Apparmor blocks usb devices in libvirt in Saucy

Bug #1245251 reported by Claude Durocher
24
This bug affects 4 people
Affects Status Importance Assigned to Milestone
libvirt (Ubuntu)
Fix Released
High
Unassigned
Saucy
Fix Released
High
Unassigned

Bug Description

=============================
SRU Justification:
1. Impact: usb devices can't be used under libvirt kvm guests
2. Development fix: allow libvirt to have read access to some information it now insists on having.
3. Stable fix: cherrypick of dev fix
4. Test case: create a libvirt VM with a usb device passed from the host
5. Regression potential: This allows libvirt to see a bit more host system information, however the security team sees no problem with it.
==============================

Upgraded a 13.04 64 bit to 13.10. Before the upgrade, I had KVM guest with usb devices working well. Since the upgrade, apparmor blocks access to usb devices with the following errors :

Oct 27 14:17:29 laptop kernel: [ 5771.844806] type=1400 audit(1382897849.445:339): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/bus/usb/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0
Oct 27 14:17:29 laptop kernel: [ 5771.844847] type=1400 audit(1382897849.445:340): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0
Oct 27 14:17:29 laptop kernel: [ 5771.844947] type=1400 audit(1382897849.445:341): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/bus/usb/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0
Oct 27 14:17:29 laptop kernel: [ 5771.844967] type=1400 audit(1382897849.445:342): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0
Oct 27 14:17:29 laptop kernel: [ 5771.845051] type=1400 audit(1382897849.445:343): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/bus/usb/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0
Oct 27 14:17:29 laptop kernel: [ 5771.845069] type=1400 audit(1382897849.445:344): apparmor="DENIED" operation="open" parent=1 profile="libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee" name="/dev/" pid=12253 comm="qemu-system-x86" requested_mask="r" denied_mask="r" fsuid=119 ouid=0

The profile looks fine :

/etc/apparmor.d/libvirt/libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee:

#
# This profile is for the domain whose UUID matches this file.
#

#include <tunables/global>

profile libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee {
  #include <abstractions/libvirt-qemu>
  #include <libvirt/libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee.files>

}

/etc/apparmor.d/libvirt/libvirt-655920dd-7b6f-f20b-bb77-b5bbaa133eee.files:

# DO NOT EDIT THIS FILE DIRECTLY. IT IS MANAGED BY LIBVIRT.
  "/var/log/libvirt/**/windows-xp.log" w,
  "/var/lib/libvirt/**/windows-xp.monitor" rw,
  "/var/run/libvirt/**/windows-xp.pid" rwk,
  "/run/libvirt/**/windows-xp.pid" rwk,
  "/var/run/libvirt/**/*.tunnelmigrate.dest.windows-xp" rw,
  "/run/libvirt/**/*.tunnelmigrate.dest.windows-xp" rw,
  "/home/vm/windowsxp.img" rw,
  "/dev/bus/usb/002/012" rw,
  "/dev/bus/usb/002/011" rw,
  "/dev/bus/usb/002/007" rw,

Revision history for this message
Claude Durocher (claude-d) wrote :

I found a workaround by adding the following to /etc/apparmor.d/abstractions/libvirt-qemu :

  /dev/bus/usb/ r,
  /etc/udev/udev.conf r,
  /sys/bus/ r,
  /sys/class/ r,

Revision history for this message
Jamie Strandboge (jdstrand) wrote :

I have not tested this myself, but the accesses listed in comment #1 look ok to add to /etc/apparmor.d/abstractions/libvirt-qemu. Seems like qemu changed the way it finds things and needs these new accesses.

Revision history for this message
Serge Hallyn (serge-hallyn) wrote : Re: [Bug 1245251] Re: Apparmor blocks usb devices in libvirt in Saucy

Quoting Jamie Strandboge (<email address hidden>):
> I have not tested this myself, but the accesses listed in comment #1
> look ok to add to /etc/apparmor.d/abstractions/libvirt-qemu. Seems like
> qemu changed the way it finds things and needs these new accesses.

Thanks, Jamie. I've not yet tested myself (kvm module not loading)
but I'll add the permissions after testing.

 importance: high
 status: confirmed

Changed in libvirt (Ubuntu):
importance: Undecided → High
status: New → Confirmed
Revision history for this message
Serge Hallyn (serge-hallyn) wrote :

@ClaudeD,

could you please list either a kvm command line or a virsh dumpxml output (or both) for a domain that has trouble? I tried to reproduce jsut passing in a yubikey, but had no permission problems.

(in my test, the kvm command was started with specific hostbus=2,hostaddr=4 information, so presumably you're passing in something more generic which requires qemu to look for host info...)

description: updated
Changed in libvirt (Ubuntu Saucy):
importance: Undecided → High
status: New → Confirmed
Revision history for this message
Claude Durocher (claude-d) wrote :
Download full text (3.2 KiB)

<domain type='kvm'>
   <name>windows-xp</name>
   <uuid>655920dd-7b6f-f20b-bb77-b5bbaa133eee</uuid>
   <memory unit='KiB'>1048576</memory>
   <currentMemory unit='KiB'>1048576</currentMemory>
   <vcpu placement='static' current='1'>2</vcpu>
   <os>
     <type arch='x86_64' machine='pc-i440fx-1.4'>hvm</type>
     <boot dev='hd'/>
   </os>
   <features>
     <acpi/>
     <apic/>
     <pae/>
   </features>
   <clock offset='localtime'/>
   <on_poweroff>destroy</on_poweroff>
   <on_reboot>restart</on_reboot>
   <on_crash>restart</on_crash>
   <devices>
     <emulator>/usr/bin/kvm-spice</emulator>
     <disk type='file' device='disk'>
       <driver name='qemu' type='raw' cache='writeback'/>
       <source file='/home/vm/windowsxp.img'/>
       <target dev='vda' bus='virtio'/>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x07'
function='0x0'/>
     </disk>
     <disk type='file' device='cdrom'>
       <driver name='qemu' type='raw'/>
       <target dev='hdc' bus='ide'/>
       <readonly/>
       <address type='drive' controller='0' bus='1' target='0' unit='0'/>
     </disk>
     <controller type='usb' index='0'>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x01'
function='0x2'/>
     </controller>
     <controller type='ide' index='0'>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x01'
function='0x1'/>
     </controller>
     <controller type='virtio-serial' index='0'>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x05'
function='0x0'/>
     </controller>
     <controller type='pci' index='0' model='pci-root'/>
     <interface type='network'>
       <mac address='52:54:00:4f:de:62'/>
       <source network='default'/>
       <model type='virtio'/>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x03'
function='0x0'/>
     </interface>
     <serial type='pty'>
       <target port='0'/>
     </serial>
     <console type='pty'>
       <target type='serial' port='0'/>
     </console>
     <channel type='spicevmc'>
       <target type='virtio' name='com.redhat.spice.0'/>
       <address type='virtio-serial' controller='0' bus='0' port='1'/>
     </channel>
     <input type='tablet' bus='usb'/>
     <input type='mouse' bus='ps2'/>
     <graphics type='spice' autoport='yes'/>
     <sound model='ich6'>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x04'
function='0x0'/>
     </sound>
     <video>
       <model type='qxl' ram='65536' vram='65536' heads='1'/>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x02'
function='0x0'/>
     </video>
     <hostdev mode='subsystem' type='usb' managed='yes'>
       <source>
         <vendor id='0x03f0'/>
         <product id='0x7504'/>
       </source>
     </hostdev>
     <hostdev mode='subsystem' type='usb' managed='yes'>
       <source>
         <vendor id='0x0fcf'/>
         <product id='0x1004'/>
       </source>
     </hostdev>
     <hostdev mode='subsystem' type='usb' managed='yes'>
       <source>
         <vendor id='0x9710'/>
         <product id='0x7780'/>
       </source>
     </hostdev>
     <memballoon model='virtio'>
       <address type='pci' domain='0x0000' bus='0x00' slot='0x06'
function='0x0'/>
...

Read more...

Revision history for this message
David Medberry (med) wrote :

Using the #1 work around, got my USB devices working again.
(Installed USBDeview in Windows to have a view of the USB tree in Windows guest.)

Revision history for this message
David Medberry (med) wrote :

Oh, and I had been using kvm-qemu via the virtual machine manager.
Here's what my qemu command looks like:

medberry@handsofblue:~$ sudo cat /proc/3378/cmdline |xargs -0
qemu-system-x86_64 -machine accel=kvm:tcg -name Fitbit -S -machine pc-i440fx-1.5,accel=kvm,usb=off -m 1024 -realtime mlock=off -smp 1,sockets=1,cores=1,threads=1 -uuid 162efb1b-c7c6-e1ee-ee57-cadf20930861 -no-user-config -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/Fitbit.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=localtime -no-shutdown -device piix3-usb-uhci,id=usb,bus=pci.0,addr=0x1.0x2 -drive file=/home/medberry/Documents/Win7.qcow,if=none,id=drive-ide0-0-0,format=raw -device ide-hd,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0,bootindex=1 -netdev tap,fd=27,id=hostnet0 -device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:8c:87:ac,bus=pci.0,addr=0x3 -chardev pty,id=charserial0 -device isa-serial,chardev=charserial0,id=serial0 -device usb-tablet,id=input0 -vnc 127.0.0.1:0 -vga std -device intel-hda,id=sound0,bus=pci.0,addr=0x4 -device hda-duplex,id=sound0-codec0,bus=sound0.0,cad=0 -device usb-host,hostbus=3,hostaddr=9,id=hostdev0 -device usb-host,hostbus=3,hostaddr=10,id=hostdev1 -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5

Revision history for this message
David Medberry (med) wrote :

xml for my domain

Revision history for this message
David Medberry (med) wrote :

@Serge

I was passing in specific devices all along and this just doesn't work in Saucy.

medberry@handsofblue:~$ dpkg -l libvirt-bin qemu-system
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Architecture Description
+++-===============================================-============================-============================-===================================================================================================
ii libvirt-bin 1.1.1-0ubuntu8 amd64 programs for the libvirt library
ii qemu-system 1.5.0+dfsg-3ubuntu5 amd64 QEMU full system emulation binaries

(and yes I was able to setup my new Fitbit Force after all the hoop jumping... Yay)

Revision history for this message
David Medberry (med) wrote :

Updates should be made to:
https://help.ubuntu.com/community/KVM/Managing
for the USB Passthrough method.

(It also may have some errors--are we still limited to USB 1.1?)

Revision history for this message
Serge Hallyn (serge-hallyn) wrote :

@med,

I'm confused. What is "the #1 workaround" in comment #6, and what are you saying in comment #9 that does not work?

My intent is to push the apparmor lines into policy as soon as 1.1.4 hits trusty, then SRU to saucy. Though if you're saying that does not work, then not :)

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package libvirt - 1.1.4-0ubuntu2

---------------
libvirt (1.1.4-0ubuntu2) trusty; urgency=low

  * debian/patches/9002-better_default_uri_virsh.patch: Update to fix the
    FTBFS.
 -- Chuck Short <email address hidden> Wed, 13 Nov 2013 11:04:29 -0500

Changed in libvirt (Ubuntu):
status: Confirmed → Fix Released
Revision history for this message
Brian Murray (brian-murray) wrote : Please test proposed package

Hello ClaudeD, or anyone else affected,

Accepted libvirt into saucy-proposed. The package will build now and be available at http://launchpad.net/ubuntu/+source/libvirt/1.1.1-0ubuntu8.2 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, and change the tag from verification-needed to verification-done. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed. In either case, details of your testing will help us make a better decision.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

Changed in libvirt (Ubuntu Saucy):
status: Confirmed → Fix Committed
tags: added: verification-needed
Revision history for this message
David Groos (djgroos) wrote :

Hello libvirt people,

I've been experimenting with kvm with saucy host and guests in the last couple days. I ran into the problem where, while the desktop vm launched correctly, I was unable to type in my password or click into that field--usb seemed to not work. So, I followed the directions in #13 and it's links and did the EnabledProposed. I installed the libvirt/1.1.1-0ubuntu8.2 package, but it didn't work at first. Now, I don't know if it would have at that point if I had added the USB controller (I think it was, I'm not in front of that computer now) at that time and set it to USB 2, but anyway, at first it didn't work. So then I installed all of the regular kvm packages in the proposed group that had the -0ubuntu8.2 numbering, I think there were 5 altogether. Anyway, after I added these 5 proposed packages, added the USB controller and set to USB 2, it work great. I wish I had more precise details. I'm now using one vm for a saucy desktop with spice and also have a 12.04 vm server using a .vdi that I imported from vbox (after converting to .gcow2). Thanks for your work. Let me know if I can provide further info.

Revision history for this message
Claude Durocher (claude-d) wrote :

Tested 1.1.1-0ubuntu8.2 on my workstation : usb devices were correctly detected and activated in my Windows XP guest.

tags: added: verification-done
removed: verification-needed
Revision history for this message
Leendert Keus (lj-keus) wrote :

I have installed the update 1.1.1-0ubuntu8.2 in a test environment and it's looking good now. Is it possible to upgrade our servers with apt-get upgrade (how long will it take before it is available?) or must I use the proposed package?

Revision history for this message
Stéphane Graber (stgraber) wrote : Update Released

The verification of this Stable Release Update has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regresssions.

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package libvirt - 1.1.1-0ubuntu8.2

---------------
libvirt (1.1.1-0ubuntu8.2) saucy-proposed; urgency=low

  * add d/p/util_use_w_flag_when_calling_iptables.patch (LP: #1245322)
  * debian/apparmor/libvirt-qemu: allow access to usb info (LP: #1245251)
  * debian/apparmor/libvirt-qemu: allow access to hugepages mounts
    (LP: #1250216)
 -- Serge Hallyn <email address hidden> Thu, 14 Nov 2013 10:09:24 -0600

Changed in libvirt (Ubuntu Saucy):
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.