On 2024/12/5 11:40, wangzijie wrote:
On 2024/12/4 20:23, wangzijie wrote:
[...]
Hi, yong
Thanks for sharing what you proposed to do. By the way, I noticed that
when setattr, read_all_xattr will set xattr header's magic and refcount,
but it seems we don't check these values in header(kerne
>On 2024/12/4 20:23, wangzijie wrote:
>> When we check inode which just has inline xattr data, we copy
>> inline xattr data from inode, check it(maybe fix it) and copy
>> it again to inode. We can check and fix xattr inplace for this
>> kind of inode to reduce memcpy times.
>>
>> Signed-off-by: wa
On 2024/12/4 20:23, wangzijie wrote:
When we check inode which just has inline xattr data, we copy
inline xattr data from inode, check it(maybe fix it) and copy
it again to inode. We can check and fix xattr inplace for this
kind of inode to reduce memcpy times.
Signed-off-by: wangzijie
---
When we check inode which just has inline xattr data, we copy
inline xattr data from inode, check it(maybe fix it) and copy
it again to inode. We can check and fix xattr inplace for this
kind of inode to reduce memcpy times.
Signed-off-by: wangzijie
---
fsck/fsck.c | 18 ++
1 fil
