-
Notifications
You must be signed in to change notification settings - Fork 20.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
core,eth: call frame tracing #23087
core,eth: call frame tracing #23087
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Generally LGTM, it's in the right direction, but a few nitpicky comments
core/vm/evm.go
Outdated
} | ||
return ret, address, contract.Gas, err | ||
} | ||
|
||
// Create creates a new contract using code as deployment code. | ||
func (evm *EVM) Create(caller ContractRef, code []byte, gas uint64, value *big.Int) (ret []byte, contractAddr common.Address, leftOverGas uint64, err error) { | ||
contractAddr = crypto.CreateAddress(caller.Address(), evm.StateDB.GetNonce(caller.Address())) | ||
return evm.create(caller, &codeAndHash{code: code}, gas, value, contractAddr) | ||
return evm.create(caller, &codeAndHash{code: code}, gas, value, contractAddr, CreateType) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Doesn't this also trigger for a regular create-transaction?
this.callstack[len-1].calls.push(call) | ||
} | ||
}, | ||
// finalize recreates a call object using the final desired field oder for json |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
// finalize recreates a call object using the final desired field oder for json | |
// finalize recreates a call object using the final desired field order for json |
35f0b78
to
416a747
Compare
All tests are passing now! I came across two points:
|
One remaining item (if we want to), rename "call_tracer.js" -> "call_tracer_legacy.js" and "callframe_tracer.js" -> "call_tracer.js". When a user asks to use |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
SGTM, I could not meaningfully review the js and the test cases though
// about internal messages of a transaction. | ||
{ | ||
callstack: [{}], | ||
step: function(log, db) {}, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do you need to declare this function here? Wouldn't this defeat the purpose of not having to call back into js for every step?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oof good catch
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ooof! I need to rerun the benches now!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Turns out the speedup is much more, it's only that I had forgotten to remove
|
|
|
* eth/tracers: implement debug.intermediateRoots (ethereum#23594) 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, rpc: disable memory output by default in traces (ethereum#23558) * 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> * eth/tracers: abort evm execution when trace is aborted (ethereum#23580) Signed-off-by: wenbiao <delweng@gmail.com> * eth/tracers: avoid unsyncronized mutations on trie database (ethereum#23632) 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> * core,eth: call frame tracing (ethereum#23087) 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: re-write of 4byte tracer using enter/exit (ethereum#23622) * eth/tracers: add re-write of 4byte tracer using enter/exit * eth/tracers: fix 4byte indent Signed-off-by: wenbiao <delweng@gmail.com> * eth/tracers: tx.BaseFee not implemented Signed-off-by: wenbiao <delweng@gmail.com> * eth/tracers: do the JSON serialization via .js to capture C faults Signed-off-by: wenbiao <delweng@gmail.com> * eth/tracers: fix callTracer fault handling (ethereum#23667) * eth/tracers: fix calltracer fault handling * eth/tracers: fix calltracer indentation Signed-off-by: wenbiao <delweng@gmail.com> * eth/tracers: invoke enter/exit on 0-value calls to inex accounts (ethereum#23828) Signed-off-by: wenbiao <delweng@gmail.com> * eth: make traceChain avoid OOM on long-running tracing (ethereum#23736) 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> * eth/tracers: expose contextual infos (block hash, tx hash, tx index) Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: redefine Context Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: support for golang tracers + add golang callTracer (ethereum#23708) * eth/tracers: add basic native loader * eth/tracers: add GetResult to tracer interface * eth/tracers: add native call tracer * eth/tracers: fix call tracer json result * eth/tracers: minor fix * eth/tracers: fix * eth/tracers: fix benchTracer * eth/tracers: test native call tracer * eth/tracers: fix * eth/tracers: rm extra make Co-authored-by: Martin Holst Swende <martin@swende.se> * eth/tracers: rm extra make * eth/tracers: make callFrame private * eth/tracers: clean-up and comments * eth/tracers: add license * eth/tracers: rework the model a bit * eth/tracers: move tracecall tests to subpackage * cmd/geth: load native tracers * eth/tracers: minor fix * eth/tracers: impl stop * eth/tracers: add native noop tracer * renamings Co-authored-by: Martin Holst Swende <martin@swende.se> * eth/tracers: more renamings * eth/tracers: make jstracer non-exported, avoid cast * eth/tracers, core/vm: rename vm.Tracer to vm.EVMLogger for clarity * eth/tracers: minor comment fix * eth/tracers/testing: lint nitpicks * core,eth: cancel evm on nativecalltracer stop * Revert "core,eth: cancel evm on nativecalltracer stop" This reverts commit 01bb908. * eth/tracers: linter nits * eth/tracers: fix output on err Co-authored-by: Martin Holst Swende <martin@swende.se> Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: make native calltracer default (ethereum#23867) Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: package restructuring (ethereum#23857) * eth/tracers: restructure tracer package * core/vm/runtime: load js tracers * eth/tracers: mv bigint js code to own file * eth/tracers: add method docs for native tracers * eth/tracers: minor doc fix * core,eth: cancel evm on nativecalltracer stop * core/vm: fix failing test Co-authored-by: Sina Mahmoodi <itz.s1na@gmail.com> Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: ethapi.TransactionArgs was not merged Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: fix the api_test with ErrInsufficientFunds to ErrInsufficientFundsForTransfer Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: check posa before statedb.Prepare in IntermiateRoots api Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: make js calltracer default, compatible with old version Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: fix the default callTrace name of callTracerJs Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * Revert "eth/tracers: fix the default callTrace name of callTracerJs" This reverts commit 62a3bc215d9f07e422a4c659289bb3ba4f9ed2fa. Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * Revert "eth/tracers: make js calltracer default, compatible with old version" This reverts commit 85ef42c0ea651f0b228d4209b1b2598b24e12f1f. Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> * eth/tracers: fix the variable race condition Signed-off-by: wenbiao <wenbiao.zheng@ambergroup.io> Co-authored-by: Martin Holst Swende <martin@swende.se> Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de> Co-authored-by: Sina Mahmoodi <1591639+s1na@users.noreply.github.com> Co-authored-by: Péter Szilágyi <peterke@gmail.com> Co-authored-by: Sina Mahmoodi <itz.s1na@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.
core,eth: call frame tracing (ethereum#23087)
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.
Introduces 2 new optional methods
enter()
andexit()
for js tracers which are invoked on stepping in and out a call frame. These methods are not invoked for depth 0. Currently these are the data fields passed to each of them:If these methods are not present in the JS script geth will default to only invoking
step()
. On the other hand, you have the option of only providingenter()
andexit()
, and not providingstep()
if call frames are the only thing you care about. This should yield a ~5x tracing speedup because it avoids the go-js barrier on every opcode.The PR also comes with a re-write of the
callTracer
. As a backup we keep the previous tracing script under the namecallTracerLegacy
. Behaviour of both tracers are equivalent for the most part, although there are some small differences (improvements):CALL*/CREATE*
fails without having executed any code, i.e. in the initial conditions like whether sender has enough balance, then no call frame will be recorded.gas
andgasUsed
values are set to0
instead ofnil
where they are unavailable. E.g. a simple transfer has agasUsed
of0
. ASELFDESTRUCT
has bothgas
andgasUsed
set to0
.internal failure