gtkpod and Filesystem

gtkpod and Filesystem

[Alejandro Bonilla]

Hi,

I have Debian Sid with 2.6.15-rc5, I wonder if this could be either with a bug
in gtkpod or the kernel (FS Panic).

Whenever I try to sync my Ipod it comes up and says the FS is Write-Procteted
or Read-Only. (This is on the gtkpod interface)
"Error opening '/media/IPOD/iPod_Control/Music/F15/gtkpod108107.mp3' for
writing (Read-only file system)."

drivers/usb/core/inode.c: creating file '003'
Initializing USB Mass Storage driver...
usb-storage 1-4:1.0: usb_probe_interface
usb-storage 1-4:1.0: usb_probe_interface - got id
scsi1 : SCSI emulation for USB Mass Storage devices
usb-storage: device found at 3
usb-storage: waiting for device to settle before scanning
usbcore: registered new driver usb-storage
USB Mass Storage support registered.
  Vendor: Apple     Model: iPod              Rev: 1.62
  Type:   Direct-Access                      ANSI SCSI revision: 00
SCSI device sdb: 3999743 512-byte hdwr sectors (2048 MB)
sdb: Write Protect is off
sdb: Mode Sense: 68 00 00 08
sdb: assuming drive cache: write through
SCSI device sdb: 3999743 512-byte hdwr sectors (2048 MB)
sdb: Write Protect is off
sdb: Mode Sense: 68 00 00 08
sdb: assuming drive cache: write through
 sdb: sdb1 sdb2
sd 1:0:0:0: Attached scsi removable disk sdb
sd 1:0:0:0: Attached scsi generic sg1 type 0
usb-storage: device scan complete
FAT: Filesystem panic (dev sdb2)
    fat_get_cluster: invalid cluster chain (i_pos 120196)
    File system has been set read-only
FAT: Filesystem panic (dev sdb2)
    fat_get_cluster: invalid cluster chain (i_pos 120196)
FAT: Filesystem panic (dev sdb2)
    fat_get_cluster: invalid cluster chain (i_pos 120196)
FAT: Filesystem panic (dev sdb2)
    fat_get_cluster: invalid cluster chain (i_pos 120196)
FAT: Filesystem panic (dev sdb2)
    fat_get_cluster: invalid cluster chain (i_pos 120196)


--
Open WebMail Project (http://openwebmail.org)

Re: gtkpod and Filesystem

[Gunter Ohrner]

Alejandro Bonilla wrote:
> I have Debian Sid with 2.6.15-rc5, I wonder if this could be either with a
> bug in gtkpod or the kernel (FS Panic).

Maybe an FS error on your iPod? Did you try to reformat or dosfsck it?

Greetings,

  Gunter

Re: gtkpod and Filesystem

[Jan Engelhardt]

>> I have Debian Sid with 2.6.15-rc5, I wonder if this could be either with a
>> bug in gtkpod or the kernel (FS Panic).
>
>Maybe an FS error on your iPod? Did you try to reformat or dosfsck it?

Even then, the filesystem code should handle corrupt filesystems more 
gracefully.


Jan Engelhardt
-- 

Re: gtkpod and Filesystem

[Gunter Ohrner]

Jan Engelhardt wrote:
>>> a bug in gtkpod or the kernel (FS Panic).
>>Maybe an FS error on your iPod? Did you try to reformat or dosfsck it?
> Even then, the filesystem code should handle corrupt filesystems more
> gracefully.

Mh, what's "more gracefully" in the light of fs corruption? The driver just
blocked write access to avoid further damage caused by writing to an
inconsistent file system which sound perfectly reasonable to me. Writing to
a corrupted fs could cause anything to it, depending on the corruption, so
better act safe than sorry...

Greetings,

  Gunter

Re: gtkpod and Filesystem

[Jan Engelhardt]

>>>> a bug in gtkpod or the kernel (FS Panic).
>>>Maybe an FS error on your iPod? Did you try to reformat or dosfsck it?
>> Even then, the filesystem code should handle corrupt filesystems more
>> gracefully.
>
>Mh, what's "more gracefully" in the light of fs corruption?

	return -EIO;

through all instances back to userspace and keep returning EIO for all 
future requests. But let the user still umount the device.

>The driver just
>blocked write access to avoid further damage caused by writing to an
>inconsistent file system which sound perfectly reasonable to me. Writing to
>a corrupted fs could cause anything to it, depending on the corruption, so
>better act safe than sorry...

The interesting part comes when the filesystem corrupts itself
 (= the code corrupts the on-disk data), and/plus it does not 
notice quickly enough.


Jan Engelhardt
-- 

Re: gtkpod and Filesystem

[Gunter Ohrner]

Jan Engelhardt wrote:
>>Mh, what's "more gracefully" in the light of fs corruption?
> return -EIO;

In this case you wouldn't even be able to read data from the disk which is
not affected by the inconsistencies. If the FS driver switches to readonly
mode, that's possible at least. Personally I like such a kind of "Best
effort" behaviour. Well, but I'm not the vfat32 driver developer, so who am
I to argue? :-)

Greetings,

  Gunter

Re: gtkpod and Filesystem

[Gunter Ohrner]

[-- Attachment #1: Type: text/plain, Size: 1270 bytes --]

Am Freitag, 16. Dezember 2005 18:19 schrieben Sie:
> |> 	Actually. Issue fixed. It is really odd that a dosfsck fixed it.

I'm glad it helped. :-)

> |i dont see how that is odd.. if the filesystem was somehow corrupted
> |dosfsck would have corrected it.
> Odd how it got "corrupted" and odd on why it would work on Windows and
> not in Linux.

Mh, that's just the same phenomenon one always sees with web sites. If 
some data structure, be it a file system or an HTML-like tag soup, 
doesn't conform to the specification, the processors (fs driber or web 
browser) behavoiur will be implementation dependant... One specific error 
will break one implementation but not the other, while another error will 
just cause the opposite, there may be implementations which are more 
sensitive to specification violation and so on...

So the behaviour you observed is not that surprising, after all. ;)

> Anyway, it works now. I was more worried on the FS Panic than anything
> else.

Well, as I understood it this message just stated that the fs driver was 
confused and blocked write access for safety reasons, to avoid further 
damage.

It had nothing to do with a "kernel panic" or similar.

Greetings and happy music listening,

  Gunter

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