On Thu, Nov 17, 2011 at 05:49:10AM -0500, Laine Stump wrote:
On 11/15/2011 06:14 AM, Daniel P. Berrange wrote:
>From: "Daniel P. Berrange"<berrange(a)redhat.com>
>
>Rename the macvtap.c file to virnetdevmacvlan.c to reflect its
>functionality. Move the port profile association code out into
>virnetdevvportprofile.c. Make the APIs available unconditionally
>to callers
One concern I have after reading this sentence - do you changes have
any implications for systems that don't have the correct version of
libnl, and thus cannot have macvtap enabled? (e.g. RHEL5)
Assuming that the answer to the above is "no", and going on faith
(and the fact that make && make check passed) that the gigantic diff
really is just moving identical code between files (except for
removing the #ifdef MACVTAPs in several places) - ACK.
The goal of this entire series, is that all *header* files definitions
should be unconditionally available. The idea is that callers (eg the
QEMU driver) can then avoid be littered with conditionals like
#ifdef MACVTAP
The implementation source files will be the sole place where the
conditionals live, in the #else clauses providing stubs for every
API which simply raise a libvirt error with ENOSYS
Overall this should make our code more robust wrt the various different
conditional compile options
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 :|