Merge lwaftr to next #1047
Merge lwaftr to next #1047
Conversation
'lwaftr query' improvements
Use a fixed amount of memory for reassembling IPv6 fragments, on a ctable. It only allocates once, on initialization, and never frees memory. When the backing table is full, it randomly ejects fragments. Tests include out-of-order reassembly. Bounds checking is done to avoid buffer overflows, regardles of input offsets. This includes a to-memory clone for Igalia's struct packet implementation. This allows IPv6 reassembly to avoid a subtle packet corruption bug on implementations where packet.data is an abstraction rather than a raw buffer.
This requires new tests to only be added to one file, rather than two, helping maintainability.
Normalize the vlan and non-vlan test suite
Merge ipv6 hardening
This is ctables-based, and similar to hardened IPv6 reassembly.
Hardened IPv4 reassembly. All tests pass.
Our codebase now relies on vlan tags being stripped/inserted nearer the edges of the app network.
Remove obsolete l2_size approach
This also includes a facility to regenerate test counter files.
Add a flag to allow monitor all packets in on-a-stick mode
This allows more accurate reporting in cases where apps process the packets before the lwaftr can (and after leaving it), including: a) Fragmentation/reassembly b) Control-plane messages, including ARP/NDP/pinging the lwaftr itself
| @@ -75,6 +70,32 @@ function PodHashMap.new(key_type, value_type, hash_fn) | |||
| return phm | |||
| end | |||
|
|
|||
| -- FIXME: There should be a library to help allocate anonymous | |||
| -- hugepages, not this code. | |||
There was a problem hiding this comment.
Still true. :-)
(Not a blocker, no action required for this PR).
| @@ -70,6 +70,7 @@ end | |||
|
|
|||
| function selftest () | |||
| print("selftest: timer") | |||
|
|
|||
There was a problem hiding this comment.
This seems like a spurious new line. (Not a blocker)
|
There is a new “nh_fwd” app that seems to be very lwaftr centric, or at least the README is vague to me. Why does it not belong into |
|
The nh_fwd (next-hop forward) app is part of snabb vmx. It sits outside (around, really) a snabb data plane and forwards control traffic to a virtualized control plane, such as an Ubuntu image or a Juniper vMX switch. That way it learns the next hop and can stamp on the L2 address of where the packets should go next. It is not lwaftr-specific. |
| if not args.counter:match(".counter$") then | ||
| args.counter = args.counter..".counter" | ||
| end | ||
| if not shm.exists("/"..S.getpid().."/"..args.counter) then |
There was a problem hiding this comment.
I think this should be shm.exists(args.counter). SHM paths are no longer resolved based on a changing shm.path.
|
I haven't spotted anything else that @eugeneia hasn't. |
This also addresses the review comments from upstream #1047, noting that the shm.exists() API changed.
Merge upstream ingress drop
If we are recording ingress drops to a counter, use the right check for the counter's existence. Addresses snabbco#1047 (review).
|
The three new commits look ok to me. |
Fix shm.exists() check in ingress drop monitor
|
Added a couple patches to |
Remove duplicated code in SnabbVMX selftests
|
LGTM, thanks for all your work on this. |
|
Originally |
|
So I made the rookie mistake of just making this a PR from lwaftr instead of branching off lwaftr and PR'ing from that. I will merge 3d33fb6 to wingo-next and pr that to next. |
|
@dpino thanks for the clarification on that, let's fix the readme issue in a followup |
* Remove the 'lwaftr generator' command * Move the 'packetblaster lwaftr' code from src/apps/ to the command directory * Add docs for using 'packetblaster lwaftr' in place of 'lwaftr generator'
Remove hardcoded constants in nh_fwd
…lename as in the other commands (#514)
|
Closing this PR as it's already merged and I made the rookie mistake of PRing this from a live branch which continues to get updates and thus false-alarm notifications. Thank you all! |
Merge v2018.04 prerelease from next to lwaftr
This PR merges in the last 5 months or so of lwaftr development to upstream Snabb. It also adds Snabb vMX, the system that joins up a Snabb dataplane to an external virtualized control/management plane, also adding "on-a-stick" operation for the lwaftr in particular but in a way that can apply to any Snabb dataplane.
With a few minor exceptions, this diff includes only changes to
src/apps/lwaftr,src/programs/snabbvmx, andsrc/programs/lwaftr. Review is very welcome on any of it, but I expect that you will want to focus on the non-lwaftr, non-vmx bits.