A mutex around our persistent creation alone didn't solve this problem. Removing v8::Persistent and replacing them with v8::Global is not ideal, since we reuse these Persistent objects across multiple contexts to cache common function/object templates.
On Friday, August 26, 2022 at 9:17:00 AM UTC-7 loude...@gmail.com wrote: > Oh, single-threaded. I would have expected v8 to use a mutex around their > persistent (global handle) management. I'll try a mutex on my side and see > if that resolves the issue. Thanks for your help. 🤞 > > On Fri, Aug 26, 2022 at 12:52 AM dinf...@chromium.org < > dinf...@chromium.org> wrote: > >> Hi, >> >> Those persistent/global handles are single-threaded only, could it be >> that you allocate such handles from multiple threads? >> >> There is also v8::Global which Reset()s itself in the destructor >> automatically. We would even like to remove v8::Persistent entirely at some >> point (see >> https://bugs.chromium.org/p/v8/issues/detail?id=12915&q=v8%3A%3APersistent&can=2 >> ). >> >> Cheers, >> Dominik >> >> On Friday, August 26, 2022 at 12:23:12 AM UTC+2 loude...@gmail.com wrote: >> >>> Yeah, we are out of date and I will attempt to upgrade soon. But, I do >>> think there's a threading issue here that's very intermittent. I did read >>> other posts that describe problems if you don't call Reset on the >>> persistent object. See any problems here? >>> >>> typedef std::pair<DataDeleterP*, utils::PersistentP<v8::Object> > >>> PeristentWeakDataPair; >>> typedef v8::WeakCallbackInfo<PeristentWeakDataPair> >>> PersistentWeakData; >>> >>> static >>> utils::PersistentP<v8::Object> makePersistent(v8::Isolate *isolate, >>> v8::Handle<v8::Object> object, DataDeleterP* dataDeleterPP, >>> PersistentWeakData::Callback callback) { >>> utils::PersistentP<v8::Object> persistentObjP(new >>> PERSISTENT_BASE(v8::Object)(isolate, object)); >>> persistentObjP->SetWeak(new PeristentWeakDataPair(dataDeleterPP, >>> persistentObjP), callback, v8::WeakCallbackType::kParameter); >>> return persistentObjP; >>> } >>> >>> static >>> void releasePersistent (const PersistentWeakData& data) >>> { >>> PeristentWeakDataPair* paramPairP = data.GetParameter(); >>> >>> delete paramPairP->first; >>> paramPairP->second->Reset(); >>> paramPairP->second.reset(); >>> delete paramPairP; >>> >>> #ifdef ADOBE_V8_DEBUG_V8_ALLOCATIONS >>> DecrementV8AllocCount(); >>> #endif >>> } >>> >>> On Thursday, August 25, 2022 at 6:02:02 AM UTC-7 les...@chromium.org >>> wrote: >>> >>>> This is unfortunately too little detail to go on, my best guess would >>>> be that you're passing in an invalid Isolate pointer (something to do with >>>> lifetimes in the embedder?). Also note that you're using a 1-year old >>>> version of V8. >>>> >>>> On Wednesday, August 24, 2022 at 10:42:51 PM UTC+2 loude...@gmail.com >>>> wrote: >>>> >>>>> Anyone?? >>>>> >>>>> On Wednesday, June 15, 2022 at 5:25:20 PM UTC-7 loude...@gmail.com >>>>> wrote: >>>>> >>>>>> Any ideas why I'm seeing an intermittent crash here? It happens >>>>>> during a stress test for Adobe Character Animator (v8 >>>>>> version: 9.4.146.24). This call is made 1000s of times in our app and >>>>>> occasionally there's a crash. >>>>>> >>>>>> Character Animator >>>>>> (Beta).exe!v8::internal::GlobalHandles::Create(class >>>>>> v8::internal::Object) >>>>>> Unknown >>>>>> Character Animator >>>>>> (Beta).exe!v8::internal::GlobalHandles::Create(unsigned __int64) >>>>>> Unknown >>>>>> Character Animator (Beta).exe!v8::V8::GlobalizeReference(class >>>>>> v8::internal::Isolate *,unsigned __int64 *) Unknown >>>>>> > [Inline Frame] Character Animator >>>>>> (Beta).exe!v8::PersistentBase<v8::Object>::New(v8::Isolate *) Line 10971 >>>>>> >>>>>> C++ >>>>>> [Inline Frame] Character Animator >>>>>> (Beta).exe!v8::Persistent<v8::Object,adobe_v8::utils::PersistentTraitsT<v8::Object>>::{ctor}(v8::Isolate >>>>>> >>>>>> *) Line 682 C++ >>>>>> Character Animator >>>>>> (Beta).exe!adobe_v8::makePersistent(v8::Isolate * isolate, >>>>>> v8::Local<v8::Object> object, boost::shared_ptr<void> * dataDeleterPP, >>>>>> void(*)(const v8::WeakCallbackInfo<std::pair<boost::shared_ptr<void> >>>>>> *,adobe_v8::utils::PersistentP<v8::Object>>> &) callback) Line 1508 >>>>>> C++ >>>>>> Character Animator >>>>>> (Beta).exe!adobe_v8::CreatePersistentData(v8::Isolate * inIsolateP, >>>>>> v8::Local<v8::Object> obj, boost::shared_ptr<void> dataDeleterP) Line >>>>>> 1548 >>>>>> C++ >>>>>> Character Animator >>>>>> (Beta).exe!adobe_v8::AttachPersistentDataToInstance(v8::Isolate * >>>>>> inIsolateP, v8::Local<v8::Object> instance, void * dataP, >>>>>> boost::shared_ptr<void> dataDeleterP) Line 1583 C++ >>>>>> >>>>>> The thread 0x8c4 has exited with code 0 (0x0). >>>>>> <31172> <ExportStageVideoTask> <5> Number of export render threads: 4 >>>>>> Exception thrown at 0x000000014105BACA in Character Animator >>>>>> (Beta).exe: 0xC0000005: Access violation reading location >>>>>> 0xFFFFFFFFFFFFFFFF. >>>>>> >>>>> -- >> -- >> v8-users mailing list >> v8-u...@googlegroups.com >> http://groups.google.com/group/v8-users >> --- >> You received this message because you are subscribed to a topic in the >> Google Groups "v8-users" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/v8-users/Pg5OPm7uPFY/unsubscribe. >> To unsubscribe from this group and all its topics, send an email to >> v8-users+u...@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/v8-users/a95e8264-8f5a-4b5c-9e14-410982a61683n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/v8-users/a95e8264-8f5a-4b5c-9e14-410982a61683n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- -- v8-users mailing list v8-users@googlegroups.com http://groups.google.com/group/v8-users --- You received this message because you are subscribed to the Google Groups "v8-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to v8-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/v8-users/f8b001af-e05b-4202-8f43-46ab13fd626en%40googlegroups.com.