Re: [Xen-devel] capturing SIGKILL in DomU

["Srujan D. Kotikela" <ksrujandas@xxxxxxxxx>]

Hi Michal,

Thanks for those suggestions. Will look into xs_watch() and get back to you. Any pointers for the documentation would be appreciated.

thanks again.

--
Srujan D. Kotikela

On Tue, Oct 5, 2010 at 8:49 AM, Michal Novotny <minovotn@xxxxxxxxxx> wrote:

Hi Srujan,
basically this would be the trigger-like mechanism using the xenstore since it's using the xenstore watch facility itself if you implement it as I recommend. I'm not suggesting using the polling via xs_read() but you may study how to establish the xs_watch. This is being done, for example, by qemu-dm for looking for pci_ins/pci_del �and USB hotplug stuff using the xenstore_process_dm_command_event() function being called from the xenstore_process_event for the XS_WATCH_TOKEN vector - see tools/ioemu-dir/xenstore.c for reference.

I think this method is OK since it's already used for watching the hot-plug/hot-unplug stuff by the device model itself. Nevertheless I don't know what exactly do you want to do and maybe it's better to modify the hypervisor for your purpose but like I say, I don't know the purpose why you need this. All I'm saying it that this communication can be achieved using the xenstore watch implementation of the user-space stack which is being used for hot-plugging/hot-unplugging stuff already.

Regards,
Michal

On 10/05/2010 03:36 PM, Srujan D. Kotikela wrote:

Hi Michal,

I would prefer an trigger-like mechanism for communication rather a a process where I have to constantly poll or look for an event in DomU. How computing intensive would be xenstore lookup compared to event-channels with 100s of DomUs?

How about modifying hypervisor to do an upcall for the event handler in Dom0 when a specific interrupt occurs in DomU? If this can be done can you suggest any pointers for the same?

Thanks for your help.

--
Srujan D. Kotikela

On Tue, Oct 5, 2010 at 8:22 AM, Michal Novotny <minovotn@xxxxxxxxxx <mailto:minovotn@xxxxxxxxxx>> wrote:

� �Hi Srujan,

� �I'm not that familiar with event channels themselves however I was
� �thinking about using xenstore. You can modify the device model
� �(qemu-dm) to be watching some entry in the xenstore and the
� �communication could be both way since if you establish a xenstore
� �watch in both Dom0 and DomU you could intercept the changes on
� �both sides.

� �If you would like to use interrupts instead you may have to modify
� �the HVMLoader source codes at tools/firmware/hvmloader of the
� �user-space stack but I think using the xenstore could do the job
� �since this is how it's working with PV drivers AFAIK since PV
� �drivers themselves implement xenbus to connect to host's xenstore
� �facility.

� �Hope this helps!

� �Cheers,
� �Michal


� �On 10/05/2010 03:14 PM, Srujan D. Kotikela wrote:

� � � �Hi Michal,

� � � �I have no special interest in SIGKILL. All I want to do is
� � � �notify Dom0 about an event in DomU (I don't need to pass any
� � � �data). I am trying to indicate events by signals or
� � � �interrupts. It means, if a "particular" interrupt has occurred
� � � �in DomU, the Dom0 should be notified. Is there any other way
� � � �of doing the same other than using event channels?

� � � �I am successful in establishing the event channel. But I am
� � � �not quite sure how to send a notification that an event
� � � �occurred in DomU to Dom0. Any pointers for the same would be
� � � �appreciated.

� � � �--
� � � �Srujan D. Kotikela


� � � �On Tue, Oct 5, 2010 at 3:36 AM, Michal Novotny
� � � �<minovotn@xxxxxxxxxx <mailto:minovotn@xxxxxxxxxx>

� � � �<mailto:minovotn@xxxxxxxxxx <mailto:minovotn@xxxxxxxxxx>>> wrote:

� � � � � Hi Srujan,
� � � � � what about adding a signal handler to qemu-dm in the
� � � � � tools/ioemu-dir of the user-space tools? Using the signal()
� � � �API?
� � � � � Nevertheless why would you like to catch SIGKILL? This one
� � � �(as can
� � � � � be seen using included program source and killing it using
� � � �kill -9
� � � � � pid or kill -SIGKILL pid) is not being caught at all
� � � �nevertheless
� � � � � most of the other signals can be caught.

� � � � � This is the source of the example mentioned:
� � � � � #include <stdio.h>
� � � � � #include <signal.h>
� � � � � #include <stdlib.h>

� � � � � void sig_handler(int sig) {
� � � � � � �fprintf(stderr, "Signal %d caught.\n", sig);
� � � � � � �exit(sig);
� � � � � }

� � � � � int main()
� � � � � {
� � � � � � �signal(SIGINT, sig_handler);
� � � � � � �signal(SIGKILL, sig_handler);

� � � � � � �sleep(10000);
� � � � � � �return 0;
� � � � � }

� � � � � When I did try SIGINT (Ctrl + C or kill -2 pid) it caught the
� � � � � signal well but when I did try kill -9 pid (or kill
� � � �-SIGKILL pid
� � � � � respectively) it was not working at all since it killed the
� � � � � process instead of going to the signal handler. When you
� � � �need to
� � � � � catch signals like interruption signal (Ctrl + C one) this will
� � � � � work fine.

� � � � � Michal


� � � � � On 10/04/2010 09:03 PM, Srujan D. Kotikela wrote:

� � � � � � � Hi,

� � � � � � � I am trying to capture SIGKILL through event channel.

� � � � � � � On my Dom0, the following process is running (remaining
� � � �code
� � � � � � � in attachment).


� � � � � � � � �int main(void){

� � � � � � � � � � �int ret, dom, remote_dom;

� � � � � � � � � � �//initialize domains
� � � � � � � � � � �dom=0;
� � � � � � � � � � �remote_dom=2;

� � � � � � � � � � �//create the event channel
� � � � � � � � � � �ret = create_channel(dom, remote_dom);


� � � � � � � � � � �if (0 == ret) {
� � � � � � � � � � � � �printf("\n Event Channel established
� � � �successfully \n");
� � � � � � � � � � �} else {
� � � � � � � � � � � � �return -1; � �//EVENT_CHANNEL_CREATION_FAILED
� � � � � � � � � � �}

� � � � � � � � � � �//wait 20 seconds for an event to occur in DomU
� � � � � � � � � � �wait_for_event(20);

� � � � � � � � � � �//close the opened interfaces
� � � � � � � � � � �close_channel();

� � � � � � � � � � �return 0;

� � � � � � � � �}


� � � � � � � While this process is running; I killed a process in DomU
� � � � � � � using `*kill SIGKILL pid*`

� � � � � � � How can I capture this event (occured in DomU) at the
� � � �Dom0. I
� � � � � � � watched /dev/xen/evtchn, but no notification.


� � � � � � � --
� � � � � � � Srujan D. Kotikela


� � � � � � � _______________________________________________
� � � � � � � Xen-devel mailing list
� � � �Xen-devel@xxxxxxxxxxxxxxxxxxx
� � � �<mailto:Xen-devel@xxxxxxxxxxxxxxxxxxx>
� � � �<mailto:Xen-devel@xxxxxxxxxxxxxxxxxxx
� � � �<mailto:Xen-devel@xxxxxxxxxxxxxxxxxxx>>

� � � �http://lists.xensource.com/xen-devel



� � � � � -- � � Michal Novotny<minovotn@xxxxxxxxxx

� � � �<mailto:minovotn@xxxxxxxxxx> <mailto:minovotn@xxxxxxxxxx

� � � �<mailto:minovotn@xxxxxxxxxx>>>, RHCE

� � � � � Virtualization Team (xen userspace), Red Hat




� �-- � � Michal Novotny<minovotn@xxxxxxxxxx <mailto:minovotn@xxxxxxxxxx>>, RHCE
� �Virtualization Team (xen userspace), Red Hat

--
Michal Novotny<minovotn@xxxxxxxxxx>, RHCE
Virtualization Team (xen userspace), Red Hat

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel