Possible memory leak inside emscripten_sleep

[MizantropLoL]

Anytime emscripten_sleep runs, it slightly increases usage of memory by the app (due to windows task manages). I thought it can be asynchronously calls, but memory usage kept increasing even if I remove functionality and run only

while(1)
emscripten_sleep(0));

Then I thought, It might be because of the allocator, which is deallocating memory but not letting it get to the system. Then I saw message for heap reallocation happening, I set breakpoint and at the time it was shooting, in the callstack I saw Asyncify.allocateData() inside handleSleep.
Asan detected no errors.
Would be great to hear some ideas of how to solve this issue

[kripken]

Is this application using memory growth? If not, then the Windows task manager would not be aware of allocation inside the wasm memory (it's all a single big array from the OS point of view).

If it is using growth, and the array grows, that's worth looking into. The malloc inside allocateData has a corresponding free in handleSleep. Confirming that is called with some console logging might help.

[MizantropLoL]

ALLOW_MEMORY_GROWTH is set to 0, but memory usage of corresponding Google Chrome process is still growing.

Google Chrome Canary v94.0.4589.2

[kripken]

Ok, that is unrelated to malloc/free in emscripten then.

You can use the browser's memory profiling tools to look for memory leaks on the JS side. Note that memory may grow even without a leak, and then decrease when a GC happens.

If it's not that, then it might be that the browser is using more memory over time due to caching or memory fragmentation or such. In those cases, there is nothing you can do, but also nothing to worry about probably.

[stale]

This issue has been automatically marked as stale because there has been no activity in the past year. It will be closed automatically if no further activity occurs in the next 30 days. Feel free to re-open at any time if this issue is still relevant.