Does Xen support a dynamic allocation of physical memory for a HVM domain? Like the baloon driver of a domU? If not, how can this be achieved if this is possible at all? Thanks, David. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 31/12/06 8:49 am, "David Pilger" <pilger.david@gmail.com> wrote:> Does Xen support a dynamic allocation of physical memory for a HVM > domain? Like the baloon driver of a domU? > > If not, how can this be achieved if this is possible at all?All that''s missing is the balloon driver itself (but the PV Linux one could be used as a basis), support for using the memory-reservation hypercall from HVM context, and a way to blow away qemu''s memory-mapping cache when memory is given back to Xen from the HVM guest. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 12/31/06, Keir Fraser <Keir.Fraser@cl.cam.ac.uk> wrote:> > All that''s missing is the balloon driver itself (but the PV Linux one could > be used as a basis), support for using the memory-reservation hypercall from > HVM context, and a way to blow away qemu''s memory-mapping cache when memory > is given back to Xen from the HVM guest. >If you plan to not modify the HVM, Is there a way to make portions of physical memory unusable for a HVM windows from the hypervisor in a dynamic way? Thanks, _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 31/12/06 1:03 pm, "David Pilger" <pilger.david@gmail.com> wrote:> On 12/31/06, Keir Fraser <Keir.Fraser@cl.cam.ac.uk> wrote: >> >> All that''s missing is the balloon driver itself (but the PV Linux one could >> be used as a basis), support for using the memory-reservation hypercall from >> HVM context, and a way to blow away qemu''s memory-mapping cache when memory >> is given back to Xen from the HVM guest. >> > > If you plan to not modify the HVM, Is there a way to make portions of > physical memory unusable for a HVM windows from the hypervisor in a > dynamic way?In a way that is detectable by the guest OS in a way that causes it to react in a reasonable way? Unlikely, since there is no equivalent of this on a native x86 platform. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 12/31/06, Keir Fraser <Keir.Fraser@cl.cam.ac.uk> wrote:> On 31/12/06 1:03 pm, "David Pilger" <pilger.david@gmail.com> wrote: > > > On 12/31/06, Keir Fraser <Keir.Fraser@cl.cam.ac.uk> wrote: > >> > >> All that''s missing is the balloon driver itself (but the PV Linux one could > >> be used as a basis), support for using the memory-reservation hypercall from > >> HVM context, and a way to blow away qemu''s memory-mapping cache when memory > >> is given back to Xen from the HVM guest. > >> > > > > If you plan to not modify the HVM, Is there a way to make portions of > > physical memory unusable for a HVM windows from the hypervisor in a > > dynamic way? > > In a way that is detectable by the guest OS in a way that causes it to react > in a reasonable way? Unlikely, since there is no equivalent of this on a > native x86 platform. >Pretty conclusive :) Thanks! _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel