Post by John A. WallacePost by Frank Mehnert4.3.8 in general is a stable release but due to changes in certain
aspects of machine status registers there might be problems with
certain host CPUs.
But again, without more information we can't help.
Is there something specific you can point to for us to see which "certain
host CPUs" you have in mind? I would like to understand this problem
better, but without more information it is hard to see whether I can
possibly troubleshoot this issue or whether mine might be one of those
included among the group of certain host CPUs. If it is incompatible, that
is one thing; but if it is not, then, well, that would be something else.
In VirtualBox 4.3.8 the way how machine status registers are emulated was
changed. VirtualBox up to version 4.3.6 used sensible default values for
most MSRs. These values were determined by trial and error to make most
guests happy. However, the set of MSRs depends on the CPU microarchitecture.
For instance, Intel Pention M is very different from Intel Core i7.
Certain guests or guest drivers expect different MSR behavior for different
CPUs. To fulfil this expectation, the behavior was changed in 4.3.8. This
change was actually planned for 4.3.0 but it wasn't ready in time. You
get a first impression if you look at this directory:
https://www.virtualbox.org/browser/vbox/trunk/src/VBox/VMM/VMMR3/cpus
For most CPUs, the behavior shouldn't be much different but for CPUs
which differ too much from all examples in this directory, the emulation
is probably not precise enough. After 4.3.8 we added support for Intel
SandyBridge as well as support for more recent AMD Athlon 64 X2 CPUs
which should fix some problems users saw on these CPUs.
We cannot provide a list of supported CPUs. Often the difference between
the CPU microarchitectures is that small that one can use the complete
emulation of the previous architecture. We have a number of different
CPU architectures available and did extensive tests but we surely cannot
catch all problems. Therefore it is important that users open tickets for
such 4.3.8 regressions (actually it's very important to differ between
4.3.8 regressions and bugs users already saw with 4.3.6!) on our public
bugtracker so we can investigate and provide 4.3.9 test builds so users
can check if their problem is now fixed or not. For instance such reports:
https://www.virtualbox.org/ticket/12734
https://www.virtualbox.org/ticket/12736
https://www.virtualbox.org/ticket/12744
And it is important for us to get the feedback if such problem is fixed!
But as I said in a previews mail, the normal user should not experience
any different behavior between 4.3.6 and 4.3.8. In general, the 4.3.8
release is quite stable.
Kind regards,
Frank
--
Dr.-Ing. Frank Mehnert | Software Development Director, VirtualBox
ORACLE Deutschland B.V. & Co. KG | Werkstr. 24 | 71384 Weinstadt, Germany
Hauptverwaltung: Riesstr. 25, D-80992 München
Registergericht: Amtsgericht München, HRA 95603
Geschäftsführer: Jürgen Kunz
Komplementärin: ORACLE Deutschland Verwaltung B.V.
Hertogswetering 163/167, 3543 AS Utrecht, Niederlande
Handelsregister der Handelskammer Midden-Niederlande, Nr. 30143697
Geschäftsführer: Alexander van der Ven, Astrid Kepper, Val Maher