[Anonymous]

I am using vanilla linux 2.4.22 source + xfs +
loop-AES-v1.7e.

I have found out that xfs quota does not work with
encrypted option turned on. (through the loop device)

for example:

/dev/hde11 /home xfs
encryption=AES128,encrypted,grpquota,usrquota 0 0

has the encryption, but unable to assign quota, enforce
or show quota.

without the encryption, quota just works.

I believe it is not xfs specific, do you have any ideas?

Thanks,

Logged In: YES
user_id=238645

This appears to be result of bug or limitation in quota
utilities that fail to locate the loop device between
ide disk partition /dev/hde11 and mount point /home

This workaround seems to work:

losetup -e AES128 /dev/loop2 /dev/hde11
mount -t xfs /dev/loop2 /home -o grpquota,usrquota
setquota -F xfs USER 1000 1000 300 300 /home
umount /home
losetup -d /dev/loop2

Unless you or someone else can pinpoint a bug in
loop-AES package that causes this, I am soon going to
close this 'bug' as invalid. Real fix is to use quota
utilities that are able locate loop devices between
partitions and mount points and direct their actions to
the loop device.

Logged In: NO

I am trying to reproduce what you do.
[root@cbis-dev1 ying]# losetup -e AES128 /dev/loop4 /dev/hde10
Password:
[root@cbis-dev1 ying]# mount -t xfs /dev/loop4
/var/spool/mail -o usrquota
mount: wrong fs type, bad option, bad superblock on /dev/loop4,
or too many mounted file systems

any ideas?

Thanks,

Logged In: NO

sorry for the previous post, i get it to work now.

however, once umount / reboot, there is no way to tell it to
remember it correct?

Logged In: YES
user_id=894222

just FYI, quota works with loop device.

# mkfs.xfs -dfile,size=1g,name=/tmp/loopy
meta-data=/tmp/loopy isize=256 agcount=8,
agsize=32768 blks
= sectsz=512
data = bsize=4096 blocks=262144,
imaxpct=25
= sunit=0 swidth=0 blks,
unwritten=1
naming =version 2 bsize=4096
log =internal log bsize=4096 blocks=2560,
version=1
= sectsz=512 sunit=0 blks
realtime =none extsz=65536 blocks=0,
rtextents=0
# mkdir /mnt/loopy
# mount -t xfs -o loop,usrquota /tmp/loopy /mnt/loopy
# repquota /mnt/loopy
*** Report for user quotas on device /dev/loop0
Block grace time: 7days; Inode grace time: 7days
Block limits File limits
User used soft hard grace used soft
hard grace
----------------------------------------------------------------------
root -- 0 0 0 3 0
0

--

Logged In: YES
user_id=238645

I have included a patch that fixes a bug in quota-3.09
that affects loop device mounts. Old quota code did not
work at all with loop mounts where mount program
created a loop=/dev/loop2 mount option entry to
/etc/mtab. Old quota code assumed that file or device
underneath loop device was neither block nor character
device.

If you patch and recompile quota package yourself, you
will need quota-3.09.tar.gz tarball (from linuxquota
project on sourceforge) and my quota-3.09-loopfix.diff
patch. Then do this as root:

zcat quota-3.09.tar.gz | tar xvf -
cd quota-tools
patch -p1 <../quota-3.09-loopfix.diff
./configure --prefix=/usr
make
make install

Logged In: YES
user_id=238645

I just uploaded new version of quota-3.09 patch.
Earlier one still failed on file backed loops.

Logged In: YES
user_id=238645

I just uploaded third version of quota-3.09 patch.
Earlier one handled bind mounts incorrectly.

Logged In: NO

Great, It works!

Thanks!

Logged In: YES
user_id=238645

This is not loop-AES bug.
I am now closing this as 'invalid'.

• status: open --> closed-invalid