My main OS is Ubuntu 7.10. I'm doing my blackberry dev in a VMWare virtual machine running Windows XP.

Previously connecting my blackberry to the VM worked without any problems, and the blackberry was detected as "RIM Composite Device". Now, however, the VM detects the device as "RIM Mass Storage Device", and clicking to connect it doesn't do anything. I'm not sure why the behaviour has changed.

Does anyone know how I can get my blackberry connected again?

Many thanks in advance,

Ben

Do you have BlackBerry Desktop Manager installed in the Windows session?

Devices with SD card capability initially present themselves to linux as USB Mass Storage Devices on channel 6 and typically the usb_storage kernel module captures the device and takes control. Programs such as barry's bcharge can be used to reset the USB port to make the BB expose channel 4 which is the 'RIM Composite Device' (mass storage + device control).

If it worked before but doesn't now, something has changed in your process (or in the machine setup) - what that was, though, you'll have to dig around and think about. Somehow you had the VM instance able to talk to the channel 4 composite device at one point... does that tickle/jog any memories?

Yes Desktop Manager is installed in the Windows virtual machine.

I haven't changed anything in either the Linux or Windows OSs from when it worked to when it didn't work. Linux doesn't ever automount the blackberry.

I'll try installing barry and see if that helps.

Thanks, Ben

I had the same problem running Vista as a virtual machine on an intel mac. I reinstalled the Desktop Manager and it corrected the problem.

I still have the same problem after installing and running bcharge. I've also uninstalled the desktop manager and installed the latest version from the blackberry website, without any luck.

Is there anything I can add to the vmx file to force it to detect the blackberry as a composite device?

Thanks, Ben

I've now tried connecting the blackberry using VMWare workstation instead of VMPlayer, and now I get the error message:

The specified device appears to be claimed by another driver (usb-storage) on the host operating system which means that the device may be in use. To continue, the device will first be disconnected from its current driver.

This was obviously the problem in VMPlayer, it just didn't report it. The blackberry isn't mounted in Linux. Any advice on how I can disconnect the driver?

Thanks, Ben

Adding usb.generic.skipsetconfig = "true" to my vmx file, and doing an rmmod usb_storage allows me to connect the blackberry to the VM, but only as a mass storage device. The desktop manager still cannot connect to the device.

Any ideas?

Thanks again, Ben

OK I've got a little bit further. By following the "Cause 2" instructions at:

BlackBerry Search Results

and removing the usb.generic.skipsetconfig = "true" line from the vmx file I can now connect to the blackberry, and it's detected as a composite device. However, the Desktop Manager software can't detect the blackberry! :(

Anyone know where I can go from here?

Thanks yet again, Ben

I've got it working! The "Safely remove hardware" icon was showing in the taskbar. It listed the blackberry. I removed it and then it was detected by the desktop software!

Oh no - the same problem has happened again. It was all working OK. I did a couple of builds of my blackberry software, loading it onto the device using JavaLoader.exe. After the blackberry rebooted I can't connect again.

If I then close VMWare, edit the vmx to remove the usb autoload line, start vmware, unload usb_storage - then I can connect again.

Is there any way around this?

Wow! I was never able to get my desktop manager to work, but I use Innotek Virtualbox though. Feisty only recognizes the mass storage device, but I just drag and drop for most of my stuff anyways.

OK I got a little lost (it's early, drinking coffee still ) - can you outline the steps in a 1-2-3 fashion for us on how, from scratch, you can get it to work? Once I have a good picture in the brain then we can start to look for holes in the logic and make it work easier...

(if we can)

PS: did you disable that USB_SUSPEND stuff via one of the methods in that post from the Barry guys? (you can pass it on the cmdline or use a /proc call I believe)