RPL fixes and enhancements #1034
Conversation
tbursztyka
left a comment
tbursztyka
left a comment
There was a problem hiding this comment.
Looks like you can drop the 2 firsts patches.
subsys/net/ip/rpl.c
Outdated
There was a problem hiding this comment.
Is this worth exposing, why would it be needed to be used from shell? (besides some low level rpl testing maybe). Repairing the rpl network is meant to be automatic, and afaik, rpl is designed to be smart enough to do so relevantly.
There was a problem hiding this comment.
This is related to #980 RPL border router sample application where user can manually trigger repair.
There was a problem hiding this comment.
Could you merge the 2 first patches in one then? I don't see why it has to be separated, both follows the same exact need.
subsys/net/ip/rpl.c
Outdated
There was a problem hiding this comment.
same comment as previous patch
subsys/net/ip/rpl.c
Outdated
There was a problem hiding this comment.
you could have the reverse test to return false directly here.
It would be easier to read, and avoid to break the line on the if () condition below.
Just a taste preference though.
Btw, (it's weird we can't comment the commit message): any rfc reference to add in commit message?
That is indeed weird. I usually create a separate comment if there is some issue in commit message. |
The commit b14586c ("net: rpl: RPL route entry was fetched too late") dropped the DAO packet too early which prevents the RPL root node functionality. Rework the earlier commit so that Coverity issues are fixed but the RPL DAO message is also properly handled. Signed-off-by: Jukka Rissanen <[email protected]>
The net_rpl_repair_root() and net_rpl_set_root_with_version() functions were static which prevented global repair when using an instance id. Making those functions global allows RPL network repairing from shell for example. Signed-off-by: Jukka Rissanen <[email protected]>
If debug level was low but debugging was activated, then some of the debug variables were left unused. Signed-off-by: Jukka Rissanen <[email protected]>
Instead of always forwarding the DAO, send DAO-ACK if we are the root node. Signed-off-by: Jukka Rissanen <[email protected]>
Create net_rpl_foreach_parent() function that will traverse through all the parents and return their information via callback. This is needed by net-shell in later commit. Signed-off-by: Jukka Rissanen <[email protected]>
Add "net rpl" command to net-shell which will return both static RPL configuration from Kconfig, and dynamic run time configuration of the nodes and parents. Signed-off-by: Jukka Rissanen <[email protected]>
Print route information by groupping them for each network
interface like this:
IPv6 routes for interface 0xa8007140
====================================
IPv6 prefix : 2001:db8::212:4b00:0:3/128
neighbor : 0xa80065e0
link addr : 00:12:4B:00:00:00:00:03
IPv6 prefix : 2001:db8::212:4b00:0:1/128
neighbor : 0xa8006660
link addr : 00:12:4B:00:00:00:00:01
Signed-off-by: Jukka Rissanen <[email protected]>
The "net nbr" command was printing the heading of neighbor list incorrectly. Signed-off-by: Jukka Rissanen <[email protected]>
|
Updated new version according to comments. |
…ephyrproject-rtos#1034) Naming the assert function accomplishes the same thing here. Signed-off-by: Geoff Gustafson <[email protected]>
3 participants
No description provided.