2.6.0-test8: scsi_sysfs.c still not detaching properly

2.6.0-test8: scsi_sysfs.c still not detaching properly

[Phil Dibowitz]

[-- Attachment #1: Type: text/plain, Size: 1915 bytes --]

I'm experiencing a bug similar to the one referenced here:

http://sourceforge.net/mailarchive/message.php?msg_id=6252693

But the patch in that thread has been applied to 2.6.0-test8, which I'm 
running, and I'm still having the same problem - albeit with a different 
device.

The device in question is a Digital Camera (Casio Exilim) that uses the 
usb-storage driver. Basically device files do not get unallocated upon 
disconnect, and thus successive connections get different device files 
associated with them.

When attached, I get the following (and it attaches to sda1):

Oct 19 12:12:20 rider kernel: hub 1-0:1.0: new USB device on port 2, 
assigned address 3
Oct 19 12:12:20 rider kernel: scsi1 : SCSI emulation for USB Mass 
Storage devices
Oct 19 12:12:20 rider kernel:   Vendor: Casio     Model: QV 
DigitalCamera  Rev: 5010
Oct 19 12:12:20 rider kernel:   Type:   Direct-Access 
    ANSI SCSI revision: 02
Oct 19 12:12:21 rider scsi.agent[9314]: how to add device type= at 
/devices/pci0000:00/0000:00:07.2/usb1/1-2/1-2:1.0/host1/1:0:0:0 ??
Oct 19 12:12:21 rider kernel: SCSI device sda: 20208 512-byte hdwr 
sectors (10 MB)
Oct 19 12:12:21 rider kernel: sda: Write Protect is off
Oct 19 12:12:21 rider kernel: sda: cache data unavailable
Oct 19 12:12:21 rider kernel:  sda: sda1
Oct 19 12:12:21 rider kernel: Attached scsi removable disk sda at scsi1, 
channel 0, id 0, lun 0


When unattached, I get:

Oct 19 12:13:14 rider kernel: usb 1-2: USB disconnect, address 3


And the next time I attach, it attaches to sdb1.

Any help would be greatly appreciated.

-- 
Phil Dibowitz                             phil@ipom.com
Freeware and Technical Pages              Insanity Palace of Metallica
http://www.phildev.net/                   http://www.ipom.com/

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
  - Benjamin Franklin, 1759


[-- Attachment #2: Type: application/pgp-signature, Size: 251 bytes --]

Re: 2.6.0-test8: scsi_sysfs.c still not detaching properly

[Patrick Mansfield]

On Sun, Oct 19, 2003 at 12:16:00PM -0700, Phil Dibowitz wrote:
> I'm experiencing a bug similar to the one referenced here:
> 
> http://sourceforge.net/mailarchive/message.php?msg_id=6252693

That page gives me "Message Not Found".

> But the patch in that thread has been applied to 2.6.0-test8, which I'm 
> running, and I'm still having the same problem - albeit with a different 
> device.

> The device in question is a Digital Camera (Casio Exilim) that uses the 
> usb-storage driver. Basically device files do not get unallocated upon 
> disconnect, and thus successive connections get different device files 
> associated with them.
> 
> When attached, I get the following (and it attaches to sda1):
> 
> Oct 19 12:12:20 rider kernel: hub 1-0:1.0: new USB device on port 2, 
> assigned address 3
> Oct 19 12:12:20 rider kernel: scsi1 : SCSI emulation for USB Mass 
> Storage devices
> Oct 19 12:12:20 rider kernel:   Vendor: Casio     Model: QV 
> DigitalCamera  Rev: 5010
> Oct 19 12:12:20 rider kernel:   Type:   Direct-Access 
>     ANSI SCSI revision: 02
> Oct 19 12:12:21 rider scsi.agent[9314]: how to add device type= at 
> /devices/pci0000:00/0000:00:07.2/usb1/1-2/1-2:1.0/host1/1:0:0:0 ??
> Oct 19 12:12:21 rider kernel: SCSI device sda: 20208 512-byte hdwr 
> sectors (10 MB)
> Oct 19 12:12:21 rider kernel: sda: Write Protect is off
> Oct 19 12:12:21 rider kernel: sda: cache data unavailable
> Oct 19 12:12:21 rider kernel:  sda: sda1
> Oct 19 12:12:21 rider kernel: Attached scsi removable disk sda at scsi1, 
> channel 0, id 0, lun 0
> 
> 
> When unattached, I get:
> 
> Oct 19 12:13:14 rider kernel: usb 1-2: USB disconnect, address 3
> 
> 
> And the next time I attach, it attaches to sdb1.
> 
> Any help would be greatly appreciated.

Can you delete it prior to physical removal via the sysfs attribute (echo
foo > /sysfs/bus/scsi/devices/1:0:0:0/delete)? If that doesn't work, you
might be running an older kernel.

Is the device umounted (not open) when you disconnect it? If not, try
Mike's recent patch (or try his patch anyway). But I would expect to see
an oops or other odd behaviour in that case.

-- Patrick Mansfield

Re: 2.6.0-test8: scsi_sysfs.c still not detaching properly

[Phil Dibowitz]

[-- Attachment #1: Type: text/plain, Size: 901 bytes --]

Patrick Mansfield wrote:
> Can you delete it prior to physical removal via the sysfs attribute (echo
> foo > /sysfs/bus/scsi/devices/1:0:0:0/delete)? If that doesn't work, you
> might be running an older kernel.
> 
> Is the device umounted (not open) when you disconnect it? If not, try
> Mike's recent patch (or try his patch anyway). But I would expect to see
> an oops or other odd behaviour in that case.

Well, this is embarassing, but it appears I didn't properly test this 
after my -test8 upgrade, because it actually works quite well.

Sorry about that.
-- 
Phil Dibowitz                             phil@ipom.com
Freeware and Technical Pages              Insanity Palace of Metallica
http://www.phildev.net/                   http://www.ipom.com/

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
  - Benjamin Franklin, 1759


[-- Attachment #2: Type: application/pgp-signature, Size: 251 bytes --]