On Thu, Jan 22, 2015 at 06:28:51PM +0300, Konstantin Khlebnikov wrote:

On 09.12.2014 08:22, Li Xi wrote:

quoted

+static int ext4_ioctl_setproject(struct file *filp, __u32 projid)
+{
+	struct inode *inode = file_inode(filp);
+	struct super_block *sb = inode->i_sb;
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	int err;
+	handle_t *handle;
+	kprojid_t kprojid;
+	struct ext4_iloc iloc;
+	struct ext4_inode *raw_inode;
+
+	struct dquot *transfer_to[EXT4_MAXQUOTAS] = { };
+
+	/* Make sure caller can change project. */
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
+	if (projid != EXT4_DEF_PROJID
+	    && !EXT4_HAS_RO_COMPAT_FEATURE(sb,
+			EXT4_FEATURE_RO_COMPAT_PROJECT))
+		return -EOPNOTSUPP;
+
+	if (!EXT4_HAS_RO_COMPAT_FEATURE(sb,
+			EXT4_FEATURE_RO_COMPAT_PROJECT)) {
+		BUG_ON(__kprojid_val(EXT4_I(inode)->i_projid)
+		       != EXT4_DEF_PROJID);
+		if (projid != EXT4_DEF_PROJID)
+			return -EOPNOTSUPP;
+		else
+			return 0;
+	}
+
+	kprojid = make_kprojid(&init_user_ns, (projid_t)projid);

Maybe current_user_ns()?
This code should be user-namespace aware from the beginning.

No, the code is correct. Project quotas have nothing to do with
UIDs and so should never have been included in the uid/gid
namespace mapping infrastructure in the first place.

Point in case: directory subtree quotas can be used as a resource
controller for limiting space usage within separate containers that
share the same underlying (large) filesystem via mount namespaces.

Cheers,

Dave.

PS: can you please trim your reply to just the sections of the
patch you are commenting to? Finding replies like this in a large
patch is like searching for a needle in a haystack...
-- 
Dave Chinner
david-FqsqvQoI3Ljby3iVrkZq2A@public.gmane.org