[R4R]Feature/backport geth native trace #581

jsvisa · 2021-11-23T15:46:38Z

Description

The Geth native trace has a
There is a significant performance improvement with Geth's native trace, release in Release Vallhallan Threshold (v1.10.12) · ethereum/go-ethereum, and the specific implementation code is in the two PR:

eth/tracers: support for golang tracers + add golang callTracer by s1na · Pull Request #23708 · ethereum/go-ethereum
eth/tracers: make native call tracer default by s1na · Pull Request #23867 · ethereum/go-ethereum
core,eth: call frame tracing by s1na · Pull Request #23087 · ethereum/go-ethereum
eth/tracers: package restructuring by holiman · Pull Request #23857 · ethereum/go-ethereum

In our BSC Archive node tests, we found that the CPU usage was reduced from 40% to 10%(20cores reduced to 2 cores)

Rationale

Follow in Geth's footsteps and speed up block synchronization

Example

add an example CLI or API response...

Changes

Notable changes:

Switch the call tracer to a fast native Go implementation
Fixed some data errors with trace in the original JS version

unclezoro · 2021-11-24T03:08:16Z

@jsvisa Thanks a lot. I will start looking into this PR.

jsvisa · 2021-11-24T03:41:01Z

@guagualvcha thanks for your reply, I will fix the Unit test failure later today

jsvisa · 2021-11-24T08:03:18Z

fixed the golang-lint, @guagualvcha PTAL

unclezoro · 2021-12-01T09:24:12Z

@yutianwu @kyrie-yl Please help review this PR.

This PR implements a new debug method, which I've talked briefly about to some other client developers. It allows the caller to obtain the intermediate state roots for a block (which might be either a canon block or a 'bad' block). Signed-off-by: wenbiao <delweng@gmail.com>

* core: cmd: invert disableMemory * core: fix missed inversion * cmd/evm: preserve Flags but change default value * Apply suggestions from code review Co-authored-by: Martin Holst Swende <martin@swende.se> Co-authored-by: Martin Holst Swende <martin@swende.se> Signed-off-by: wenbiao <delweng@gmail.com>

Signed-off-by: wenbiao <delweng@gmail.com>

This PR fixes an issue in traceChain, where the statedb Commit operation was performed asynchronously with dereference-operations agains the underlying trie.Database instance. Due to how the reference counting works within the trie database (where parent count is recursively updated when new parents are added), doing dereferencing in the middle of Commit can cause the refcount to become wrong, leading to an inconsistent state. This was fixed by doing Commit/Deref from the same routine. Signed-off-by: wenbiao <delweng@gmail.com>

This change introduces 2 new optional methods; `enter()` and `exit()` for js tracers, and makes `step()` optiona. The two new methods are invoked when entering and exiting a call frame (but not invoked for the outermost scope, which has it's own methods). Currently these are the data fields passed to each of them: enter: type (opcode), from, to, input, gas, value exit: output, gasUsed, error The PR also comes with a re-write of the callTracer. As a backup we keep the previous tracing script under the name `callTracerLegacy`. Behaviour of both tracers are equivalent for the most part, although there are some small differences (improvements), where the new tracer is more correct / has more information. Signed-off-by: wenbiao <delweng@gmail.com>

* eth/tracers: add re-write of 4byte tracer using enter/exit * eth/tracers: fix 4byte indent Signed-off-by: wenbiao <delweng@gmail.com>

Signed-off-by: wenbiao <delweng@gmail.com>

* eth/tracers: fix calltracer fault handling * eth/tracers: fix calltracer indentation Signed-off-by: wenbiao <delweng@gmail.com>

…828) Signed-off-by: wenbiao <delweng@gmail.com>

This PR changes long-running chain tracing, so that it at some points releases the memory trie db, and switch over to a fresh disk-backed trie. Signed-off-by: wenbiao <delweng@gmail.com>