[Qemu-devel] [PATCH qom-cpu 00/21 v9] target-i386: convert CPU features into properties

[Igor Mammedov <imammedo@redhat.com>]

It's reordered and rebased v8 plus CPUID feature bits conversion to properties
and cleanups that are removing unused anymore *_feature_name arrays.

dynamic => static properties conversion is still making sense as cleanup of
initfn(), consolidating all properties in one place and making uniform
property setters signatures, so it was kept.

hyperv and dynamic => static properties conversion are covered by virt-test's
qemu_cpu test group.

On top of that, CPUID feature bits conversion and cleanups it's allowed.

git for testing: https://github.com/imammedo/qemu/tree/x86-cpu-properties.v9

v8-v9:
* reorder hyperv/check/force and cpu_x86_parse_featurestr() cleanups to be the first
* when converting dynamic properties inline custom DEFINE_PROP_FOO directly in
  cpu_x86_properties property array. (req: afaerber)
* change "f-" CPUID feature bits prefix to "feat-" (req: afaerber)
* replace F_* macros with a single X86CPU_FEAT() macro and simplify it setting
default value to 0, intended for defaults to be set by initfn()
* kvmclock: fix endless loop on compat kvmclock2 append
* converted check_features_against_host() and x86_cpu_list() to use
  static properties definition for feature name lookups, removing along the way
  unused anymore *_feature_name arrays.

v7-v8:
* split out dynamic properties convertion patch into per property patches
  to simplify review
* drop feature bits convertion

v6-v7:
* convert globals check_cpuid, enforce_cpuid and  hyperv_* to fields of
  CPUState
* Make PropertyInfo-s static
* maintain legacy kvmclock semantic in cpu_x86_parse_featurestr()
* existing properties code are not moved around, just fixed signatures where
  it's needed and used visitors.

v5-v6:
* when converting feature names to property names, replace '_' with '-'
* separate patches converting existing dynamic properties into one, were
  squashed into one [1/9] and change tested with virt-test(next).
* patches that were touching +-foo features are squashed into one [9/9],
  to avoid behavior change between them(f-kvmclock property).
* the rest of conversions were basicaly rebased on top of current qom-cpu-next
  tree, with small corrections

Igor Mammedov (20):
  target-i386: Move hyperv_* static globals to X86CPU
  target-i386: convert 'hv_spinlocks' to static property
  target-i386: convert 'hv_relaxed' to static property
  target-i386: convert 'hv_vapic' to static property
  target-i386: convert 'check' and 'enforce' to static properties
  target-i386: cleanup 'foo' feature handling'
  target-i386: cleanup 'foo=val' feature handling
  target-i386: cpu: convert 'level' to static property
  target-i386: cpu: convert 'xlevel' to static property
  target-i386: cpu: convert 'family' to static property
  target-i386: cpu: convert 'model' to static property
  target-i386: cpu: convert 'stepping' to static property
  target-i386: cpu: convert 'vendor' to static property
  target-i386: cpu: convert 'model-id' to static property
  target-i386: cpu: convert 'tsc-frequency' to static property
  target-i386: set [+-]feature using static properties
  qdev: introduce qdev_prop_find_bit()
  target-i386: use static properties in check_features_against_host() to
    print CPUID feature names
  target-i386: use static properties to list CPUID features
  target-i386: remove unused *_feature_name arrays

 hw/core/qdev-properties.c    |  15 +
 include/hw/qdev-properties.h |  13 +
 target-i386/Makefile.objs    |   2 +-
 target-i386/cpu-qom.h        |   6 +
 target-i386/cpu.c            | 705 +++++++++++++++++++++----------------------
 target-i386/cpu.h            |   4 +
 target-i386/hyperv.c         |  64 ----
 target-i386/hyperv.h         |  45 ---
 target-i386/kvm.c            |  36 ++-
 9 files changed, 415 insertions(+), 475 deletions(-)
 delete mode 100644 target-i386/hyperv.c
 delete mode 100644 target-i386/hyperv.h

-- 
1.8.3.1