From: Tim Deegan Subject: Re: [PATCH][v4] PV extension of HVM(hybrid) support in Xen Date: Mon, 1 Mar 2010 10:21:27 +0000 Message-ID: <20100301102127.GA17243@whitby.uk.xensource.com> References: <201003011743.41341.sheng@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Return-path: Content-Disposition: inline In-Reply-To: <201003011743.41341.sheng@linux.intel.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Sheng Yang Cc: Ian Campbell , Ian Pratt , xen-devel , Keir Fraser List-Id: xen-devel@lists.xenproject.org At 09:43 +0000 on 01 Mar (1267436621), Sheng Yang wrote: > @@ -3109,6 +3117,36 @@ > break; > } > > + case HVMOP_enable_pv: { Why does this have to be explicitly enabled? Can't you just notice that a domain is using the evtchnop hypercalls? > + struct xen_hvm_pv_type a; > + struct domain *d; > + > + if ( copy_from_guest(&a, arg, 1) ) > + return -EFAULT; > + > + rc = rcu_lock_target_domain_by_id(a.domid, &d); Domains do this to each other? It looks like it has surprising side-effects. > + if ( rc != 0 ) > + return rc; > + > + rc = -EINVAL; > + if ( !is_hvm_domain(d) ) > + goto param_fail5; > + > + rc = xsm_hvm_param(d, op); > + if ( rc ) > + goto param_fail5; > + > + if (a.flags & HVM_PV_EVTCHN) { > + update_domain_wallclock_time(d); > + hvm_funcs.set_tsc_offset(d->vcpu[0], 0); Only vcpu 0? Doesn't this do horrible things to timekeeping in the guest? > + d->hvm_pv_enabled |= XEN_HVM_PV_EVTCHN_ENABLED; > + printk("HVM: PV featured evtchn enabled\n"); Please remove your debugging printks. > + } > +param_fail5: > + rcu_unlock_domain(d); > + break; > + } > + > default: > { > gdprintk(XENLOG_WARNING, "Bad HVM op %ld.\n", op); Tim. -- Tim Deegan Principal Software Engineer, XenServer Engineering Citrix Systems UK Ltd. (Company #02937203, SL9 0BG)