Re: timekeeping regression?

[Brad Spencer <brad%anduin.eldar.org@localhost>]

"Greg A. Woods" <woods%planix.ca@localhost> writes:

>
> At Wed, 20 Mar 2024 03:31:00 +0000, "Mathew, Cherry G.*" <c%bow.st@localhost> wrote:
> Subject: Re: timekeeping regression?
>>
>> I think the problem here is that there's no clarity in the API about
>> ownership. So we'll probably need to have a look at what linux and/or
>> FreeBSD does, to understand what the current consensus is - this seems
>> to be the linux way of documentation - a sort of mob rules situation.
>
> Hmmm...  FreeBSD (14.0) is now the only domU system I have that's been
> keeping time properly, even with the dom0 running on one pinned CPU.
>
> All the NetBSD systems have experienced large time drifts after about 7
> days of uptime.
>
> So, yeah, FreeBSD is doing something right/better!
>
> --
> 					Greg A. Woods <gwoods%acm.org@localhost>
>
> Kelowna, BC     +1 250 762-7675           RoboHack <woods%robohack.ca@localhost>
> Planix, Inc. <woods%planix.com@localhost>     Avoncote Farms <woods%avoncote.ca@localhost>
>


I have lost track... is your FreeBSD domain a pure PVH or a normal PV or
PVSHIM??  All of my multi-vcpu pure PVH guests can keep time pretty
well, but nearly all multi-vcpu PV guest (pure PV or PVSHIM) will tend
to lose their minds after a while, although not in the same way you have
been describing.  Mine tend to want to have the monatomic clock go
backwards, but can keep the clock synced until the backwards event
happens.  The single-vcpu guests seem to be able to keep time ok in all
cases.  I have never pined cpus, but do mess with the scheduling credit
some.








-- 
Brad Spencer - brad%anduin.eldar.org@localhost - KC8VKS - http://anduin.eldar.org