gnrc/RPL: instance ID not updated on reconnect #14407
Description
Expected results
When debugging something unrelated I just stumbled over the following behavior:
- start two RIOT nodes in the same network, node A as RPL root: let
instance ID := 11, node B as RPL node - node B selects node A as parent and updates its internal state:
node A:
2020-07-01 10:06:31,591 # rpl
2020-07-01 10:06:31,593 # instance table: [X]
2020-07-01 10:06:31,595 # parent table: [ ] [ ] [ ]
2020-07-01 10:06:31,595 #
2020-07-01 10:06:31,601 # instance [11 | Iface: 8 | mop: 2 | ocp: 0 | mhri: 256 | mri 0]
2020-07-01 10:06:31,608 # dodag [affe::1 | R: 256 | OP: Router | PIO: on | TR(I=[8,20], k=10, c=1, TC=6s)]
node B:
2020-07-01 10:06:33,161 # rpl
2020-07-01 10:06:33,163 # instance table: [X]
2020-07-01 10:06:33,166 # parent table: [X] [ ] [ ]
2020-07-01 10:06:33,166 #
2020-07-01 10:06:33,171 # instance [11 | Iface: 8 | mop: 2 | ocp: 0 | mhri: 256 | mri 0]
2020-07-01 10:06:33,179 # dodag [affe::1 | R: 512 | OP: Router | PIO: on | TR(I=[8,20], k=10, c=0, TC=3s)]
2020-07-01 10:06:33,183 # parent [addr: fe80::d18c:56ff:fe24:3ef | rank: 256]
- now we reboot the RPL root (node A) and re-initialize it with a new instance ID (
22):
node A:
2020-07-01 10:07:03,783 # instance table: [X]
2020-07-01 10:07:03,786 # parent table: [ ] [ ] [ ]
2020-07-01 10:07:03,786 #
2020-07-01 10:07:03,791 # instance [22 | Iface: 8 | mop: 2 | ocp: 0 | mhri: 256 | mri 0]
2020-07-01 10:07:03,799 # dodag [affe::1 | R: 256 | OP: Router | PIO: on | TR(I=[8,20], k=10, c=0, TC=0s)]
Problem: node B will does re-selects node A as parent after the reboot, but it will not update its instance ID:
node B:
2020-07-01 10:07:14,090 # rpl send dis
2020-07-01 10:07:14,094 # success: send a DIS
2020-07-01 10:07:14,094 #
2020-07-01 10:07:20,075 # rpl
2020-07-01 10:07:20,077 # instance table: [X]
2020-07-01 10:07:20,080 # parent table: [X] [ ] [ ]
2020-07-01 10:07:20,080 #
2020-07-01 10:07:20,085 # instance [11 | Iface: 8 | mop: 2 | ocp: 0 | mhri: 256 | mri 0]
2020-07-01 10:07:20,093 # dodag [affe::1 | R: 512 | OP: Router | PIO: on | TR(I=[8,20], k=10, c=0, TC=8s)]
2020-07-01 10:07:20,097 # parent [addr: fe80::d18c:56ff:fe24:3ef | rank: 256]
I would expect node B updating its parent table to the new instance ID, right?!
Actual results
see above
Versions
Observed in current master, using two nrf52dk nodes running IP-over-BLE.