|
|
|
|
|
|
|
|
|
|
xen-devel
RE: [Xen-devel] Clock interrupts + hypervisor calls
My understanding is that
1) There will be only one hypercall at any point of time( there is no
queue). Is this right?
No. In an SMP system there can be more than one Hypercall active at
any given time - one per CPU.
2) This hypercall can poll for events/work to do in case it is
preemptable. Is there any way that the hypercall can be made
non-preemptable?
Yes, you remove the poll-call. This will of course make the
hypercall non-preemptable and it was probably made preemptable for a very
good reason [lile somone found that it's working a lot better if you allow
preemption at this point], so it's probably not such a great idea to do
this.
--
Mats
Thanks,
Sowmya.
On 2/28/06, Keir
Fraser <Keir.Fraser@xxxxxxxxxxxx>
wrote:
On
28 Feb 2006, at 19:22, sowmya dayanand wrote:
> Is there a flag
that the long-lived hypercalls set to indicate that > they can be
preempted? Also is the flag per-domain to indicate that > the
hypercall is in progress in the particular domain?
Preemption is
voluntary. Long-lived hypercalls poll for pending events / work to
do.
-- Keir
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- RE: [Xen-devel] Clock interrupts + hypervisor calls,
Petersson, Mats <=
|
|
|
|
|