[libvirt] [PATCHv2] util: fix libvirtd startup failure due to netlink error

[Laine Stump]

On 05/04/2012 11:58 AM, Eric Blake wrote:
> On 05/04/2012 06:58 AM, Gerhard Stenzel wrote:
>> On Thu, 2012-05-03 at 11:10 -0400, Laine Stump wrote:
>>> Even with ACK, I will wait to push this until I have verification that
>>> it does not break lldpad<-->libvirtd communication (if it does, I may
>>> need to use the nl_handle allocated during virNetlinkStartup() for
>>> virNetlinkEventServiceStart()). 
>> libvirtd->lldpad communication is still working, but lldpad->libvirtd
>> not anymore (CONNECTION_REFUSED).
> On the surface, that makes it sound like lldpad has the same bug as
> libnl, of blindly assuming that it knows which address will be bound,
> rather than realizing that netlink sockets might be used elsewhere and
> thus the bound address is not guaranteed.  Is it something that can be
> fixed quickly in lldpad?
>

Actually it turns out that lldpad is just using the address send as the
source nl_pid in messages it receives from libvirtd, and
virNetlinkCommand has that field hardcoded to be set to getpid(). I've
just made a set of patches that (I hope) remedies that:

  https://www.redhat.com/archives/libvir-list/2012-May/msg00340.html

and also made some builds for Gerhard to test (since he's lucky enough
to have access to the proper hardware :-)

(Roopa, you may also want to take a look and make sure that I didn't
mess up any of the non-lldpad netlink communications - I don't have the
hardware to test those either...)