I have found it in drafts and decided to give it one more try.
> > Although, I still get the following hang on shutdown:
> >
> > {HIGH TONE BEEP HERE}
> > usb.c: USB disconnect on device
> > usb-ohci.c: bogus NDP=255 for OHCI usb - 09:00.1
> > (the above two statements are repeated ~4x's)
> >
> > usb-ohci.c: USB HC Takeover failed!
> > usb.c: USB bus 3 deregistered
> > usb.c: USB disconnect on device
> > usb-ohci.c: USB HDC TakeOver failed!
> > usb.c: USB bus 4 deregistered
> > hcd.c: remove: 09:00.2, state 3
> > usb.c: USB disconnect on device
> >
> > {TOTAL SYSTEM HANG HERE!}
There are two kernel bugs here.
1. ehci driver does not tolerate well when hardware is removed
during operation and totally hangs attempting to stop controller. Driver
maintainer takes care of it (at least so he said).
2. cs driver is using wrong shutdown sequence - it first powers down
slot (thus effectively removing hardware) and then calling subordinate driver
cleanup. See bug 1 :-) It seems pretty easy to fix but I do not have hardware
and cannot test it.
So if anybody who has this problem is willing to help he will need to recompile
kernel to test a patch. Two times actually but second time it is really fast.
-andrej
