VFAT mount (bug or feature?)

VFAT mount (bug or feature?)

[Udo A. Steinberg]

[-- Attachment #1: Type: text/plain, Size: 660 bytes --]


Hello,

In my /etc/fstab I have the following entry:

/dev/hda1  /win   vfat   defaults,umask=022  1 1

Why does 2.5.47 have user/group restricted permissions on the mount
point and all its subdirectories, despite the umask setting?

uas@Corona:~> uname -a
Linux Corona 2.4.20-rc1 #1 Tue Oct 29 23:39:44 CET 2002 i686 unknown
uas@Corona:~> ls -lad /win
drwxr-xr-x   22 root     root        16384 Jan  1  1970 /win


uas@Corona:~> uname -a
Linux Corona 2.5.47 #3 Mon Nov 11 20:02:05 CET 2002 i686 unknown
uas@Corona:~> ls -lad /win
drwxr--r--   22 root     root        16384 Jan  1  1970 /win

uas@Corona:~> mount --version
mount: mount-2.11w


Regards,
-Udo.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: VFAT mount (bug or feature?)

[OGAWA Hirofumi]

"Udo A. Steinberg" <us15@os.inf.tu-dresden.de> writes:

> Hello,
> 
> In my /etc/fstab I have the following entry:
> 
> /dev/hda1  /win   vfat   defaults,umask=022  1 1
> 
> Why does 2.5.47 have user/group restricted permissions on the mount
> point and all its subdirectories, despite the umask setting?

The dmask option was added at 2.5.43. It's umask for directory, and
default is umask of process when mounting.  Please use it.

eg.
    # mount -t vfat /dev/xxx /xxx -o dmask=022

Regards
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

Re: VFAT mount (bug or feature?)

[Andries Brouwer]

On Wed, Nov 13, 2002 at 01:47:04AM +0100, Udo A. Steinberg wrote:

> In my /etc/fstab I have the following entry:
> 
> /dev/hda1  /win   vfat   defaults,umask=022  1 1
> 
> Why does 2.5.47 have user/group restricted permissions on the mount
> point and all its subdirectories, despite the umask setting?

Yes. This is due to a somewhat buggy change in 2.5.43.

(In the good old days umask had a well-defined meaning;
today on recent 2.5 and with vfat, but not with ntfs,
it means something else.)

The correct change would have been to add both dmask and fmask,
just like ntfs did, and leave the meaning of umask alone.

(I made a patch a few days ago, but have not yet found
the time to submit it to Linus. Maybe tomorrow,
if nobody else does it first.)

Andries

re: VFAT mount (bug or feature?

[Jeff Chua]

On Wed, Nov 13, 2002 at 01:47:04AM +0100, Udo A. Steinberg wrote:

>> In my /etc/fstab I have the following entry:
>>
>> /dev/hda1  /win   vfat   defaults,umask=022  1 1
>>
>> Why does 2.5.47 have user/group restricted permissions on the mount
>> point and all its subdirectories, despite the umask setting?

> Yes. This is due to a somewhat buggy change in 2.5.43.

This is buggy in 2.4.20-rc1 as well.

drwxr--r--   40 root     root         4096 Jan  1  1970 /dos

Any patch for 2.4.20-rcx?


Thanks,
Jeff
[ jchua@fedex.com ]

Re: VFAT mount (bug or feature?

[Andries Brouwer]

On Sat, Nov 16, 2002 at 11:52:42PM +0800, Jeff Chua wrote:

> > Yes. This is due to a somewhat buggy change in 2.5.43.
> 
> This is buggy in 2.4.20-rc1 as well.
> 
> drwxr--r--   40 root     root         4096 Jan  1  1970 /dos
> 
> Any patch for 2.4.20-rcx?

Hm. Why do you think this is wrong? Didnt you ask for it?

Re: VFAT mount (bug or feature?

[Jeff Chua]

On Sat, 16 Nov 2002, Andries Brouwer wrote:

> > drwxr--r--   40 root     root         4096 Jan  1  1970 /dos
> >
> > Any patch for 2.4.20-rcx?
>
> Hm. Why do you think this is wrong? Didnt you ask for it?

No, I didn't ask for it.

But, 2.4.20-rc2 seems ok.

'mount -o umask=022' now works under 2.4.20-rc2.

Thanks,
Jeff.

Re: VFAT mount (bug or feature?

[Jeff Chua]

On Sun, 17 Nov 2002, Jeff Chua wrote:

> But, 2.4.20-rc2 seems ok.
>
> 'mount -o umask=022' now works under 2.4.20-rc2.

"remount" doesn't work.

mount -o umask=022 -t vfat /dev/hda1 /dos
mount -o umask=000 -o remount -t vfat /dev/hda1 /dos

The above didn't work.

umount /dos
mount -o umask=022 -t vfat /dev/hda1 /dos
umount /dos
mount -o umask=000 -t vfat /dev/hda1 /dos


Jeff