Closed
Bug 1275448
Opened 8 years ago
Closed 8 years ago
Intermittent e10s PROCESS-CRASH | Main app process exited normally | application crashed [@ JS::Zone::~Zone()] after Assertion failure: typeDescrObjects.empty(), at /builds/slave/m-cen-m64-d-000000000000000000/build/src/js/src/gc/Zone.cpp:61
Categories
(Core :: JavaScript: GC, defect, P3)
Core
JavaScript: GC
Tracking
()
RESOLVED
FIXED
mozilla50
People
(Reporter: philor, Assigned: jonco)
References
(Blocks 1 open bug)
Details
(Keywords: assertion, intermittent-failure)
Attachments
(1 file)
4.16 KB,
patch
|
sfink
:
review+
|
Details | Diff | Splinter Review |
https://treeherder.mozilla.org/logviewer.html#?job_id=3937992&repo=mozilla-central https://treeherder.mozilla.org/logviewer.html#?job_id=9498026&repo=fx-team
Updated•8 years ago
|
Blocks: e10s-tests
tracking-e10s:
--- → +
Comment hidden (Intermittent Failures Robot) |
Assignee | ||
Comment 2•8 years ago
|
||
From the logs we're leaking a whole bunch of stuff at shutdown (see logs) which is also causing this assert to go off.
Comment hidden (Intermittent Failures Robot) |
Comment hidden (Intermittent Failures Robot) |
Assignee | ||
Comment 6•8 years ago
|
||
This doesn't fix the underlying problem which is that we sometimes leak, it just means that we don't crash due to an assertion failure in that case. The assertion still runs if we don't leak which is most of the time.
Assignee: nobody → jcoppeard
Attachment #8773307 -
Flags: review?(sphink)
Comment 7•8 years ago
|
||
Comment on attachment 8773307 [details] [diff] [review] bug1275448-dont-crash-on-leak Review of attachment 8773307 [details] [diff] [review]: ----------------------------------------------------------------- ::: js/src/jsgc.cpp @@ +3599,5 @@ > > // We are about to delete the Zone; this will leave the Zone* > // in the arena header dangling if there are any arenas > // remaining at this point. > + mozilla::DebugOnly<bool> arenasEmpty = zone->arenas.checkEmptyArenaLists(); Can you name this allArenasEmpty?
Attachment #8773307 -
Flags: review?(sphink) → review+
Assignee | ||
Comment 8•8 years ago
|
||
(In reply to Steve Fink [:sfink] [:s:] from comment #7) > Can you name this allArenasEmpty? I don't want to do this because it's not just a predicate but prints information about any cells that are still live (in debug builds).
Pushed by jcoppeard@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/c9abfc5925a2 Don't crash on shutdown due to non-empty type descriptors set if we leak r=sfink
Comment 10•8 years ago
|
||
(In reply to Jon Coppeard (:jonco) from comment #8) > (In reply to Steve Fink [:sfink] [:s:] from comment #7) > > Can you name this allArenasEmpty? > > I don't want to do this because it's not just a predicate but prints > information about any cells that are still live (in debug builds). I just meant the DebugOnly<bool> local variable, not the method name, which I'm fine with. But it's fine.
Comment hidden (Intermittent Failures Robot) |
Assignee | ||
Comment 12•8 years ago
|
||
(In reply to Steve Fink [:sfink] [:s:] from comment #10) Ah, sorry, I misunderstood your comment.
Comment 13•8 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/c9abfc5925a2
Status: NEW → RESOLVED
Closed: 8 years ago
status-firefox50:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla50
You need to log in
before you can comment on or make changes to this bug.
Description
•