Syncing stucked and can't be gracefully shutdown

[bitrocks]

System information

Geth version: 1.10.14-stable
OS & Version: Linux

Expected behaviour

1. sync to the latest height
2. gracefully shutdown when kill with SIGTERM signal

Actual behaviour

1. the node stopped syncing in block height 13842763, (latest: 13885049), also with warn like Failed to decode block body like I reported in keeps showing Failed to decode block body, error=EOF #24158

INFO [12-26|12:33:56.111] Deep froze chain segment                 blocks=554 elapsed=241.616ms number=13,752,279 hash=2e59b3..f96a84
INFO [12-26|12:33:58.431] Imported new chain segment               blocks=73  txs=17086 mgas=1164.054 elapsed=8.387s    mgasps=138.779 number=13,842,299 hash=342930..98e0ef age=5d23h4m   dirty=2.50GiB
INFO [12-26|12:34:06.553] Imported new chain segment               blocks=71  txs=13764 mgas=1099.929 elapsed=8.122s    mgasps=135.415 number=13,842,370 hash=14483f..2641fd age=5d22h49m  dirty=2.50GiB
INFO [12-26|12:34:14.642] Imported new chain segment               blocks=61  txs=11946 mgas=939.153  elapsed=8.089s    mgasps=116.100 number=13,842,431 hash=f9d338..04b378 age=5d22h34m  dirty=2.50GiB
INFO [12-26|12:34:22.731] Imported new chain segment               blocks=63  txs=11259 mgas=945.075  elapsed=8.088s    mgasps=116.838 number=13,842,494 hash=229347..6716fb age=5d22h23m  dirty=2.50GiB
INFO [12-26|12:34:30.758] Imported new chain segment               blocks=64  txs=11538 mgas=1013.797 elapsed=8.027s    mgasps=126.296 number=13,842,558 hash=8b58b7..6b6467 age=5d22h9m   dirty=2.50GiB
INFO [12-26|12:34:38.848] Imported new chain segment               blocks=71  txs=13084 mgas=1109.530 elapsed=8.089s    mgasps=137.161 number=13,842,629 hash=7050b9..e01c99 age=5d21h55m  dirty=2.50GiB
INFO [12-26|12:34:46.864] Imported new chain segment               blocks=68  txs=12332 mgas=1040.302 elapsed=8.016s    mgasps=129.776 number=13,842,697 hash=7bf296..8d08e2 age=5d21h42m  dirty=2.50GiB
INFO [12-26|12:34:54.922] Imported new chain segment               blocks=60  txs=11309 mgas=929.221  elapsed=8.058s    mgasps=115.316 number=13,842,757 hash=f210aa..664ab5 age=5d21h29m  dirty=2.49GiB
INFO [12-26|12:34:55.780] Imported new chain segment               blocks=6   txs=1647  mgas=118.826  elapsed=858.637ms mgasps=138.389 number=13,842,763 hash=608650..5b1793 age=5d21h26m  dirty=2.50GiB
INFO [12-26|12:34:55.787] Downloader queue stats                   receiptTasks=0 blockTasks=32901 itemSize=71.65KiB  throttle=3659
WARN [12-26|12:34:56.108] Failed to decode block body              block=11,305,880 error=EOF
INFO [12-26|12:44:55.200] New local node record                    seq=17107 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=60493 tcp=30303
WARN [12-26|12:45:13.302] Checkpoint challenge timed out, dropping id=e9b288e40fe09256 conn=dyndial addr=212.102.60.119:30303 type=Geth/v1.10.9-stable/...
INFO [12-26|13:06:51.372] New local node record                    seq=17108 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=1158  tcp=30303
INFO [12-26|13:12:04.523] New local node record                    seq=17109 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=60493 tcp=30303
INFO [12-26|13:20:20.199] New local node record                    seq=17110 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=1158  tcp=30303
INFO [12-26|13:23:30.618] Writing clean trie cache to disk         path=/data/chain-data/eth/geth/triecache threads=1
INFO [12-26|13:23:30.618] Regenerated local transaction journal    transactions=0 accounts=0
INFO [12-26|13:23:31.912] Persisted the clean trie cache           path=/data/chain-data/eth/geth/triecache elapsed=1.294s

2. I try to gracefully shut it down with SIGTERM signal, but it shows:

WARN [12-27|03:38:37.882] Already shutting down, interrupt more to panic. times=9
INFO [12-27|03:46:35.320] New local node record                    seq=17209 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=51291 tcp=30303
INFO [12-27|03:52:10.772] New local node record                    seq=17210 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=1158  tcp=30303
INFO [12-27|03:54:48.611] New local node record                    seq=17211 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=51291 tcp=30303
INFO [12-27|04:02:42.190] New local node record                    seq=17212 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=1158  tcp=30303
WARN [12-27|04:02:57.649] Already shutting down, interrupt more to panic. times=8
INFO [12-27|04:05:15.901] New local node record                    seq=17213 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=51291 tcp=30303
INFO [12-27|04:09:01.191] New local node record                    seq=17214 id=9ed7c028b0b3fb36 ip=125.34.62.110 udp=1158  tcp=30303

similar issue #21273 , is it some new kind of deadlock?

Steps to reproduce the behaviour

Backtrace

https://gist.github.com/bitrocks/77db0474aa319839520b3808832574f5

When submitting logs: please submit them as text and not screenshots.

[fjl]

This happens because of double locking of freezer.writeLock in function rawdb.ReadCanonicalBodyRLP.