Bug 1770140

Summary: Don't perform tseg validation when reading configs from disk
Product: Red Hat Enterprise Linux Advanced Virtualization Reporter: Yanqiu Zhang <yanqzhan>
Component: libvirtAssignee: Daniel Berrangé <berrange>
Status: CLOSED ERRATA QA Contact: Meina Li <meili>
Severity: low Docs Contact:
Priority: low    
Version: 8.1CC: chhu, dyuan, jdenemar, jsuchane, knoel, lmen, pkrempa, xuzhang, yafu, yanqzhan
Target Milestone: rcKeywords: Triaged
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: libvirt-7.3.0-1.el8 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2021-11-16 07:49:56 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version: 7.1.0
Embargoed:
Attachments:
Description Flags
libvirtd.log none

Description Yanqiu Zhang 2019-11-08 09:46:56 UTC 
Description of problem:
Define a q35 guest with smm tseg, remove qemu-kvm-rhev and restart libvirtd, the vm disappears in virsh list.

Should not perform this validation when reading configs from disk. 

This causes user confusion :"say qemu is accidentally uninstalled, and libvirtd is restarted. the qemu driver capabilities are now empty, and any pre-existing configs will fail to load. virsh list --all is now empty, and users are confused." per bz1043572.


Version-Release number of selected component (if applicable):
libvirt-4.5.0-28.el7.x86_64
qemu-kvm-rhev-2.12.0-38.el7.x86_64

How reproducible:
100%

Steps to Reproduce:
1. start a q35 guest with smm tseg:
    <type arch='x86_64' machine='pc-q35-rhel7.6.0'>hvm</type>
...
    <smm state='on'>
      <tseg unit='MiB'>48</tseg>
    </smm>

# virsh list --all
 Id    Name                           State
----------------------------------------------------
 -     rhel7.7-q35                    shut off

2. # rpm -e --nodeps qemu-kvm-rhev
# virsh list --all
 Id    Name                           State
----------------------------------------------------
 -     rhel7.7-q35                    shut off

3. # systemctl restart libvirtd
# systemctl restart libvirtd
# virsh list --all
 Id    Name                           State
----------------------------------------------------

# cat libvirtd.log |grep error
2019-11-07 07:17:51.450+0000: 97241: debug : virEventPollRunOnce:643 : Poll got error event 4
2019-11-07 07:17:52.428+0000: 97351: error : virQEMUCapsNewForBinaryInternal:4555 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-07 07:17:52.428+0000: 97351: error : qemuDomainDefTsegPostParse:3670 : unsupported configuration: Setting TSEG size is not supported with this QEMU binary
2019-11-07 07:17:52.429+0000: 97351: error : virDomainObjListLoadAllConfigs:630 : Failed to load config for domain 'rhel7.7-q35'


Actual results:
Q35 guest with smm tseg disappears in virsh list if remove qemu-kvm-rhev and restart libvirtd

Expected results:
Should not perform this validation when reading configs from disk. 

Additional info:
1. If remove tseg, will not reproduce.
2. Reinstall qemu-kvm-rhev, restart libvirtd, the vm shows in list again.
3. The vm with smm tseg can be started successfully when qemu-kvm-rhev installed there.
Comment 2 Yanqiu Zhang 2019-11-08 10:38:26 UTC 
Created attachment 1633951 [details]
libvirtd.log
Comment 3 Yanqiu Zhang 2019-11-22 03:33:40 UTC 
Reproduces on rhel-av8.1.1

Pkgs info:
libvirt-daemon-5.6.0-8.module+el8.1.1+4828+0f2d78eb.x86_64
qemu-kvm-4.1.0-15.module+el8.1.1+4700+209eec8f.x86_64

# virsh list --all|grep rhel8.1-ovmf
 -    rhel8.1-ovmf     shut off

# rpm -e --nodeps qemu-kvm-core
# systemctl restart libvirtd

# virsh list --all|grep rhel8.1-ovmf

# cat /var/log/libvirt/libvirtd.log |grep "error :"
2019-11-22 04:19:29.583+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.583+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.583+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.584+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.584+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.585+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.585+0000: 12438: error : qemuDomainDefTsegPostParse:4271 : unsupported configuration: Setting TSEG size is not supported with this QEMU binary
2019-11-22 04:19:29.585+0000: 12438: error : virDomainObjListLoadAllConfigs:638 : Failed to load config for domain 'rhel8.1-ovmf'
2019-11-22 04:19:29.586+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.587+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.587+0000: 12438: error : qemuDomainDefTsegPostParse:4271 : unsupported configuration: Setting TSEG size is not supported with this QEMU binary
2019-11-22 04:19:29.588+0000: 12438: error : virDomainObjListLoadAllConfigs:638 : Failed to load config for domain 'full-rhel8.0'
2019-11-22 04:19:29.588+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.588+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.589+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.589+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
2019-11-22 04:19:29.590+0000: 12438: error : virCapabilitiesDomainDataLookupInternal:754 : invalid argument: could not find capabilities for arch=x86_64 domaintype=kvm 
2019-11-22 04:19:29.590+0000: 12438: error : virQEMUCapsNewForBinaryInternal:4668 : Cannot check QEMU binary /usr/libexec/qemu-kvm: No such file or directory
...
Comment 4 Peter Krempa 2020-02-19 11:08:17 UTC 
qemuDomainDefTsegPostParse does not skip the validation if qemuCaps is NULL. This is necessary for exactly this reason.
Comment 7 Peter Krempa 2021-03-10 17:17:05 UTC 
As of upstream commit:

commit 8d157c13ccd5201385e8038669e24c3c47d4f302
Author: Daniel P. Berrangé <berrange>
Date:   Tue Dec 10 11:35:43 2019 +0000

    qemu: honour parseOpaque instead of refetching caps
    
    The use of the parseOpaque parameter was mistakenly removed in
    
      commit 4a4132b4625778cf80acb9c92d06351b44468ac3
      Author: Daniel P. Berrangé <berrange>
      Date:   Tue Dec 3 10:49:49 2019 +0000
    
        conf: don't use passed in caps in post parse method
    
    causing the method to re-fetch qemuCaps that were already just
    fetched and put into parseOpaque.
    
    This is inefficient when parsing incoming XML, but for live
    XML this is more serious as it means we use the capabilities
    for the current QEMU binary on disk, rather than the running
    QEMU.
    
    That commit, however, did have a useful side effect of fixing
    a crasher bug in the qemu post parse callback introduced by
    
      commit 5e939cea896fb3373a6f68f86e325c657429ed3d
      Author: Jiri Denemark <jdenemar>
      Date:   Thu Sep 26 18:42:02 2019 +0200
    
        qemu: Store default CPU in domain XML
    
    The qemuDomainDefSetDefaultCPU() method in that patch did not
    allow for the possibility that qemuCaps would be NULL and thus
    resulted in a SEGV.
    
    This shows a risk in letting each check in the post parse
    callback look for qemuCaps == NULL. The safer option is to
    check once upfront and immediately stop (postpone) further
    validation.

The post-parse callbacks are postponed if qemuCaps are not present, so the VM will not vanish.

v5.10.0-224-g8d157c13cc
Comment 11 Meina Li 2021-05-18 10:45:21 UTC 
Verified Version:
libvirt-7.3.0-1.module+el8.5.0+11004+f4810536.x86_64
qemu-kvm-6.0.0-16.module+el8.5.0+10848+2dccc46d.x86_64


Verified Steps:
1. Prepare a guest with smm.
# virsh list --all
 Id   Name             State
---------------------------------
 -    lmn              shut off
# virsh dumpxml lmn | grep /smm -B2
    <smm state='on'>
      <tseg unit='MiB'>48</tseg>
    </smm>
2. Uninstall qemu-kvm package.
# rpm -e --nodeps qemu-kvm
# virsh list --all
 Id   Name             State
---------------------------------
 -    lmn              shut off
3. Restart libvirtd and check virsh list.
# systemctl restart libvirtd
# virsh list --all
 Id   Name             State
---------------------------------
 -    lmn              shut off
Comment 13 errata-xmlrpc 2021-11-16 07:49:56 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 (virt:av bug fix and enhancement update), 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/RHBA-2021:4684