ospfd: interface speed change during intf add

ospfd: interface speed change during intf add

The problem is seen where speed mismatch caused ECMP route
not being reflected with correct number paths (NHs).

During cold boot, some interface speed updated by zebra as
part of one shot timer and triggers interface add to clients.
In this case, ospf already have created interface (bond interface),
but speed was not updated, trigger to do interface speed change
as part of interface add, which will trigger all Router LSA to
use updated speed into cost calculation.

Ticket:CM-22170
Testing Done:
Bring up CLOS config with Spine and leafs. Leaf have CLAG pair,
with same VRR ip address.
At spine one of the bond connecting to leaf node was having
higher speed than the paired device, With this fix, at spine (DUT)
bond interface speed is equal from all peer nodes.

Signed-off-by: Chirag Shah <chirag@cumulusnetworks.com>