Re: [Bitpim-devel] linux debugging questions (really USB questions)

[Roger B. <ro...@ro...>]

> I then tried /dev/usb/acm/0 as a hope of accessing the diagnostic port,

http://bitpim.sourceforge.net/testhelp/hoto-usbintro.htm

The VX6000 has 3 interfaces.  0 and 1 are the modem interface and
are used by the acm driver.  Interface 2 is the diagnositcs
interface and is what BitPim needs to talk to.  It can only 
be accessed via libusb on Linux as there is no driver.

usbscan should list every USB device connected to your machine
(not just phones).  If you aren't getting that then something
is wrong.  Note that if you are working from the source code,
you need to build the USB module in native/usb which also
requires SWIG amongst other things.

The prebuilt rpm has all the right stuff built and links 
dynamically to libusb:

  $ ldd _libusb.so 
  libusb-0.1.so.4 => /usr/lib/libusb-0.1.so.4 (0x400e6000)
  libc.so.6 => /lib/i686/libc.so.6 (0x400ed000)
  /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x80000000)

> but the bit about not reading
> the vendor string seems like it might be an issue?

None of the LG phones will give up a vendor string except when
they very first get plugged in, and then they tell the OS as
it interrogates them.

> However I know that we should just be looking for the IDs?

Yes, that is what happens.  The strings from the device itself
are shown in the port browser if present.

We also have our own USB ids to string database to augment
the standard one.

> This could simply be that I don't know where to get access to the damn
> thing in my /dev (devfs) tree.

libusb accesses the usbfs tree.

> Any suggestions as to what the right
> major (and minor) numbers for the diag port device should be?  Then I
> could just run a find command to figure out what debian called it?

It doesn't work like that (except for the modem interface).  The
devices and interfaces appear as nodes in the usb filesystem.

Roger