Data Race in dialScheduler

[SecTechTool]

System information

Geth version: 1.10.9-unstable-9ada4a2e-20210910
OS & Version: MacOS
Network: Private test net

Expected behaviour

node run normally

Actual behaviour

Data race in dialScheduler

Steps to reproduce the behaviour

1. setup a 10-node private geth nodes lcoally
2. setup a fuzzing node continually sending fuzzed messages to other 10 normal geth nodes.
3. After more than 24 hours fuzzing experiment, one of the geth node who is run in fast mode crashed down.
   The running command for the node is ./build/bin/geth --identity "ETH-node10" --datadir "node10" --ethash.dagdir "node10" --port "30312" --maxpeers 15 --networkid 10086 --syncmode "fast" --bootnodes "enode://e71bec68f09c4b9567bd4575d855ea61b179b1d64e6f78c861ebddf3783178f95edaaf39647c1f792bc654d0931ad25415d50c25c437787183c0b0a32a76da85@127.0.0.1:0?discport=30301" --mine --miner.etherbase 0xd192415624a039b24ad571f96cb438de9f0556a7 --miner.threads 1 console

Backtrace

[holiman]

Yup, but this is a duplicate of #23430