[libvirt-users] Fwd: Haswell 4770 misidentified as Sandy Bridge

[Martin Kletzander]

On 06/13/2013 10:11 PM, Michael Giardino wrote:
> Hi,
> 
> I'm running libvert on a Debian 7 system. I have upgraded libvert and qemu
> from source (v1.06 and 1.5.0 respectively) and the problem persists. The
> guest OS is also a Debian 7 system running a non-SMP kernel. The error
> message from virt-manager is
> 
> Error starting domain: unsupported configuration: guest and host CPU are
> not compatible: Host CPU does not provide required features: rtm, invpcid,
> erms, bmi2, smep, avx2, hle, bmi1, fsgsbase, pcid
> 
> Traceback (most recent call last):
>   File "/usr/share/virt-manager/virtManager/asyncjob.py", line 45, in
> cb_wrapper
>     callback(asyncjob, *args, **kwargs)
>   File "/usr/share/virt-manager/virtManager/asyncjob.py", line 66, in tmpcb
>     callback(*args, **kwargs)
>   File "/usr/share/virt-manager/virtManager/domain.py", line 1114, in
> startup
>     self._backend.create()
>   File "/usr/local/lib/python2.7/dist-packages/libvirt.py", line 698, in
> create
>     if ret == -1: raise libvirtError ('virDomainCreate() failed', dom=self)
> libvirtError: unsupported configuration: guest and host CPU are not
> compatible: Host CPU does not provide required features: rtm, invpcid,
> erms, bmi2, smep, avx2, hle, bmi1, fsgsbase, pcid
> 
> but the host /proc/cpuinfo gives the correct flags as
> 
> processor : 7
> vendor_id : GenuineIntel
> cpu family : 6
> model : 60
> model name : Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz
> stepping : 3
> microcode : 0x7
> cpu MHz : 800.000
> cache size : 8192 KB
> physical id : 0
> siblings : 8
> core id : 3
> cpu cores : 4
> apicid : 7
> initial apicid : 7
> 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 fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt
> tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm ida arat epb
> xsaveopt pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase
> tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm
> bogomips : 6795.91
> clflush size : 64
> cache_alignment : 64
> address sizes : 39 bits physical, 48 bits virtual
> power management:
> 
> Is there a place in the code I can look where the CPU flags are checked? I
> did create this VM on the Debian 7 versions of libvirt and qemu, but I
> don't see that it should be much of a problem since you can swap CPUs in
> virt-manager.
> 
> The end goal is to read Sandy Bridge+ RAPL MSRs from the guest OS.
> perf_events are now being exported, which is great for my work, but if I
> could read some of those RAPL MSRs, It would make my life orders of
> magnitude easier.
> 
> Any help would be much appreciated.
> 

This really looks like a bug, but before saying that for sure, could you
dump the whole /proc/cpuinfo for all the cores (cpus)?  Or make sure
they are really the same?  We currently don't support getting cpuid info
from all cpus, because we don't control what cpus the cpuid instruction
is being ran on.

Thanks,
Martin