refactor(client): simplify keyupdate testcase implementation

[mxinden]

The QUIC Interop Runner keyupdate testcase has a client establish a connection to the server and then trigger a key update.

https://github.com/quic-interop/quic-interop-runner/blob/2a2534a1284d50d99ff92884d4f1ecf98fb41e4c/testcases.py#L889

This testcase always uses the http09 client and server implementation.

This commit simplifies the testcase implementation:

• Given that it is only used with http09, move it to http09.rs.
• Reduce the KeyUpdateState struct to a single bool.
• Mark the --key-update command line argument as hidden, given that it is only set indirectly through the -t keyupdate flag.
• Try to run client.initiate_key_update on events only, not on every new received datagram.

In addition it enables the keyupdate test on the Neqo qns.yml CI workflow.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.13%. Comparing base (166b84c) to head (a83b972).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1808   +/-   ##
=======================================
  Coverage   93.13%   93.13%           
=======================================
  Files         117      117           
  Lines       36363    36339   -24     
=======================================
- Hits        33865    33843   -22     
+ Misses       2498     2496    -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mxinden]

keyupdate testcase succeeding (as it did before) across our test matrix.

	neqo-latest	quic-go	ngtcp2	neqo	msquic
neqo-latest	✓(H,U)	✓(H,U)	✓(H,U)	✓(H,U)	✓(H,U)
	?()	?()	?()	?()	?()
	✕()	✕()	✕()	✕()	✕()
:---------:	:---------:	:-----:	:----:	:----:	:----:
quic-go	✓(H,U)
	?()
	✕()
:---------:	:---------:	:-----:	:----:	:----:	:----:
ngtcp2	✓(H,U)
	?()
	✕()
:---------:	:---------:	:-----:	:----:	:----:	:----:
neqo	✓(H,U)
	?()
	✕()
:---------:	:---------:	:-----:	:----:	:----:	:----:
msquic	✓(H,U)
	?()
	✕()
:---------:	:---------:	:-----:	:----:	:----:	:----:

https://github.com/mozilla/neqo/actions/runs/8615252271/job/23610385581?pr=1808

[github-actions]

Benchmark results

Performance differences relative to 329af2f.

• drain a timer quickly time: [360.09 ns 367.82 ns 375.25 ns]
  change: [-1.8724% +0.0209% +2.0240%] (p = 0.98 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 1+1 entries
  time: [194.65 ns 195.10 ns 195.57 ns]
  change: [-0.0091% +2.7422% +10.069%] (p = 0.25 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 3+1 entries
  time: [238.64 ns 239.22 ns 239.87 ns]
  change: [-0.3720% +0.3155% +1.0455%] (p = 0.41 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 10+1 entries
  time: [236.67 ns 237.54 ns 238.55 ns]
  change: [-0.6077% -0.0707% +0.5377%] (p = 0.81 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 1000+1 entries
  time: [218.86 ns 219.05 ns 219.25 ns]
  change: [-1.2930% -0.6098% +0.0891%] (p = 0.08 > 0.05)
  No change in performance detected.

• RxStreamOrderer::inbound_frame()
  time: [118.61 ms 118.76 ms 119.00 ms]
  change: [-0.2877% -0.0509% +0.1811%] (p = 0.71 > 0.05)
  No change in performance detected.

• transfer/Run multiple transfers with varying seeds
  time: [119.71 ms 120.00 ms 120.29 ms]
  thrpt: [33.253 MiB/s 33.334 MiB/s 33.414 MiB/s]
  change:
  time: [+0.9814% +1.2866% +1.6137%] (p = 0.00 < 0.05)
  thrpt: [-1.5881% -1.2702% -0.9718%]
  Change within noise threshold.

• transfer/Run multiple transfers with the same seed
  time: [120.45 ms 120.65 ms 120.86 ms]
  thrpt: [33.097 MiB/s 33.154 MiB/s 33.209 MiB/s]
  change:
  time: [+0.4748% +0.7020% +0.9136%] (p = 0.00 < 0.05)
  thrpt: [-0.9053% -0.6971% -0.4725%]
  Change within noise threshold.

• 1-conn/1-100mb-resp (aka. Download)/client
  time: [1.1129 s 1.1167 s 1.1211 s]
  thrpt: [89.198 MiB/s 89.553 MiB/s 89.855 MiB/s]
  change:
  time: [-0.0832% +0.5697% +1.1667%] (p = 0.11 > 0.05)
  thrpt: [-1.1532% -0.5664% +0.0833%]
  No change in performance detected.

• 1-conn/10_000-parallel-1b-resp (aka. RPS)/client
  time: [427.55 ms 429.78 ms 432.06 ms]
  thrpt: [23.145 Kelem/s 23.268 Kelem/s 23.389 Kelem/s]
  change:
  time: [-0.5981% +0.1057% +0.7970%] (p = 0.77 > 0.05)
  thrpt: [-0.7907% -0.1056% +0.6017%]
  No change in performance detected.

• 1-conn/1-1b-resp (aka. HPS)/client
  time: [51.780 ms 52.084 ms 52.487 ms]
  thrpt: [19.052 elem/s 19.200 elem/s 19.312 elem/s]
  change:
  time: [+1.6380% +2.7697% +3.9125%] (p = 0.00 < 0.05)
  thrpt: [-3.7652% -2.6950% -1.6116%]
  💔 Performance has regressed.

Client/server transfer results

Transfer of 134217728 bytes over loopback.

Client	Server	CC	Pacing	Mean [ms]	Min [ms]	Max [ms]	Relative
msquic	msquic			773.9 ± 258.0	531.2	1218.3	1.00
neqo	msquic	reno	on	852.1 ± 83.8	742.6	1017.1	1.00
neqo	msquic	reno		909.3 ± 229.2	735.2	1376.2	1.00
neqo	msquic	cubic	on	930.0 ± 260.8	736.6	1434.0	1.00
neqo	msquic	cubic		860.3 ± 261.8	766.3	1604.7	1.00
msquic	neqo	reno	on	4498.9 ± 182.6	4230.4	4859.0	1.00
msquic	neqo	reno		4400.8 ± 297.7	4054.7	5104.2	1.00
msquic	neqo	cubic	on	4552.2 ± 212.2	4286.7	4986.7	1.00
msquic	neqo	cubic		4501.3 ± 273.6	4133.6	5090.6	1.00
neqo	neqo	reno	on	3617.8 ± 282.0	3176.7	4158.7	1.00
neqo	neqo	reno		3898.0 ± 250.5	3588.5	4351.2	1.00
neqo	neqo	cubic	on	4368.4 ± 311.1	4016.3	4950.9	1.00
neqo	neqo	cubic		4360.9 ± 223.1	4029.3	4794.3	1.00

⬇️ Download logs