[libvirt] [PATCHv3] conf: Add the cpu duplicate use check for vm numa settings

[Michal Privoznik]

On 05.05.2015 12:13, Luyao Huang wrote:
> https://bugzilla.redhat.com/show_bug.cgi?id=1176020
> 
> We had a check for the vcpu count total number in <numa>
> before, however this check is not good enough. There are
> some examples:
> 
> 1. one of cpu id is out of maxvcpus, can set success(cpu count = 5 < 10):
> 
> <vcpu placement='static'>10</vcpu>
> <cell id='0' cpus='0-3,100' memory='512000' unit='KiB'/>
> 
> 2. use the same cpu in 2 cell, can set success(cpu count = 8 < 10):
> <vcpu placement='static'>10</vcpu>
> <cell id='0' cpus='0-3' memory='512000' unit='KiB'/>
> <cell id='1' cpus='0-3' memory='512000' unit='KiB'/>
> 
> 3. use the same cpu in 2 cell, cannot set success(cpu count = 11 > 10):
> <vcpu placement='static'>10</vcpu>
> <cell id='0' cpus='0-6' memory='512000' unit='KiB'/>
> <cell id='1' cpus='0-3' memory='512000' unit='KiB'/>
> 
> Add a check for numa cpus, check if duplicate use one cpu in more
> than one cell.
> 
> Signed-off-by: Luyao Huang <lhuang at redhat.com>
> ---
> v3: remove the max vcpu check, just add a check when
> parse the numa settings, and change the subject.
> 
>  src/conf/numa_conf.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/src/conf/numa_conf.c b/src/conf/numa_conf.c
> index 7ad3f66..e25d7a8 100644
> --- a/src/conf/numa_conf.c
> +++ b/src/conf/numa_conf.c
> @@ -738,6 +738,15 @@ virDomainNumaDefCPUParseXML(virDomainNumaPtr def,
>          }
>          VIR_FREE(tmp);
>  
> +        for (j = 0; j < i; j++) {
> +            if (virBitmapOverlaps(def->mem_nodes[j].cpumask,
> +                                  def->mem_nodes[i].cpumask)) {
> +                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
> +                               _("NUMA cells %zu and %zu have overlapping vCPU ids"), i, j);
> +                goto cleanup;
> +            }
> +        }
> +
>          ctxt->node = nodes[i];
>          if (virDomainParseMemory("./@memory", "./@unit", ctxt,
>                                   &def->mem_nodes[cur_cell].mem, true, false) < 0)
> 

Even though this does not compile, I'm squashing in the obvious fix,
ACKing and pushing.

Michal