On Oct 22, 2014 11:28 AM, "Antti Kantee" <[email protected]> wrote:
>
> On 21/10/14 18:58, Justin Cormack wrote:
> > On Tue, Oct 21, 2014 at 7:34 PM, Antti Kantee <[email protected]> wrote:
> >>> Signal handling does work if you only use kqueue to receive signals, I
> >>> certainly don't have any inclination to add any more signal handling
> >>> than that...
> >>
> >> Does receiving signals from a rump kernel via rump_sys_kqueue really
> >> work? At least by default I hugely doubt it.
> >
> > Oh no I misremembered, it gives not supported.
>
> Signal delivery via kqueue, a concept akin to signalfd, probably could
> be supported with relatively little effort. However, I doubt many real
> programs use it. In fact, it's most useful when you do not want signals
> to mess up processing.
>
> Relevant code:
> http://nxr.netbsd.org/xref/src/sys/rump/librump/rumpkern/signals.c#43
>
> (note, that code was written and "designed" with the assumption that
> rump kernels are running in userspace. The sigmodel stuff was "just an
> idea")
I think it is reasonably common to use kqueue for everything including
signals in modern code. But signals are not used much, but it would be nice
to support it so there is a solution for them in place. Even if it is
largely for eg NetBSD tests...
------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://p.sf.net/sfu/Zoho
_______________________________________________
rumpkernel-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rumpkernel-users
