This PR enables running the new discv5 protocol in both LES client
and server mode. In client mode it mixes discv5 and dnsdisc iterators
(if both are enabled) and filters incoming ENRs for "les" tag and fork ID.
The old p2p/discv5 package and all references to it are removed.

Co-authored-by: Felix Lange <fjl@twurst.com>

This moves the tracing RPC API implementation to package eth/tracers.
By doing so, package eth no longer depends on tracing and the duktape JS engine.

The change also enables tracing using the light client. All tracing methods work with the
light client, but it's a lot slower compared to using a full node.

This fixes #22131 and adds a test reproducing the issue.

This PR introduces a new config field SyncFromCheckpoint for light client.

In some special scenarios, it's required to start synchronization from some
arbitrary checkpoint or even from the scratch. So this PR offers this
flexibility to users so that the synchronization start point can be configured.

There are two relevant configs: SyncFromCheckpoint and Checkpoint.

- If the SyncFromCheckpoint is true, the light client will try to sync from the
  specified checkpoint.

- If the Checkpoint is not configured, then the light client will sync from the
  scratch(from the latest header if the database is not empty)

Additional notes: these two configs are not visible in the CLI flags but only
accessable in the config file.

Example Usage:

[Eth]
SyncFromCheckpoint = true

[Eth.Checkpoint]
SectionIndex = 100
SectionHead = "0xabc"
CHTRoot = "0xabc"
BloomRoot = "0xabc"

PS. Historical checkpoint can be retrieved from the synced full node or light
client via les_getCheckpoint API.

This PR has two changes in the les protocol:

- the auxRoot is not supported. See ethereum/devp2p#171 for more information
- the empty response will be returned in GetHelperTrieProofsMsg request if the merkle
   proving is failed. note, for backward compatibility, the empty merkle proof as well as
   the request auxiliary data will still be returned in  les2/3 protocol no matter the proving
   is successful or not. the proving failure can happen e.g. request the proving for a
   non-included entry in helper trie (unstable header).

During the snap and eth refactor, the net_version rpc call was falsely deprecated.
This restores the net_version RPC handler as most eth2 nodes and other software
depend on it.

* les: allow tx unindexing in les/4 light server mode

* les: minor fixes

* les: more small fixes

* les: add meaningful constants for recentTxIndex handshake field

This commit splits the eth package, separating the handling of eth and snap protocols. It also includes the capability to run snap sync (https://github.com/ethereum/devp2p/blob/master/caps/snap.md

) , but does not enable it by default. 

Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
Co-authored-by: Martin Holst Swende <martin@swende.se>

This PR implements unclean shutdown marker. Every time geth boots, it adds a timestamp to a list of timestamps in the database. This list is capped at 10. At a clean shutdown, the timestamp is removed again. 
Thus, when geth exits unclean, the marker remains, and at boot up we show the most recent unclean shutdowns to the user, which makes it easier to diagnose root-causes to certain problems. 

Co-authored-by: Nagy Salem <me@muhnagy.com>

The previous fix #21960 converted the float to an intermediate signed int, before attempting the uint conversion. Although this works, this doesn't guarantee that other architectures will work the same.

* les: introduce forkID

* les: address comment

* les, light: remove untrusted header retrieval in ODR

* les: polish

* light: check the hash equality in odr

This commit enables users to specify which signer they want to use while creating their transactOpts.
Previously all contract interactions used the homestead signer. Now a user can specify whether they
want to sign with homestead or EIP155 and specify the chainID which adds another layer of security.

Closes #16484

* les: revert arm float bug workaround to check go 1.15

* add traces to reproduce outside travis

* simpler workaround

Changes:

    Simplify nested complexity
    If an if blocks ends with a return statement then remove the else nesting.

Most of the changes has also been reported in golint https://goreportcard.com/report/github.com/ethereum/go-ethereum#golint

Also fixes a bug in les/flowcontrol that caused the overflow.

* all: core: split vm.Config into BlockConfig and TxConfig

* core: core/vm: reset EVM between tx in block instead of creating new

* core/vm: added docs

This adds a few tiny fixes for les and the p2p simulation framework:

LES Parts

- Keep the LES-SERVER connection even it's non-synced

  We had this idea to reject the connections in LES protocol if the les-server itself is
  not synced. However, in LES protocol we will also receive the connection from another
  les-server. In this case even the local node is not synced yet, we should keep the tcp
  connection for other protocols(e.g. eth protocol).

- Don't count "invalid message" for non-existing GetBlockHeadersMsg request

  In the eth syncing mechanism (full sync, fast sync, light sync), it will try to fetch
  some non-existent blocks or headers(to ensure we indeed download all the missing chain).
  In this case, it's possible that the les-server will receive the request for
  non-existent headers. So don't count it as the "invalid message" for scheduling
  dropping.

- Copy the announce object in the closure

  Before the les-server pushes the latest headers to all connected clients, it will create
  a closure and queue it in the underlying request scheduler. In some scenarios it's
  problematic. E.g, in private networks, the block can be mined very fast. So before the
  first closure is executed, we may already update the latest_announce object. So actually
  the "announce" object we want to send is replaced.

  The downsize is the client will receive two announces with the same td and then drop the
  server.

P2P Simulation Framework

- Don't double register the protocol services in p2p-simulation "Start".

  The protocols upon the devp2p are registered in the "New node stage". So don't reigster
  them again when starting a node in the p2p simulation framework

- Add one more new config field "ExternalSigner", in order to use clef service in the
  framework.

* les: move NodeStateMachine from clientPool to LesServer

* les: new header broadcaster

* les: peerCommons.headInfo always contains last announced head

* les: remove clientPeerSet and serverSet

* les: fixed panic

* les: fixed --nodiscover option

* les: disconnect all peers at ns.Stop()

* les: added comments and fixed signed broadcasts

* les: removed unused parameter, fixed tests

* les, les/lespay/server: refactor client pool

* les: use ns.Operation and sub calls where needed

* les: fixed tests

* les: removed active/inactive logic from peerSet

* les: removed active/inactive peer logic

* les: fixed linter warnings

* les: fixed more linter errors and added missing metrics

* les: addressed comments

* cmd/geth: fixed TestPriorityClient

* les: simplified clientPool state machine

* les/lespay/server: do not use goroutine for balance callbacks

* internal/web3ext: fix addBalance required parameters

* les: removed freeCapacity, always connect at minCapacity initially

* les: only allow capacity change with priority status

Co-authored-by: rjl493456442 <garyrong0905@gmail.com>

This PR adds an extra guarantee to NodeStateMachine: it ensures that all
immediate effects of a certain change are processed before any subsequent
effects of any of the immediate effects on the same node. In the original
version, if a cascaded change caused a subscription callback to be called
multiple times for the same node then these calls might have happened in a
wrong chronological order.

For example:

- a subscription to flag0 changes flag1 and flag2
- a subscription to flag1 changes flag3
- a subscription to flag1, flag2 and flag3 was called in the following order:

   [flag1] -> [flag1, flag3]
   [] -> [flag1]
   [flag1, flag3] -> [flag1, flag2, flag3]

This happened because the tree of changes was traversed in a "depth-first
order". Now it is traversed in a "breadth-first order"; each node has a
FIFO queue for pending callbacks and each triggered subscription callback
is added to the end of the list. The already existing guarantees are
retained; no SetState or SetField returns until the callback queue of the
node is empty again. Just like before, it is the responsibility of the
state machine design to ensure that infinite state loops are not possible.
Multiple changes affecting the same node can still happen simultaneously;
in this case the changes can be interleaved in the FIFO of the node but the
correct order is still guaranteed.

A new unit test is also added to verify callback order in the above scenario.

* les: close all connected les-server when shutdown

* les: linter nitpick

Co-authored-by: Martin Holst Swende <martin@swende.se>

This PR significantly changes the APIs for instantiating Ethereum nodes in
a Go program. The new APIs are not backwards-compatible, but we feel that
this is made up for by the much simpler way of registering services on
node.Node. You can find more information and rationale in the design
document: https://gist.github.com/renaynay/5bec2de19fde66f4d04c535fd24f0775.

There is also a new feature in Node's Go API: it is now possible to
register arbitrary handlers on the user-facing HTTP server. In geth, this
facility is used to enable GraphQL.

There is a single minor change relevant for geth users in this PR: The
GraphQL API is no longer available separately from the JSON-RPC HTTP
server. If you want GraphQL, you need to enable it using the
./geth --http --graphql flag combination.

The --graphql.port and --graphql.addr flags are no longer available.

* cmd, consensus, eth, les: implement light fetcher

* les: address comment

* les: address comment

* les: address comments

* les: check td after delivery

* les: add linearExpiredValue for error counter

* les: fix import

* les: fix dead lock

* les: order announces by td

* les: encapsulate invalid counter

* les: address comment

* les: add more checks during the delivery

* les: fix log

* eth, les: fix lint

* eth/fetcher: address comment

This change introduces garbage collection for the light client. Historical
chain data is deleted periodically. If you want to disable the GC, use
the --light.nopruning flag.

* les/checkpointoracle: don't lookup checkpoint more than once per second

* les/checkpoint/oracle: change oracle checktime to 1 minute

* rpc, internal/ethapi: default rpc gascap at 50M + better error message

* eth,internal: make globalgascap uint64

* core/tests: fix compilation failure

* eth/config: gascap at 25M + minor review concerns

* cmd, eth, internal, les: add gasprice cap

* cmd/utils, eth: add default value for gasprice cap

* all: use txfee cap

* cmd, eth: add fix

* cmd, internal: address comments

This PR reimplements the light client server pool. It is also a first step
to move certain logic into a new lespay package. This package will contain
the implementation of the lespay token sale functions, the token buying and
selling logic and other components related to peer selection/prioritization
and service quality evaluation. Over the long term this package will be
reusable for incentivizing future protocols.

Since the LES peer logic is now based on enode.Iterator, it can now use
DNS-based fallback discovery to find servers.

This document describes the function of the new components:
https://gist.github.com/zsfelfoldi/3c7ace895234b7b345ab4f71dab102d4