On Wed, 18 Apr 2012 16:05:11 -0400 (EDT), Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
> On Wed, 18 Apr 2012, Felipe Balbi wrote:
>
> > > You'd be right except for one thing: You listed the code from
> > > udc-core.c:usb_gadget_remove_driver() wrongly. In my tree it actually
> > > looks like this:
> > >
> > > if (udc_is_newstyle(udc)) {
> > > udc->driver->disconnect(udc->gadget);
> > > udc->driver->unbind(udc->gadget);
> > > usb_gadget_udc_stop(udc->gadget, udc->driver);
> > > usb_gadget_disconnect(udc->gadget);
> > > } else {
> > > usb_gadget_stop(udc->gadget, udc->driver);
> > > }
> > >
> > > Notice the relative order of the usb_gadget_disconnect() and
> > > usb_gadget_udc_stop() calls. This hasn't been changed since 3.4-rc2,
> > > has it? I hope not -- if it has changed then it is now wrong.
> >
> > you need to fetch Greg's usb-linus branch ;-)
>
> In fact I have it -- the code listed above is the most recent version
> in the usb-linus branch.
>
> (Argh! It turns out one of my scripts needs to be fixed. It depends
> on there not being any tags reachable from the main linus branch other
> than the vX.Y and vX.Y-rcN tags created by Linus himself. But now
> Greg's usb-X.Y-rcN tags are there too. That fooled me into thinking I
> was using 3.4-rc2 with only slight additions when in fact I was using
> usb-3.4-rc2 with slight additions. I'll have to restrict the script to
> match only tag names beginning with 'v'.)
>
> Anyway, apparently Alexander hasn't gotten commit
> 8ae8090c82eb407267001f75b3d256b3bd4ae691 (usb: gadget: udc-core: fix
> asymmetric calls in remove_driver) yet. Presumably that will take care
> of his problem.
I have, and that's how I found the initial problem. Before this patch,
udc_stop() gets called first, sets the dum->driver = NULL (also calls
pullup(), which with the dum->driver being NULL doesn't do any harm), so
that when usb_gadget_disconnect() calls pullup(), it again won't do any
harm due to dum->driver being NULL.
But after this patch, there's a problem.
Regards,
--
Alex
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
