[WIP] Restructure module for thread safety using context aware initialization #15
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…lization Previously, this module would cause an error if loaded in two separate contexts, e.g. the main thread and a worker thread. We can use the NODE_MODULE_INIT macro to establish our module as being context aware. Next, we need to remove our reliance upon static shared data, especially any v8 objects. Accessing a v8 object across isolates causes a cryptic runtime error. In accordance with the v8 native extensions docs, we create a separate class to store mutable data per thread, and shift all shared v8 objects into this structure (ProfilerData).
Is there a reproduce? |
const {Worker, isMainThread, parentPort} = require('worker_threads');
const profiler = require('v8-profiler-node8');
const workerCode = `
const profiler = require('v8-profiler-node8');
const {parentPort} = require('worker_threads');
profiler.startProfiling('worker');
function work() {
for (let i = 0; i < 1e7; i++) {}
}
work();
const profile = profiler.stopProfiling('worker');
parentPort.postMessage(JSON.stringify(profile));
`;
profiler.startProfiling('main');
const worker = new Worker(workerCode, {eval: true});
worker.on('message', (m) => {
const profile = profiler.stopProfiling('main');
console.log('main thread profile', profile);
console.log('worker thread profile', m);
});
worker.on('error', (e) => {
throw e;
});Using Node v12.18.2, results in this error: With this PR applied, it successfully generates traces on both threads. |
Thanks, I'll take a look and get back to you :) |
|
Heads up in my testing it looks like this PR can cause segfaults right now, still some work to do before it's stable. Will look into that tomorrow. |
|
This is because I think |
|
Adding in a segfault handler made it clear that the existing segfault remains in the currently untouched heap profiler, and disabling it fixes the segfault. I'm going to have a go at making the same sort of changes to the heap profiler. |
Feels much cleaner to have explicitly typed arguments.
hyj1991
approved these changes
Jul 3, 2020
| using v8::Isolate; | ||
|
|
||
| ProfilerData::ProfilerData(v8::Local<Context> newContext, Isolate* isolate) { | ||
| context = newContext; |
|
This feature landed in 7330e00 with the npm i v8-profiler-node8@next --save |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously, this module would cause an error if loaded in two separate
contexts, e.g. the main thread and a worker thread.
We can use the NODE_MODULE_INIT macro to establish our module as being
context aware.
Next, we need to remove our reliance upon static shared data, especially
any v8 objects. Accessing a v8 object across isolates causes a cryptic
runtime error.
In accordance with the Node native extensions docs, we create a separate
class to store mutable data per thread, and shift all shared v8 objects
into this structure (ProfilerData).
TODO:
ifdefs are still correctI'm not a C++ dev by trade so apologies if I've made any super obvious mistakes. I'm happy to make changes based on feedback if I've taken a wrong turn somewhere.