Re: [Qemu-devel] [PATCH 04/10] spapr: improve ibm, architecture-vec-5 property handling

[David Gibson <david@gibson.dropbear.id.au>]

[-- Attachment #1: Type: text/plain, Size: 4226 bytes --]

On Wed, Oct 26, 2016 at 10:58:09AM +1100, David Gibson wrote:
> On Mon, Oct 24, 2016 at 11:47:30PM -0500, Michael Roth wrote:
> > ibm,architecture-vec-5 is supposed to encode all option vector 5 bits
> > negotiated between platform/guest. Currently we hardcode this property
> > in the boot-time device tree to advertise a single negotiated
> > capability, "Form 1" NUMA Affinity, regardless of whether or not CAS
> > has been invoked or that capability has actually been negotiated.
> > 
> > Improve this by generating ibm,architecture-vec-5 based on the full
> > set of option vector 5 capabilities negotiated via CAS.
> > 
> > Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
> > ---
> >  hw/ppc/spapr.c              | 23 +++++++++++++++++------
> >  include/hw/ppc/spapr_ovec.h |  1 +
> >  2 files changed, 18 insertions(+), 6 deletions(-)
> > 
> > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> > index 3b64580..828072a 100644
> > --- a/hw/ppc/spapr.c
> > +++ b/hw/ppc/spapr.c
> > @@ -659,14 +659,28 @@ static int spapr_dt_cas_updates(sPAPRMachineState *spapr, void *fdt,
> >                                  sPAPROptionVector *ov5_updates)
> >  {
> >      sPAPRMachineClass *smc = SPAPR_MACHINE_GET_CLASS(spapr);
> > -    int ret = 0;
> > +    int ret = 0, offset;
> >  
> >      /* Generate ibm,dynamic-reconfiguration-memory node if required */
> >      if (spapr_ovec_test(ov5_updates, OV5_DRCONF_MEMORY)) {
> >          g_assert(smc->dr_lmb_enabled);
> >          ret = spapr_populate_drconf_memory(spapr, fdt);
> > +        if (ret) {
> > +            goto out;
> > +        }
> >      }
> >  
> > +    offset = fdt_path_offset(fdt, "/chosen");
> > +    if (offset < 0) {
> > +        offset = fdt_add_subnode(fdt, 0, "chosen");
> > +        if (offset < 0) {
> > +            return offset;
> > +        }
> 
> Just asserting offset >= 0 would be fine here.  We always create a
> /chosen node.

Duh.  Realised during testing that of course this *is* necessary for
the case where we're just making a CAS patch to the tree, rather than
building the whole tree.  I've reverted my ill-considered change in my
tree back to your original patch.

> 
> > +    }
> > +    ret = spapr_ovec_populate_dt(fdt, offset, spapr->ov5_cas,
> > +                                 "ibm,architecture-vec-5");
> > +
> > +out:
> >      return ret;
> >  }
> >  
> > @@ -792,14 +806,9 @@ static void spapr_dt_chosen(sPAPRMachineState *spapr, void *fdt)
> >      char *stdout_path = spapr_vio_stdout_path(spapr->vio_bus);
> >      size_t cb = 0;
> >      char *bootlist = get_boot_devices_list(&cb, true);
> > -    unsigned char vec5[] = {0x0, 0x0, 0x0, 0x0, 0x0, 0x80};
> >  
> >      _FDT(chosen = fdt_add_subnode(fdt, 0, "chosen"));
> >  
> > -    /* Set Form1_affinity */
> > -    _FDT(fdt_setprop(fdt, chosen, "ibm,architecture-vec-5",
> > -                     vec5, sizeof(vec5)));
> > -
> >      _FDT(fdt_setprop_string(fdt, chosen, "bootargs", machine->kernel_cmdline));
> >      _FDT(fdt_setprop_cell(fdt, chosen, "linux,initrd-start",
> >                            spapr->initrd_base));
> > @@ -1778,6 +1787,8 @@ static void ppc_spapr_init(MachineState *machine)
> >          spapr_validate_node_memory(machine, &error_fatal);
> >      }
> >  
> > +    spapr_ovec_set(spapr->ov5, OV5_FORM1_AFFINITY);
> > +
> >      /* init CPUs */
> >      if (machine->cpu_model == NULL) {
> >          machine->cpu_model = kvm_enabled() ? "host" : smc->tcg_default_cpu;
> > diff --git a/include/hw/ppc/spapr_ovec.h b/include/hw/ppc/spapr_ovec.h
> > index 09afd59..47fa04c 100644
> > --- a/include/hw/ppc/spapr_ovec.h
> > +++ b/include/hw/ppc/spapr_ovec.h
> > @@ -44,6 +44,7 @@ typedef struct sPAPROptionVector sPAPROptionVector;
> >  
> >  /* option vector 5 */
> >  #define OV5_DRCONF_MEMORY       OV_BIT(2, 2)
> > +#define OV5_FORM1_AFFINITY      OV_BIT(5, 0)
> >  
> >  /* interfaces */
> >  sPAPROptionVector *spapr_ovec_new(void);
> 



-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]