Christian Limpach wrote: >> From: virtualization-bounces at lists.osdl.org >> [mailto:virtualization-bounces at lists.osdl.org] On Behalf Of >> Rusty Russell >> >> 2) The hypervisor type is handed through %ebx to the startup_paravirt >> function at boot. Currently 0 = Xen 3.0, 1 = VMI. >> > > I think this is a really really bad idea because you're requiring that > all operating systems agree on these values or you impose it upon the > hypervisor's domain builder to be operating system specific. I really > don't think either of those are very appealing, especially considering > that there's a much simpler solution using hypervisor specific entry > points. > > That said, %ebx = 0 works for us, so maybe I should just not care... > I've been arguing against hypervisor specific entry points, not just because we don't need one (in fact, setting %ebx on entry is nearly an impossible amount of work for us), but because the ELF file format just doesn't support that. Maybe I'm out of place in trying to suggest an alternative mechanism for the Xen entry point into Linux, but I really would like to see the kernel entry point play nice with kexec and future bootloaders / domain builders, which is why I think it is important. We need a probing type interface instead of a magic start interface. We can probe cleanly without side effects on normal hardware platforms, which makes it nicer. In the end, we would very much like to have a nicer and clean start of day, like Xen does, but unfortunately for us today that is very far off in the future. Zach
