[Qemu-devel] [PATCH v2 0/9] x86 CPU model versioning

[Eduardo Habkost]

Changes v1 -> v2:

* Patch "i386: Infrastructure for versioned CPU models" was
  rewritten and split in two:
  * i386: Register versioned CPU models
  * i386: Make unversioned CPU models be aliases
* -IBRS, -noTSX, -IBPB CPU models are now aliases
* Enable rdctl-no, ibrs-all, skip-l1dfl-vmentry in
  Cascadelake-Server-v2
* New patch added:
  * i386: Get model-id from CPU object on "-cpu help"

---
Original description in v1:

This series implements basic infrastructure for CPU model
versioning, as discussed before[1][2][3].  This will finally
allow us to update CPU models in ways that introduce new software
or hardware requirements.

My original plan was to use "query-cpu-model-expansion
mode=static" to resolve aliases, but I dropped that plan because
it would increase complexity for management software a lot.
static CPU models are documented as not being affected by the
machine type and accelerator at all, which would make the
versioned CPU models very inconvenient to use in the command
line.  e.g.: users would be forced to replace:

  -cpu Haswell

with:

  -cpu 
Haswell-4.1,+2apic,+monitor,+kvmclock,+kvm-nopiodelay,+kvm-asyncpf,+kvm-steal-time,+kvm-pv-eoi,+kvmclock-stable-bit,+x2apic,-acpi,-monitor,-svm

In the end, making the versioned CPU models static is not a
requirement at all: what we really need is to drop the
runnability guarantees from unversioned CPU model names, and
require management software to resolve the unversioned alias
before saving the VM configuration.

Guest ABI compatibility and live migration guarantees are going
to be kept: unversioned CPU models will still be usable with live
migration.  Only runnability guarantees when updating the machine
type will be dropped.  This means unversioned CPU models are
still reported as migration-safe in query-cpu-definitions.

The last patch in the series demonstrates how the new feature can
be used to update a CPU model: it adds a Cascadelake-Server-4.1.1
CPU model, including "arch-capabilities=on" and "stepping=5".
Unfortunately we can't enable arch-capabilities in the -4.1
version of Cascadelake-Server because it would break our existing
runnability guarantees.

[1] https://www.mail-archive.com/address@hidden/msg167342.html
[2] https://www.mail-archive.com/address@hidden/msg590034.html
[3] https://www.mail-archive.com/address@hidden/msg611244.html

---
Cc: Paolo Bonzini <address@hidden>
Cc: Pavel Hrdina <address@hidden>
Cc: Jiri Denemark <address@hidden>
Cc: "Hu, Robert" <address@hidden>
Cc: Tao Xu <address@hidden>
Cc: Richard Henderson <address@hidden>
Cc: "Dr. David Alan Gilbert" <address@hidden>
Cc: "Daniel P. Berrange" <address@hidden>
Cc: Igor Mammedov <address@hidden>

Eduardo Habkost (9):
  qmp: Add "alias-of" field to query-cpu-definitions
  i386: Add x-force-features option for testing
  i386: Get model-id from CPU object on "-cpu help"
  i386: Register versioned CPU models
  i386: Define -IBRS, -noTSX, -IBRS versions of CPU models
  i386: Replace -noTSX, -IBRS, -IBPB CPU models with aliases
  i386: Make unversioned CPU models be aliases
  docs: Deprecate CPU model runnability guarantees
  i386: Add Cascadelake-Server-v2 CPU model

 qapi/target.json                           |    9 +-
 include/hw/i386/pc.h                       |    3 +
 target/i386/cpu-qom.h                      |   10 +-
 target/i386/cpu.h                          |   28 +
 hw/i386/pc.c                               |    3 +
 hw/i386/pc_piix.c                          |    4 +
 hw/i386/pc_q35.c                           |    4 +
 target/i386/cpu.c                          | 1028 +++++++++-----------
 qemu-deprecated.texi                       |   19 +
 tests/acceptance/x86_cpu_model_versions.py |  304 ++++++
 10 files changed, 831 insertions(+), 581 deletions(-)
 create mode 100644 tests/acceptance/x86_cpu_model_versions.py

-- 
2.18.0.rc1.1.g3f1ff2140