VMware Communities
shuras
Contributor
Contributor
Jump to solution

/build/mts/release/bora-2249910/bora/vmcore/vmx/main/timeTracker_user.c:234 bugNr=148722

  After Ubuntu upgrade I've got this kind of error. It is caused by kernel.

With old 3.13 kernels everything works as it should. New kernel 3.16 and I can't start any of VM.

Tried to reinstall, same thing. Same Ubuntu 14.10 working fine with the same Vmware Wks 10.0.4 on my other PC.

Log file attached.

Reply
0 Kudos
1 Solution

Accepted Solutions
dariusd
VMware Employee
VMware Employee
Jump to solution

If you run the command

   cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq

does it output

   6300000

?  If so, the problem is between your OS kernel, your firmware (BIOS/EFI) and your CPU.  We're being given bogus data and there is nothing we can do about it.  You may be able to fix it with a firmware update for your motherboard.  It might be a defect in the new OS kernel itself, or it might be that the more recent OS kernel is newly using some new feature that is being misconfigured by your host's firmware.

If you need to override it, edit (as root) your /etc/vmware/config and add the line:

   host.cpukHz = 3467779

and it should get you going.

Cheers,

--

Darius

View solution in original post

Reply
0 Kudos
3 Replies
dariusd
VMware Employee
VMware Employee
Jump to solution

Your CPU is reporting that it will run at up to 6.3 GHz.  There is a hardcoded "sanity limit" in Workstation of 6.0 GHz.  Have you tweaked your BIOS settings to allow your CPU to run with such a high multiplier?  If so, you might want to reduce the multiplier a little bit, to somewhere closer to the real system capabilities.  If not, you might want to look for BIOS updates to see if they set the multiplier to a more reasonable value.

(By way of additional explanation: In order to maintain a sane notion of guest time, Workstation needs to know the maximum *possible* CPU speed that a system can encounter.  When the maximum possible speed differs significantly from the regular CPU speed, it will become more difficult for us to keep accurate guest time, so it's not a good idea to set the maximum Turbo Boost multiplier to a very high level that your processor will never really be able to achieve.  Or is this running in liquid nitrogen? :smileygrin:  Anyway, speeds above 6.0 GHz are assumed to be a configuration problem, so we bail out rather than end up in a situation where tracking guest time becomes exceedingly difficult for no good reason.)

Hope this helps!

--

Darius

Reply
0 Kudos
shuras
Contributor
Contributor
Jump to solution

  No I don't use O/C. Got myself quite ordinary "Intel(R) Core(TM) i5-3570K CPU @ 3.40GHz" with default settings.

And this problem come and go with new/old kernel switching. In case of insane overclocking, old kernel should not help me to solve this problem.

It must be something wrong with clock measurement code. There was similar cases before. On Windows build there was a possibility to manually enter clock rate in the registry.

Is there anything like that but for Linux?

Reply
0 Kudos
dariusd
VMware Employee
VMware Employee
Jump to solution

If you run the command

   cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq

does it output

   6300000

?  If so, the problem is between your OS kernel, your firmware (BIOS/EFI) and your CPU.  We're being given bogus data and there is nothing we can do about it.  You may be able to fix it with a firmware update for your motherboard.  It might be a defect in the new OS kernel itself, or it might be that the more recent OS kernel is newly using some new feature that is being misconfigured by your host's firmware.

If you need to override it, edit (as root) your /etc/vmware/config and add the line:

   host.cpukHz = 3467779

and it should get you going.

Cheers,

--

Darius

Reply
0 Kudos