[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH RFC v1 06/74] tools/libelf: fix elf notes check for PVH guest
On Thu, Jan 04, 2018 at 07:37:20AM -0700, Jan Beulich wrote: > >>> On 04.01.18 at 14:05, <wei.liu2@xxxxxxxxxx> wrote: > > PVH only requires PHYS32_ENTRY to be set. Return immediately if that's > > the case. > > So I guess the bug(?) being fixed is that so far loader or guest_os, > and xen_ver settings are also required. However, you fail to mention > _why_ you think they're not required. I can sort of see this for > loader and maybe guest_os, but for the Xen version this isn't as > obvious, mainly because any arguments I can think of right now > would equally apply to PV. Got it from docs/misc/pvh.markdown. It doesn't state other notes are required. I'm not sure if xen_version (always "xen-3.0"?) will be meaningful or useful. It is not the end of the world if we check it but we do need to be careful to not break existing OSes (mini-os for one only sets PHYS32_ENTRY for PVH mode but that's easy to fix). > > > --- a/xen/common/libelf/libelf-dominfo.c > > +++ b/xen/common/libelf/libelf-dominfo.c > > @@ -381,6 +381,13 @@ static elf_errorstatus elf_xen_note_check(struct > > elf_binary *elf, > > return 0; > > } > > > > + /* PVH only requires one ELF note to be set */ > > + if ( parms->phys_entry != UNSET_ADDR32 ) > > + { > > + elf_msg(elf, "ELF: Found PVH image\n"); > > + return 0; > > + } > > If the other entries are of no interest for PVH, I think that this > then calls for dropping their logging from pvh_load_kernel(). Sure. > I'm also surprised that I can't find any use of any of the three > values checked in libxc. Libxc delegates the work to libelf AIUI. Wei. _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxxx https://lists.xenproject.org/mailman/listinfo/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |