On Tue 29-03-22 10:48:49, Amir Goldstein wrote:
> The inotify mask flags IN_ONESHOT and IN_EXCL_UNLINK are not "internal
> to kernel" and should be exposed in procfs fdinfo so CRIU can restore
> them.
>
> Fixes: 6933599697c9 ("inotify: hide internal kernel bits from fdinfo")
> Signed-off-by: Amir Goldstein <amir73il@xxxxxxxxx>
Good catch!
> diff --git a/fs/notify/fdinfo.c b/fs/notify/fdinfo.c
> index 57f0d5d9f934..3451708fd035 100644
> --- a/fs/notify/fdinfo.c
> +++ b/fs/notify/fdinfo.c
> @@ -83,16 +83,9 @@ static void inotify_fdinfo(struct seq_file *m, struct fsnotify_mark *mark)
> inode_mark = container_of(mark, struct inotify_inode_mark, fsn_mark);
> inode = igrab(fsnotify_conn_inode(mark->connector));
> if (inode) {
> - /*
> - * IN_ALL_EVENTS represents all of the mask bits
> - * that we expose to userspace. There is at
> - * least one bit (FS_EVENT_ON_CHILD) which is
> - * used only internally to the kernel.
> - */
> - u32 mask = mark->mask & IN_ALL_EVENTS;
> - seq_printf(m, "inotify wd:%x ino:%lx sdev:%x mask:%x ignored_mask:%x ",
> + seq_printf(m, "inotify wd:%x ino:%lx sdev:%x mask:%x ignored_mask:0 ",
> inode_mark->wd, inode->i_ino, inode->i_sb->s_dev,
> - mask, mark->ignored_mask);
> + inotify_mark_user_mask(mark));
I think inotify_mark_user_mask() helper is overdoing it a bit. Just using
INOTIFY_USER_MASK here directly is IMHO fine.
Honza
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
