock the mutex.
>>
>> Al, is it right?
>
> no.. kern_path_create() returns with i_mutex held.
yeah, I hasn't seen the *return* when *dentry* is gotten successfully.
Sorry for the noise
>
>> Thanks
>>
>> Signed-off-by: Wanlong Gao<gaowanlong@cn.fujitsu.com>
>> ---
>>   drivers/base/devtmpfs.c |    1 -
>>   1 files changed, 0 insertions(+), 1 deletions(-)
>>
>> diff --git a/drivers/base/devtmpfs.c b/drivers/base/devtmpfs.c
>> index b89fffc..5384883 100644
>> --- a/drivers/base/devtmpfs.c
>> +++ b/drivers/base/devtmpfs.c
>> @@ -157,7 +157,6 @@ static int dev_mkdir(const char *name, mode_t mode)
>>   		/* mark as kernel-created inode */
>>   		dentry->d_inode->i_private =&thread;
>>   	dput(dentry);
>> -	mutex_unlock(&path.dentry->d_inode->i_mutex);
>>   	path_put(&path);
>>   	return err;
>>   }
>


-- 
Thanks
Best Regards
Wanlong Gao