Re: [PATCH v2 1/2] Introduce (x86) CPU model deprecation API

[Eduardo Habkost]

On Fri, Sep 11, 2020 at 02:22:51PM +0800, Robert Hoo wrote:
> On Wed, 2020-09-09 at 14:15 -0400, Eduardo Habkost wrote:
> > Hi,
> > 
> > Thanks for the patch, and sorry for taking so long to review
> > this.  I'm finally getting to the patches that were postponed to
> > 5.2.
> > 
> > Comments and questions below:
> > 
> > On Thu, Jun 11, 2020 at 10:47:55AM +0800, Robert Hoo wrote:
> > > Complement versioned CPU model framework with the ability of
> > > marking some
> > > versions deprecated. When that CPU model is chosen, get some
> > > warning. The
> > > warning message is customized, e.g. telling in which future QEMU
> > > version will
> > > it be obsoleted.
> > > The deprecation message will also appear by x86_cpu_list_entry(),
> > > e.g. '-cpu
> > > help'.
> > > QMP 'query-cpu-definitions' will also return a bool value
> > > indicating the
> > > deprecation status.
> > > 
> > > Changes in v2:
> > > Move deprecation check from parse_cpu_option() to
> > > machine_run_board_init(), so
> > > that it can cover implicit cpu_type assignment cases.
> > > Add qapi new member documentation. Thanks Eric for comment and
> > > guidance on qapi.
> > > 
> > > Signed-off-by: Robert Hoo <robert.hu@linux.intel.com>
> > > ---
> > >  hw/core/machine.c        | 11 +++++++++--
> > >  include/hw/core/cpu.h    |  1 +
> > >  qapi/machine-target.json |  7 ++++++-
> > >  target/i386/cpu.c        | 45
> > > +++++++++++++++++++++++++++++++++++++++++++--
> > >  4 files changed, 59 insertions(+), 5 deletions(-)
> > > 
> > > diff --git a/hw/core/machine.c b/hw/core/machine.c
> > > index bb3a7b1..9318964 100644
> > > --- a/hw/core/machine.c
> > > +++ b/hw/core/machine.c
> > > @@ -1083,6 +1083,8 @@ MemoryRegion
> > > *machine_consume_memdev(MachineState *machine,
> > >  void machine_run_board_init(MachineState *machine)
> > >  {
> > >      MachineClass *machine_class = MACHINE_GET_CLASS(machine);
> > > +    ObjectClass *oc = object_class_by_name(machine->cpu_type);
> > > +    CPUClass *cc;
> > >  
> > >      if (machine->ram_memdev_id) {
> > >          Object *o;
> > > @@ -1102,11 +1104,10 @@ void machine_run_board_init(MachineState
> > > *machine)
> > >       * specified a CPU with -cpu check here that the user CPU is
> > > supported.
> > >       */
> > >      if (machine_class->valid_cpu_types && machine->cpu_type) {
> > > -        ObjectClass *class = object_class_by_name(machine-
> > > >cpu_type);
> > >          int i;
> > >  
> > >          for (i = 0; machine_class->valid_cpu_types[i]; i++) {
> > > -            if (object_class_dynamic_cast(class,
> > > +            if (object_class_dynamic_cast(oc,
> > >                                            machine_class-
> > > >valid_cpu_types[i])) {
> > >                  /* The user specificed CPU is in the valid field,
> > > we are
> > >                   * good to go.
> > > @@ -1129,6 +1130,12 @@ void machine_run_board_init(MachineState
> > > *machine)
> > >          }
> > >      }
> > >  
> > > +    /* Check if CPU type is deprecated and warn if so */
> > > +    cc = CPU_CLASS(oc);
> > > +    if (cc->deprecation_check) {
> > > +        cc->deprecation_check(oc);
> > > +    }
> > 
> > Why do we need target-specific code here?  A CPUClass::deprecated
> > field would be much simpler.
> > 
> Because the Warning message composing is target-specific, using
> X86CPUVersionDefinition.note.
> Other targets can have their own warning message composing approaches.

I think I understand what you were trying to do, but having each
target with a different warning message would be a bad thing, not
a desirable feature.  The warning message can be generic.

> 
> > > +
> > >      machine_class->init(machine);
> > >  }
> > >  
> > > diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
> > > index 497600c..1ca47dc 100644
> > > --- a/include/hw/core/cpu.h
> > > +++ b/include/hw/core/cpu.h
> > > @@ -218,6 +218,7 @@ typedef struct CPUClass {
> > >      void (*disas_set_info)(CPUState *cpu, disassemble_info *info);
> > >      vaddr (*adjust_watchpoint_address)(CPUState *cpu, vaddr addr,
> > > int len);
> > >      void (*tcg_initialize)(void);
> > > +    void (*deprecation_check)(ObjectClass *oc);
> > >  
> > >      /* Keep non-pointer data at the end to minimize holes.  */
> > >      int gdb_num_core_regs;
> > > diff --git a/qapi/machine-target.json b/qapi/machine-target.json
> > > index f2c8294..c24f506 100644
> > > --- a/qapi/machine-target.json
> > > +++ b/qapi/machine-target.json
> > > @@ -285,6 +285,10 @@
> > >  #            in the VM configuration, because aliases may stop
> > > being
> > >  #            migration-safe in the future (since 4.1)
> > >  #
> > > +# @deprecated: If true, this CPU model is deprecated and may be
> > > removed in
> > > +#              in some future version of QEMU according to the
> > > QEMU deprecation
> > > +#              policy. (since 5.1)
> > 
> > Next version needs to say "since 5.2".
> 
> Sure.
> > 
> > > +#
> > >  # @unavailable-features is a list of QOM property names that
> > >  # represent CPU model attributes that prevent the CPU from
> > > running.
> > >  # If the QOM property is read-only, that means there's no known
> > > @@ -309,7 +313,8 @@
> > >              'static': 'bool',
> > >              '*unavailable-features': [ 'str' ],
> > >              'typename': 'str',
> > > -            '*alias-of' : 'str' },
> > > +            '*alias-of' : 'str',
> > > +            'deprecated' : 'bool' },
> > >    'if': 'defined(TARGET_PPC) || defined(TARGET_ARM) ||
> > > defined(TARGET_I386) || defined(TARGET_S390X) ||
> > > defined(TARGET_MIPS)' }
> > >  
> > >  ##
> > > diff --git a/target/i386/cpu.c b/target/i386/cpu.c
> > > index ba05da3..0d8638a 100644
> > > --- a/target/i386/cpu.c
> > > +++ b/target/i386/cpu.c
> > > @@ -1599,6 +1599,7 @@ typedef struct X86CPUVersionDefinition {
> > >      const char *alias;
> > >      const char *note;
> > >      PropValue *props;
> > > +    bool       deprecated;
> > >  } X86CPUVersionDefinition;
> > >  
> > >  /* Base definition for a CPU model */
> > > @@ -1638,6 +1639,11 @@ struct X86CPUModel {
> > >       * This matters only for "-cpu help" and query-cpu-definitions
> > >       */
> > >      bool is_alias;
> > > +    /*
> > > +     * If true, this model is deprecated, and may be removed in
> > > the future.
> > > +     * Trying to use it now will cause a warning.
> > > +     */
> > > +    bool deprecated;
> > >  };
> > >  
> > >  /* Get full model name for CPU version */
> > > @@ -4128,8 +4134,7 @@ static X86CPUVersion
> > > x86_cpu_model_resolve_version(const X86CPUModel *model)
> > >      X86CPUVersion v = model->version;
> > >      if (v == CPU_VERSION_AUTO) {
> > >          v = default_cpu_version;
> > > -    }
> > > -    if (v == CPU_VERSION_LATEST) {
> > > +    } else if (v == CPU_VERSION_LATEST) {
> > 
> > Why is this change necessary?
> 
> Just kind of compulsion of avoiding unnecessary if() :-). 'v' can only
> be one of CPU_VERSION_AUTO and CPU_VERSION_LATEST, unnecessary to judge
> twice.

I think this breaks the case where default_cpu_version is set to
CPU_VERSION_LATEST

> > 
> > >          return x86_cpu_model_last_version(model);
> > >      }
> > >      return v;
> > > @@ -4975,6 +4980,7 @@ static void x86_cpu_definition_entry(gpointer
> > > data, gpointer user_data)
> > >      info->migration_safe = cc->migration_safe;
> > >      info->has_migration_safe = true;
> > >      info->q_static = cc->static_model;
> > > +    info->deprecated = cc->model ? cc->model->deprecated : false;
> > >      /*
> > >       * Old machine types won't report aliases, so that alias
> > > translation
> > >       * doesn't break compatibility with previous QEMU versions.
> > > @@ -5411,6 +5417,7 @@ static void
> > > x86_register_cpudef_types(X86CPUDefinition *def)
> > >          m->cpudef = def;
> > >          m->version = vdef->version;
> > >          m->note = vdef->note;
> > > +        m->deprecated = vdef->deprecated;
> > >          x86_register_cpu_model_type(name, m);
> > >  
> > >          if (vdef->alias) {
> > > @@ -5418,6 +5425,8 @@ static void
> > > x86_register_cpudef_types(X86CPUDefinition *def)
> > >              am->cpudef = def;
> > >              am->version = vdef->version;
> > >              am->is_alias = true;
> > > +            am->note = vdef->note;
> > 
> > Is this extra line related to the deprecation feature?
> > 
> > It doesn't seem related, and it doesn't seem necessary as the
> > `note` field is already ignored for alias CPU models.
> 
> Because it is unused by other features, I use it to store model
> specific deprecation message.
> > 
> > > +            am->deprecated = vdef->deprecated;
> > >              x86_register_cpu_model_type(vdef->alias, am);
> > >          }
> > >      }
> > > @@ -7233,6 +7242,37 @@ static Property x86_cpu_properties[] = {
> > >      DEFINE_PROP_END_OF_LIST()
> > >  };
> > >  
> > > +static void x86_cpu_deprecation_check(ObjectClass *oc)
> > > +{
> > > +    X86CPUClass *xcc = X86_CPU_CLASS(oc);
> > > +    X86CPUVersion effective_version;
> > > +    const X86CPUVersionDefinition *vdef;
> > > +
> > > +    if (xcc->model == NULL) {
> > > +        return;
> > > +    }
> > > +
> > > +    if (xcc->model->version == CPU_VERSION_LEGACY) {
> > > +        /* Treat legacy version as v1 */
> > > +        effective_version = 1;
> > > +    } else {
> > > +        effective_version = x86_cpu_model_resolve_version(xcc-
> > > >model);
> > > +    }
> > > +
> > > +    vdef = xcc->model->cpudef->versions;
> > > +
> > > +    if (vdef == NULL) {
> > > +        return;
> > > +    } else {
> > > +        if (vdef[effective_version - 1].deprecated) {
> > > +            warn_report("Effective CPU model '%s' -- %s",
> > > +                    x86_cpu_versioned_model_name(xcc->model-
> > > >cpudef,\
> > > +                                                effective_version)
> > > ,
> > > +                    vdef[effective_version - 1].note);
> > > +        }
> > > +    }
> > 
> > Why do we need this extra logic?  Isn't it simpler to just add a
> > bool CPUClass::deprecated field, and set:
> > 
> >    cpu->deprecated = model->deprecated;
> > 
> > inside x86_cpu_cpudef_class_init()?
> > 
> All these are to fulfill the target you expected earlier:
> 
> "We need a proper CPU model deprecation API.  Deprecation info
> should appear on query-cpu-definitions and should trigger a
> warning when using the CPU model."
> 
> So I think each deprecated model shall have its own deprecation
> message, e.g. by which version it's going to be deprecation, etc.

There's nothing x86-specific about having deprecated CPU models,
so I don't understand the reason for the x86-specific hook.

If the .note field is the reason you added the arch-specific
hook, you can just add a CPUClass::deprecation_note field and
make the feature generic.

> 
> > > +}
> > > +
> > >  static void x86_cpu_common_class_init(ObjectClass *oc, void *data)
> > >  {
> > >      X86CPUClass *xcc = X86_CPU_CLASS(oc);
> > > @@ -7291,6 +7331,7 @@ static void
> > > x86_cpu_common_class_init(ObjectClass *oc, void *data)
> > >      cc->tlb_fill = x86_cpu_tlb_fill;
> > >  #endif
> > >      cc->disas_set_info = x86_disas_set_info;
> > > +    cc->deprecation_check = x86_cpu_deprecation_check;
> > >  
> > >      dc->user_creatable = true;
> > >  }
> > > -- 
> > > 1.8.3.1
> > > 
> > 
> > 
> 

-- 
Eduardo