GeezerVisonThe World Wide Web, restored to 800 pixels of glory. | ||||||||||||
| Home|About|Gaming cabinet|Archives|How it works|Starter sites | ||||||||||||
GeezerVison is showing github.com — vm: fix leak in vm.compileFunction when importModuleDynamically is used by joyeecheung · Pull Request #46785 · nodejs/node Conversation
|
|
Draft for now because I need to do some tests to see if the lifetime management is still safe..but it does fix #42080 locally for me. |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
20a2568 to
f77f70e
Compare
There was an error while loading. Please reload this page.
|
I did some analysis in #44211 (comment) - so I think this should fix the leak in const preBakedModule = new vm.SyntheticModule(); vm.compileFunction(code, [], { importModuleDynamically: () => { return preBakedModule; }, }); Then the small leakage of that single vm.compileFunction(code, [], { importModuleDynamically: (specifier) => { return new vm.SyntheticModule(...); }, }); Then they are going to leak much more. To fix the modules properly we'd need something like https://chromium-review.googlesource.com/c/v8/v8/+/3172764 so that's probably going to take much longer. But anyway this PR at least fixes the leak of |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Previously in the implementation there was a cycle that V8 could not
detect:
Strong global reference to CompiledFnEntry (JS wrapper)
-> strong reference to callback setting (through the
callbackMap key-value pair)
-> importModuleDynamically (wrapper in internalCompileFunction())
-> Strong reference to the compiled function (through closure in
internalCompileFunction())
The CompiledFnEntry only gets GC'ed when the compiled function is GC'ed.
Since the compiled function is always reachable as described above,
there is a leak.
We only needed the first strong global reference because we didn't want
the function to outlive the CompiledFnEntry. In this case it can be
solved by using a private symbol instead of going with the global
reference + destruction in the weak callback, which V8's GC is not
going to understand.
Sorry, something went wrong.
There was an error while loading. Please reload this page.
f77f70e to
219c771
Compare
Sorry, something went wrong.
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
The reason will be displayed to describe this comment to others. Learn more.
Great analysis, thanks!
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Commit Queue failed- Loading data for nodejs/node/pull/46785 ✔ Done loading data for nodejs/node/pull/46785 ----------------------------------- PR info ------------------------------------ Title vm: fix leak in vm.compileFunction when importModuleDynamically is used (#46785) Author Joyee Cheung (@joyeecheung) Branch joyeecheung:fix-compile-function-leak -> nodejs:main Labels c++, lib / src, needs-ci Commits 1 - vm: fix leak in vm.compileFunction when importModuleDynamically is used Committers 1 - Joyee Cheung PR-URL: https://github.com/nodejs/node/pull/46785 Fixes: https://github.com/nodejs/node/issues/42080 Reviewed-By: Anna Henningsen Reviewed-By: James M Snell Reviewed-By: Chengzhong Wu ------------------------------ Generated metadata ------------------------------ PR-URL: https://github.com/nodejs/node/pull/46785 Fixes: https://github.com/nodejs/node/issues/42080 Reviewed-By: Anna Henningsen Reviewed-By: James M Snell Reviewed-By: Chengzhong Wu -------------------------------------------------------------------------------- ℹ This PR was created on Wed, 22 Feb 2023 19:47:29 GMT ✔ Approvals: 3 ✔ - Anna Henningsen (@addaleax): https://github.com/nodejs/node/pull/46785#pullrequestreview-1310473405 ✔ - James M Snell (@jasnell) (TSC): https://github.com/nodejs/node/pull/46785#pullrequestreview-1312336429 ✔ - Chengzhong Wu (@legendecas) (TSC): https://github.com/nodejs/node/pull/46785#pullrequestreview-1312758385 ✖ Last GitHub CI failed ℹ Last Full PR CI on 2023-02-24T08:36:57Z: https://ci.nodejs.org/job/node-test-pull-request/49973/ - Querying data for job/node-test-pull-request/49973/ ✔ Last Jenkins CI successful -------------------------------------------------------------------------------- ✔ Aborted `git node land` session in /home/runner/work/node/node/.ncuhttps://github.com/nodejs/node/actions/runs/4283087915 |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
|
Landed in 986498b |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Previously in the implementation there was a cycle that V8 could not
detect:
Strong global reference to CompiledFnEntry (JS wrapper)
-> strong reference to callback setting (through the
callbackMap key-value pair)
-> importModuleDynamically (wrapper in internalCompileFunction())
-> Strong reference to the compiled function (through closure in
internalCompileFunction())
The CompiledFnEntry only gets GC'ed when the compiled function is GC'ed.
Since the compiled function is always reachable as described above,
there is a leak.
We only needed the first strong global reference because we didn't want
the function to outlive the CompiledFnEntry. In this case it can be
solved by using a private symbol instead of going with the global
reference + destruction in the weak callback, which V8's GC is not
going to understand.
PR-URL: #46785
Fixes: #42080
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Previously in the implementation there was a cycle that V8 could not
detect:
Strong global reference to CompiledFnEntry (JS wrapper)
-> strong reference to callback setting (through the
callbackMap key-value pair)
-> importModuleDynamically (wrapper in internalCompileFunction())
-> Strong reference to the compiled function (through closure in
internalCompileFunction())
The CompiledFnEntry only gets GC'ed when the compiled function is GC'ed.
Since the compiled function is always reachable as described above,
there is a leak.
We only needed the first strong global reference because we didn't want
the function to outlive the CompiledFnEntry. In this case it can be
solved by using a private symbol instead of going with the global
reference + destruction in the weak callback, which V8's GC is not
going to understand.
PR-URL: #46785
Fixes: #42080
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Previously in the implementation there was a cycle that V8 could not
detect:
Strong global reference to CompiledFnEntry (JS wrapper)
-> strong reference to callback setting (through the
callbackMap key-value pair)
-> importModuleDynamically (wrapper in internalCompileFunction())
-> Strong reference to the compiled function (through closure in
internalCompileFunction())
The CompiledFnEntry only gets GC'ed when the compiled function is GC'ed.
Since the compiled function is always reachable as described above,
there is a leak.
We only needed the first strong global reference because we didn't want
the function to outlive the CompiledFnEntry. In this case it can be
solved by using a private symbol instead of going with the global
reference + destruction in the weak callback, which V8's GC is not
going to understand.
PR-URL: #46785
Fixes: #42080
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: James M Snell <[email protected]>
Reviewed-By: Chengzhong Wu <[email protected]>
Reviewed-By: Benjamin Gruenbaum <[email protected]>
Sorry, something went wrong.
There was an error while loading. Please reload this page.
|
Since this is reverted on 19.8.1 is there a follow up issue I can track for a follow up diff that isn't crashy? |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
|
You can track the original issue: #42080. |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
There was an error while loading. Please reload this page.
|
I have another fix that passes the test here locally and also locally no longer crashes the reproduction in #47096: #48510 |
Sorry, something went wrong.
There was an error while loading. Please reload this page.
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it.
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it.
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it.
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in #46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. PR-URL: #48510 Refs: #44211 Refs: #42080 Refs: #47096 Refs: #43205 Refs: #38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. PR-URL: nodejs#48510 Refs: nodejs#44211 Refs: nodejs#42080 Refs: nodejs#47096 Refs: nodejs#43205 Refs: nodejs#38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in #46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. Backport-PR-URL: #49874 PR-URL: #48510 Refs: #44211 Refs: #42080 Refs: #47096 Refs: #43205 Refs: #38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. PR-URL: nodejs#48510 Refs: nodejs#44211 Refs: nodejs#42080 Refs: nodejs#47096 Refs: nodejs#43205 Refs: nodejs#38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in nodejs#46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. PR-URL: nodejs#48510 Refs: nodejs#44211 Refs: nodejs#42080 Refs: nodejs#47096 Refs: nodejs#43205 Refs: nodejs#38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
Previously when managing the importModuleDynamically callback of vm.compileFunction(), we use an ID number as the host defined option and maintain a per-Environment ID -> CompiledFnEntry map to retain the top-level referrer function returned by vm.compileFunction() in order to pass it back to the callback, but it would leak because with how we used v8::Persistent to maintain this reference, V8 would not be able to understand the cycle and would just think that the CompiledFnEntry was supposed to live forever. We made an attempt to make that reference known to V8 by making the CompiledFnEntry weak and using a private symbol to make CompiledFnEntry strongly references the top-level referrer function in #46785, but that turned out to be unsound, because the there's no guarantee that the top-level function must be alive while import() can still be initiated from that function, since V8 could discard the top-level function and only keep inner functions alive, so relying on the top-level function to keep the CompiledFnEntry alive could result in use-after-free which caused a revert of that fix. With this patch we use a symbol in the host defined options instead of a number, because with the stage-3 symbol-as-weakmap-keys proposal we could directly use that symbol to keep the referrer alive using a WeakMap. As a bonus this also keeps the other kinds of referrers alive as long as import() can still be initiated from that Script/Module, so this also fixes the long-standing crash caused by vm.Script being GC'ed too early when its importModuleDynamically callback still needs it. PR-URL: #48510 Backport-PR-URL: #51004 Refs: #44211 Refs: #42080 Refs: #47096 Refs: #43205 Refs: #38695 Reviewed-By: Antoine du Hamel <[email protected]> Reviewed-By: Benjamin Gruenbaum <[email protected]> Reviewed-By: Stephen Belanger <[email protected]>
There was an error while loading. Please reload this page.
Successfully merging this pull request may close these issues.
Memory leak invm.compileFunction when using importModuleDynamically
There was an error while loading. Please reload this page.