Re: [Xen-devel] [PATCH v4 15/18] xen/mem_sharing: VM forking

[Tamas K Lengyel <tamas@xxxxxxxxxxxxx>]

On Thu, Jan 9, 2020 at 8:10 AM Roger Pau Monné <roger.pau@xxxxxxxxxx> wrote:
>
> On Thu, Jan 09, 2020 at 06:41:12AM -0700, Tamas K Lengyel wrote:
> > On Thu, Jan 9, 2020 at 3:29 AM Julien Grall <julien@xxxxxxx> wrote:
> > >
> > > Hi Tamas,
> > >
> > > On 08/01/2020 17:14, Tamas K Lengyel wrote:
> > > > +static int mem_sharing_fork(struct domain *d, struct domain *cd)
> > > > +{
> > > > +    int rc;
> > > > +
> > > > +    if ( !d->controller_pause_count &&
> > > > +         (rc = domain_pause_by_systemcontroller(d)) )
> > >
> > > AFAIU, the parent domain will be paused if it wasn't paused before and
> > > this will not be unpaused by the same hypercall. Right?
> >
> > Yes, it needs to remain paused as long as there are forks active from
> > it. Afterwards it can be unpaused.
>
> If you want the parent domain to remain paused for as long as the
> forks are active, shouldn't each fork increment the pause count on
> creation and decrement it when the fork is destroyed?

That would work.

>
> How can you assure no other operation or entity has incremented
> controller_pause_count temporary and is likely to decrement it at some
> point while forks are still active?

Right now we don't do sanity checks. It is just expected that the user
is not doing anything insane like that - I would argue that for an
experimental system (one that is even hidden behind CONFIG_EXPERT) an
assumption like that is safe to make. But doing the pause/unpause
combo you describe above is pretty simple and should get the job done.

Tamas

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