1160754 - Firefox crash [@ XPCWrappedNative::FlatJSObjectFinalized() ]

Status: RESOLVED INCOMPLETE

(Core :: JavaScript Engine, defect)

Description

[Igor Manturov Jr.]

Attachment: crash.txt

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:37.0) Gecko/20100101 Firefox/37.0
Build ID: 20150415140819



Actual results:

0.0.0 Linux 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt9-3~deb8u1 (2015-04-24) x86_64
XPCWrappedNative::FlatJSObjectFinalized()

Comment 1

[Nickolay_Ponomarev]

Please post the crash ID from about:crashes next time - the JSON is hard to read.

The crash looks the same as the one in bug 608217. Can you reproduce? Do you have any additional information?

Where did "XPCWrappedNativeScope::XPCWrappedNativeScope(JSContext*, JS::Handle<JSObject*>)" in summary come from?

Comment 2

[Igor Manturov Jr.]

I can provide the following info https://crash-stats.mozilla.com/report/index/38116a61-b5b6-4678-acdb-a268f2150501

Comment 3

[Igor Manturov Jr.]

. This crash began to happen after updating to Firefox 37.0.2. A browser closes unexpectedly,it might happen in anytime, even if you do nothing.

Comment 4

[Nickolay_Ponomarev]

Thanks for the link!

>A browser closes unexpectedly,it might happen in anytime, even if you do nothing.
That observation matches the crash stack: the crash happens during GC - background bookkeeping process.

Does this happen regularly? I can only see two recent Linux reports with this signature, one of which is yours - did you submit the rest of reports?

Comment 5

[Igor Manturov Jr.]

I had at least 20 or more crashes beore I decided to submit the report.

Comment 6

[Nickolay_Ponomarev]

Igor, thanks! Could you list a few other crash reports you have to see what do they have in common.

Pinging dmajor, who looked into this (bug 835798 comment 8). Anything else we want the reporter to do?

Comment 7

[(Away)]

I don't have any expertise with Linux minidumps. Redirecting to the JS team.

Comment 8

[Terrence Cole [:terrence]]

The crash appears to be from a broken C++ pointer in an XPConnect native wrapper. Unless we can get STR, I don't know of any way to debug this.

Bobby, the bustage is showing up in XPConnect: do you have any insight that might make this report actionable without STR?

Comment 9

[Bobby Holley (:bholley)]

(In reply to Terrence Cole [:terrence] from comment #8)
> Bobby, the bustage is showing up in XPConnect: do you have any insight that
> might make this report actionable without STR?

The only thing I notice is that there's an anonymous extension installed. Igor, can you try disabling that extension in about:addons and seeing if the crash continues?

Comment 10

[Igor Manturov Jr.]

Bobby, if you mean Ember Inspector, I was trying to disable this extension, but it does't have any effect.

Comment 11

[Wayne Mery (:wsmwk)]

Igor, can you still reproduce when using a current version?
And if so, do you have steps to reproduce?

Comment 12

[Wayne Mery (:wsmwk)]

(In reply to Wayne Mery (:wsmwk, NI for questions) from comment #11)
> Igor, can you still reproduce when using a current version?
> And if so, do you have steps to reproduce?

Comment 13

[Phoenix]

Resolved per whiteboard