1263601 - crash in mozilla::ipc::FatalError | mozilla::dom::PContentParent::FatalError | mozilla::dom::PContentParent::Write

Status: RESOLVED WORKSFORME

(Core :: DOM: Content Processes, defect, P3)

Description

[Mihai Boldan, QA [:mboldan]]

This bug was filed from the Socorro interface and is 
report bp-ad8dbd55-09f3-41f3-8038-6d70f2160411.
=============================================================

Steps to reproduce:
1. Launch Firefox.
2. Set devtools.responsive.html.enabled pref to true.
3. Go to any web page (Eg. ebay.com).
4. Enable RDM tool.
5. Take a series of screenshots.
6. Go to the Download folder and delete the screenshots.

Expected results:
The screenshots are deleted.

Actual results:
The screenshots are deleted and the FF crashes.

Notes:
1. More reports: https://crash-stats.mozilla.com/report/list?product=Firefox&signature=mozilla%3A%3Aipc%3A%3AFatalError+%7C+mozilla%3A%3Adom%3A%3APContentParent%3A%3AFatalError+%7C+mozilla%3A%3Adom%3A%3APContentParent%3A%3AWrite#tab-reports
2. I was unable to reproduce this issue on Windows 7 x 64 or on Mac OS X 10.10.4.
3. Also, I tried to crash FF by downloading files from http://www.thinkbroadband.com/download/ without success.

Comment 1

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

I have a hunch this might be fixed by the changes in bug 1263515 (or a related IPC bug).  :mboldan, could you retest with tomorrow's build (20160412) which should hopefully include the changes from bug 1263515?

Comment 2

[Mihai Boldan, QA [:mboldan]]

Attachment: crash RDM.mp4

(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #1)
> I have a hunch this might be fixed by the changes in bug 1263515 (or a
> related IPC bug).  :mboldan, could you retest with tomorrow's build
> (20160412) which should hopefully include the changes from bug 1263515?

The crash is still reproducible on FF 48.0a1 (2016-04-12) and on Ubuntu 14.04 x86.
As you can see in the attached video, the crash is not reproducible 100%.
Let me know if I can provide you any more information.

Comment 3

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

I think that's all we need for now.  Thanks for the video, it makes the steps quite clear!

Comment 4

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

I think the IPC team would have a better chance of making progress here.  Let's try moving this bug over there.

Comment 5

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

To clarify for IPC team, when comment 0 step 4 says "4. Enable RDM tool.", you can do this by going to Tools menu -> Web Developer -> Responsive Design Mode.

Comment 6

[Andrew McCreight [:mccr8]]

This is happening inside a call to mozilla::dom::ContentParent::MaybeInvokeDragSession().

Comment 7

[Andrew McCreight [:mccr8]]

IPCFatalErrorMsg is "NULL actor value passed to non-nullable param", but I'm not sure how that would happen from SendInvokeDragSession.

Comment 8

[Neil Deakin]

The stack suggests that this is a regression from bug 1256952 which has today been backed out. So this should be fixed now. Can someone confirm?

Comment 9

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

:mboldan, can you retest in tomorrow's Nightly?

Comment 10

[Mihai Boldan, QA [:mboldan]]

I've tested this issue on Firefox 49.0a1 (2016-05-06) and on Ubuntu 14.04 x86.
The crash is no longer reproducible.

Comment 11

[Neil Deakin]

Andrew, from your comments above, were you able to reproduce this? I am unable to with the regressing patch. There's a new patch in 1256952 which fixes other regressions but I'm not sure if it fixes this crash.

If it does still crash, would you be able to provide more details as to what is happening?

Comment 12

[Andrew McCreight [:mccr8]]

Sorry, I was just looking at crash reports.