[Bug report] Unable to remove 'default' routes using the CLI tools

[Thomas Graf]

On 03/22/13 at 05:21pm, Roopa Prabhu wrote:
> On 3/21/13 11:09 AM, Thomas Graf wrote:
> >On 03/19/13 at 01:37pm, Roopa Prabhu wrote:
> >>Fix libnl to include additional attributes in the hash key depending
> >>on 'message type + msg flags'.
> >>ie In this case, use next hop info during lookup when ('msgtype ==
> >>RTM_NEWROUTE&&  (msgflag == NLM_F_CREATE || 'msgflag&
> >>NLM_F_APPEND' || 'and more if needed'))
> >
> >Why do we want to only regard the nexthop info if the append flag
> >is set? What's wrong with always comparing nexthops?
> 
> I pointed out nexthop because in tony's problem, only nexthop
> attribute was different between the two routes. It could have been
> any other attribute like scope, type etc.
> 
> I think our current key attributes are correct. Kernel also seems to
> use the same during lookup. However, I was trying to say that after
> lookup we should probably use more attributes to compare the objects
> than simply nl_object identical. We could use nl_object_diff or a
> variant depending on whether kernel appended, replaced or created a
> fresh route. At this point i am afraid adding more attributes in the
> lookup might break something else.
> I will think about this some more and also look at the kernel code again.

OK, I now understand what you have been saying before. Let's look
into how we can improve the kernel situation and be more verbose
on what action was taken.