Re: usb driver and probe function (is never called)
Ulrich Eckhardt wrote:
>
> On Monday 24 December 2001 13:32, Erik Steffl wrote:
> > I am trying to make the USB work on my debian unstable system, using
> > kernel 2.4.14 and it just doesn't work. the specific problem is that the
> > function 'probe' (specified when registering driver) is NEVER called and
> > therefore driver doesn't manage any usb devices. as far as I can tell
> > the system is set up correctly, see below for modules, kernel settings
> > etc.
> >
> have you also printk-mined another usb-driver ? I remember a bit from hacking
> the pegasus-driver. There you had to register the usb-driver with a set of
> vendor/device-IDs so the usb-core could find the proper driver without asking
> _every_ driver to handle the device. However, that was with 2.2.18 or so.
>
> > so far I have tried the sample usb driver from linux device driver,
> > 2nd edition (o'reilly) and personal jukebox (www.pjbox.com) drivers,
> > both behave in exactly the same way - I get a message in syslog that the
> > driver was registered and that's it, the message from probe function is
> > never printed and the driver doesn't work (cat says No such device)
> >
> If it is available on paper, it isn't up to date, especially not with
> 2.4.something :-)
>
> >
> > Dec 22 01:46:55 localhost kernel: hub.c: port 1 connection change
> > Dec 22 01:46:55 localhost kernel: hub.c: port 1, portstatus 301, change
> > 1, 1.5 Mb/s
> > Dec 22 01:46:55 localhost kernel: hub.c: port 1, portstatus 300, change
> > 0, 1.5 Mb/s
> >
> I use USB but have neither 'portstatus' nor 'connection change' in my kernel
> logs. On thing I would ask you to do is to try eg a mouse that is known to
> work or something like that. Just to see that it is the driver, not the setup.
at this point I am quite sure it's the setup, the driver is confirmed
to work with 2.4.14 (cpqpjb) and the usbsample (from the book) is for
2.4 too (not sure if for 2.4.14). what should usb syslog messages look
like?
2.4.14 has support for both old and new drivers (old have 2-parameter
probe, new ones have 3 parameter probe function)
I also tried to plug in mouse (logitech mouse man, wireless, optical)
but still got the same messages, no attempt to load driver (I don't have
driver for mouse though, I'll probably try that as next step (even
though I think the problem is that kernel does not look for ANY drivers,
the existence of drivers seem to be irrelevant).
> > lsusb (usbview show the same info) output:
> >
> > Bus 002 Device 001: ID 0000:0000 Virtual Hub
> [snip]
> > Bus 001 Device 001: ID 0000:0000 Virtual Hub
> [snap]
> EOF ? Whether your device is handled by any driver or not, it should appear
> in here! Try another device and see if that gives you anything.
well, yes, that's all there is, there are no further usb related
messages in syslog. obviously it doesn't even try to look for device
drivers even though it should, I have looked (briefly) at the code of
usb_register and some other functions in usb.c - whenever new device is
connected it should look at the modules and run probe function (that was
registered) but for some reason it does not happen on my system. I
haven't put any extra debug info into usb.c yet...
is uhci or altarnative uhci better (for 2.4.14)? I though both should
work... is there any documentation about setting up usb devices for
linux? so far I haven't found any that would list any step that I
missed.
are any other modules besides usbcore and usb-uhci meeded fpr uusb
support (aside from actual device drivers for particular devices)?
is it possible it's hw problem? it works to certain extend, I get at
least some messages when I plug in/unplug devices so it's not completely
disconnected but is it possible that there's some other hw problem?
erik
Reply to: