on 2022/4/18 11:04, Xiubo Li wrote:
>
> On 4/15/22 7:02 PM, Yang Xu wrote:
>> Since vfs has stripped S_ISGID, we don't need this code any more.
>>
>> Signed-off-by: Yang Xu <xuyang2018.jy@xxxxxxxxxxx>
>> ---
>> fs/ceph/file.c | 4 ----
>> 1 file changed, 4 deletions(-)
>>
>> diff --git a/fs/ceph/file.c b/fs/ceph/file.c
>> index 6c9e837aa1d3..8e3b99853333 100644
>> --- a/fs/ceph/file.c
>> +++ b/fs/ceph/file.c
>> @@ -651,10 +651,6 @@ static int ceph_finish_async_create(struct inode
>> *dir, struct dentry *dentry,
>> /* Directories always inherit the setgid bit. */
>> if (S_ISDIR(mode))
>> mode |= S_ISGID;
>> - else if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP) &&
>> - !in_group_p(dir->i_gid) &&
>> - !capable_wrt_inode_uidgid(&init_user_ns, dir, CAP_FSETID))
>> - mode &= ~S_ISGID;
>
> Could you point me where has done this for ceph ?
You can see the 6th patch, it added prepare_mode for tmpfile, open,
mknodat, mkdirat in vfs. The prepare_mode does inode sgid strip and
umask strip.
Best Regards
Yang Xu
>
> -- Xiubo
>
>
>> } else {
>> in.gid = cpu_to_le32(from_kgid(&init_user_ns, current_fsgid()));
>> }
>
