Re: [RFC 07/12] target-i386/cpu: Introduce apic_id_for_cpu() function

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Jan 11, 2013 at 12:31:23AM +0100, Igor Mammedov wrote:
> On Wed,  9 Jan 2013 16:53:47 -0200
> Eduardo Habkost <ehabkost@xxxxxxxxxx> wrote:
> 
> > This function will be used by both the CPU initialization code and the
> > fw_cfg table initialization code.
> > 
> > Later this function will be updated to generate APIC IDs according to
> > the CPU topology.
> > 
> > Signed-off-by: Eduardo Habkost <ehabkost@xxxxxxxxxx>
> > ---
> >  target-i386/cpu.c | 17 ++++++++++++++++-
> >  target-i386/cpu.h |  2 ++
> >  2 files changed, 18 insertions(+), 1 deletion(-)
> > 
> > diff --git a/target-i386/cpu.c b/target-i386/cpu.c
> > index 492656c..33787dc 100644
> > --- a/target-i386/cpu.c
> > +++ b/target-i386/cpu.c
> > @@ -2192,6 +2192,21 @@ void x86_cpu_realize(Object *obj, Error **errp)
> >      cpu_reset(CPU(cpu));
> >  }
> >  
> > +/* Calculates initial APIC ID for a specific CPU index
> > + *
> > + * Currently we need to be able to calculate the APIC ID from the CPU index
> > + * alone (without requiring a CPU object), as the QEMU<->Seabios interfaces have
> > + * no concept of "CPU index", and the NUMA tables on fw_cfg need the APIC ID of
> > + * all CPUs up to max_cpus.
> > + */
> > +uint32_t apic_id_for_cpu(unsigned int cpu_index)
> > +{
> > +    /* right now APIC ID == CPU index. this will eventually change to use
> > +     * the CPU topology configuration properly
> > +     */
> > +    return cpu_index;
> > +}
> > +
> >  static void x86_cpu_initfn(Object *obj)
> >  {
> >      CPUState *cs = CPU(obj);
> > @@ -2226,7 +2241,7 @@ static void x86_cpu_initfn(Object *obj)
> >                          x86_cpuid_get_tsc_freq,
> >                          x86_cpuid_set_tsc_freq, NULL, NULL, NULL);
> >  
> > -    env->cpuid_apic_id = cs->cpu_index;
> > +    env->cpuid_apic_id = apic_id_for_cpu(cs->cpu_index);
> APIC ID is set by board so CPU shouldn't do it by itself.
> 
> Could you create static property for it and set it from board using property
> instead?

I had a version that had the PC CPU creation code set the apic-id using
a apic-id property, but I am using this method to keep the bug fix
simpler and less intrusive, as I we don't have much time to remodel the
CPU creation code before soft freeze.

The previous version can be seen at:
  https://github.com/ehabkost/qemu-hacks/tree/apic-id-topology-using-apic-id-prop
The relevant apic-id-setting code is at:
  https://github.com/ehabkost/qemu-hacks/commit/a79156d0ad99ee89c0def535ce4cafcc4aa9a486

I believe the final solution for calculating the APIC IDs should instead
involve passing the package ID and bit width of core+thread IDs to a
package/socket object (so we don't need to expose the APIC IDs outside
QEMU). But whatever design we're going to choose to allow CPU hotplug, I
didn't want to introduce complexity that will get into our way when we
remodel the CPU creation/hotplug interfaces later.


> 
> >  
> >      /* init various static tables used in TCG mode */
> >      if (tcg_enabled() && !inited) {
> > diff --git a/target-i386/cpu.h b/target-i386/cpu.h
> > index f3c9df5..dbd9899 100644
> > --- a/target-i386/cpu.h
> > +++ b/target-i386/cpu.h
> > @@ -1238,4 +1238,6 @@ void disable_kvm_pv_eoi(void);
> >  /* Return name of 32-bit register, from a R_* constant */
> >  const char *get_register_name_32(unsigned int reg);
> >  
> > +uint32_t apic_id_for_cpu(unsigned int cpu_index);
> > +
> >  #endif /* CPU_I386_H */
> > -- 
> > 1.7.11.7
> > 
> 
> 
> -- 
> Regards,
>   Igor

-- 
Eduardo
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux