RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Bug 1404952 - udev rewrites permissions set by libvirt on block devices that are closed
Summary: udev rewrites permissions set by libvirt on block devices that are closed
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Linux 7
Classification: Red Hat
Component: libvirt
Version: 7.4
Hardware: x86_64
OS: Linux
high
urgent
Target Milestone: rc
: ---
Assignee: Michal Privoznik
QA Contact: yafu
URL:
Whiteboard:
: 1401575 (view as bug list)
Depends On:
Blocks: 1404990 1404992 1446211
TreeView+ depends on / blocked
 
Reported: 2016-12-15 08:21 UTC by Jaroslav Suchanek
Modified: 2020-09-10 10:02 UTC (History)
8 users (show)

Fixed In Version: libvirt-3.0.0-1.el7
Doc Type: If docs needed, set a value
Doc Text:
Clone Of: 1354251
: 1404990 1404992 (view as bug list)
Environment:
Last Closed: 2017-08-01 17:19:14 UTC
Target Upstream Version:
Embargoed:

Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Knowledge Base (Solution) 2431201 0 None None None 2016-12-16 01:01:39 UTC
Red Hat Product Errata RHEA-2017:1846 0 normal SHIPPED_LIVE libvirt bug fix and enhancement update 2017-08-01 18:02:50 UTC

Description Jaroslav Suchanek 2016-12-15 08:21:25 UTC 
+++ This bug was initially created as a clone of Bug #1354251 +++

Description of problem:

A VM is unable to migrate from Host A to B due to EACCESS on a block device.

1. Destination (B) qemu process is created, opens block device DEV
2. In final stages of migration, qemu process closes device DEV
3. udev event is triggered rewriting the permissions of DEV
4. qemu process opens again device DEV before actually running the VM
5. qemu process fails to open DEV, receiving EACCESS as the permissions changed.

Version-Release number of selected component (if applicable):
udev-147-2.51.el6.x86_64
libvirt-0.10.2-29.el6_5.12.x86_64
qemu-kvm-0.12.1.2-2.415.el6_5.15.x86_64 (in function drives_reopen)

How reproducible:
100% - thank you Roman Hodain/Gordon Watson

Steps to Reproduce:
# cat /etc/udev/rules.d/51-qemu.rules
KERNEL=="sd*", GROUP="qemu"

# chown root:root /dev/sd*; ls -l /dev/sd*; ./test.out /dev/sda; sleep 1; ls -l /dev/sd*;
brw-rw----. 1 root root 8,  0 Jul 10 03:12 /dev/sda
brw-rw----. 1 root root 8, 16 Jul 10 03:14 /dev/sdb
Opening device /dev/sda
Closing device /dev/sda
brw-rw----. 1 root qemu 8,  0 Jul 10 03:29 /dev/sda
brw-rw----. 1 root root 8, 16 Jul 10 03:14 /dev/sdb

------------------------------------------------------------------------------

#include <stdio.h>
#include <fcntl.h>

int main ( int argc, char **argv ) 
{
    printf("Opening device %s\n", argv[1]);
    int fd = open(argv[1], O_RDWR);
    if (fd<0) {
        printf("Unable to open %s", argv[1]);
        return(1);
    }
    printf("Closing device %s\n", argv[1]);
    close(fd);
    return 0;
}

Actual results:
Permissions changed

Expected results:
Permissions not changed

Additional info:

Are RHEL6 and RHEL7 missing this rule in 50-udev-default.rules?
> ACTION!="add", GOTO="default_end"

https://cgit.freedesktop.org/systemd/systemd/commit/?id=22582bb2cbe85b40de5f561589e0468dac769515

The rule which is causing trouble in this case is:
SUBSYSTEM=="block", GROUP="disk"

From our tests until now, this can be easily worked around by adding the qemu process to the disk group. However, what is the correct behavior here? Should udev rewrite it or not?

I don't see how libvirt can prevent it from happening. Perhaps there is, but I cannot see any libudev mechanism to prevent this. And since this happens within qemu (it's really a close followed by an open) I am not sure how libvirt could be modified to reapply these permissions mid-flight.

Similar issues and related information:
https://www.redhat.com/archives/libvir-list/2011-April/msg00113.html
https://bugzilla.redhat.com/show_bug.cgi?id=588348

--- Additional comment from Michal Sekletar on 2016-07-11 14:11:49 CEST ---

There is an earlier commit that actually introduced this behaviour, i.e. not changing permissions on change events [1]. Commit you linked in bug description is only a rework of that. Conceptually, it made possible to change permissions of a device node on other event types as well, but introduced check in the rules file that prevents resetting permissions on change events. This was previously handled directly in the C code.

That being said, I think we can also introduce such change in rules. Basically we would only add 'ACTION!="add", GOTO="default_end"' at appropriate place and jump label LABEL="default_end" at the end of file. We don't need part of the patch that touches C code because we don't have that code in RHEL6 version of udev.

[1] https://github.com/systemd/systemd/commit/48a849ee17fb25e0001bfcc0f28a4aa633d016a1


--- Additional comment from Michal Privoznik on 2016-10-26 14:38:12 CEST ---

First round of the patches proposed on the upstream list:

https://www.redhat.com/archives/libvir-list/2016-October/msg01151.html

--- Additional comment from Michal Privoznik on 2016-11-24 15:50:17 CET ---

Patches proposed upstream:

https://www.redhat.com/archives/libvir-list/2016-November/msg01208.html

However, the change is fairly invasive, therefore I don't think we can guarantee stable package if they are backported.

--- Additional comment from Michal Privoznik on 2016-12-07 09:40:33 CET ---

v2 proposed upstream:

https://www.redhat.com/archives/libvir-list/2016-December/msg00231.html
Comment 1 Michal Privoznik 2016-12-15 08:34:00 UTC 
Moving to POST:

commit f444faa94a0e30f7dfdd47dce18b526abb0aaa9f
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Dec 6 17:35:12 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Enable mount namespace
    
    https://bugzilla.redhat.com/show_bug.cgi?id=1404952
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 661887f558208074169b0d3340c457108b6a023d
Author:     Michal Privoznik <mprivozn>
AuthorDate: Fri Nov 18 16:34:45 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Let users opt-out from containerization
    
    Given how intrusive previous patches are, it might happen that
    there's a bug or imperfection. Lets give users a way out: if they
    set 'namespaces' to an empty array in qemu.conf the feature is
    suppressed.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit f95c5c48d416dcdab2f3ee7718f12a02833c9339
Author:     Michal Privoznik <mprivozn>
AuthorDate: Fri Nov 18 15:19:12 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Manage /dev entry on RNG hotplug
    
    When attaching a device to a domain that's using separate mount
    namespace we must maintain /dev entries in order for qemu process
    to see them.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit f5fdf23a68d5c9838890451c1c50b4ae1062d8d2
Author:     Michal Privoznik <mprivozn>
AuthorDate: Fri Nov 18 14:53:27 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Manage /dev entry on chardev hotplug
    
    When attaching a device to a domain that's using separate mount
    namespace we must maintain /dev entries in order for qemu process
    to see them.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 6e57492839c0f644ade6b4174993f33f72b66ba3
Author:     Michal Privoznik <mprivozn>
AuthorDate: Wed Nov 16 15:27:47 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Manage /dev entry on hostdev hotplug
    
    When attaching a device to a domain that's using separate mount
    namespace we must maintain /dev entries in order for qemu process
    to see them.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 81df21507bef94ae53a056156e4aa6661f29237a
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 16:53:04 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Manage /dev entry on disk hotplug
    
    When attaching a device to a domain that's using separate mount
    namespace we must maintain /dev entries in order for qemu process
    to see them.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit eadaa975480d10eb057eb72bf833888e88e948e8
Author:     Michal Privoznik <mprivozn>
AuthorDate: Wed Nov 23 11:52:57 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Enter the namespace on relabelling
    
    Instead of trying to fix our security drivers, we can use a
    simple trick to relabel paths in both namespace and the host.
    I mean, if we enter the namespace some paths are still shared
    with the host so any change done to them is visible from the host
    too.
    Therefore, we can just enter the namespace and call
    SetAllLabel()/RestoreAllLabel() from there. Yes, it has slight
    overhead because we have to fork in order to enter the namespace.
    But on the other hand, no complexity is added to our code.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 2160f338a74543634e26aeddef1e4c63184660da
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 16:10:23 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare RNGs when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 8ec8a8c5ffa0a4b662013313116b4f166bfe989e
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 16:03:02 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare inputs when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 2c654490f355c8d8e7ad4748952008391299b411
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 15:25:15 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare TPM when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 4e4451019cb2e6dea355e93e946e0169023753c6
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 15:17:05 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare chardevs when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 73267cec46e08e74f6297c44b8f47c68180b3712
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 14:37:52 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare hostdevs when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 054202d02062c313e01e6c8b0084b91a738d13aa
Author:     Michal Privoznik <mprivozn>
AuthorDate: Mon Nov 14 17:36:45 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Prepare disks when starting a domain
    
    When starting a domain and separate mount namespace is used, we
    have to create all the /dev entries that are configured for the
    domain.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit bb4e529664a6e1ef08030aefc96f21f14eba2aea
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 11:30:18 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu: Spawn qemu under mount namespace
    
    Prime time. When it comes to spawning qemu process and
    relabelling all the devices it's going to touch, there's inherent
    race with other applications in the system (e.g. udev). Instead
    of trying convincing udev to not touch libvirt managed devices,
    we can create a separate mount namespace for the qemu, and mount
    our own /dev there. Of course this puts more work onto us as we
    have to maintain /dev files on each domain start and device
    hot(un-)plug. On the other hand, this enhances security also.
    
    From technical POV, on domain startup process the parent
    (libvirtd) creates:
    
      /var/lib/libvirt/qemu/$domain.dev
      /var/lib/libvirt/qemu/$domain.devpts
    
    The child (which is going to be qemu eventually) calls unshare()
    to create new mount namespace. From now on anything that child
    does is invisible to the parent. Child then mounts tmpfs on
    $domain.dev (so that it still sees original /dev from the host)
    and creates some devices (as explained in one of the previous
    patches). The devices have to be created exactly as they are in
    the host (including perms, seclabels, ACLs, ...). After that it
    moves $domain.dev mount to /dev.
    
    What's the $domain.devpts mount there for then you ask? QEMU can
    create PTYs for some chardevs. And historically we exposed the
    host ends in our domain XML allowing users to connect to them.
    Therefore we must preserve devpts mount to be shared with the
    host's one.
    
    To make this patch as small as possible, creating of devices
    configured for domain in question is implemented in next patches.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit a5896e8ca404a2e975808728328e44efd49a7960
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 15 11:28:51 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    qemu_cgroup: Expose defaultDeviceACL
    
    This is a list of devices that qemu needs for its run (apart from
    what's configured for domain). The devices on the list are
    enabled in the CGroups by default so they will be good candidates
    for initial /dev for new qemu.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 5ac52bd0fe80d1741071250f485ae54375508e48
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Dec 6 16:06:02 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virscsivhost: Introduce virSCSIVHostDeviceGetPath
    
    We will need this function in near future so that we know what
    /dev device corresponds to the SCSI device.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 6bcacd55e537c0fc3b793949637197e82a9dffcb
Author:     Michal Privoznik <mprivozn>
AuthorDate: Wed Nov 16 15:27:20 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virscsi: Introduce virSCSIDeviceGetPath
    
    We will need this function in near future so that we know what
    /dev device corresponds to the SCSI device.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit c4237d8e0ca090c7db76e5a226efa0ed2305835d
Author:     Michal Privoznik <mprivozn>
AuthorDate: Wed Nov 16 15:26:59 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virusb: Introduce virUSBDeviceGetPath
    
    We will need this function in near future so that we know what
    /dev device corresponds to the USB device.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 654b4d48bcdeeaf31df131644544bb1277f0f8bb
Author:     Michal Privoznik <mprivozn>
AuthorDate: Tue Nov 22 11:14:08 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virfile: Introduce ACL helpers
    
    Namely, virFileGetACLs, virFileSetACLs, virFileFreeACLs and
    virFileCopyACLs. These functions are going to be required when we
    are creating /dev for qemu. We have copy anything that's in
    host's /dev exactly as is. Including ACLs.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 1a7c9a5d5087d562fabfd7b7ff3cd1b9b19e9419
Author:     Michal Privoznik <mprivozn>
AuthorDate: Thu Nov 10 16:17:48 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virfile: Introduce virFileSetupDev
    
    This part of code that LXC currently uses will be reused so move
    to a generic function.
    
    Signed-off-by: Michal Privoznik <mprivozn>

commit 48a12d3b2554cc7a4255ef9ff8564c0a3ef7c1b3
Author:     Michal Privoznik <mprivozn>
AuthorDate: Thu Nov 10 14:55:48 2016 +0100
Commit:     Michal Privoznik <mprivozn>
CommitDate: Thu Dec 15 09:25:16 2016 +0100

    virprocess: Introduce virProcessSetupPrivateMountNS
    
    This part of code that LXC currently uses will be reused so move
    to a generic function.
    
    Signed-off-by: Michal Privoznik <mprivozn>

v2.5.0-121-gf444faa94a
Comment 2 Jaroslav Suchanek 2016-12-15 08:37:59 UTC 
*** Bug 1401575 has been marked as a duplicate of this bug. ***
Comment 5 yafu 2017-05-18 10:45:01 UTC 
Hi,Michal,

I tried to verify this bug and found a issue. If there is a running guest, When I disable creating namespace for qemu process in the qemu.conf and restart libvirtd, then destroy the guest and start the guest again. Libvirt still creates the namespace for the guest. 
Would you help to check the issue please? Thanks a lot.

Steps to reproduce:
1.Start a guest:
#virsh start rhel7.3-min 

2.Disable the namespace in qemu.conf
#vim /etc/libvirt/qemu.conf
namespaces = [  ]
 
3.Restart libvirtd service:
#systemctl restart libvirtd

4.Destroy and start the guest again:
#virsh destroy rhel7.3-min

5.Start the guest again:
#virsh start rhel7.3-min

6.libvirt still create namespace for the guest:
#lsns | grep qemu
lsns | grep -i qemu
4026532522 mnt        1 31993 qemu   /usr/libexec/qemu-kvm -name guest=rhel7.3-min ...

7.Need to restart libvirtd service again after step 4 to not create namespace for the guest.
Comment 6 Michal Privoznik 2017-05-22 08:41:04 UTC 
Yes(In reply to yafu from comment #5)
> Hi,Michal,
> 

Oh yes. It is a bug. However, given how big the feature is I think it can be tracked in a separate bug instead of this one.
Comment 7 yafu 2017-05-22 09:16:01 UTC 
(In reply to Michal Privoznik from comment #6)
> Yes(In reply to yafu from comment #5)
> > Hi,Michal,
> > 
> 
> Oh yes. It is a bug. However, given how big the feature is I think it can be
> tracked in a separate bug instead of this one.

Thanks Michal. File a separate bug to track the issue in comment 5:
https://bugzilla.redhat.com/show_bug.cgi?id=1453142
Comment 8 yafu 2017-06-22 02:48:34 UTC 
Reproduced with:
qemu-kvm-rhev-0.12.1.2-2.491.el6_8.3.x86_64
libvirt-0.10.2-60.el6.x86_64
udev-147-2.72.el6.x86_64
Test steps are as https://bugzilla.redhat.com/show_bug.cgi?id=1354690#c41.
Comment 9 yafu 2017-06-22 04:28:15 UTC 
Verify with:
qemu-kvm-rhev-2.9.0-12.el7.x86_64
libvirt-3.2.0-14.el7.x86_64
systemd-219-41.el7.x86_64

Test steps:
1.Start a guest with block device:
  <disk type='block' device='lun' sgio='unfiltered'>
      <driver name='qemu' type='raw' cache='none'/>
      <source dev='/dev/disk/by-path/ip-10.66.71.72:3260-iscsi-iqn.2016-03.com.virttest:logical-pool.target-lun-0'/>
      <target dev='vdb' bus='scsi'/>
      <shareable/>
      <alias name='scsi0-0-0-1'/>
      <address type='drive' controller='0' bus='0' target='0' unit='1'/>
    </disk>

2.Check the dac and selinux context of the block device after guest start:
# ll /dev/disk/by-path/ip-10.66.71.72:3260-iscsi-iqn.2016-03.com.virttest:logical-pool.target-lun-0
lrwxrwxrwx. 1 root root 9 Jun 22 11:07 /dev/disk/by-path/ip-10.66.71.72:3260-iscsi-iqn.2016-03.com.virttest:logical-pool.target-lun-0 -> ../../sdb
# ll -Z /dev/sdb
brw-rw----. root disk system_u:object_r:fixed_disk_device_t:s0 /dev/sdb

3.Check the dac and selinux context in the qemu process namespace:
#lsns | grep -i qemu
# lsns | grep -i qemu
4026532392 mnt        1 15678 qemu   /usr/libexec/qemu-kvm -name guest=rhel7.3-min,debug-threads=on ...
#nsenter -t 15678 -m
## ll -Z /dev/sdb
brw-rw----. qemu qemu system_u:object_r:svirt_image_t:s0 /dev/sdb

According to the test results of step 2 and step 3, could see libvirt does not change the dac/selinux context of the block device on the host os when starting a guest, it changes the dac and selinux context in the qemu process's namespace insteadly. So the dac/selinux context changing during guest migration will have no effect on the guest starting in the target host.
Comment 10 yafu 2017-06-22 11:30:44 UTC 
Verify the bug about devices used in the guest when enable namespaces with:
qemu-kvm-rhev-2.9.0-12.el7.x86_64
libvirt-3.2.0-14.el7.x86_64
systemd-219-41.el7.x86_64

Test steps:
1.Start a guest with iscsi disk:
#virsh dumpxml rhel7.3-min
...
   <disk type='block' device='disk'>
      <driver name='qemu' type='raw'/>
      <source dev='/dev/sdb'/>
      <backingStore/>
      <target dev='vdb' bus='virtio'/>
      <alias name='virtio-disk1'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/>
    </disk>
...

2.Check the dac/selinux context in both host os and qemu namespaces:
(1)In the host os:
# ll -Z /dev/sdb
brw-rw----. root disk system_u:object_r:fixed_disk_device_t:s0 /dev/sdb
(2)In the qemu process namespace:
#nsenter -t 15678 -m
## ll -Z /dev/sdb
brw-rw----. qemu qemu system_u:object_r:svirt_image_t:s0 /dev/sdb 

2.Hotunplug the iscsi disk:
#virsh detach-disk rhel7.3 vdb
Disk detached successfully

3.Check the dac/selinux context in both host os and qemu namespaces:
(1)In the host os:
# ll -Z /dev/sdb
brw-rw----. root disk system_u:object_r:fixed_disk_device_t:s0 /dev/sdb
(2)In the qemu process namespace:
#nsenter -t 15678 -m
#ll -Z /dev/sdb
brw-rw----. root root system_u:object_r:fixed_disk_device_t:s0 /dev/sdb

4.Hotplug the iscsi disk:
#virsh attach-disk rhel7.3-min /dev/sdb vdb
Disk attached successfully

5.Check the dac/selinux context in both host os and qemu namespaces:
(1)In the host os:
# ll -Z /dev/sdb
brw-rw----. root disk system_u:object_r:fixed_disk_device_t:s0 /dev/sdb
(2)In the qemu process namespace:
#nsenter -t 15678 -m
## ll -Z /dev/sdb
brw-rw----. qemu qemu system_u:object_r:svirt_image_t:s0 /dev/sdb 

6.Check the dac/selinux context in the host os after guest destroy:
#virsh destroy rhel7.3-min
# ll -Z /dev/sdb
brw-rw----. root root system_u:object_r:fixed_disk_device_t:s0 /dev/sdb

Also test with the rng/chardev/hostdev/input devices, libvirt can manage the source file/dev in the qemu process's namespaces correctly.
Comment 11 errata-xmlrpc 2017-08-01 17:19:14 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2017:1846
Comment 12 errata-xmlrpc 2017-08-01 23:59:00 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2017:1846
Note You need to log in before you can comment on or make changes to this bug.