Created attachment 25058 [details]
output of lsusb -v from Lenny

(guess attachments need to be added one-by-one...)

Created attachment 25059 [details]
output of lsusb -v from Sid (2.6.32)

(switched to email.  Please respond via emailed reply-to-all, not via the
bugzilla web interface).

(It's a regression)

On Mon, 15 Feb 2010 21:54:25 GMT
bugzilla-daemon@bugzilla.kernel.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=15321
> 
>            Summary: vfat usb device cannot be mounted on 2.6.32, works on
>                     2.6.26
>            Product: Drivers
>            Version: 2.5
>     Kernel Version: 2.6.32 (debian 2.6.32-trunk-686 ver. 2.6.32-5)
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: normal
>           Priority: P1
>          Component: USB
>         AssignedTo: greg@kroah.com
>         ReportedBy: shai@platonix.com
>         Regression: Yes
> 
> 
> Created an attachment (id=25057)
>  --> (http://bugzilla.kernel.org/attachment.cgi?id=25057)
> Output of dmesg after connecting device to a Lenny (2.6.26) machine
> 
> Hi,
> 
> This is my first kernel bug, I hope the report is filled correctly. I am
> filing
> it on advice from  Moritz Muehlenhoff of the Debian kernel team, see
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=569906
> 
> Repeating the story for your convenience: The device is a so-called MP4
> Player.
> I can mount it on a machine running Debian Stable, that is, kernel 2.6.26, by
> just specifying the vfat filesystem type. I cannot mount it on Debian
> unstable,
> running kernel 2.6.32, nor on Ubuntu 9.10, which uses 2.6.31-19.
> 
> I am copying below the relevant part of dmesg output from 2.6.32 (as
> collected
> by Debian's reportbug utility), and attaching three files: dmesg from 2.6.26,
> and "lsusb -v" from both systems when the device is attached. These inputs
> (and
> perhaps others which are on the Debian bug) led mr. Muehlenhoff to the
> conclusion that the problem is related to the "list of USB mass storage
> devices, which are not fully standards compliant and to which several
> workarounds are applied".
> 
> Thanks,
> 
>   Shai.
> 
> [1815555.504019] usb 1-1: new high speed USB device using ehci_hcd and
> address
> 33
> [1815555.638436] usb 1-1: New USB device found, idVendor=10d6, idProduct=1101
> [1815555.638441] usb 1-1: New USB device strings: Mfr=1, Product=2,
> SerialNumber=3
> [1815555.638445] usb 1-1: Product: media player
> [1815555.638447] usb 1-1: Manufacturer: actions
> [1815555.638450] usb 1-1: SerialNumber: _________K
> [1815555.638563] usb 1-1: configuration #1 chosen from 1 choice
> [1815555.640305] scsi43 : SCSI emulation for USB Mass Storage devices
> [1815555.640535] usb-storage: device found at 33
> [1815555.640538] usb-storage: waiting for device to settle before scanning
> [1815560.640360] usb-storage: device scan complete
> [1815560.681606] scsi 43:0:0:0: Direct-Access     GENERIC  USB DISK DEVICE 
> 1.00 PQ: 0 ANSI: 0 CCS
> [1815560.682225] scsi 43:0:0:1: Direct-Access     GENERIC  USB DISK DEVICE 
> 1.00 PQ: 0 ANSI: 0 CCS
> [1815560.682705] sd 43:0:0:0: Attached scsi generic sg2 type 0
> [1815560.682876] sd 43:0:0:1: Attached scsi generic sg3 type 0
> [1815560.686844] sd 43:0:0:0: [sdb] 3833344 512-byte logical blocks: (1.96
> GB/1.82 GiB)
> [1815560.687712] sd 43:0:0:1: [sdc] Attached SCSI removable disk
> [1815560.688335] sd 43:0:0:0: [sdb] Write Protect is off
> [1815560.688339] sd 43:0:0:0: [sdb] Mode Sense: 00 12 00 00
> [1815560.688342] sd 43:0:0:0: [sdb] Assuming drive cache: write through
> [1815560.691708] sd 43:0:0:0: [sdb] Assuming drive cache: write through
> [1815560.691712]  sdb:
> [1815560.695333] sd 43:0:0:0: [sdb] Assuming drive cache: write through
> [1815560.695337] sd 43:0:0:0: [sdb] Attached SCSI removable disk
>

On Fri, 19 Feb 2010, Andrew Morton wrote:

> > http://bugzilla.kernel.org/show_bug.cgi?id=15321
> > 
> >            Summary: vfat usb device cannot be mounted on 2.6.32, works on
> >                     2.6.26

...

> > Hi,
> > 
> > This is my first kernel bug, I hope the report is filled correctly. I am
> filing
> > it on advice from  Moritz Muehlenhoff of the Debian kernel team, see
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=569906
> > 
> > Repeating the story for your convenience: The device is a so-called MP4
> Player.
> > I can mount it on a machine running Debian Stable, that is, kernel 2.6.26,
> by
> > just specifying the vfat filesystem type. I cannot mount it on Debian
> unstable,
> > running kernel 2.6.32, nor on Ubuntu 9.10, which uses 2.6.31-19.

What do you mean exactly?  That is, what happens when you try to mount 
it?

> > I am copying below the relevant part of dmesg output from 2.6.32 (as
> collected
> > by Debian's reportbug utility), and attaching three files: dmesg from
> 2.6.26,
> > and "lsusb -v" from both systems when the device is attached. These inputs
> (and
> > perhaps others which are on the Debian bug) led mr. Muehlenhoff to the
> > conclusion that the problem is related to the "list of USB mass storage
> > devices, which are not fully standards compliant and to which several
> > workarounds are applied".

Muehlenhoff wasn't entirely correct; 2.6.26 does include an entry with
idVendor=10d6, but it has idProduct=2200 instead of 1101.

The best way to find the cause of the problem is for you to collect a
usbmon trace under 2.6.32, showing what happens when you plug in the
drive and try to mount it.  See the kernel source file
Documentation/usb/usbmon.txt for instructions.  Attach the trace to the
bug report and then reply to this email to let us know it's ready.

It wouldn't hurt to collect an analogous usbmon trace under 2.6.26; the 
differences between the two traces might be informative.

Alan Stern

Created attachment 25146 [details]
usbmon output from Sid (2.6.32)

Here is the usbmon output on 2.6.32, while connecting, trying to mount, and finally disconnecting the device. The mount attempt was:

$ sudo mount -t vfat -o readonly /dev/sdb  /media/mp4/
mount: wrong fs type, bad option, bad superblock on /dev/sdb,
       missing codepage or helper program, or other error
       In some cases useful info is found in syslog - try
       dmesg | tail  or so

$ echo $?
32

Created attachment 25147 [details]
usbmon output from Lenny (2.6.26)

This is usbmon output from 2.6.26. The output was collected while I connected the device, mounted it (essentially same command as before), listed its root directory to make sure everything succeeded, unmounted and disconnected.

Created attachment 25148 [details]
no, 2.6.32 does mount it after all...

I'm sorry -- I've probably done something wrong earlier, and then tried to 
mount the device as "readonly" to be on the safe side, and that seems to have 
been the source of my problems later. After generating the usbmon reports 
(relying on shell history), I noticed the "-o readonly" in the command and the 
"...bad option, " in the error message, tried without it, and it worked.

There still is something weird here -- HAL doesn't seem to realize the device 
is there, but that's probably not a kernel issue. I've attached the usbmon 
output of successful connect-mount-ls-umount-disconnect on 2.6.32 for 
completeness.

I apologize sincerely for wasting your time,

Shai.

PS: I'm closing the bug. I hope that's ok, and I hope I'll be allowed to reopen if it turns out that there's a kernel issue after all.

Thanks, and again, sorry,

Shai.

I'm sorry -- I've probably done something wrong earlier, and then tried to 
mount the device as "readonly" to be on the safe side, and that seems to have 
been the source of my problems later. After generating the usbmon reports 
(relying on shell history), I noticed the "-o readonly" in the command and the 
"...bad option, " in the error message, tried without it, and it worked.

There still is something weird here -- HAL doesn't seem to realize the device 
is there, but that's probably not a kernel issue. I've attached the usbmon 
output of successful connect-mount-ls-umount-disconnect on 2.6.32 for 
completeness.

I apologize sincerely for wasting your time,

Shai.