[libvirt] [PATCH v2 13/17] qemu: assume support for all migration protocols except rdma

[Daniel P. Berrange]

On Mon, Nov 09, 2015 at 06:27:02PM -0500, John Ferlan wrote:
> 
> 
> On 11/09/2015 11:24 AM, Daniel P. Berrange wrote:
> > Since we require QEMU 0.12.0, we can assume that QEMU supports
> > all of the fd, tcp, unix and exec migration protocols.
> > 
> > Signed-off-by: Daniel P. Berrange <berrange at redhat.com>
> > ---
> >  src/qemu/qemu_capabilities.c                       | 34 +---------
> >  src/qemu/qemu_capabilities.h                       | 10 +--
> >  src/qemu/qemu_command.c                            | 44 +------------
> >  src/qemu/qemu_migration.c                          | 77 ++++++----------------
> >  tests/qemuargv2xmltest.c                           |  1 -
> >  tests/qemucapabilitiesdata/caps_1.2.2-1.caps       |  4 --
> >  tests/qemucapabilitiesdata/caps_1.3.1-1.caps       |  4 --
> >  tests/qemucapabilitiesdata/caps_1.4.2-1.caps       |  4 --
> >  tests/qemucapabilitiesdata/caps_1.5.3-1.caps       |  4 --
> >  tests/qemucapabilitiesdata/caps_1.6.0-1.caps       |  4 --
> >  tests/qemucapabilitiesdata/caps_1.6.50-1.caps      |  4 --
> >  tests/qemucapabilitiesdata/caps_2.1.1-1.caps       |  4 --
> >  tests/qemucaps2xmldata/all_1.6.0-1.caps            |  4 --
> >  tests/qemucaps2xmldata/nodisksnapshot_1.6.0-1.caps |  4 --
> >  tests/qemuhelptest.c                               | 32 ---------
> >  .../qemuxml2argv-migrate-numa-unaligned.args       |  2 +-
> >  .../qemuxml2argvdata/qemuxml2argv-restore-v1.args  | 23 -------
> >  tests/qemuxml2argvdata/qemuxml2argv-restore-v1.xml | 28 --------
> >  tests/qemuxml2argvtest.c                           | 13 ++--
> >  19 files changed, 33 insertions(+), 267 deletions(-)
> >  delete mode 100644 tests/qemuxml2argvdata/qemuxml2argv-restore-v1.args
> >  delete mode 100644 tests/qemuxml2argvdata/qemuxml2argv-restore-v1.xml
> > 
> 
> Similar to 8/17 - should the "-incoming" be removed from
> qemuParseCommandLine?

No, we're leaving cli parsing in place.
 
> Also, Coverity complained about one thing - although I'm not quite sure
> "how" it used the definition it did... I think this is a false positive,
> but could easily be avoided.

> > +    int fds[2];
> 
> (1) Event var_decl: 	Declaring variable "fds" without initializer.
> Also see events:
> 

[snip]

> (4) Event uninit_use_in_call: 	Using uninitialized element of array
> "fds" when calling "rpl_pipe2". [details]
> Also see events: 	[var_decl]
> 
> Not sure why this shows up - guess I'd have to understand the build a
> bit better... The only place where I found rpl_pipe2 is in unistd.h.
> 
> Initializing to {-1, -1}; avoids the error

Yeah that seems fine todo.


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 :|