[Kernel-packages] [Bug 1849720] Re: KVM with e1000e and WinGuest Host OS on kernel 5.3 (ok with 5.0)
Hi I've faced a similar problem on Debian. My system is a Ryzen 1700 and I use windows and OSX VMs for work, and after I upgraded to Debian 10 I started getting random host freezes some time after booting the VMs. Strangely, I have a linux VM with almost identical config as the Windows 10 one and it doesn't cause a host crash. Since my processor is a rather troublesome one (I had hardware bugs in the past, which were worked around in the kernel), I assumed this to be another such case, especially since the bug didn't occur if I disabled SMT on the BIOS. However, since my system was very stable under Debian 9, I spent quite a few hours trying to root out what change could have cause it, and at least in my specific case (which I'm not sure is the same reported here), was traced back to a certain libvirt commit, more specifically this one: https://github.com/libvirt/libvirt/commit/3527f9dde67460e9f2d50ce52b8dade8c0848e86 So a suggestion to anyone affected: try to explicitly disable seccomp by setting `seccomp_sandbox = 0` in /etc/libvirt/qemu.conf -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1849720 Title: KVM with e1000e and WinGuest Host OS on kernel 5.3 (ok with 5.0) Status in linux package in Ubuntu: Confirmed Status in qemu package in Ubuntu: In Progress Bug description: After upgrading to Ubuntu 19.10 I noticed that all of my Windows Servers VMs would cause a hard crash on the Host OS shortly after starting up the VM. I tracked it down to the Guest OS attempting to use the Network Connection, and if I disabled the virtual NIC for the VM, everything runs OK (albeit without a working network connection for the Guest OS). Note that I'm just using the built in virtual network called "default" that get's installed by default and uses NAT forwarding. I believe the problem is related to AppArmor, as I noticed some errors present in various log files. Unfortunately, due to a time critical project I had to roll back to Ubuntu 19.04 and didn't capture any of the log files. I did, however, find another user on Reddit with the exact same problems that I encountered and he agreed to let me post the log files. Here are what are think are the relevant pieces from the log files: === Oct 22 22:59:23 brian-pc dnsmasq[2178]: exiting on receipt of SIGTERM Oct 22 22:59:23 brian-pc kernel: [ 67.001284] device virbr0-nic left promiscuous mode Oct 22 22:59:23 brian-pc kernel: [ 67.001298] virbr0: port 1(virbr0-nic) entered disabled state Oct 22 22:59:23 brian-pc NetworkManager[3557]: [1571799563.3862] device (virbr0-nic): released from master device virbr0 Oct 22 22:59:23 brian-pc gnome-shell[4401]: Removing a network device that was not added Oct 22 22:59:23 brian-pc gnome-shell[2463]: Removing a network device that was not added Oct 22 22:59:23 brian-pc avahi-daemon[1621]: Interface virbr0.IPv4 no longer relevant for mDNS. Oct 22 22:59:23 brian-pc avahi-daemon[1621]: Leaving mDNS multicast group on interface virbr0.IPv4 with address 192.168.122.1. Oct 22 22:59:23 brian-pc avahi-daemon[1621]: Withdrawing address record for 192.168.122.1 on virbr0. Oct 22 22:59:23 brian-pc NetworkManager[3557]: [1571799563.6859] device (virbr0): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') Oct 22 22:59:23 brian-pc gnome-shell[2463]: Removing a network device that was not added Oct 22 22:59:23 brian-pc gnome-shell[4401]: Removing a network device that was not added Oct 22 22:59:23 brian-pc dbus-daemon[1610]: [system] Activating via systemd: service name='org.freedesktop.nm_dispatcher' unit='dbus-org.freedesktop.nm-dispatcher.service' requested by ':1.192' (uid=0 pid=3557 comm="/usr/sbin/NetworkManager --no-daemon " label="unconfined") Oct 22 22:59:23 brian-pc systemd[1]: Starting Network Manager Script Dispatcher Service... Oct 22 22:59:23 brian-pc gnome-shell[2463]: Object NM.ActiveConnection (0x55ccfb376e50), has been already deallocated — impossible to get any property from it. This might be caused by the object having been destroyed from C code using something such as destroy(), dispose(), or remove() vfuncs. Oct 22 22:59:23 brian-pc gnome-shell[2463]: == Stack trace for context 0x55ccfb8d15f0 == Oct 22 22:59:23 brian-pc gnome-shell[2463]: #0 55ccfbc736c0 i resource:///org/gnome/shell/ui/status/network.js:1329 (7f52226be550 @ 56) Oct 22 22:59:23 brian-pc gnome-shell[2463]: #1 55ccfbc73628 i resource:///org/gnome/shell/ui/status/network.js:1346 (7f52226be5e0 @ 113) Oct 22 22:59:23 brian-pc gnome-shell[2463]: #2 55ccfbc73588 i resource:///org/gnome/shell/ui/status/network.js:2049 (7f52226c1940 @ 216) Oct 22 22:59:23 brian-pc gnome-shell[2463]: #3 55ccfbc734f0 i
[Kernel-packages] [Bug 1586195] Re: Realtek 8153-based ethernet adapter on usb3 eventually stops working requiring unplug/replug
The following workaround fixes for me: https://ubuntuforums.org/showthread.php?t=2226734 Here are the steps adapted for systemd: - Check the output of `ls /sys/bus/pci/drivers/xhci_hcd`. In my laptop I see ":00:14.0 bind new_id remove_id uevent unbind". The goal is to get the id of the xhci_hcd, for me it is ":00:14.0"(which I will use in the following steps). Also verify that the output shows "bind" and "unbind" as directory entries. The following commands should be executed in a root shell - Create a script to unbind the xhci: # cat > /usr/local/bin/unbind-xhci << "EOF" #!/bin/sh echo -n ':00:14.0' | tee /sys/bus/pci/drivers/xhci_hcd/unbind EOF # chmod +x /usr/local/bin/unbind-xhci - Create a script to rebind the xhci: # cat > /usr/local/bin/rebind-xhci << "EOF" #!/bin/sh echo -n ':00:14.0' | tee /sys/bus/pci/drivers/xhci_hcd/bind EOF # chmod +x /usr/local/bin/rebind-xhci - Create a systemd service that runs before sleep: # cat > /etc/systemd/system/unbind-xhci.service << "EOF" [Unit] Description=Unbind xHCI Host Controller Driver before sleep Before=sleep.target [Service] Type=oneshot ExecStart=/usr/local/bin/unbind-xhci [Install] WantedBy=sleep.target EOF - Create a systemd service that runs after wakeup: # cat > /etc/systemd/system/rebind-xhci.service << "EOF" [Unit] Description=Unbind xHCI Host Controller Driver after wakeup After=suspend.target After=hibernate.target After=hybrid-sleep.target [Service] Type=oneshot ExecStart=/usr/local/bin/rebind-xhci [Install] WantedBy=suspend.target WantedBy=hibernate.target WantedBy=hybrid-sleep.target EOF - Reload systemd # systemctl daemon-reload -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1586195 Title: Realtek 8153-based ethernet adapter on usb3 eventually stops working requiring unplug/replug Status in linux package in Ubuntu: Expired Status in linux source package in Xenial: Expired Bug description: Continued in Bug #1622322. Seems to be paired with dmesg entries like this on the trusty kernel: [24763.731054] usb 4-4.2: Disable of device-initiated U1 failed. [24763.734507] usb 4-4.2: Disable of device-initiated U2 failed. [24763.734594] r815x 4-4.2:2.0 eth3: unregister 'r815x' usb-:00:14.0-4.2, RTL8153 ECM Device [24763.802963] usb 4-4.2: Set SEL for device-initiated U1 failed. [24763.806450] usb 4-4.2: Set SEL for device-initiated U2 failed. [24763.809953] usb 4-4.2: usb_reset_and_verify_device Failed to disable LTM [24763.809953] . [24763.810579] usb 4-4.2: USB disconnect, device number 7 [24763.940765] userif-1: sent link down event. [24763.940769] userif-1: sent link up event. And like this on mainline: [ 259.731142] usb 4-3: new SuperSpeed USB device number 4 using xhci_hcd [ 259.752915] usb 4-3: New USB device found, idVendor=0bda, idProduct=8153 [ 259.752918] usb 4-3: New USB device strings: Mfr=1, Product=2, SerialNumber=6 [ 259.752920] usb 4-3: Product: USB 10/100/1000 LAN [ 259.752921] usb 4-3: Manufacturer: Realtek [ 259.752922] usb 4-3: SerialNumber: 0100 [ 259.909368] usb 4-3: reset SuperSpeed USB device number 4 using xhci_hcd [ 259.993254] r8152 4-3:1.0 eth0: v1.08.3 [ 260.079151] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready [ 296.266600] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready [ 300.785814] r8152 4-3:1.0 eth0: Stop submitting intr, status -71 [ 302.832697] usb 4-3: usb_reset_and_verify_device Failed to disable LTM [ 302.832697] . [ 302.833086] usb 4-3: USB disconnect, device number 4 Mainline version fyi: [0.00] Linux version 4.6.0-040600-generic (kernel@gomeisa) (gcc version 5.3.1 20160509 (Ubuntu 5.3.1-19ubuntu1) ) #201605151930 SMP Sun May 15 23:32:59 UTC 2016 I also get a similar message about 'Failed to disable LTM' if I just unplug the device before the breakage occurs. I've attached lsusb before the breakage occurs (so you can see the device) from the mainline kernel. Interestingly it appears the drivers on trusty and wily/mainline are different (r815x vs r8152) yet still exhibit similar troubles. I'm still digging around to try and figure out why this doesn't seem to be working, but I was wondering if you all had any idea what might be going pear-shaped here. Thanks! To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1586195/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
