On 16/11/2018 10:11, Jan Beulich wrote: >>>> On 15.11.18 at 22:47, <andrew.coop...@citrix.com> wrote: >> Boris has confirmed that noone appears to be using PVRDTSCP any more, and in >> the decade since it was introduced, guest kernel / hardware support has >> provided a better alternative. > Doesn't removal of functionality require knowing that it was never used > at all, rather than just knowing that nothing uses it anymore? What if > some old guest somewhere relies on it?
Its an all-or-nothing feature. The entirety of your VM userspace need to support it, or timing will go wrong on migrate. We already established that it appears to be a vestigial Oracle-ism for which no consumer side code ever appeared, and that isn't used. What is unacceptable is PVRDTSCP's implementation causing breakages in architectural behaviour for non-PVRDTSCP configurations, and one way or another, this needs fixing. Please can we make a decision, because I don't have to time (or indeed, the want) to and fix this a 3rd different way if that's going to run into a similar reaction. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel