refactor trie_prefetcher to be similar to upstream structurally

[darioush]

Why this should be merged

The current trie_prefetcher is modified significantly compared to upstream, which makes it difficult to merge changes from upstream.
This PR aims to preserve the existing behavior but only with minor inline modifications to upstream code.

How this works

Uses a wrapper for DB to track the central worker pool.
Adds a couple hooks for stopping the background processes.

How this was tested

CI, benchmark for trie size ~50m k/vs and adding 100, 200, 500 kvs/ block

Should be added to release notes?

Added

[darioush]

                                                 │ master-50m-2account-run-cc.txt │    alt2-50m-2account-run-cc.txt     │
                                                 │             sec/op             │    sec/op     vs base               │
PrefetcherDatabase/updates_100_prefetchers_0-16                       1.522 ± 23%    1.622 ± 23%       ~ (p=0.481 n=10)
PrefetcherDatabase/updates_100_prefetchers_1-16                       1.237 ±  6%    1.320 ±  8%  +6.64% (p=0.043 n=10)
PrefetcherDatabase/updates_100_prefetchers_4-16                      274.6m ± 38%   290.8m ± 39%       ~ (p=0.684 n=10)
PrefetcherDatabase/updates_100_prefetchers_16-16                     73.52m ±  5%   71.74m ±  8%       ~ (p=0.315 n=10)
PrefetcherDatabase/updates_200_prefetchers_0-16                      682.6m ±  5%   702.1m ±  1%       ~ (p=0.089 n=10)
PrefetcherDatabase/updates_200_prefetchers_1-16                      659.6m ±  3%   680.8m ±  3%  +3.22% (p=0.003 n=10)
PrefetcherDatabase/updates_200_prefetchers_4-16                      208.8m ±  3%   213.9m ±  4%       ~ (p=0.190 n=10)
PrefetcherDatabase/updates_200_prefetchers_16-16                     122.9m ±  5%   116.5m ±  1%  -5.19% (p=0.007 n=10)
PrefetcherDatabase/updates_500_prefetchers_0-16                       1.161 ±  2%    1.185 ±  1%       ~ (p=0.089 n=10)
PrefetcherDatabase/updates_500_prefetchers_1-16                       1.172 ±  2%    1.167 ±  1%       ~ (p=0.853 n=10)
PrefetcherDatabase/updates_500_prefetchers_4-16                      389.2m ±  2%   391.1m ±  2%       ~ (p=0.393 n=10)
PrefetcherDatabase/updates_500_prefetchers_16-16                     271.3m ±  1%   266.7m ±  1%  -1.69% (p=0.000 n=10)
geomean                                                              446.7m         454.0m        +1.64%

[darioush]

                                                 │ master-50m-2account-run-cc.txt │    alt2-50m-2account-run-cc.txt     │
                                                 │           allocs/op            │  allocs/op    vs base               │
PrefetcherDatabase/updates_100_prefetchers_0-16                      1.255M ±  0%   1.254M ±  0%       ~ (p=0.165 n=10)
PrefetcherDatabase/updates_100_prefetchers_1-16                      1.267M ±  0%   1.267M ±  0%       ~ (p=0.579 n=10)
PrefetcherDatabase/updates_100_prefetchers_4-16                      391.1k ± 25%   390.9k ± 25%       ~ (p=0.912 n=10)
PrefetcherDatabase/updates_100_prefetchers_16-16                     180.6k ±  3%   178.0k ±  2%       ~ (p=0.912 n=10)
PrefetcherDatabase/updates_200_prefetchers_0-16                      756.1k ±  0%   755.6k ±  0%       ~ (p=0.436 n=10)
PrefetcherDatabase/updates_200_prefetchers_1-16                      776.3k ±  0%   777.3k ±  0%       ~ (p=0.052 n=10)
PrefetcherDatabase/updates_200_prefetchers_4-16                      422.8k ±  0%   423.1k ±  0%       ~ (p=0.393 n=10)
PrefetcherDatabase/updates_200_prefetchers_16-16                     325.5k ±  3%   317.6k ±  0%       ~ (p=0.289 n=10)
PrefetcherDatabase/updates_500_prefetchers_0-16                      1.582M ±  0%   1.582M ±  0%       ~ (p=0.436 n=10)
PrefetcherDatabase/updates_500_prefetchers_1-16                      1.630M ±  0%   1.633M ±  0%  +0.19% (p=0.019 n=10)
PrefetcherDatabase/updates_500_prefetchers_4-16                      840.1k ±  0%   843.2k ±  0%  +0.37% (p=0.000 n=10)
PrefetcherDatabase/updates_500_prefetchers_16-16                     743.5k ±  0%   744.8k ±  0%  +0.18% (p=0.002 n=10)
geomean                                                              703.2k         701.3k        -0.27%

[darioush]

Let's wait for Etna to activate then we can ship this in the following version

[qdm12]

NIce 👍
I'm just worried about func (p *subfetcherPool) execute(fn func(Trie)) which looks a bit strange

[qdm12]

LGTM but with my limited knowledge especially on geth original and modified files though