Re: Blackberry support: Barry vs. usb_storage, round 2

[Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>]

On Thu, 29 Nov 2007, Chris Frey wrote:

> Hi folks,
> 
> Recent Blackberry devices are a rather unique breed, since they provide
> multiple product ID's in one device, and a vendor-specific method
> to change between them.
> 
> The usual way to perform this change is to connect to the device,
> send some vendor specific control messages, and then do a
> usb_set_configuration().  For Classic devices, this causes the
> device to reset itself.  For Pearl devices, it waits for the host
> to call usb_reset().
> 
> The problem is that the new 8830 devices will reset on
> usb_set_configuration(), and a usb_reset() call will just undo the
> intended mode change and go back to the state it was in when you
> first plugged it in.
> 
> This causes the following race conditions between Barry and usb_storage
> on most Linux systems:
> 
> 	- plug in the 8830
> 	- usb_storage sees a Mass Storage interface, and connects with it
> 	- udev sees the device and runs bcharge at the same time
> 	- bcharge connects and sends the vendor specific control messages
> 	- bcharge calls usb_set_configuration() and nothing happens
> 		since usb_storage has an interface claimed
> 	- normally this is ok, for the Pearl, and bcharge calls
> 		usb_reset(), which sets the mode and everyone is happy...
> 
> But...
> 
> 	- for the 8830, usb_reset() goes back to the start, the mode is
> 		not changed, and there is no way to change it while
> 		usb_storage has Mass Storage claimed, since the mode
> 		change occurs on usb_set_configuration().
> 
> Questions:
> 
> 	- is there a safe way to delay usb_storage, perhaps from udev,
> 		so that when Barry is installed the mode is changed
> 		before usb_storage looks for a drive?

No.

> 	- is there a better way to handle this?  perhaps force usb_storage
> 		to release the interface somehow, since this only happens
> 		when the device is first plugged in?

Yes.  The bcharge program can unbind usb-storage from the mass-storage 
interface, if needed.  The libusb call is 
usb_detach_kernel_driver_np().  Or it can be accomplished through 
usbfs directly (USBDEVFS_DISCONNECT ioctl), or through sysfs.

Note that there's no point in doing this if Barry decides that it 
doesn't need to change the device's mode.

Alan Stern


-------------------------------------------------------------------------
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell.  From the desktop to the data center, Linux is going
mainstream.  Let it simplify your IT future.
http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4
_______________________________________________
linux-usb-devel@xxxxxxxxxxxxxxxxxxxxx
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel