On 28/09/15 11:56, Jan Beulich wrote: >>>> On 28.09.15 at 12:35, <roger....@citrix.com> wrote: >> El 21/09/15 a les 16.47, Jan Beulich ha escrit: >>>>>> On 04.09.15 at 14:09, <roger....@citrix.com> wrote: >>>> --- a/xen/include/public/elfnote.h >>>> +++ b/xen/include/public/elfnote.h >>>> @@ -200,9 +200,18 @@ >>>> #define XEN_ELFNOTE_SUPPORTED_FEATURES 17 >>>> >>>> /* >>>> + * Physical entry point into the kernel. >>>> + * >>>> + * 32bit entry point into the kernel. Xen will use this entry point >>>> + * in order to launch the guest kernel in 32bit protected mode >>>> + * with paging disabled. >>>> + */ >>>> +#define XEN_ELFNOTE_PHYS32_ENTRY 18 >>> The comment reads as if this was the case for all kinds of guests, >>> yet I suppose it doesn't apply to PV ones. This should be made >>> explicit if so. >> Yes, what about the following: >> >> 32bit entry point into the kernel. Xen will use this entry point >> in order to launch the guest kernel in 32bit protected mode with paging >> disabled inside of an HVM container. > Depends: If the note's presence means this and only this entry point > will be used, then okay. If, however, normal PV and/or HVM operation > of such a guest is still intended to be possible, then I think this is still > too vague. Perhaps > > 32bit entry point into the kernel. When requested to launch the > guest kernel in a HVM container, Xen will use this entry point to > launch the guest in 32bit protected mode with paging disabled. > Ignored otherwise.
A multi-mode binary seems very likely, certainly for the near future. As such, this param is an indication of supporting DMLite. The guest kernel knows it was started in DMLite mode if this is the entry point used. If another entry point is used, the guest was not started in DMLite mode. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel
