On Fri, Jun 03, 2016 at 03:26:59AM -0600, Jan Beulich wrote: > >>> On 02.06.16 at 18:12, <daniel.ki...@oracle.com> wrote: > > On Thu, Jun 02, 2016 at 02:34:28AM -0600, Jan Beulich wrote: > >> >>> On 01.06.16 at 21:03, <daniel.ki...@oracle.com> wrote: > >> > Hence, we are not able to get anything from EFI. Latter > >> > means that we were booted via older multiboot2 version which shutdown > >> > boot services. > >> > >> Hmm, that's certainly one of the possibilities, yes. I wonder then > >> whether we wouldn't better do away with all of those output > >> attempts then. > > > > This is one option. However, IMO, not nice. Maybe we should stay with > > serial plus VGA on legacy BIOS platforms. Serials looks quite well > > defined even on EFI platforms. However, if it is not available x86 > > out instruction should not do any harm. Though if we wish to increase > > chance of reaching interested parties I would send error messages to > > 0x3f8 and 0x2f8 at once. AFAICT, on many machines usually at least one > > if not both standard/legacy serial ports are available. > > On many _older_ machines you mean. Any new machines I got > hold of in the last couple of years didn't have any serial ports > anymore (some still have them implemented in the LPC, but not > wired through, but legacy free systems wouldn't have them at all > anymore).
And there are also new EFI machines that have no VGA whatsoever but only serial. > > Jan > _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel
