On 12 Nov 2009, George Spelvin wrote: > >> I spoke too soon. It's more stable, but it can still be made to hang > >> up in the same way. I'm starting to suspect a PCIe interrupt > >> problem. I need to read the ehci spec and figure out if an interrupt > >> getting lost could cause such a hang... :-( > > > Sure it could. In the most recent kernels anyway; earlier kernels had > > an I/O watchdog timer which would kick in. But the timer is disabled > > on Intel hardware, because the hardware is supposed to be reliable. > > I was just RTFSing and seeing that. I was thinking that my diagnosis > was obviously wrong, since the code clearly says it can run without > any interrupt at all, but clearly that was written assuming the > watchdog was enabled. > > Is it easy to re-enable? Yes. Get rid of the ehci->need_io_watchdog = 0; line in drivers/usb/host/ehci-pci.c. Alan Stern -- 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
