On Mon, Apr 02, 2012 at 02:10:38PM -0700, Jason Helfman wrote:
On Mon, Apr 02, 2012 at 08:48:11PM +0100, Daniel P. Berrange thus
spake:
>>>>http://lorie.secnap.net/tb/index.php?action=display_markup_log&build=9-jgh_AMD64&id=1289
>>>>http://lorie.secnap.net/tb/index.php?action=display_markup_log&build=9-jgh_I386&id=1289
>
>The test suite doesn't depend on anything that the main code doesn't
>already require. So you should be able to just run 'make check' and
>it'll compile & run the tests.
>
>Of course reality may differ from theory :-)
I've enabled testing, now, if you want to take a look at the above urls.
Any comments?
>
>>>I'd hope to should pass, and would be interested in any failures
Let me know if you would like to see anything more here.
>>>
>>>
>I strongly expect that the QEMU driver has a large number of Linux-isms
>in it. That said, I also expect all the Linux-isms to be solveable. So
>my recommendation is probably to enable it, then insert enough #ifdef's
>to get just the basics compiling. Then post the #ifdef details to the
>list so we can figure out an optimal way to handle the Linux-isms long
>term, and make suggestions on how to provide BSD alternative impls.
>
>The kernel-headers check is an example of a Linux-ism we should look
>to solve.
I agree with removing the Linux-ism's, as it will only make it more
portable. Do you have an example of header files that I should be looking
for?
Here are two obvious candidates:
$ grep linux/ src/qemu/qemu*.c
src/qemu/qemu_driver.c:# include <linux/kvm.h>
src/qemu/qemu_driver.c:/* add definitions missing in older linux/kvm.h */
src/qemu/qemu_process.c:#include <linux/capability.h>
Regards,
Daniel
--
|:
http://berrange.com -o-
http://www.flickr.com/photos/dberrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|:
http://entangle-photo.org -o-
http://live.gnome.org/gtk-vnc :|