Closed Bug 1724031 Opened 4 years ago Closed 4 years ago

Consider using reserved slots for JSFunction data

Categories

(Core :: JavaScript Engine, task, P3)

Product:
Core
Component:
JavaScript Engine
Type:
task

Tracking

()

Status:
RESOLVED FIXED
Milestone:
93 Branch
Tracking Flags:
Tracking Status
firefox93 --- fixed

People

(Reporter: jonco, Assigned: jonco)

References

Details

Attachments

(15 files)

Bug 1724031 - Part 0: Make set() method public on GC pointer wrappers r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 1: Make FunctionFlags setters return *this r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 2: Store a functions flags and argument count in a single 32-bit field r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 3: Store JSFunction flags and arg count as a JS::Value r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 4: Store JSFunction's native function / environment object field as a JS::Value r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 5: Store JSFunction's JIT data / script field as a JS::Value r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 6: Store JSFunction's atom field as a JS::Value r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 7: Factor out JIT test for whether an object is a function r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 8: Remove some unused JSFunction methods r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 9: Add a separate JSClass for extended functions r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 10: Use fixed slots to store function data r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 12: Fix GDB scripts to work with new function layout r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 9.5: Move JSFunction::class_ to avoid confusion r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 11: Prevent gecko profiler from observing moved function pointers r?jandem
48 bytes, text/x-phabricator-request
Details | Review
Bug 1724031 - Part 13: Fix race writing recover info r?nbp
48 bytes, text/x-phabricator-request
Details | Review

JSFunction is one of a few JSObject classes that uses its own custom data layout:

https://searchfox.org/mozilla-central/source/js/src/vm/JSFunction.h#43-131

This adds complexity to object allocation and in the GC (and probably other places too). It would be great if this could store its data in reserved slots instead.

Priority: -- → P3

There's no reason for these to be private when they all have operator=
implemented in terms of set() anyway.

Assignee: nobody → jcoppeard
Status: NEW → ASSIGNED

This is useful later on so we can update a function's flags with a single
expression rather than reading, updating and setting the flags separately.

Depends on D123080

In preparation for storing this in a slot, this stores the two values together
in the same field.

Depends on D123081

In preparation for storing them in a slot, store them in a Value. The value is
encoded as a private uint32 value.

Depends on D123082

In preparation for storing this in a slot, store it as a Value.

The native function is stored as a private value. This requires casting it to
void* which AFAIU only conditionally supported in the C++ spec but in practice
supported everwhere. We used to do this in other places, although we have
removed them now.

Depends on D123083

In preparation for storing this in a slot, store it as a JS::Value.

All values including JSScript pointers are stored as private values becuase the
JIT assumes it can access SelfHostedLazyScript pointers and JSScript pointers
in the same way. This means we will still need a trace hook when we store this
data in slots.

Depends on D123084

We will need to change this later as extended functions will have a different
class, so factor out the check.

Depends on D123086

GetClonedSelfHostedFunctionNameOffMainThread doesn't seem to be used any more,
so the JSFunction methods it used can go too.

Depends on D123087

Move class fields to fixed slots. I was originally going to remove the function
alloc kinds but they are used to iterate all functions for relazification.

Another complication is that the JIT assumes that the prototype property for
function constructors is always stored in the first dynamic slot, which
requires extended functions have a special size class with six fixed slots.

Depends on D123089

Still going through this, but do you know what the effect is on for instance "Base Content JS" (the SY(ab) task on Treeherder)?

Flags: needinfo?(jcoppeard)

(In reply to Jan de Mooij [:jandem] from comment #13)

Still going through this, but do you know what the effect is on for instance "Base Content JS" (the SY(ab) task on Treeherder)?

Whoops only 32-bit, of course. That's probably acceptable.

Flags: needinfo?(jcoppeard)

Functions can now have one of two classes so it's best if JSFunction::class_
doesn't exist or people may use that without thinking.

Depends on D123089

(In reply to Jan de Mooij [:jandem] from comment #14)
I ran the awsy tests and the results don't show any conclusive regressions although Base Content JS for 32 bit Windows shows a 3.88% regression at low confidence.

https://treeherder.mozilla.org/perfherder/compare?originalProject=try&originalRevision=4c9189f3600845caf007338bba792308b1a84ff9&newProject=try&newRevision=1a10cc2f41ba86e7b9c16ae849a1466eb1bc25d5&framework=4&page=1&filter=32

Pushed by jcoppeard@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/9f61d500b30c Part 0: Make set() method public on GC pointer wrappers r=jandem https://hg.mozilla.org/integration/autoland/rev/0ed5698bb080 Part 1: Make FunctionFlags setters return *this r=jandem https://hg.mozilla.org/integration/autoland/rev/b384e6621405 Part 2: Store a functions flags and argument count in a single 32-bit field r=jandem https://hg.mozilla.org/integration/autoland/rev/720ceec410a3 Part 3: Store JSFunction flags and arg count as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/672daa49ab88 Part 4: Store JSFunction's native function / environment object field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/01dd672b7e27 Part 5: Store JSFunction's JIT data / script field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/b5229cf51aa4 Part 6: Store JSFunction's atom field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/f420e87d84c8 Part 7: Factor out JIT test for whether an object is a function r=jandem https://hg.mozilla.org/integration/autoland/rev/70db288e80aa Part 8: Remove some unused JSFunction methods r=jandem https://hg.mozilla.org/integration/autoland/rev/a88d9295e473 Part 9: Add a separate JSClass for extended functions r=jandem https://hg.mozilla.org/integration/autoland/rev/a022c9b1a62c Part 9.5: Move JSFunction::class_ to avoid confusion r=jandem https://hg.mozilla.org/integration/autoland/rev/933063e7ecdd Part 10: Use fixed slots to store function data r=jandem https://hg.mozilla.org/integration/autoland/rev/5a3ee7a0dcc4 Part 11: Fix GDB scripts to work with new function layout r=jandem

there is also a spider monkey tsan build bustage: failure log

Changing functions to use fixed slots creates a problem when the profiler
tries to get the script for on-stack functions. Unfortunately this can happen
during minor GC after the functions have been moved but before the profiler
data pointing to them has been updated.

The function data itself is mostly intact, but the object header (used for the
shape) is not valid. As far as I can see this only affects the debug assertions
around checking that slot numbers are valid. However the simplest fix is to
suppress the profiler during the intial part of minor GC, until the roots have
been traced (which will update the profiler data).

The patch moves the code to trace the roots in a minor GC to a separate method
and adds a guard to suppress the profiler.œ

What do you think? Alternatively we would have to add a load of 'unchecked'
methods to a buch of accessors on JSFunction, and to NativeObject to get fixed
slots without touching the shape.

Depends on D123094

Attachment #9237021 - Attachment description: Bug 1724031 - Part 11: Fix GDB scripts to work with new function layout r?jandem → Bug 1724031 - Part 12: Fix GDB scripts to work with new function layout r?jandem
Pushed by jcoppeard@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/0cca6901539d Part 0: Make set() method public on GC pointer wrappers r=jandem https://hg.mozilla.org/integration/autoland/rev/d0fe0b0b2b84 Part 1: Make FunctionFlags setters return *this r=jandem https://hg.mozilla.org/integration/autoland/rev/a7f3f8eb7ac0 Part 2: Store a functions flags and argument count in a single 32-bit field r=jandem https://hg.mozilla.org/integration/autoland/rev/e4d4251e86a3 Part 3: Store JSFunction flags and arg count as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/58dd7d7a733e Part 4: Store JSFunction's native function / environment object field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/1a4a5c11b9f1 Part 5: Store JSFunction's JIT data / script field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/bb4c2384f676 Part 6: Store JSFunction's atom field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/ae06621c5728 Part 7: Factor out JIT test for whether an object is a function r=jandem https://hg.mozilla.org/integration/autoland/rev/3327c1d290b7 Part 8: Remove some unused JSFunction methods r=jandem https://hg.mozilla.org/integration/autoland/rev/75f1f67429e1 Part 9: Add a separate JSClass for extended functions r=jandem https://hg.mozilla.org/integration/autoland/rev/5673e5c4f996 Part 9.5: Move JSFunction::class_ to avoid confusion r=jandem https://hg.mozilla.org/integration/autoland/rev/7991bfb259ae Part 10: Use fixed slots to store function data r=jandem https://hg.mozilla.org/integration/autoland/rev/3a79fb5a7a07 Part 11: Fix GDB scripts to work with new function layout r=jandem

This fixes an intermittent TSAN failure due to accessing JSFunction off main thread.

Pushed by jcoppeard@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/cc120b96f315 Part 0: Make set() method public on GC pointer wrappers r=jandem https://hg.mozilla.org/integration/autoland/rev/3875b61767f4 Part 1: Make FunctionFlags setters return *this r=jandem https://hg.mozilla.org/integration/autoland/rev/632caa336ddc Part 2: Store a functions flags and argument count in a single 32-bit field r=jandem https://hg.mozilla.org/integration/autoland/rev/56cafd086270 Part 3: Store JSFunction flags and arg count as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/60a5118996fa Part 4: Store JSFunction's native function / environment object field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/f8d97ab88f6e Part 5: Store JSFunction's JIT data / script field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/0d15d0423ffe Part 6: Store JSFunction's atom field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/f5fbec033eab Part 7: Factor out JIT test for whether an object is a function r=jandem https://hg.mozilla.org/integration/autoland/rev/f5a9492b8db5 Part 8: Remove some unused JSFunction methods r=jandem https://hg.mozilla.org/integration/autoland/rev/00b6255dfd99 Part 9: Add a separate JSClass for extended functions r=jandem https://hg.mozilla.org/integration/autoland/rev/731989a668e5 Part 9.5: Move JSFunction::class_ to avoid confusion r=jandem https://hg.mozilla.org/integration/autoland/rev/5795d473a641 Part 10: Use fixed slots to store function data r=jandem https://hg.mozilla.org/integration/autoland/rev/b956c169063f Part 11: Prevent gecko profiler from observing moved function pointers r=jandem https://hg.mozilla.org/integration/autoland/rev/2afd98d56716 Part 12: Fix GDB scripts to work with new function layout r=jandem https://hg.mozilla.org/integration/autoland/rev/6e4eb9e09774 Part 13: Fix race writing recover info r=tcampbell

Backed out for assertion failures on JSFunction.h

[task 2021-08-26T16:37:42.403Z] 16:37:42     INFO - GECKO(6400) | Assertion failure: getFixedSlot(AtomSlot).isUndefined(), at /builds/worker/checkouts/gecko/js/src/vm/JSFunction.h:338
[task 2021-08-26T16:37:42.415Z] 16:37:42     INFO -  Initializing stack-fixing for the first stack frame, this may take a while...
[task 2021-08-26T16:37:48.091Z] 16:37:48     INFO - GECKO(6400) | #01: JSFunction::initAtom(JSAtom*) [js/src/vm/JSFunction.h:338]
[task 2021-08-26T16:37:48.092Z] 16:37:48     INFO - GECKO(6400) | #02: js::GlobalObject::getSelfHostedFunction(JSContext*, JS::Handle<js::GlobalObject*>, JS::Handle<js::PropertyName*>, JS::Handle<JSAtom*>, unsigned int, JS::MutableHandle<JS::Value>) [js/src/vm/GlobalObject.cpp:1003]
[task 2021-08-26T16:37:48.092Z] 16:37:48     INFO - GECKO(6400) | #03: JS::NewFunctionFromSpec(JSContext*, JSFunctionSpec const*, JS::Handle<JS::PropertyKey>) [js/src/jsapi.cpp:2220]
[task 2021-08-26T16:37:48.093Z] 16:37:48     INFO - GECKO(6400) | #04: js::DefineFunctions(JSContext*, JS::Handle<JSObject*>, JSFunctionSpec const*, js::DefineAsIntrinsic) [js/src/vm/JSObject.cpp:2535]
[task 2021-08-26T16:37:48.094Z] 16:37:48     INFO - GECKO(6400) | #05: JS_DefineFunctions(JSContext*, JS::Handle<JSObject*>, JSFunctionSpec const*) [js/src/vm/PropertyAndElement.cpp:951]
[task 2021-08-26T16:37:48.095Z] 16:37:48     INFO - GECKO(6400) | #06: js::GlobalObject::resolveConstructor(JSContext*, JS::Handle<js::GlobalObject*>, JSProtoKey, js::GlobalObject::IfClassIsDisabled) [js/src/vm/GlobalObject.cpp:351]
[task 2021-08-26T16:37:48.095Z] 16:37:48     INFO - GECKO(6400) | #07: JS_ResolveStandardClass(JSContext*, JS::Handle<JSObject*>, JS::Handle<JS::PropertyKey>, bool*) [js/src/jsapi.cpp:948]
[task 2021-08-26T16:37:48.096Z] 16:37:48     INFO - GECKO(6400) | #08: js::LookupProperty(JSContext*, JS::Handle<JSObject*>, JS::Handle<JS::PropertyKey>, JS::MutableHandle<JSObject*>, js::PropertyResult*) [js/src/vm/JSObject.cpp:1857]
[task 2021-08-26T16:37:48.096Z] 16:37:48     INFO - GECKO(6400) | #09: js::LookupName(JSContext*, JS::Handle<js::PropertyName*>, JS::Handle<JSObject*>, JS::MutableHandle<JSObject*>, JS::MutableHandle<JSObject*>, js::PropertyResult*) [js/src/vm/JSObject.cpp:1867]
[task 2021-08-26T16:37:48.097Z] 16:37:48     INFO - GECKO(6400) | #10: bool js::GetEnvironmentName<(js::GetNameMode)0>(JSContext*, JS::Handle<JSObject*>, JS::Handle<js::PropertyName*>, JS::MutableHandle<JS::Value>) [js/src/vm/Interpreter-inl.h:182]
[task 2021-08-26T16:37:48.098Z] 16:37:48     INFO - GECKO(6400) | #11: Interpret(JSContext*, js::RunState&) [js/src/vm/Interpreter.cpp:3386]
[task 2021-08-26T16:37:48.098Z] 16:37:48     INFO - GECKO(6400) | #12: js::RunScript(JSContext*, js::RunState&) [js/src/vm/Interpreter.cpp:370]
[task 2021-08-26T16:37:48.099Z] 16:37:48     INFO - GECKO(6400) | #13: js::InternalCallOrConstruct(JSContext*, JS::CallArgs const&, js::MaybeConstruct, js::CallReason) [js/src/vm/Interpreter.cpp:520]
[task 2021-08-26T16:37:48.099Z] 16:37:48     INFO - GECKO(6400) | #14: InternalCall(JSContext*, js::AnyInvokeArgs const&, js::CallReason) [js/src/vm/Interpreter.cpp:548]
[task 2021-08-26T16:37:48.100Z] 16:37:48     INFO - GECKO(6400) | #15: js::Call(JSContext*, JS::Handle<JS::Value>, JS::Handle<JS::Value>, js::AnyInvokeArgs const&, JS::MutableHandle<JS::Value>, js::CallReason) [js/src/vm/Interpreter.cpp:565]
[task 2021-08-26T16:37:48.100Z] 16:37:48     INFO - GECKO(6400) | #16: js::SpreadCallOperation(JSContext*, JS::Handle<JSScript*>, unsigned char*, JS::Handle<JS::Value>, JS::Handle<JS::Value>, JS::Handle<JS::Value>, JS::Handle<JS::Value>, JS::MutableHandle<JS::Value>) [js/src/vm/Interpreter.cpp:4994]
[task 2021-08-26T16:37:48.100Z] 16:37:48     INFO - GECKO(6400) | #17: Interpret(JSContext*, js::RunState&) [js/src/vm/Interpreter.cpp:3192]
[task 2021-08-26T16:37:48.101Z] 16:37:48     INFO - GECKO(6400) | #18: js::RunScript(JSContext*, js::RunState&) [js/src/vm/Interpreter.cpp:370]
[task 2021-08-26T16:37:48.101Z] 16:37:48     INFO - GECKO(6400) | #19: js::InternalCallOrConstruct(JSContext*, JS::CallArgs const&, js::MaybeConstruct, js::CallReason) [js/src/vm/Interpreter.cpp:520]
[task 2021-08-26T16:37:48.101Z] 16:37:48     INFO - GECKO(6400) | #20: InternalCall(JSContext*, js::AnyInvokeArgs const&, js::CallReason) [js/src/vm/Interpreter.cpp:548]
[task 2021-08-26T16:37:48.102Z] 16:37:48     INFO - GECKO(6400) | #21: js::Call(JSContext*, JS::Handle<JS::Value>, JS::Handle<JS::Value>, js::AnyInvokeArgs const&, JS::MutableHandle<JS::Value>, js::CallReason) [js/src/vm/Interpreter.cpp:565]
[task 2021-08-26T16:37:48.102Z] 16:37:48     INFO - GECKO(6400) | #22: JS::Call(JSContext*, JS::Handle<JS::Value>, JS::Handle<JS::Value>, JS::HandleValueArray const&, JS::MutableHandle<JS::Value>) [js/src/vm/CallAndConstruct.cpp:117]
[task 2021-08-26T16:37:48.102Z] 16:37:48     INFO - GECKO(6400) | #23: mozilla::dom::MessageListener::ReceiveMessage(mozilla::dom::BindingCallContext&, JS::Handle<JS::Value>, mozilla::dom::ReceiveMessageArgument const&, JS::MutableHandle<JS::Value>, mozilla::ErrorResult&) [s3:gecko-generated-sources:5e0870a806400d903ec078bd402348078738a9b76a8e02612e63fa33085794714a96a39abab4a7f1def6794939e411a8edd18572a818b613371d45dc6342cc5b/dom/bindings/MessageManagerBinding.cpp::6101]
[task 2021-08-26T16:37:48.103Z] 16:37:48     INFO - GECKO(6400) | #24: mozilla::dom::MessageListener::ReceiveMessage(mozilla::dom::ReceiveMessageArgument const&, JS::MutableHandle<JS::Value>, mozilla::ErrorResult&, char const*, mozilla::dom::CallbackObject::ExceptionHandling, JS::Realm*) [s3:gecko-generated-sources:4b9f933fd91b8763e1a7ae00fb950ff7d40a1ccba0acac7d1976885142e0c018aa3880cfe4c7fa0c2986616f6e7c25b07b9793b6ff5bdeb403f6b3f2c116b16a/dist/include/mozilla/dom/MessageManagerBinding.h::783]
[task 2021-08-26T16:37:48.103Z] 16:37:48     INFO - GECKO(6400) | #25: mozilla::dom::JSActor::CallReceiveMessage(JSContext*, mozilla::dom::JSActorMessageMeta const&, JS::Handle<JS::Value>, JS::MutableHandle<JS::Value>, mozilla::ErrorResult&) [dom/ipc/jsactor/JSActor.cpp:274]
[task 2021-08-26T16:37:48.104Z] 16:37:48     INFO - GECKO(6400) | #26: mozilla::dom::JSActor::ReceiveQuery(JSContext*, mozilla::dom::JSActorMessageMeta const&, JS::Handle<JS::Value>, mozilla::ErrorResult&) [dom/ipc/jsactor/JSActor.cpp:306]
[task 2021-08-26T16:37:48.105Z] 16:37:48     INFO - GECKO(6400) | #27: mozilla::dom::JSActorManager::ReceiveRawMessage(mozilla::dom::JSActorMessageMeta const&, mozilla::Maybe<mozilla::dom::ipc::StructuredCloneData>&&, mozilla::Maybe<mozilla::dom::ipc::StructuredCloneData>&&) [dom/ipc/jsactor/JSActorManager.cpp:198]
[task 2021-08-26T16:37:48.106Z] 16:37:48     INFO - GECKO(6400) | #28: mozilla::detail::RunnableFunction<mozilla::dom::JSActor::SendRawMessageInProcess(mozilla::dom::JSActorMessageMeta const&, mozilla::Maybe<mozilla::dom::ipc::StructuredCloneData>&&, mozilla::Maybe<mozilla::dom::ipc::StructuredCloneData>&&, std::function<already_AddRefed<mozilla::dom::JSActorManager> ()>&&)::$_0>::Run() [xpcom/threads/nsThreadUtils.h:532]
[task 2021-08-26T16:37:48.107Z] 16:37:48     INFO - GECKO(6400) | #29: mozilla::RunnableTask::Run() [xpcom/threads/TaskController.cpp:503]
[task 2021-08-26T16:37:48.107Z] 16:37:48     INFO - GECKO(6400) | #30: mozilla::TaskController::DoExecuteNextTaskOnlyMainThreadInternal(mozilla::detail::BaseAutoLock<mozilla::Mutex&> const&) [xpcom/threads/TaskController.cpp:805]
[task 2021-08-26T16:37:48.108Z] 16:37:48     INFO - GECKO(6400) | #31: mozilla::TaskController::ExecuteNextTaskOnlyMainThreadInternal(mozilla::detail::BaseAutoLock<mozilla::Mutex&> const&) [xpcom/threads/TaskController.cpp:641]
[task 2021-08-26T16:37:48.109Z] 16:37:48     INFO - GECKO(6400) | #32: mozilla::TaskController::ProcessPendingMTTask(bool) [xpcom/threads/TaskController.cpp:425]
[task 2021-08-26T16:37:48.110Z] 16:37:48     INFO - GECKO(6400) | #33: mozilla::detail::RunnableFunction<mozilla::TaskController::InitializeInternal()::$_0>::Run() [xpcom/threads/nsThreadUtils.h:533]
[task 2021-08-26T16:37:48.110Z] 16:37:48     INFO - GECKO(6400) | #34: nsThread::ProcessNextEvent(bool, bool*) [xpcom/threads/nsThread.cpp:1152]
[task 2021-08-26T16:37:48.111Z] 16:37:48     INFO - GECKO(6400) | #35: NS_ProcessNextEvent(nsIThread*, bool) [xpcom/threads/nsThreadUtils.cpp:466]
[task 2021-08-26T16:37:48.112Z] 16:37:48     INFO - GECKO(6400) | #36: mozilla::ipc::MessagePump::Run(base::MessagePump::Delegate*) [ipc/glue/MessagePump.cpp:85]
[task 2021-08-26T16:37:48.112Z] 16:37:48     INFO - GECKO(6400) | #37: MessageLoop::RunInternal() [ipc/chromium/src/base/message_loop.cc:331]
[task 2021-08-26T16:37:48.113Z] 16:37:48     INFO - GECKO(6400) | #38: MessageLoop::Run() [ipc/chromium/src/base/message_loop.cc:307]
[task 2021-08-26T16:37:48.113Z] 16:37:48     INFO - GECKO(6400) | #39: nsBaseAppShell::Run() [widget/nsBaseAppShell.cpp:139]
[task 2021-08-26T16:37:48.114Z] 16:37:48     INFO - GECKO(6400) | #40: nsAppStartup::Run() [toolkit/components/startup/nsAppStartup.cpp:285]
[task 2021-08-26T16:37:48.114Z] 16:37:48     INFO - GECKO(6400) | #41: XREMain::XRE_mainRun() [toolkit/xre/nsAppRunner.cpp:5283]
[task 2021-08-26T16:37:48.115Z] 16:37:48     INFO - GECKO(6400) | #42: XREMain::XRE_main(int, char**, mozilla::BootstrapConfig const&) [toolkit/xre/nsAppRunner.cpp:5475]
[task 2021-08-26T16:37:48.115Z] 16:37:48     INFO - GECKO(6400) | #43: XRE_main(int, char**, mozilla::BootstrapConfig const&) [toolkit/xre/nsAppRunner.cpp:5534]
[task 2021-08-26T16:37:48.116Z] 16:37:48     INFO - GECKO(6400) | #44: ??? [/builds/worker/workspace/build/application/firefox/firefox + 0x41d95]
[task 2021-08-26T16:37:48.116Z] 16:37:48     INFO - GECKO(6400) | #45: __libc_start_main [/lib/x86_64-linux-gnu/libc.so.6 + 0x21b97]
[task 2021-08-26T16:37:48.117Z] 16:37:48     INFO - GECKO(6400) | #46: ??? [/builds/worker/workspace/build/application/firefox/firefox + 0x41979]
[task 2021-08-26T16:37:48.117Z] 16:37:48     INFO - GECKO(6400) | #47: ??? (???:???)
[task 2021-08-26T16:37:48.118Z] 16:37:48     INFO - GECKO(6400) | ExceptionHandler::GenerateDump cloned child 6494
[task 2021-08-26T16:37:48.118Z] 16:37:48     INFO - GECKO(6400) | ExceptionHandler::SendContinueSignalToChild sent continue signal to child
[task 2021-08-26T16:37:48.119Z] 16:37:48     INFO - GECKO(6400) | ExceptionHandler::WaitForContinueSignal waiting for continue signal...
[task 2021-08-26T16:37:48.119Z] 16:37:48     INFO - TEST-INFO | Main app process: exit 11
[task 2021-08-26T16:37:48.120Z] 16:37:48     INFO - Buffered messages finished
Flags: needinfo?(jcoppeard)
Flags: needinfo?(jcoppeard)
Pushed by jcoppeard@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/7ed4db9eadf8 Part 0: Make set() method public on GC pointer wrappers r=jandem https://hg.mozilla.org/integration/autoland/rev/8dc4e71e2a87 Part 1: Make FunctionFlags setters return *this r=jandem https://hg.mozilla.org/integration/autoland/rev/7b1d9df0326d Part 2: Store a functions flags and argument count in a single 32-bit field r=jandem https://hg.mozilla.org/integration/autoland/rev/3fd30df11e67 Part 3: Store JSFunction flags and arg count as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/01975b4520f3 Part 4: Store JSFunction's native function / environment object field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/1415c1d5b485 Part 5: Store JSFunction's JIT data / script field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/a14e1a409342 Part 6: Store JSFunction's atom field as a JS::Value r=jandem https://hg.mozilla.org/integration/autoland/rev/6f6342f5827e Part 7: Factor out JIT test for whether an object is a function r=jandem https://hg.mozilla.org/integration/autoland/rev/f93209b0b5cc Part 8: Remove some unused JSFunction methods r=jandem https://hg.mozilla.org/integration/autoland/rev/4ce1541a6d9c Part 9: Add a separate JSClass for extended functions r=jandem https://hg.mozilla.org/integration/autoland/rev/aae629cbe232 Part 9.5: Move JSFunction::class_ to avoid confusion r=jandem https://hg.mozilla.org/integration/autoland/rev/ec528ee761f8 Part 10: Use fixed slots to store function data r=jandem https://hg.mozilla.org/integration/autoland/rev/3aa2aa420e24 Part 11: Prevent gecko profiler from observing moved function pointers r=jandem https://hg.mozilla.org/integration/autoland/rev/907f6eeb9f3b Part 12: Fix GDB scripts to work with new function layout r=jandem https://hg.mozilla.org/integration/autoland/rev/6f1eb9f44fd8 Part 13: Fix race writing recover info r=tcampbell
Flags: needinfo?(jcoppeard)
Regressions: 1755871
Regressions: 1762582
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: