* Event Channel Semantics
@ 2006-11-09 11:03 Julian Stecklina
2006-11-09 15:34 ` Mark Williamson
0 siblings, 1 reply; 4+ messages in thread
From: Julian Stecklina @ 2006-11-09 11:03 UTC (permalink / raw)
To: xen-devel
Hello,
I am developing some userspace tools on Xen/Linux and have problems
regarding the semantics of /dev/linux/evtchn. The situation is as follows:
Suppose client A allocates an unbound port.
Client B connects to this via bind_interdomain.
Client A reads on /dev/linux/evtchn to see whether it has been notified.
If client B now notifies client A, the latter gets the correct port from
the call to read(). But all subsequent notifications get lost, and
client A just sits there blocking on read().
Is this intended? If not, where is the problem?
Regards,
Julian Stecklina
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Event Channel Semantics
2006-11-09 11:03 Event Channel Semantics Julian Stecklina
@ 2006-11-09 15:34 ` Mark Williamson
2006-11-16 1:06 ` Julian Stecklina
0 siblings, 1 reply; 4+ messages in thread
From: Mark Williamson @ 2006-11-09 15:34 UTC (permalink / raw)
To: xen-devel; +Cc: Julian Stecklina
> I am developing some userspace tools on Xen/Linux and have problems
> regarding the semantics of /dev/linux/evtchn. The situation is as follows:
>
> Suppose client A allocates an unbound port.
> Client B connects to this via bind_interdomain.
> Client A reads on /dev/linux/evtchn to see whether it has been notified.
>
> If client B now notifies client A, the latter gets the correct port from
> the call to read(). But all subsequent notifications get lost, and
> client A just sits there blocking on read().
>
> Is this intended? If not, where is the problem?
Are you clearing the event channel after receiving the first notification?
Cheers,
Mark
--
Dave: Just a question. What use is a unicyle with no seat? And no pedals!
Mark: To answer a question with a question: What use is a skateboard?
Dave: Skateboards have wheels.
Mark: My wheel has a wheel!
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Event Channel Semantics
2006-11-09 15:34 ` Mark Williamson
@ 2006-11-16 1:06 ` Julian Stecklina
2006-11-16 1:08 ` John Levon
0 siblings, 1 reply; 4+ messages in thread
From: Julian Stecklina @ 2006-11-16 1:06 UTC (permalink / raw)
To: xen-devel
Mark Williamson wrote:
> Are you clearing the event channel after receiving the first notification?
No, but is not obvious how to do that using the ioctls provided in
xen/linux/evtchn.h. IOCTL_EVTCHN_RESET seems to apply to all event channels.
Regards,
Julian
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Re: Event Channel Semantics
2006-11-16 1:06 ` Julian Stecklina
@ 2006-11-16 1:08 ` John Levon
0 siblings, 0 replies; 4+ messages in thread
From: John Levon @ 2006-11-16 1:08 UTC (permalink / raw)
To: Julian Stecklina; +Cc: xen-devel
On Thu, Nov 16, 2006 at 02:06:34AM +0100, Julian Stecklina wrote:
> Mark Williamson wrote:
> > Are you clearing the event channel after receiving the first notification?
>
> No, but is not obvious how to do that using the ioctls provided in
> xen/linux/evtchn.h. IOCTL_EVTCHN_RESET seems to apply to all event channels.
Write the event channel number to the device.
regards
john
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2006-11-16 1:08 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-11-09 11:03 Event Channel Semantics Julian Stecklina
2006-11-09 15:34 ` Mark Williamson
2006-11-16 1:06 ` Julian Stecklina
2006-11-16 1:08 ` John Levon
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.