Re: [Xen-devel] clock source in PV Linux

[Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>]

On 12/18/19 12:36 AM, Roman Shaposhnik wrote:
> On Wed, Dec 11, 2019 at 12:41 AM Jan Beulich <jbeulich@xxxxxxxx> wrote:
> > On 11.12.2019 09:16, Jürgen Groß wrote:
> > > On 11.12.19 08:28, Jan Beulich wrote:
> > > > Jürgen, Boris,
> > > >
> > > > I've noticed
> > > >
> > > > <6>clocksource: Switched to clocksource tsc
> > > >
> > > > as the final clocksource related boot message in a PV Dom0's
> > > > log with 5.4.2. Is it intentional that it's not the "xen" one
> > > > that gets used by default?
> > > I think this is fine. I just tested it and I'm seeing the same in dom0,
> > > while in a PV domU "xen" is used per default.
> > >
> > > In dom0 "tsc" should be okay in case it is stable. Or are you expecting
> > > problems with that setting?
> > Well, first of all I found this surprising. Whether there are problems to
> > be expected largely depends on the reliability of the "stable" detection
> > in PV Dom0.
> Related question: does this mean that tsc is now default for PVH as well?
>
> The reason I'm asking is because I'm still a bit worried about the
> clock drift with tsc.

dom0 will use TSC for either PV or PVH:

xen_time_init():
       /* As Dom0 is never moved, no penalty on using TSC there */
        if (xen_initial_domain())
                xen_clocksource.rating = 275;

But as far as TSC stability I'd think it should be sufficiently checked 
by generic TSC init code?


-boris

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel