On Thu, Jul 23, 2015 at 12:01 AM, Ed White <edmund.h.wh...@intel.com> wrote:
> +int p2m_destroy_altp2m_by_id(struct domain *d, unsigned int idx)
> +{
> + struct p2m_domain *p2m;
> + int rc = -EINVAL;
> +
> + if ( !idx || idx > MAX_ALTP2M )
> + return rc;
> +
> + domain_pause_except_self(d);
> +
> + altp2m_list_lock(d);
> +
> + if ( d->arch.altp2m_eptp[idx] != INVALID_MFN )
> + {
> + p2m = d->arch.altp2m_p2m[idx];
> +
> + if ( !_atomic_read(p2m->active_vcpus) )
> + {
> + p2m_flush_table(d->arch.altp2m_p2m[idx]);
> + /* Uninit and reinit ept to force TLB shootdown */
> + ept_p2m_uninit(d->arch.altp2m_p2m[idx]);
> + ept_p2m_init(d->arch.altp2m_p2m[idx]);
> + d->arch.altp2m_eptp[idx] = INVALID_MFN;
> + rc = 0;
> + }
> + }
> +
> + altp2m_list_unlock(d);
> +
> + domain_unpause_except_self(d);
Can you put on your list of clean-ups to send before the 4.6 release
to have this return -EBUSY rather than -EINVAL if there are active
vcpus?
Thanks -- see my response to 00/15 for the rest.
-George
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
