[Libvirt-cim] [PATCH] (#2) Fix device_parsing of processors

[Kaitlin Rupert]

> @@ -346,15 +347,24 @@ static int parse_vcpu_device(xmlNode *no
> 
>          free(count_str);
> 
> -        list = calloc(1, sizeof(*list));
> +        list = calloc(count, sizeof(*list));
>          if (list == NULL)
>                  goto err;
> -        
> -        list->dev.vcpu.quantity = count;
> +
> +        for (i = 0; i < count; i++) {
> +                struct virt_device *vdev = &list[i];
> +                struct vcpu_device *cdev = &vdev->dev.vcpu;
> +
> +                cdev->number = i;
> +
> +                vdev->type = CIM_RES_TYPE_PROC;
> +                if (asprintf(&vdev->id, "%i", i) == -1)
> +                        vdev->id = NULL;
> +        }
> 
>          *vdevs = list;
> 
> -        return 1;
> +        return count;

You could do return count + 1 here and then use the quantity instead of 
number.  This would save you a step in the rasd_from_vdev() call.


But either way is fine by me.

> 
>  static void print_dev_emu(struct virt_device *dev,
> diff -r 01b0a24fc16f -r f0153bd85ec1 src/Virt_RASD.c
> --- a/src/Virt_RASD.c	Mon Apr 14 15:17:08 2008 -0700
> +++ b/src/Virt_RASD.c	Tue Apr 15 11:50:05 2008 -0400
> @@ -169,8 +169,10 @@ static CMPIInstance *rasd_from_vdev(cons
>                  CMSetProperty(inst, "Limit",
>                                (CMPIValue *)&dev->dev.mem.maxsize, CMPI_uint64);
>          } else if (dev->type == CIM_RES_TYPE_PROC) {
> +                /* This would be the place to set the virtualquantity. */
> +                int quantity = dev->dev.vcpu.number + 1;

Probably doesn't matter for this case, but the type here should match 
the type you set in CMSetProperty.

>                  CMSetProperty(inst, "VirtualQuantity",
> -                              (CMPIValue *)&dev->dev.vcpu.quantity, CMPI_uint32);
> +                              (CMPIValue *)&quantity, CMPI_uint32);

This needs to be CMPI_uint64 (per the mof) - otherwise, the attribute 
might not set correctly in all cases.

-- 
Kaitlin Rupert
IBM Linux Technology Center
kaitlin at linux.vnet.ibm.com