On Tue 07-10-14 13:46:20, Andreas Dilger wrote:

On Oct 7, 2014, at 1:29 PM, Jan Kara [off-list ref] wrote:

quoted

On Tue 07-10-14 07:30:28, Dave Chinner wrote:

quoted

On Wed, Oct 01, 2014 at 09:31:25PM +0200, Jan Kara wrote:

quoted

We support user, group, and project quotas. Tell VFS about it.

CC: xfs@oss.sgi.com
CC: Dave Chinner <david@fromorbit.com>
Signed-off-by: Jan Kara <jack@suse.cz>
---
fs/xfs/xfs_super.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index b194652033cd..b32e998e8cbc 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c

@@ -1419,6 +1419,8 @@ xfs_fs_fill_super(

	sb->s_export_op = &xfs_export_operations;
#ifdef CONFIG_XFS_QUOTA
	sb->s_qcop = &xfs_quotactl_operations;
+	sb->s_dquot.allowed_types = (1 << USRQUOTA) | (1 << GRPQUOTA) |
+				    (1 << PRJQUOTA);

Would it be better to define masks for these rather than open
coding these shifts everywhere?

 I can do that. Any suggestion for a name? I was thinking about it for a
while and couldn't come up with anything satisfactory...

Better to have QUOTA at the start, and TYPE in the name, so maybe:

enum quota_types {
	QUOTA_TYPE_USR = 1 << USRQUOTA,
	QUOTA_TYPE_GRP = 1 << GRPQUOTA,
	QUOTA_TYPE_PRJ = 1 << PRJQUOTA,
};

or maybe "enum quota_type_mask" or similar.

I prefer named enums over #defines since this makes it more clear
when declaring variables like "allowed_types" what valid values are
instead of just "int" that someone might mistakenly set to USRQUOTA
directly or something.

  OK, QUOTA_TYPE_USR etc. looks good. I'm not so sure about named enum. I'd
be fine with named enum for USRQUOTA, GRPQUOTA, etc. - i.e., types which
should have one of the named values but for a bitmask with arbitrary
combinations of flags it looks confusing to me (although technically
it would work).

								Honza
-- 
Jan Kara [off-list ref]
SUSE Labs, CR