libvirt users - Jan 2017 - Re: LibVirt query CPU Model support and restore operation

hi,

yes we are talking about same host here.

You could've just used -cpu host-model-only
>> I dont want to host model ( it is working with host model) .
AS KVM seems to allow support Haswell with native commands . Does this mean
in my case KVM is allowing software emulated cpu model.

Can libvirt allow  software-emulated for non host model ?  if yes how .

thanks





On Tue, Jan 24, 2017 at 1:32 PM, Martin Kletzander <mkletzan@redhat.com>
wrote:
> On Tue, Jan 24, 2017 at 10:08:12AM +0530, akhilesh rawat wrote:
>
>> Hello ,
>>
>> It was working good working with kvm management tools using libvirt .
>> virsh/virt-manager .
>>
>> But then i got annoyed when management tool did not allow me to change
the
>> CPU model while creating new virtual machine .
>> error
>>
>>
> Are we talking about the same host?
>
> root@kvm-server qemu]# virt-install --virt-type kvm --name compute-2 --cpu
>> Haswell-noTSX --ram=61440 --vcpus=20 --os-type=linux
--os-variant=generic
>> --disk
>> compute2-disk0.qcow2,device=disk,bus=ide,size=300,sparse=tru
>> e,format=qcow2
>> --pxe --network bridge=virbr0,model=e1000 --network
>> bridge=virbr0,model=e1000 --network bridge=virbr0,model=virtio
--network
>> bridge=virbr0,model=virtio --graphics vnc,port=5906 --noautoconsole
>>
>> Starting install...
>> ERROR    unsupported configuration: guest and host CPU are not
compatible:
>> Host CPU does not provide required features: invpcid, erms, bmi2, smep,
>> avx2, bmi1, fsgsbase, movbe, fma
>> Domain installation does not appear to have been successful.
>>
>> I could not resolve this error .
>>
>>
> You could've just used -cpu host-model-only.
>
> Finally i came across the post which say that libvirt cross verifies  flags
>> under CPU model present in
>> /usr/share/libvirt/cpu_map.xml and cat /proc/cpuinfo .
>>
>> An yes the flag complained were not present in host proc/cpuinfo.
>>
>> Question 1: Is it really needed for Libvirt to do this check .
>>
>>
> It is, otherwise users might not get what they asked for.  Or it might
> be software-emulated.
>
> AS KVM seems to allow support Haswell with native commands . Why libvirt is
>> doing this check then ?
>>
>> As i was not sure where the problem lies and i was using libvirt 2.0.0
>> version i upgraded to 2.5.0 by compiling form source .
>> But then i faces issue using virsh/virt-manager as seem it was not
>> compatilble with upgarded version of libvirt.
>>
>> Now my system is little messed  around libvirt . I did reinstall/reboot
>> Quite many things not working .
>>
>> Question 2 : How can i restore original Libvirt 2.0.0 .
>>
>>
> Depends on your distro.  You can just uninstall and install the older
> version if your package managed doesn't support downgrading.  But try
to
> keep on the newer one and fix your workflow.
>
>
>> br aki
>>
>
> _______________________________________________
>> libvirt-users mailing list
>> libvirt-users@redhat.com
>> https://www.redhat.com/mailman/listinfo/libvirt-users
>>
>

On Tue, Jan 24, 2017 at 04:21:14PM +0530, akhilesh rawat
wrote:
>hi,
>
>yes we are talking about same host here.
>
>You could've just used -cpu host-model-only
>>> I dont want to host model ( it is working with host model) .
>
>AS KVM seems to allow support Haswell with native commands . Does this mean
>in my case KVM is allowing software emulated cpu model.
>
I don't remember if it gets emulated or just dropped.  Do you remember
if you saw those in /proc/cpuinfo in the guest?  Maybe QEMU just dropped
them on startup.  I believe that was why we added the code that checks
for it.
>Can libvirt allow  software-emulated for non host model ?  if yes how .
>
You can force each feature, that way it could work, but I'm not sure
with all the refactoring that was there.
>thanks
>
>
>
>
>
>On Tue, Jan 24, 2017 at 1:32 PM, Martin Kletzander
<mkletzan@redhat.com>
>wrote:
>
>> On Tue, Jan 24, 2017 at 10:08:12AM +0530, akhilesh rawat wrote:
>>
>>> Hello ,
>>>
>>> It was working good working with kvm management tools using libvirt
.
>>> virsh/virt-manager .
>>>
>>> But then i got annoyed when management tool did not allow me to
change the
>>> CPU model while creating new virtual machine .
>>> error
>>>
>>>
>> Are we talking about the same host?
>>
>> root@kvm-server qemu]# virt-install --virt-type kvm --name compute-2
--cpu
>>> Haswell-noTSX --ram=61440 --vcpus=20 --os-type=linux
--os-variant=generic
>>> --disk
>>> compute2-disk0.qcow2,device=disk,bus=ide,size=300,sparse=tru
>>> e,format=qcow2
>>> --pxe --network bridge=virbr0,model=e1000 --network
>>> bridge=virbr0,model=e1000 --network bridge=virbr0,model=virtio
--network
>>> bridge=virbr0,model=virtio --graphics vnc,port=5906 --noautoconsole
>>>
>>> Starting install...
>>> ERROR    unsupported configuration: guest and host CPU are not
compatible:
>>> Host CPU does not provide required features: invpcid, erms, bmi2,
smep,
>>> avx2, bmi1, fsgsbase, movbe, fma
>>> Domain installation does not appear to have been successful.
>>>
>>> I could not resolve this error .
>>>
>>>
>> You could've just used -cpu host-model-only.
>>
>> Finally i came across the post which say that libvirt cross verifies 
flags
>>> under CPU model present in
>>> /usr/share/libvirt/cpu_map.xml and cat /proc/cpuinfo .
>>>
>>> An yes the flag complained were not present in host proc/cpuinfo.
>>>
>>> Question 1: Is it really needed for Libvirt to do this check .
>>>
>>>
>> It is, otherwise users might not get what they asked for.  Or it might
>> be software-emulated.
>>
>> AS KVM seems to allow support Haswell with native commands . Why
libvirt is
>>> doing this check then ?
>>>
>>> As i was not sure where the problem lies and i was using libvirt
2.0.0
>>> version i upgraded to 2.5.0 by compiling form source .
>>> But then i faces issue using virsh/virt-manager as seem it was not
>>> compatilble with upgarded version of libvirt.
>>>
>>> Now my system is little messed  around libvirt . I did
reinstall/reboot
>>> Quite many things not working .
>>>
>>> Question 2 : How can i restore original Libvirt 2.0.0 .
>>>
>>>
>> Depends on your distro.  You can just uninstall and install the older
>> version if your package managed doesn't support downgrading.  But
try to
>> keep on the newer one and fix your workflow.
>>
>>
>>> br aki
>>>
>>
>> _______________________________________________
>>> libvirt-users mailing list
>>> libvirt-users@redhat.com
>>> https://www.redhat.com/mailman/listinfo/libvirt-users
>>>
>>

On Tue, Jan 24, 2017 at 16:21:14 +0530, akhilesh rawat
wrote:
> hi,
> 
> yes we are talking about same host here.
> 
> You could've just used -cpu host-model-only
> >> I dont want to host model ( it is working with host model) .
> 
> AS KVM seems to allow support Haswell with native commands . Does this mean
> in my case KVM is allowing software emulated cpu model.
No, QEMU will just drop the features which are not supported on your
host. Although there are some rare exceptions, such as x2apic, which can
be emulated even with KVM.

Anyway, starting QEMU directly should print several warnings about
features which are dropped. And when you start QEMU with
-cpu Haswell,enforce it should refuse to start.
> Can libvirt allow  software-emulated for non host model ?  if yes how .
You can bypass the libvirt check for specific features by setting the
feature policy to "force" (no idea if this can be done with
virt-install), but QEMU will still do the check itself and either drop
the features or refuse to start (with enforce) if they cannot be used on
the host.

Jirka

hello ,

thanks for comments .

I tried now with force options for CPU flag which were not supported . Now
the command with non fully supported CPU model gets executed ,  But i am
surprised to see that still Guest  cpu model is not changed and still same
as host cpu model(SAndy Bridge)

Why don't i see the model as HAswell now , could you please comment.

Command used :

virt-install --virt-type kvm --name controller-0  --cpu
Haswell-noTSX,+fma,+movbe,+fsgsbase,+bmi1,+avx2,+smep,+bmi2,+erms,+invpcid
--ram=8120 --vcpus=4,sockets=1

Linux process for KVM :

7479     1 37 10:13 ?        01:13:21 /usr/libexec/qemu-kvm -name
controller-0 -S -machine pc-i440fx-rhel7.0.0,accel=kvm,usb=off *-cpu
SandyBridge,*+fma,+pcid,+movbe,+fsgsbase,+bmi1,+avx2,+smep,+bmi2,+erms,+invpcid
-m 8120 -realtime mlock=off -smp 4,sockets=1,cores=4,threads=1



cpu guest :
processor    : 0
vendor_id    : GenuineIntel
cpu family    : 6
model        : 42
model name    : Intel Xeon E312xx (Sandy Bridge)
stepping    : 1
microcode    : 0x1
cpu MHz        : 1994.450
cache size    : 4096 KB
physical id    : 0
siblings    : 4
core id        : 0
cpu cores    : 4
apicid        : 0
initial apicid    : 0
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush mmx fxsr sse sse2 ht syscall nx rdtscp lm constant_tsc
rep_good nopl eagerfpu pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic
popcnt tsc_deadline_timer aes xsave avx hypervisor lahf_lm xsaveopt
bogomips    : 3988.90
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:


CPU Host :

processor    : 0
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz
stepping    : 7
microcode    : 0x710
cpu MHz        : 1200.000
cache size    : 15360 KB
physical id    : 0
siblings    : 12
core id        : 0
cpu cores    : 6
apicid        : 0
initial apicid    : 0
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca
cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx
pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl
xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor
ds_cpl vmx smx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic
popcnt tsc_deadline_timer aes xsave avx lahf_lm ida arat epb pln pts dtherm
tpr_shadow vnmi flexpriority ept vpid xsaveopt
bogomips    : 3990.17
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:


Thanks
Aki


On Wed, Jan 25, 2017 at 2:36 PM, Jiri Denemark <jdenemar@redhat.com>
wrote:
> On Tue, Jan 24, 2017 at 16:21:14 +0530, akhilesh rawat wrote:
> > hi,
> >
> > yes we are talking about same host here.
> >
> > You could've just used -cpu host-model-only
> > >> I dont want to host model ( it is working with host model) .
> >
> > AS KVM seems to allow support Haswell with native commands . Does this
> mean
> > in my case KVM is allowing software emulated cpu model.
>
> No, QEMU will just drop the features which are not supported on your
> host. Although there are some rare exceptions, such as x2apic, which can
> be emulated even with KVM.
>
> Anyway, starting QEMU directly should print several warnings about
> features which are dropped. And when you start QEMU with
> -cpu Haswell,enforce it should refuse to start.
>
> > Can libvirt allow  software-emulated for non host model ?  if yes how
.
>
> You can bypass the libvirt check for specific features by setting the
> feature policy to "force" (no idea if this can be done with
> virt-install), but QEMU will still do the check itself and either drop
> the features or refuse to start (with enforce) if they cannot be used on
> the host.
>
> Jirka
>