On Mon, 25 Apr 2005, David Brownell wrote: > On Sunday 24 April 2005 2:26 pm, Alan Stern wrote: > > ... > > > > At this point there was a normal "device suspended" debugging message, > > followed by a message that the root hub was suspending. However both > > power/state and *0/power/state contained 0, so as far as usbcore was > > concerned neither the root hub device nor its driver was suspended. > > The sysfs state _should_ be irrelevant.
Yes. In fact, this should have been an AUTO_STOP sort of suspend, so the sysfs and usbcore states both should have reflected that the root hub was running. > In practice, the sticking point has been the root hub timer. Paths that > affect the sysfs power/state file were also -- last I looked -- the > ones that would disable or enable the root hub timer. The root-hub timer should not present a problem now that Greg has accepted the root-hub IRQ patch. Obviously there hasn't yet been time for you to update the OHCI driver to take advantage of it. > > Then I unplugged the device. This caused a complete system lockup. > > Further testing showed that the (ints & OHCI_INTR_RD) case in ohci_irq() > > was being executed over and over again. Apparently the interrupt source > > was not getting cleared properly. > > > > This is very reminiscent of the problem you found with the original > > version of the root-hub-IRQ patch. Is it correct to assume that both > > problems will be fixed by the same change to the driver? > > The failure mode is reminiscent, but I've never seen that happen > with INTR_RD before. The OHCI spec says INTR_RD happens along > with transitioning to OHCI_USB_RESUME. What did you change > to make the hardware do that transition more than once? As far as I know, nothing. To be certain, I need to run the test again using an unpatched 2.6.12-rc3 kernel. It'll have to wait a few days until I have time... Alan Stern ------------------------------------------------------- SF.Net email is sponsored by: Tell us your software development plans! Take this survey and enter to win a one-year sub to SourceForge.net Plus IDC's 2005 look-ahead and a copy of this survey Click here to start! http://www.idcswdc.com/cgi-bin/survey?id=105hix _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
