Re: [PATCH 5/7] spapr/xics: Configure number of servers in KVM

[Greg Kurz]

On Thu, 3 Oct 2019 14:29:46 +0200
Cédric Le Goater <address@hidden> wrote:

> On 03/10/2019 14:01, Greg Kurz wrote:
> > The XICS-on-XIVE KVM devices now has an attribute to configure the number
> > of interrupt servers. This allows to greatly optimize the usage of the VP
> > space in the XIVE HW, and thus to start a lot more VMs.
> > 
> > Only set this attribute if available in order to support older POWER9 KVM
> > and pre-POWER9 XICS KVM devices.
> > 
> > The XICS-on-XIVE KVM device now reports the exhaustion of VPs upon the
> > connection of the first VCPU. Check that in order to have a chance to
> > provide an hint to the user.
> 
> That part would have been better in its own patch. Simpler to review.
>  

Yeah, possibly... on the other hand, it's only two fairly simple
hunks. :)

> > Signed-off-by: Greg Kurz <address@hidden>
> 
> Anyhow, if you split or not :
> 
> Reviewed-by: Cédric Le Goater <address@hidden>
> 
> > ---
> >  hw/intc/xics_kvm.c |   24 ++++++++++++++++++++++--
> >  1 file changed, 22 insertions(+), 2 deletions(-)
> > 
> > diff --git a/hw/intc/xics_kvm.c b/hw/intc/xics_kvm.c
> > index ba90d6dc966c..12d9524cc432 100644
> > --- a/hw/intc/xics_kvm.c
> > +++ b/hw/intc/xics_kvm.c
> > @@ -165,8 +165,15 @@ void icp_kvm_realize(DeviceState *dev, Error **errp)
> >  
> >      ret = kvm_vcpu_enable_cap(cs, KVM_CAP_IRQ_XICS, 0, kernel_xics_fd, 
> > vcpu_id);
> >      if (ret < 0) {
> > -        error_setg(errp, "Unable to connect CPU%ld to kernel XICS: %s", 
> > vcpu_id,
> > -                   strerror(errno));
> > +        Error *local_err = NULL;
> > +
> > +        error_setg(&local_err, "Unable to connect CPU%ld to kernel XICS: 
> > %s",
> > +                   vcpu_id, strerror(errno));
> > +        if (errno == ENOSPC) {
> > +            error_append_hint(&local_err, "Try -smp maxcpus=N with N < 
> > %u\n",
> > +                              MACHINE(qdev_get_machine())->smp.max_cpus);
> > +        }
> > +        error_propagate(errp, local_err);
> >          return;
> >      }
> >      enabled_icp = g_malloc(sizeof(*enabled_icp));
> > @@ -344,6 +351,7 @@ void ics_kvm_set_irq(ICSState *ics, int srcno, int val)
> >  
> >  int xics_kvm_connect(SpaprMachineState *spapr, Error **errp)
> >  {
> > +    ICSState *ics = spapr->ics;
> >      int rc;
> >      CPUState *cs;
> >      Error *local_err = NULL;
> > @@ -397,6 +405,18 @@ int xics_kvm_connect(SpaprMachineState *spapr, Error 
> > **errp)
> >          goto fail;
> >      }
> >  
> > +    /* Tell KVM about the # of VCPUs we may have (POWER9 and newer only) */
> > +    if (kvm_device_check_attr(rc, KVM_DEV_XICS_GRP_CTRL,
> > +                              KVM_DEV_XICS_NR_SERVERS)) {
> > +        uint32_t nr_servers = xics_nr_servers(ics->xics);
> > +
> > +        if (kvm_device_access(rc, KVM_DEV_XICS_GRP_CTRL,
> > +                              KVM_DEV_XICS_NR_SERVERS, &nr_servers, true,
> > +                              &local_err)) {
> > +            goto fail;
> > +        }
> > +    }
> > +
> >      kernel_xics_fd = rc;
> >      kvm_kernel_irqchip = true;
> >      kvm_msi_via_irqfd_allowed = true;
> > 
>