Re: [Qemu-devel] [RFC PATCH v0 2/6] spapr: CPU core device

[Bharata B Rao <bharata@linux.vnet.ibm.com>]

On Fri, Feb 26, 2016 at 12:13:39PM -0600, Michael Roth wrote:
> Quoting Bharata B Rao (2016-02-25 10:22:38)
> > Add sPAPR specific CPU core device that is based on generic CPU core device.
> > Creating this core device will result in creation of all the CPU thread
> > devices that are part of this core.
> > 
> > Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
> > ---
> >  hw/ppc/Makefile.objs            |   1 +
> >  hw/ppc/spapr_cpu_core.c         | 210 ++++++++++++++++++++++++++++++++++++++++
> >  include/hw/ppc/spapr_cpu_core.h |  32 ++++++
> >  3 files changed, 243 insertions(+)
> >  create mode 100644 hw/ppc/spapr_cpu_core.c
> >  create mode 100644 include/hw/ppc/spapr_cpu_core.h
> > 
> > diff --git a/hw/ppc/Makefile.objs b/hw/ppc/Makefile.objs
> > index c1ffc77..5cc6608 100644
> > --- a/hw/ppc/Makefile.objs
> > +++ b/hw/ppc/Makefile.objs
> > @@ -4,6 +4,7 @@ obj-y += ppc.o ppc_booke.o
> >  obj-$(CONFIG_PSERIES) += spapr.o spapr_vio.o spapr_events.o
> >  obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o
> >  obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o spapr_rng.o
> > +obj-$(CONFIG_PSERIES) += spapr_cpu_core.o
> >  ifeq ($(CONFIG_PCI)$(CONFIG_PSERIES)$(CONFIG_LINUX), yyy)
> >  obj-y += spapr_pci_vfio.o
> >  endif
> > diff --git a/hw/ppc/spapr_cpu_core.c b/hw/ppc/spapr_cpu_core.c
> > new file mode 100644
> > index 0000000..c44eb61
> > --- /dev/null
> > +++ b/hw/ppc/spapr_cpu_core.c
> > @@ -0,0 +1,210 @@
> > +/*
> > + * sPAPR CPU core device, acts as container of CPU thread devices.
> > + *
> > + * Copyright (C) 2016 Bharata B Rao <bharata@linux.vnet.ibm.com>
> > + *
> > + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> > + * See the COPYING file in the top-level directory.
> > + */
> > +#include "hw/cpu/core.h"
> > +#include "hw/ppc/spapr_cpu_core.h"
> > +#include "hw/ppc/spapr.h"
> > +#include "hw/boards.h"
> > +#include "qemu/error-report.h"
> > +#include "qapi/visitor.h"
> > +#include <sysemu/cpus.h>
> > +
> > +static int spapr_cpu_core_realize_child(Object *child, void *opaque)
> > +{
> > +    Error **errp = opaque;
> > +
> > +    object_property_set_bool(child, true, "realized", errp);
> > +    if (*errp) {
> > +        return 1;
> > +    }
> > +    return 0;
> > +}
> > +
> > +static void spapr_cpu_core_realize(DeviceState *dev, Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(OBJECT(dev));
> > +    sPAPRMachineState *spapr = SPAPR_MACHINE(qdev_get_machine());
> > +    Error *local_err = NULL;
> > +
> > +    if (!core->nr_threads) {
> > +        error_setg(errp, "nr_threads property can't be 0");
> > +        return;
> > +    }
> > +
> > +    if (!core->cpu_model) {
> > +        error_setg(errp, "cpu_model property isn't set");
> > +        return;
> > +    }
> > +
> > +    /*
> > +     * TODO: If slot isn't specified, plug this core into
> > +     * an existing empty slot.
> > +     */
> > +    if (!core->slot) {
> > +        error_setg(errp, "slot property isn't set");
> > +        return;
> > +    }
> > +
> > +    object_property_set_link(OBJECT(spapr), OBJECT(core), core->slot,
> > +                             &local_err);
> > +    if (local_err) {
> > +        error_propagate(errp, local_err);
> > +        return;
> > +    }
> > +
> > +    object_child_foreach(OBJECT(dev), spapr_cpu_core_realize_child, errp);
> > +}
> > +
> > +/*
> > + * This creates the CPU threads for a given @core.
> > + *
> > + * In order to create the threads, we need two inputs - number of
> > + * threads and the cpu_model. These are set as core object's properties.
> > + * When both of them become available/set, this routine will be called from
> > + * either property's set handler to create the threads.
> > + *
> > + * TODO: Dependence of threads creation on two properties is resulting
> > + * in this not-so-clean way of creating threads from either of the
> > + * property setters based on the order in which they get set. Check if
> > + * this can be handled in a better manner.
> > + */
> > +static void spapr_cpu_core_create_threads(sPAPRCPUCore *core, Error **errp)
> > +{
> > +    int i;
> > +
> > +    for (i = 0; i < core->nr_threads; i++) {
> > +        char id[32];
> > +        char type[32];
> > +
> > +        snprintf(type, sizeof(type), "%s-%s", core->cpu_model,
> > +                 TYPE_POWERPC_CPU);
> > +        object_initialize(&core->threads[i], sizeof(core->threads[i]), type);
> > +
> > +        snprintf(id, sizeof(id), "thread[%d]", i);
> > +        object_property_add_child(OBJECT(core), id, OBJECT(&core->threads[i]),
> > +                                  errp);
> > +    }
> > +}
> > +
> > +static char *spapr_cpu_core_prop_get_slot(Object *obj, Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +
> > +    return core->slot;
> > +}
> > +
> > +static void spapr_cpu_core_prop_set_slot(Object *obj, const char *val,
> > +                                              Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +
> > +    core->slot = g_strdup(val);
> > +}
> > +
> > +static char *spapr_cpu_core_prop_get_cpu_model(Object *obj, Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +
> > +    return core->cpu_model;
> > +}
> > +
> > +static void spapr_cpu_core_prop_set_cpu_model(Object *obj, const char *val,
> > +                                              Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +    MachineState *machine = MACHINE(qdev_get_machine());
> > +
> > +    /*
> > +     * cpu_model can't be different from what is specified with -cpu
> > +     */
> > +    if (strcmp(val, machine->cpu_model)) {
> > +       error_setg(errp, "cpu_model should be %s", machine->cpu_model);
> > +       return;
> > +    }
> > +
> > +    core->cpu_model = g_strdup(val);
> > +    if (core->nr_threads && core->cpu_model) {
> > +        spapr_cpu_core_create_threads(core, errp);
> > +    }
> > +}
> > +
> > +static void spapr_cpu_core_prop_get_nr_threads(Object *obj, Visitor *v,
> > +                                               const char *name, void *opaque,
> > +                                               Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +    int64_t value = core->nr_threads;
> > +
> > +    visit_type_int(v, name, &value, errp);
> > +}
> > +
> > +static void spapr_cpu_core_prop_set_nr_threads(Object *obj, Visitor *v,
> > +                                               const char *name, void *opaque,
> > +                                               Error **errp)
> > +{
> > +    sPAPRCPUCore *core = SPAPR_CPU_CORE(obj);
> > +    Error *local_err = NULL;
> > +    int64_t value;
> > +
> > +    visit_type_int(v, name, &value, &local_err);
> > +    if (local_err) {
> > +        error_propagate(errp, local_err);
> > +        return;
> > +    }
> > +
> > +    /* Allow only homogeneous configuration */
> > +    if (value != smp_threads) {
> > +        error_setg(errp, "nr_threads should be %d", smp_threads);
> > +        return;
> > +    }
> > +
> > +    core->nr_threads = value;
> > +    core->threads = g_malloc0(core->nr_threads * sizeof(PowerPCCPU));
> > +
> > +    if (core->nr_threads && core->cpu_model) {
> > +        spapr_cpu_core_create_threads(core, errp);
> > +    }
> > +}
> > +
> > +static void spapr_cpu_core_instance_init(Object *obj)
> > +{
> > +    object_property_add(obj, "nr_threads", "int",
> > +                        spapr_cpu_core_prop_get_nr_threads,
> > +                        spapr_cpu_core_prop_set_nr_threads,
> > +                        NULL, NULL, NULL);
> > +    object_property_add_str(obj, "cpu_model",
> > +                            spapr_cpu_core_prop_get_cpu_model,
> > +                            spapr_cpu_core_prop_set_cpu_model,
> > +                            NULL);
> > +    object_property_add_str(obj, "slot",
> > +                            spapr_cpu_core_prop_get_slot,
> > +                            spapr_cpu_core_prop_set_slot,
> > +                            NULL);
> 
> "slot" seems intended to be a machine-agnostic of mapping device
> types discovered from qmp_query_cpu_slots() to an appropriate
> "bus" location, but here's it a field specific to TYPE_SPAPR_CPU_CORE.
> It seems like maybe TYPE_CPU_CORE is a better place, but then on
> x86 I suppose it might be TYPE_CPU_SOCKET or something instead...

Correct.

> 
> It almost seems like a TYPE_INTERFACE_SLOTABLE would be the
> right approach, but I don't know how we could expose that as
> a property. I guess it's somewhat implied that this "interface"
> exists if qmp_query_cpu_slots() returns the type, but I wonder
> if something a bit more formal should be modeled to make the
> implementation requirements a bit clearer.
> 
> Maybe have TYPE_CPU_{CORE,SOCKET} classes have a get_slot/set_slot
> class method, expose them via "slot" property, then have the
> defaults generate "not implemented" errors?

Yes makes sense. In fact David has often times said that generic
properties/routines should be pushed to base class wherever possible.

I didn't do that in this first iteration to keep the generic changes
as minimum as possible, but yes slot should be a property of the
base class of core or socket.

Regards,
Bharata.