From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:46498) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TRqSW-0000mE-Pu for qemu-devel@nongnu.org; Fri, 26 Oct 2012 16:24:50 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TRqSV-0000Dg-MQ for qemu-devel@nongnu.org; Fri, 26 Oct 2012 16:24:48 -0400 Received: from mx1.redhat.com ([209.132.183.28]:51970) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TRqSV-0000DO-E8 for qemu-devel@nongnu.org; Fri, 26 Oct 2012 16:24:47 -0400 Date: Fri, 26 Oct 2012 16:24:44 -0400 From: Jason Baron Message-ID: <20121026202444.GA25597@redhat.com> References: <20121024145527.GB2107@hedwig.ini.cmu.edu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20121024145527.GB2107@hedwig.ini.cmu.edu> Subject: Re: [Qemu-devel] q35: usb keyboard trouble List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Gabriel L. Somlo" Cc: pbonzini@redhat.com, kraxel@redhat.com, qemu-devel@nongnu.org, hdegoede@redhat.com On Wed, Oct 24, 2012 at 10:55:28AM -0400, Gabriel L. Somlo wrote: > Jason, > > Commit d8b0dbdba325773469733222a167b54aca74de55 in the q35 tree breaks > '-usbdevice keyboard' for me. Instead of being able to type at the VM, > none of the keypresses make it through, and qemu stderr soon starts > logging this error: "usb-kbd: warning: key event queue full". > > Specifically, it's the introduction of ich9-usb-ehci1 that causes the > issue (in hw/pc_q35.c, line 197): > > + pci_create_simple_multifunction( > + host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_EHCI1_FUNC), > + true, "ich9-usb-ehci1"); > > Commenting out the above makes the problem go away. > > Thanks much, > --Gabriel > > Thanks for the bug report! What OS are you running? I'm not seeing any issue with -usbdevice keyboard on Fedora. Paolo pointed out that I was missing some ich9 specific initialization from docs/ich9-ehci-uhci.cfg. I've added that in the patch below. I have no idea if that will resolve this issue for you. Thanks, -Jason diff --git a/hw/pc_q35.c b/hw/pc_q35.c index a72ad36..22b72ed 100644 --- a/hw/pc_q35.c +++ b/hw/pc_q35.c @@ -200,20 +200,27 @@ static void pc_q35_init_late(BusState **idebus, ISADevice *rtc_state, idebus[1] = qdev_get_child_bus(&ahci->qdev, "ide.1"); if (usb_enabled) { + int i; + PCIDevice *usb; + DeviceState *usb_qdev; + /* Should we create 6 UHCI according to ich9 spec? */ - pci_create_simple_multifunction( - host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_UHCI1_FUNC), - true, "ich9-usb-uhci1"); - pci_create_simple_multifunction( - host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_UHCI2_FUNC), - true, "ich9-usb-uhci2"); - pci_create_simple_multifunction( - host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_UHCI3_FUNC), - true, "ich9-usb-uhci3"); - pci_create_simple_multifunction( + usb = pci_create_multifunction( host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_EHCI1_FUNC), true, "ich9-usb-ehci1"); - /* XXX: EHCI */ + usb_qdev = &usb->qdev; + usb_qdev->id = g_strdup("ich9-usb-bus"); + qdev_init_nofail(usb_qdev); + + for (i = 0; i < 3; i++) { + usb = pci_create_multifunction( + host_bus, PCI_DEVFN(ICH9_USB_DEV, ICH9_USB_UHCI1_FUNC + i), + true, "ich9-usb-uhci1"); + usb_qdev = &usb->qdev; + qdev_prop_set_string(usb_qdev, "masterbus", "ich9-usb-bus.0"); + qdev_prop_set_uint32(usb_qdev, "firstport", i * 2); + qdev_init_nofail(usb_qdev); + } } /* TODO: Populate SPD eeprom data. */