2003-05-16 05:29:56

by esp

[permalink] [raw]
Subject: FAT32 problems with kernel 2.4.19

(I'm forwarding this from comp.os.linux.help - I'm having the same issue
with FAT32)

hey all,

I have a removable, USB disk (ikebana) that I'm trying to use as a
portable storage device (ie: developing programs for Win32, macosx and linux
simultaneously)

Anyways, everything is sort of working except for execute privileges
on linux (win32 works, macosx works). When I try to execute something, I get:

bash-: ./a.out: Permission Denied

When I try to execute a shell script (even as root), I get:

-bash: ./aa.sh: /bin/bash: bad interpreter: Permission denied

Now, I understand that FAT32 has no execute bit.. however I explicitly set
(or an automatic process on SuSe linux set the fstab entry for the drive
to be:

/dev/sda1 /windows/D vfat
users,gid=users,umask=0002,iocharset=iso8859-1,code=437 0 0

which as I understand it sets the permissions to be 775.

Argh! so I would expect this to work transparently.

So... is there a bug in the kernel, default settings, is something
missing in the above entries, or what? I'll escalate this to
linux-devel if it looks like there is nothing wrong with the above..

forgot to mention -

When I'm writing to FAT32 partition, there seems to be a 300% incurred
size penalty over the equivalent files on ext2 (when unpacking a
source distribution like boost, gcc, etc)

Is this an artifact of the FAT32 file system, a bug in the linux IO to
the FAT32 file system, or a local bug in my system?

I'd like to get this resolved, but understand if its a microsoft
issue...
however it would be really, really nice if NTFS was supported (ie: if
linux could write files on NTFS..) I have a sneaky suspicion that
FAT32 isn't the greatest and only used on these drives because it is
the lowest common denominator.

jon



2003-05-16 06:04:45

by Marc Giger

[permalink] [raw]
Subject: Re: FAT32 problems with kernel 2.4.19

> /dev/sda1 /windows/D vfat
> users,gid=users,umask=0002,iocharset=iso8859-1,code=437 0 0

try with
/dev/sda1 /windows/D vfat users,exec,gid=users,umask=0002,iocharset=iso8859-1,code=437 0 0

greets

Marc

2003-05-16 19:33:40

by Tomas Szepe

[permalink] [raw]
Subject: Re: FAT32 problems with kernel 2.4.19

> [[email protected]]
>
> When I'm writing to FAT32 partition, there seems to be a 300% incurred
> size penalty over the equivalent files on ext2 (when unpacking a
> source distribution like boost, gcc, etc)

I don't understand what you're trying to say. Can you elaborate?

> however it would be really, really nice if NTFS was supported (ie: if
> linux could write files on NTFS..) I have a sneaky suspicion that
> FAT32 isn't the greatest and only used on these drives because it is
> the lowest common denominator.

http://linux-ntfs.sourceforge.net/info/ntfs.html

--
Tomas Szepe <[email protected]>

2003-05-16 21:09:39

by Carl-Daniel Hailfinger

[permalink] [raw]
Subject: Re: FAT32 problems with kernel 2.4.19

Tomas Szepe wrote:
> [[email protected]]
>
>> When I'm writing to FAT32 partition, there seems to be a 300% incurred
>> size penalty over the equivalent files on ext2 (when unpacking a
>> source distribution like boost, gcc, etc)
>
> I don't understand what you're trying to say. Can you elaborate?

If I understand him correctly, he is happy that ext2 has not as much
overhead as FAT32.

[email protected]: If you think the FAT32 overhead is a linux problem,
please unpack the same source tree under windows on the same partition
and report back if the space used is less than when unpacking this
source tree under linux.


HTH,
Carl-Daniel
--
http://www.hailfinger.org/