On Fri, 2010-11-05 at 10:05 -0400, Sarah Sharp wrote: > On Fri, Oct 22, 2010 at 10:40:30PM +0200, Lawrence Rust wrote: > > Hi, > > > > Just built a 2.6.36 kernel and on re-booting got this error: > > > > Oct 22 10:13:05 gagarin kernel: [ 5.157428] xhci_hcd 0000:01:00.0: Failed to enable MSI-X > > Oct 22 10:13:05 gagarin kernel: [ 5.157448] xhci_hcd 0000:01:00.0: failed to allocate MSI entry > > > > followed by a stream of these... > > > > Oct 22 10:13:05 gagarin kernel: [ 5.891099] xhci_hcd 0000:01:00.0: Spurious interrupt. > > Oct 22 10:13:05 gagarin kernel: [ 9.038074] xhci_hcd 0000:01:00.0: Spurious interrupt. > > Oct 22 10:13:05 gagarin kernel: [ 9.088027] xhci_hcd 0000:01:00.0: Spurious interrupt. > > Oct 22 10:13:05 gagarin kernel: [ 9.137971] xhci_hcd 0000:01:00.0: Spurious interrupt. > > > > repeated until re-boot. > > Lawrence, > > Does the xHCI host controller still work, despite the warning messages? Sarah, Yes, the system is totally functional. I don't have a USB3 peripheral to test at the moment, but USB2 peripherals are working fine with the USB3 controller. > Do USB devices plugged into the blue ports still work? (I assume you > have some of those, if not, you may need to figure out which ports are > connected to the xHCI host controller by watching to see if xHCI > messages like "WARN short control transfer" show up in the dmesg when > you plug a USB device into a port.) Yes, both USB2 and the blue USB3 ports are working normally. > I think what's happening is the i915 driver registers on the interrupt > later than the xHCI driver, so whenever the i915 hardware interrupts the > system, the xHCI driver has to check whether the host generated the > interrupt. It hasn't, so that printk gets triggered. > > I think the fix is just to remove the printk. Can you test the attached > patch with CONFIG_PCI_MSI=n? Tried and tested. The "Spurious interrupt" message is gone. but this warning still remains: Nov 5 09:22:15 gagarin kernel: [ 5.157235] xhci_hcd 0000:01:00.0: Failed to enable MSI-X Nov 6 09:22:15 gagarin kernel: [ 5.157235] xhci_hcd 0000:01:00.0: failed to allocate MSI entry Shouldn't this be skipped if MSI's are disabled, either by config or via the nomsi kernel parameter? -- Lawrence Rust -- 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
