We ignore silently the value of smp_threads when we set
the default VSMT value, and if smp_threads is greater than VSMT
kernel is going into trouble later.

Fixes: 8904e5a750
("spapr: Adjust default VSMT value for better migration compatibility")

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---

Notes:
    v2: display a specific error message when the default VSMT is used
        fix subject

 hw/ppc/spapr.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 32a876be56..11de276cd5 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -2311,6 +2311,12 @@ static void spapr_set_vsmt_mode(sPAPRMachineState *spapr, Error **errp)
          * overwhelmingly common case in production systems.
          */
         spapr->vsmt = 8;
+        if (spapr->vsmt < smp_threads) {
+            error_setg(&local_err, "Cannot support %d threads/core"
+                         " because it must be <= to default VSMT mode (%d)",
+                         smp_threads, spapr->vsmt);
+            goto out;
+        }
     }
 
     /* KVM: If necessary, set the SMT mode: */
-- 
2.14.3

On Thu,  8 Feb 2018 10:42:41 +0100
Laurent Vivier <lvivier@redhat.com> wrote:

> We ignore silently the value of smp_threads when we set
> the default VSMT value, and if smp_threads is greater than VSMT
> kernel is going into trouble later.
> 
> Fixes: 8904e5a750
> ("spapr: Adjust default VSMT value for better migration compatibility")
> 
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> ---
> 

Reviewed-by: Greg Kurz <groug@kaod.org>

> Notes:
>     v2: display a specific error message when the default VSMT is used
>         fix subject
> 
>  hw/ppc/spapr.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index 32a876be56..11de276cd5 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -2311,6 +2311,12 @@ static void spapr_set_vsmt_mode(sPAPRMachineState *spapr, Error **errp)
>           * overwhelmingly common case in production systems.
>           */
>          spapr->vsmt = 8;
> +        if (spapr->vsmt < smp_threads) {
> +            error_setg(&local_err, "Cannot support %d threads/core"
> +                         " because it must be <= to default VSMT mode (%d)",
> +                         smp_threads, spapr->vsmt);
> +            goto out;
> +        }
>      }
>  
>      /* KVM: If necessary, set the SMT mode: */

On Thu, Feb 08, 2018 at 10:42:41AM +0100, Laurent Vivier wrote:
> We ignore silently the value of smp_threads when we set
> the default VSMT value, and if smp_threads is greater than VSMT
> kernel is going into trouble later.
> 
> Fixes: 8904e5a750
> ("spapr: Adjust default VSMT value for better migration compatibility")
> 
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>

I don't think this is quite the rigth appoach.  Instead, I think we
want:

	spapr->vsmt = MAX(8, smp_threads);

The real point of 8904e5a750 was to replace the host-derived kvm_smt
with a fixed value (8).  I only removed the MAX, because I mistakenly
thought that checks elsewhere would prevent smp_threads > 8.

> ---
> 
> Notes:
>     v2: display a specific error message when the default VSMT is used
>         fix subject
> 
>  hw/ppc/spapr.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> index 32a876be56..11de276cd5 100644
> --- a/hw/ppc/spapr.c
> +++ b/hw/ppc/spapr.c
> @@ -2311,6 +2311,12 @@ static void spapr_set_vsmt_mode(sPAPRMachineState *spapr, Error **errp)
>           * overwhelmingly common case in production systems.
>           */
>          spapr->vsmt = 8;
> +        if (spapr->vsmt < smp_threads) {
> +            error_setg(&local_err, "Cannot support %d threads/core"
> +                         " because it must be <= to default VSMT mode (%d)",
> +                         smp_threads, spapr->vsmt);
> +            goto out;
> +        }
>      }
>  
>      /* KVM: If necessary, set the SMT mode: */

-- 
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