Re: [Xen-devel] [PATCH 15/25] argo: implement the sendv op

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 20.12.18 at 06:58, <christopher.w.clark@xxxxxxxxx> wrote:
> On Wed, Dec 12, 2018 at 3:53 AM Jan Beulich <JBeulich@xxxxxxxx> wrote:
>> >>> On 01.12.18 at 02:32, <christopher.w.clark@xxxxxxxxx> wrote:
>> > +static struct argo_ring_info *
>> > +argo_ring_find_info_by_match(const struct domain *d, uint32_t port,
>> > +                             domid_t partner_id, uint64_t partner_cookie)
>> > +{
>> > +    argo_ring_id_t id;
>> > +    struct argo_ring_info *ring_info;
>> > +
>> > +    ASSERT(rw_is_locked(&d->argo->lock));
>> > +
>> > +    id.addr.port = port;
>> > +    id.addr.domain_id = d->domain_id;
>> > +    id.partner = partner_id;
>> > +
>> > +    ring_info = argo_ring_find_info(d, &id);
>> > +    if ( ring_info && (partner_cookie == ring_info->partner_cookie) )
>> > +        return ring_info;
>>
>> Such a cookie makes mismatches unlikely, but it doesn't exclude
>> them. If there are other checks, is the cookie useful at all?
> 
> Yes, I think so and it's proved useful elsewhere in the second
> version of the series: it helps avoid sending signals to incorrect
> domains that may not be argo-enabled.

"It helps avoid" still isn't "it allows to avoid", i.e. it still sounds like
an approach reducing likelihood instead of one excluding mistakes
altogether.

Jan



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