[R4R]fix concurrent write seen of subfetcher

[unclezoro]

Description

Fix the concurrent write of seen of subfetcher.

Rationale

Currently, we do fetcher.abort in abortLoop.
abort will cause subfetcher.loop exit.

The triePrefetcher must wait for the subfetcher exit before it changing the seen of subfetcher.

Example

Fix issue

fatal error: concurrent map writes

goroutine 172 [running]:
runtime.throw({0x1548ace, 0xc02e4e0640})
	runtime/panic.go:1198 +0x71 fp=0xc01631bcd0 sp=0xc01631bca0 pc=0x44c491
runtime.mapassign_faststr(0xc02aa8db60, 0xc01cdfe140, {0xc0334c21e0, 0x20})
	runtime/map_faststr.go:211 +0x39c fp=0xc01631bd38 sp=0xc01631bcd0 pc=0x4297fc
github.com/ethereum/go-ethereum/core/state.(*subfetcher).loop(0xc01a0fcdd0)
	github.com/ethereum/go-ethereum/core/state/trie_prefetcher.go:361 +0x6cf fp=0xc01631bf70 sp=0xc01631bd38 pc=0xac89cf
github.com/ethereum/go-ethereum/core/state.newSubfetcher.func1()
	github.com/ethereum/go-ethereum/core/state/trie_prefetcher.go:261 +0x1d fp=0xc01631bf88 sp=0xc01631bf70 pc=0xac7f7d
github.com/panjf2000/ants/v2.(*goWorker).run.func1()
	github.com/panjf2000/ants/v2@v2.4.5/worker.go:70 +0x97 fp=0xc01631bfe0 sp=0xc01631bf88 pc=0x5f6b37
runtime.goexit()
	runtime/asm_amd64.s:1581 +0x1 fp=0xc01631bfe8 sp=0xc01631bfe0 pc=0x47fcc1
created by github.com/panjf2000/ants/v2.(*goWorker).run
	github.com/panjf2000/ants/v2@v2.4.5/worker.go:48 +0x68

Changes

Notable changes:

• add each change in a bullet point here
• ...

Preflight checks

• build passed (make build)
• tests passed (make test)
• manual transaction test passed

Already reviewed by

...

Related issues

... reference related issue #'s here ...