As Keir said, the probe() function gets triggered by the right things
appearing in the XenStore. The XenIntro wiki has an example of this
in section 23 (Adding a new device and triggering...):
http://wiki.xensource.com/xenwiki/XenIntro
Hopefully that will be helpful for you.
Thanks
Ryan
On 10/2/07, Michael Abd-El-Malek <mabdelmalek@xxxxxxx> wrote:
> Hello,
>
> I'm writing a split driver for XenBus. My understanding is that this is
> the "Xen way" for inter-domain discovery, bootstrapping the process by
> sharing grants and event channels.
>
> When I look at the block driver, the probe method of the front- and
> back-ends seems to be doing the bulk of the initialization work. In my
> sample driver, however, my probe function isn't called. My module gets
> inserted and the initialization function correctly installs the XenBus
> driver.
>
> Can someone please explain when the probe function is supposed to be
> called? If there's no physical device "backing" the virtual device
> (i.e., virtual block driver is a counterexample), should I be relying on
> the probe method in the first place?
>
> Thanks,
> Mike
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel
>
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|