Re: [PATCH 22/31] fs/ntfs: use kmemdup rather than duplicating its implementation

[Andrzej Hajda <a.hajda@samsung.com>]

On 09/23/2015 12:21 PM, Anton Altaparmakov wrote:
> Hi Andrzej,
>
> Thanks for your patch.  It looks fine though I don't quite see the point of it to be honest.
>
> It actually adds an additional function call (kmemdup() is not inline) just to save 1 line of source code in the driver and I don't think it improves readability or anything so why bother?  What does it gain?

kmemdup replaces combo (kmalloc + memdup) with one call.
The patch follows quite common practice of abstracting out common patterns.

Regards
Andrzej

>
> Best regards,
>
> 	Anton
>
>> On 7 Aug 2015, at 08:59, Andrzej Hajda <a.hajda@samsung.com> wrote:
>>
>> The patch was generated using fixed coccinelle semantic patch
>> scripts/coccinelle/api/memdup.cocci [1].
>>
>> [1]: http://permalink.gmane.org/gmane.linux.kernel/2014320
>>
>> Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
>> ---
>> fs/ntfs/dir.c | 7 +++----
>> 1 file changed, 3 insertions(+), 4 deletions(-)
>>
>> diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
>> index 9e38daf..2b7fef0 100644
>> --- a/fs/ntfs/dir.c
>> +++ b/fs/ntfs/dir.c
>> @@ -1172,14 +1172,13 @@ static int ntfs_readdir(struct file *file, struct dir_context *actor)
>> 	 * map the mft record without deadlocking.
>> 	 */
>> 	rc = le32_to_cpu(ctx->attr->data.resident.value_length);
>> -	ir = kmalloc(rc, GFP_NOFS);
>> +	/* Copy the index root value (it has been verified in read_inode). */
>> +	ir = kmemdup((u8 *)ctx->attr + le16_to_cpu(ctx->attr->data.resident.value_offset),
>> +		     rc, GFP_NOFS);
>> 	if (unlikely(!ir)) {
>> 		err = -ENOMEM;
>> 		goto err_out;
>> 	}
>> -	/* Copy the index root value (it has been verified in read_inode). */
>> -	memcpy(ir, (u8*)ctx->attr +
>> -			le16_to_cpu(ctx->attr->data.resident.value_offset), rc);
>> 	ntfs_attr_put_search_ctx(ctx);
>> 	unmap_mft_record(ndir);
>> 	ctx = NULL;
>> -- 
>> 1.9.1