On Fri, Oct 15, 2021 at 05:22:59PM +0200, Kashyap Chamarthy wrote: > - Remove stray texinfo syntax (remnants of texinfo to rST conversion) > - Clarify the bit about long-term stable CPU models > > TODO: In a future patch, include potential examples as discussed > here[1]. > > [1] https://lists.nongnu.org/archive/html/qemu-devel/2021-10/msg03411.html > -- On versioned CPU models, aliases, and machine types
Ping? I'd also appreciate if anyone can also answer the two questions I raised in the above thread[1]. > Signed-off-by: Kashyap Chamarthy <kcham...@redhat.com> > --- > Eduardo/DanPB: I'm not 100% sure if my wording got it right; please give > it a close reading to make sure I'm not making things worse. > --- > docs/system/cpu-models-x86.rst.inc | 25 +++++++++++++++++-------- > 1 file changed, 17 insertions(+), 8 deletions(-) > > diff --git a/docs/system/cpu-models-x86.rst.inc > b/docs/system/cpu-models-x86.rst.inc > index 6e8be7d79b..e133753920 100644 > --- a/docs/system/cpu-models-x86.rst.inc > +++ b/docs/system/cpu-models-x86.rst.inc > @@ -25,7 +25,7 @@ Two ways to configure CPU models with QEMU / KVM > typically refer to specific generations of hardware released by > Intel and AMD. These allow the guest VMs to have a degree of > isolation from the host CPU, allowing greater flexibility in live > - migrating between hosts with differing hardware. @end table > + migrating between hosts with differing hardware. > > In both cases, it is possible to optionally add or remove individual CPU > features, to alter what is presented to the guest by default. > @@ -47,11 +47,20 @@ defined. Traditionally most operating systems and > toolchains would > only target the original baseline ABI. It is expected that in > future OS and toolchains are likely to target newer ABIs. The > table that follows illustrates which ABI compatibility levels > -can be satisfied by the QEMU CPU models. Note that the table only > -lists the long term stable CPU model versions (eg Haswell-v4). > -In addition to whats listed, there are also many CPU model > -aliases which resolve to a different CPU model version, > -depending on the machine type is in use. > +can be satisfied by the QEMU CPU models. Note that the table only lists > +the long term stable CPU model versions (e.g. Haswell-v4, Haswell-v3). > +CPU models without a version tag will alias to a CPU model with a > +version tag, and the alias varies depending on the machine type. In > +addition to what is listed, there are also many CPU model aliases which > +resolve to a different CPU model version, depending on the machine type > +in use. > + > +The versioned CPU models (e.g. ``Cascadelake-Server-v4``, > +``Broadwell-v4``) are long-term stable. Further, when using a versioned > +machine type (e.g. ``pc-q35-6.0``), instead of its generic alias > +(``q35``), the CPU models that are associated with it are also long-term > +stable. This is because the CPUID features in the CPU models that are > +part of a versioned machine type do not change. > > .. _ABI compatibility levels: https://gitlab.com/x86-psABIs/x86-64-ABI/ > > @@ -185,8 +194,8 @@ features are included if using "Host passthrough" or > "Host model". > guest. Instead, the host kernel uses it to populate the MDS > vulnerability file in ``sysfs``. > > - So it should only be enabled for VMs if the host reports @code{Not > - affected} in the ``/sys/devices/system/cpu/vulnerabilities/mds`` file. > + So it should only be enabled for VMs if the host reports ``Not > + affected`` in the ``/sys/devices/system/cpu/vulnerabilities/mds`` file. > > ``taa-no`` > Recommended to inform that the guest that the host is ``not`` > -- > 2.31.1 > -- /kashyap
