>>> On 9/30/2008 at 5:45 AM, in message <20080930094546.GB7835@xxxxxxxxxxx>, Karel Zak <kzak@xxxxxxxxxx> wrote: > On Mon, Sep 29, 2008 at 03:13:27PM -0600, Ky Srinivasan wrote: >> >> >> >>> On 9/29/2008 at 5:01 PM, in message <20080929210105.GA7835@xxxxxxxxxxx>, > Karel >> Zak <kzak@xxxxxxxxxx> wrote: >> > On Mon, Sep 29, 2008 at 03:17:28PM +0200, Matthias Koenig wrote: >> >> The tool has been written by Ky Srinivasan <ksrinivasan@xxxxxxxxxx>. >> > >> > Do we really need a new tool? IMHO the hypervisor Vendor ID should >> > be exported by kernel in /sys or /proc -- or we can add this info to >> We also need a detection mechanism for fully virtualized (HVM) >> guests where we may not want to modify anything in the guest. This > > http://lkml.org/lkml/2008/9/26/351 > ^^^ It seems that the same code (based on CPUID) will be in standard > kernel. I don't think that we need to duplicate this functionality in > userspace. For example /sys/hypervisor/vendor makes more sense. This would be ideal. > > I understand that your tool is useful for arbitrary (including old) > kernels, but is it so important? We need this functionality for a number of kernels that are being deployed (either para-virtualized or fully virtualized) and kernels that will be released soon that will not have /sys/hypervisor/vendor functionality. The proposed tool addresses this problem. Even when we have all kernels of interest supporting /sys/hypervisor/vendor, it may still be useful to have a wrapper around the /sys/hypervisor/vendor information. > > BTW, I really don't like the name -- "hypervisor". What about > something less generic, "hypid", "hvid", ... Fine with me. Regards, K. Y > > Karel -- To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
