[Bug]: Keepalive 设置不生效

[iqinlongfei]

Actions I've taken before I'm here

• I've thoroughly read the documentations on this issue but still have no clue.
• I've searched the Github Issues but didn't find any duplicate issues that have been resolved.
• I've searched the internet for this issue but didn't find anything helpful.

What happened?

TCP Server设置了5秒的KeepAlive时间，gnet.WithTCPKeepAlive(5*time.Second),在压测时，Client与Server大概建立了8000+条TCP连接，Client在运行中crash掉了，但Server端并没有按照设置的KeepAlive断开连接。

看了一下源码，TCP_KEEPINTVL和TCP_KEEPIDLE设置了5s，TCP_KEEPCNT没有设置，系统设置的是9。所以总的断开时间应该是5 + 5*9 = 50s。也就是说50s后，server端会断开连接。

但我抓包看，server端并没有发送keepalive，而是每隔120s进行一次重传。65.78是server，65.79是client

Major version of gnet

v2

Specific version of gnet

v2.3.0

Operating system

Linux

OS version

Linux 5.4.56.bsk.9-amd64 x86_64

Go version

go version go1.18.2 linux/amd64

Relevant log output

没有设置logger

Code snippets (optional)

No response

How to Reproduce

1. server设置Keepalive时间5s
2. server与client建立1000条TCP连接。
3. client端crash，确保不发送REST/FIN。测试时client通过XDP+用户态TCP协议栈，在crash时不会给server端发送RESET。
4. 发现server端会保持1000条连接很长时间而且期间抓包也没有看到Keepalive报文

Does this issue reproduce with the latest release?

I haven't verified it with the latest release

[panjf2000]

能提供一下可以复现这个问题的客户端和服务端代码吗？

[iqinlongfei]

代码需要整理一下，需要提取出来

[panjf2000]

最近 gnet 对 TCP Keep-Alives 设置做了以下优化: #522，可以验证一下这个问题是否存在，我先关闭 issue 了，如果还有问题可以随时重新打开这个 issue