Date: Wed, 10 Dec 2014 17:39:46 -0500 From: Ryan Stone <rysto32@gmail.com> To: Julian Stecklina <jsteckli@os.inf.tu-dresden.de> Cc: freebsd-virtualization@freebsd.org Subject: Re: KVM Clock Message-ID: <CAFMmRNx18L829FhexZmrgmJ7kj-hi%2BXy556KPV7Nyb4qgo2A3g@mail.gmail.com> In-Reply-To: <ed327598-1439-4446-96d0-bd3b8fd24060@email.android.com> References: <lb3jnb$qo8$1@ger.gmane.org> <52D6B18D.4080204@citrix.com> <1389805150.16498.18.camel@janus.xn--pl-wia.net> <3298215.G6vx8QjdMB@pippin.baldwin.cx> <ed327598-1439-4446-96d0-bd3b8fd24060@email.android.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jan 20, 2014 at 7:41 PM, Julian Stecklina <jsteckli@os.inf.tu-dresden.de> wrote: > Ah. Thanks. This will do. Something like access_once would be perfect, though. > > I'll post an updated patch that does not duplicate code that is in xen/ soonish. Didn't get around to testing it the last days. > > Julian I'm interesting in helping to get this into the tree. I see that you've address all of the comments in this thread here: https://github.com/blitz/freebsd/commit/e130f4ba550eb905c7f56a6a26243660b4df834b I have one final question. Don't we need to run rdtsc() in the loop in pvclock_fetch_vcpu_tinfo()? If we don't, isn't there a chance that after we leave the loop, the hypervisor could migrate the guest to a different CPU where the value returned by rdtsc() will not match the parameters in the pvclock_time_info structure?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFMmRNx18L829FhexZmrgmJ7kj-hi%2BXy556KPV7Nyb4qgo2A3g>