Closed Bug 509560 Opened 15 years ago Closed 14 years ago

Breakpad doesn't send crash report when crash is caused by Flash

Categories

(Toolkit :: Crash Reporting, defect)

1.9.0 Branch
x86
Windows XP
defect
Not set
major

Tracking

()

RESOLVED FIXED

People

(Reporter: whimboo, Unassigned)

References

Details

(Whiteboard: [crashkill][crashkill-metrics][fp 10.1.53.64])

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.3pre) Gecko/20090808 Shiretoko/3.5.3pre (.NET CLR 3.5.30729) ID:20090808042506

When a crash of Firefox is caused by the Flash plugin Breakpad cannot send the report. We saw this problem while investigating bug 501304. When you are using crashme to check how Breakpad behaves for an user initiated crash the report will be sent. I use Flash version 10.0.32.18.

Ted, is this related to the Flash plugin or is it definitely our fault?
Can you explain what you see more clearly? Does the Mozilla Crash Reporter appear?
Sure. The crash reporter comes up but after hitting send and waiting 1-2 seconds the failure is shown that the crash report cannot be sent. Looking into the CrashReporter folder doesn't list any dmp file for those crashes. Similar or dependend on bug 503645?
Sounds completely unrelated. In fact, I can't see how this would be related to Flash at all. Once the crash reporter app has launched, the browser is gone, so how could Flash have any effect?
I can submit any other crash report initiated by your crashme extension. Only running the crash test on bug 501304 which is a Flash related crash causes this result. Ted, under which error conditions this mentioned error could appear? I really wonder that I don't find a dmp file in the pending or submitted folder for those crashes.
Ok, then it sounds like however this crash is happening, it's producing an empty dump file. There isn't anything we can do about this on Windows, as we use a Windows system library to generate the minidump. If the process is severely compromised, it just sometimes can't execute enough code to generate the dump. This might get better when we get to a multi-process setup, because Breakpad has support for out-of-process crash reporting, where a separate process would do the actual dump writing.
Can we reproduce this someplace else?  I'd really like to know if this is a real.
Whiteboard: [crashkill][crashkill-metrics]
Damon, you know of a current crashing bug where Flash is involved? The bug I refer to in my initial comment doesn't produce a crash anymore for me.
Trying to get more eyes on this.  Asked for help on QMO:
  http://quality.mozilla.org/forums/bugs-bugzilla/need-help-reproing-bug-509560-firefox-36
And on dev quality: 
  http://groups.google.com/group/mozilla.dev.quality/browse_thread/thread/e0eb97b627e91e98
Aakash is contacting other community members to help.

--Tim
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → DUPLICATE
Charles, please see comment 2 and comment 3. Bug 503645 is different from that one.
Status: RESOLVED → REOPENED
Resolution: DUPLICATE → ---
Okay Thanks missed that.  Sounds different but the title is very similar.  I am reading here that Flash is likely not the cause.  If the is true might consider changing the title of this bug.
A crash with Flash can be found on bug 502648. I will have to test with Windows XP again later.
(In reply to comment #11)
> Okay Thanks missed that.  Sounds different but the title is very similar.  I am
> reading here that Flash is likely not the cause.  If the is true might consider
> changing the title of this bug.

By my testing in comment 0 on one of the hard to reproduce hangs/crashes (bug 501304) i have seen that a couple of times. I wasn't able to send the report. Using crashme sending the report was no problem. So I would say it somehow depends on it or the problem was the OOM crash which happened at that time. Sadly the problem on bug 501304 is not reproducible for me anymore.
Status: REOPENED → NEW
I tried to find a page with flash content which causes a high cpu load and a crash but I'm not able to. Lucas, I know that you see a high memory consumption with the testcase on bug 501304 but does Firefox crash for you?

Charles, I don't have any experience with Flash. So I wonder if it could be possible to create a flash file which will cause such an OOM crash. I'm not sure if that can help but it would be worth a try.
This is fixed in the latest Flash Player 10.1
Status: NEW → RESOLVED
Closed: 15 years ago14 years ago
Resolution: --- → FIXED
Whiteboard: [crashkill][crashkill-metrics] → [crashkill][crashkill-metrics][fp 10.1.53.64]
You need to log in before you can comment on or make changes to this bug.