On 08/28/2012 07:28 AM, Alex Jia wrote:
* src/util/virnetdevopenvswitch.c (virNetDevOpenvswitchAddPort): avoid libvirtd crash due to derefing a NULL virtVlan->tag.
RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=852383
Signed-off-by: Alex Jia <ajia@redhat.com> --- src/util/virnetdevopenvswitch.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/src/util/virnetdevopenvswitch.c b/src/util/virnetdevopenvswitch.c index 601d79e..7d38ff8 100644 --- a/src/util/virnetdevopenvswitch.c +++ b/src/util/virnetdevopenvswitch.c @@ -100,7 +100,8 @@ int virNetDevOpenvswitchAddPort(const char *brname, const char *ifname, virBufferAsprintf(buf, "%d", virtVlan->tag[i]); } } else { - virBufferAsprintf(buf, "tag=%d", virtVlan->tag[0]); + if (virtVlan->nTags) + virBufferAsprintf(buf, "tag=%d", virtVlan->tag[0]);
While this does eliminate the crash, it shouldn't have been necessary, and doesn't fix the root cause, which was a couple levels up in the callstack. The real problem is that &net->vlan was being sent, rather than virDomainNetGetActualVlan(net) (which returns NULL if the vlan info found has no tags). I've sent a separate patch to fix that (I should have caught it during review :-( ): https://www.redhat.com/archives/libvir-list/2012-August/msg01835.html