transport: http2 server must validates header list size when early aborting stream

[joybestourous]

Modifies earlyAbortStreamHandler to include check the header list size when early aborting, and returns a RST_STREAM if the max size is exceeded.

Fixes #8766

RELEASE NOTES:

• transport: http2 server now validates header list size when early aborting stream

[codecov]

Codecov Report

❌ Patch coverage is 89.65517% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 83.38%. Comparing base (489f1f6) to head (472ec38).
⚠️ Report is 2 commits behind head on master.

Files with missing lines	Patch %	Lines
internal/transport/controlbuf.go	50.00%	1 Missing and 1 partial ⚠️
internal/transport/http2_server.go	95.74%	1 Missing and 1 partial ⚠️
test/servertester.go	71.42%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8769      +/-   ##
==========================================
- Coverage   83.42%   83.38%   -0.05%     
==========================================
  Files         419      419              
  Lines       32556    32576      +20     
==========================================
+ Hits        27161    27162       +1     
- Misses       4015     4028      +13     
- Partials     1380     1386       +6     

Files with missing lines	Coverage Δ
internal/transport/controlbuf.go	88.91% <50.00%> (+0.29%)	⬆️
internal/transport/http2_server.go	91.28% <95.74%> (+0.21%)	⬆️
test/servertester.go	61.67% <71.42%> (-0.44%)	⬇️

... and 19 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[easwars]

@eshitachandwani: I'm moving this to @arjan-bal for second review as he was the one who filed the issue in the first place.

[arjan-bal]

I believe the controlBuf.mu mutex needs to be held while accessing t.maxSendHeaderListSize since the value can be updated while handling an HTTP/2 Settings frame from the peer. For the Settings to take effect, the server need to send a Settings ACK frame to the peer. The FIFO ordering of the controlbuf ensures that last ACKed settings are always respected.

I suggest using t.controlBuf.executeAndPut with a callback that validates the header size and enqueues earlyAbortStream if validation passes. If validation fails, call t.closeStream to close the stream with a RST frame. Let me know if this adds too much complexity.