usb stick problem [Was: HEADSUP usb2 (usb4bsd) to become default in 2 weeks.]

On Thu, Dec 25, 2008 at 11:35:27AM +0100, Hans Petter Selasky wrote:
> On Wednesday 24 December 2008, Alexey Shuvaev wrote:
> > On Wed, Dec 24, 2008 at 12:12:02PM +0100, Hans Petter Selasky wrote:
> > > Hi,
> > >
> > > I have made some patches for this kind of issues which are not yet
> > > in -current.
> > >
> > > You can try to fetch the latest from:
> > >
> > > svn --username anonsvn --password anonsvn \
> > >       checkout svn://svn.turbocat.net/i4b
> > >
> > > cd i4b/trunk/i4b/src/sys/dev
> > > cp -r usb2 /sys/dev/
> > > make -C /sys/modules/usb2 clean all install
> >
> > Hello!
> >
> > During buildworld I'm getting:
> >
> > [snip]
> >
> 
> The file is not in CVS:
> 
> % less /usr/obj/usr/src/tmp/usr/include/dev/usb2/include/urio2_ioctl.h
> /usr/obj/usr/src/tmp/usr/include/dev/usb2/include/urio2_ioctl.h: No such file 
> or directory
> 
> I've removed the file from my private SVN. Maybe that is where you got the 
> files from?
> 
Exactly. I've done so trying your newest sources to see if the problem with:

Dec 25 15:29:48 localhost kernel: usb2_alloc_device:1423: set address 2 failed (ignored)
Dec 25 15:29:48 localhost kernel: usb2_alloc_device:1458: getting device descriptor at addr 2 failed!
Dec 25 15:29:49 localhost kernel: usb2_req_re_enumerate:1362: addr=2, set address failed! (ignored)
Dec 25 15:29:49 localhost kernel: usb2_req_re_enumerate:1375: getting device descriptor at addr 2 failed!
Dec 25 15:29:50 localhost kernel: usb2_req_re_enumerate:1362: addr=2, set address failed! (ignored)
Dec 25 15:29:50 localhost kernel: usb2_req_re_enumerate:1375: getting device descriptor at addr 2 failed!
Dec 25 15:29:50 localhost kernel: uhub_reattach_port:417: could not allocate new device!

is gone. Not quite yet :(

Attached are 2 logs from messages with hw.usb2.debug=15.
The successful one is when I am at the system console (where these messages
are printed too) during the plug-in of the stick.
And when I am at another virtual console, the stick normally failed to attach
(second log).
Subtle timing probem?
(Attaching logs as octet-stream with the hope that mailing lists
strip them down while you actually get them.)

FWIW, with the old stack the stick is always recognized.

Hope, this helps solve this issue,
Alexey.