x/net/http2: discrepancies in lost PING handling between Server and Transport

[mmatczuk]

Go version

x/net v0.30.0

Output of go env in your module/workspace:

Not important.

What did you do?

I configured http2.Server with

• ReadIdleTimeout
• PingTimeout
• CountError

and caused PING to be lost.

What did you see happen?

The newly added support for lost PING in server:

• only logs lost ping in verbose mode while closes the underlying connection - IMO should be logged always
• does not invoke CountError

if sc.pingSent {
    sc.vlogf("timeout waiting for PING response")
    sc.conn.Close()
    return
}

https://cs.opensource.google/go/x/net/+/refs/tags/v0.30.0:http2/server.go;l=1047

vs

func (cc *ClientConn) closeForLostPing() {
	err := errors.New("http2: client connection lost")
	if f := cc.t.CountError; f != nil {
		f("conn_close_lost_ping")
	}
	cc.closeForError(err)
}

https://cs.opensource.google/go/x/net/+/refs/tags/v0.30.0:http2/transport.go;l=1159

What did you expect to see?

I expect to see

• ErrorLog called with lost ping message
• CountError called with conn_close_lost_ping

[gabyhelp]

Related Issues and Documentation

• x/net/http2: configurable server pings #67812 (closed)
• x/net/http2: make Transport occasionally send PING frames to server #31643 (closed)
• x/net/http2: pings ineffective when connection writes block #48810 (closed)

_{(Emoji vote if this was helpful or unhelpful; more detailed feedback welcome in this discussion.)}

[gopherbot]

Change https://go.dev/cl/635555 mentions this issue: http2: improve handling of lost PING in Server