Add sctp.Association.Abort(reason string) method

[jeremija]

According to RFC 4960 Section 9.1. Also see my comment at #176, more precisely: #176 (comment)

Closes #182.

I don't think we currently do any tag verification for any packets, but we can implement that later.

I think that Close method should be modified to call Abort("close") or similar; I was unable to find any part of the spec that allows just closing the underlying connection.

[codecov]

Codecov Report

Merging #183 (5eb7e48) into master (da38088) will decrease coverage by 2.54%.
The diff coverage is 88.09%.

❗ Current head 5eb7e48 differs from pull request most recent head 50c75e6. Consider uploading reports for the commit 50c75e6 to get more accurate results

@@            Coverage Diff             @@
##           master     #183      +/-   ##
==========================================
- Coverage   80.64%   78.09%   -2.55%     
==========================================
  Files          47       47              
  Lines        3901     2648    -1253     
==========================================
- Hits         3146     2068    -1078     
+ Misses        617      441     -176     
- Partials      138      139       +1     

Flag	Coverage Δ
go	78.09% <88.09%> (-2.55%)	⬇️
wasm	64.53% <2.38%> (-3.29%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
error_cause_user_initiated_abort.go	77.77% <77.77%> (ø)
association.go	81.50% <90.32%> (-2.90%)	⬇️
error_cause.go	37.77% <100.00%> (+3.73%)	⬆️
chunk_init_ack.go	19.35% <0.00%> (-18.15%)	⬇️
chunk_init.go	19.35% <0.00%> (-17.01%)	⬇️
chunk_init_common.go	54.16% <0.00%> (-11.36%)	⬇️
packet.go	67.90% <0.00%> (-5.18%)	⬇️
chunk_cookie_ack.go	55.55% <0.00%> (-4.45%)	⬇️
chunk_payload_data.go	68.96% <0.00%> (-3.88%)	⬇️
paramtype.go	13.79% <0.00%> (-3.67%)	⬇️
... and 40 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update da38088...50c75e6. Read the comment docs.

[enobufs]

Just one minor request from me. There's merge conflict to resolve. Other changes look good to me.

[enobufs]

Are you able to add the following text just to be consistent with other error_cause_xxx.go files?

/*
   This error cause MAY be included in ABORT chunks that are sent
   because of an upper-layer request.  The upper layer can specify an
   Upper Layer Abort Reason that is transported by SCTP transparently
   and MAY be delivered to the upper-layer protocol at the peer.

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |         Cause Code=12         |      Cause Length=Variable    |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       /                    Upper Layer Abort Reason                   /
       \                                                               \
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
*/

[jeremija]

Sure, I'll try to address it over the next couple of days

[jeremija]

Done!

[enobufs]

Thanks @jeremija !

Some auto tests are failing with TestAssociation_HandlePacketBeforeInit/Abort.
From what I can see, a.handleAbort() calls a.close() which sets the state to closed. Then the test case calls a.Close() which complains "use of closed network connection".

I think calling a.close() on the receipt of Abort chunk is right. I believe we could safely skip a.Close() in the test when the test case is "Abort" in TestAssociation_HandlePacketBeforeInit.

Otherwise, it all looks good to me!

[enobufs]

This is something from before but I think this test case name should be changed to something like: "TestAssociation_HandlePacketInCookieWaitState" IMO.

[jeremija]

Thanks, I didn't realize there was a failing test. It should be fixed now. I also renamed the test case.

[enobufs]

@jeremija I am pretty sure this is a test issue. If you'd prefer, I can sort this out for you. Please let me know.

[enobufs]

Thanks @jeremija . LGTM!