Closed Bug 600150 Opened 14 years ago Closed 10 years ago

Crash [@ mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip&) ]

Categories

(Core :: Layout, defect)

x86
Windows XP
defect
Not set
critical

Tracking

()

RESOLVED WORKSFORME
Tracking Status
blocking2.0 --- ?

People

(Reporter: scoobidiver, Unassigned)

References

Details

(Keywords: crash, regression)

Crash Data

Build : Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0b7pre) Gecko/20100927
Firefox/4.0b7pre

This is a new crash signature that was introduced by b6pre/20100911 build.
It happens mainly under Windows XP.
The crash daily rate is 1-4 crashes/day in trunk builds.
It is #23 top crasher for b7pre/20100927 build.

Signature	mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip const&)
UUID	5b8761de-e67a-4e4a-9f50-459a82100925
Time 	2010-09-25 20:56:43.809952
Uptime	419
Last Crash	427 seconds (7.1 minutes) before submission
Install Age	18618 seconds (5.2 hours) since version was first installed.
Product	Firefox
Version	4.0b7pre
Build ID	20100925041313
Branch	2.0
OS	Windows NT
OS Version	5.1.2600 Service Pack 2
CPU	x86
CPU Info	GenuineIntel family 15 model 2 stepping 7
Crash Reason	EXCEPTION_ACCESS_VIOLATION_READ
Crash Address	0x0
App Notes 	AdapterVendorID: 5333, AdapterDeviceID: 8d04

Frame 	Module 	Signature [Expand] 	Source
0 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1143
1 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
2 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
3 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
4 	xul.dll 	mozilla::FrameLayerBuilder::BuildContainerLayerFor 	layout/base/FrameLayerBuilder.cpp:1410
5 	xul.dll 	nsDisplayOwnLayer::BuildLayer 	layout/base/nsDisplayList.cpp:1319
6 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1097
7 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
8 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
9 	xul.dll 	mozilla::`anonymous namespace'::ContainerState::ProcessDisplayItems 	layout/base/FrameLayerBuilder.cpp:1068
10 	xul.dll 	mozilla::FrameLayerBuilder::BuildContainerLayerFor 	layout/base/FrameLayerBuilder.cpp:1410
11 	xul.dll 	nsDisplayList::PaintForFrame 	layout/base/nsDisplayList.cpp:409
12 	xul.dll 	nsLayoutUtils::PaintFrame 	layout/base/nsLayoutUtils.cpp:1429
13 	xul.dll 	PresShell::Paint 	layout/base/nsPresShell.cpp:6114
14 	xul.dll 	nsViewManager::RenderViews 	view/src/nsViewManager.cpp:447
15 	xul.dll 	nsViewManager::Refresh 	view/src/nsViewManager.cpp:413
16 	xul.dll 	nsViewManager::DispatchEvent 	view/src/nsViewManager.cpp:913
17 	xul.dll 	AttachedHandleEvent 	view/src/nsView.cpp:193
18 	xul.dll 	nsWindow::DispatchEvent 	widget/src/windows/nsWindow.cpp:3536
19 	xul.dll 	nsWindow::DispatchWindowEvent 	widget/src/windows/nsWindow.cpp:3567
20 	xul.dll 	nsWindow::OnPaint 	widget/src/windows/nsWindowGfx.cpp:603
21 	xul.dll 	xul.dll@0x52334b 	

The regression range is :
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=79d0beec27b5&tochange=73ab2c3c5ad9

More reports at :
http://crash-stats.mozilla.com/report/list?product=Firefox&query_search=signature&query_type=exact&query=&date=09%2F27%2F2010%2023%3A58%3A31&range_value=4&range_unit=weeks&hang_type=any&process_type=any&plugin_field=&plugin_query_type=&plugin_query=&do_query=1&admin=&signature=mozilla%3A%3A%60anonymous%20namespace%27%27%3A%3AContainerState%3A%3AProcessDisplayItems%28nsDisplayList%20const%26%2C%20mozilla%3A%3AFrameLayerBuilder%3A%3AClip%20const%26%29
blocking2.0: --- → ?
Given the low volume of crashes, how confident are we in that regression range?
> Given the low volume of crashes, how confident are we in that regression range?
b6pre/20100910 : 0 crash
b6pre/20100911 : 3 crashes (first appearance)
From b6pre/20100913 to b6pre/20100915 : 0 crash
The regression range can be extended to 3 days before :
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=cf0088ff07f2&tochange=73ab2c3c5ad9
Nothing pops out in the extended regression range.

A lot of those crash reports have nsWindow::DealWithPopups in the stack before the painting. I wonder if we're painting a doomed popup or something.
The URLs associated with this crash show significant Brazilian bias; there's a possibility it's related to the Orkut talk gadget.
Based on the current crash volume and the lack of steps to reproduce, I don't think we can block on this right now.

However, I'd want to reconsider if either:
 (1) we had steps to reproduce, or,
 (2) crash volume in betas showed it was in the top 50 or so crashes
blocking2.0: ? → -
Keywords: qawanted
It is #43 top crasher on 4.0b8pre for the last 3 days.
Blocks: 633903
No longer blocks: 633903
Summary: crash mainly under Windows XP [@ mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip const&) ] → Crash [@ mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip&) ]
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:5.0a2) Gecko/20110413 Firefox/5.0a2

https://crash-stats.mozilla.com/report/index/bp-0392f265-645e-4fa8-b2ef-cb4002110414

http://cubicvr.org/CubicVR.js/bd3/BeatDetektor3HD.html

I used the time slider -> crash

reproduceable = always
(In reply to comment #6)
> However, I'd want to reconsider if either:
>  (1) we had steps to reproduce, or,
>  (2) crash volume in betas showed it was in the top 50 or so crashes
Its dupe is #5 top crasher in 4.0.1 over the last 3 days.
blocking2.0: - → ?
here is the volume profile for various releases over the last several days.
         
the bump on may 12 looks due to a site problem or malware, and not connected to any firefox release.


mozilla::.anonymous.namespace..::ContainerState::ProcessDisplayItems.nsDisplayList.const.,.mozilla::FrameLayerBuilder::Clip..
date     total    breakdown by build
         crashes  count build, count build, ...

20110501 114  	75 4.0.12011041322, 
        		32 4.02011031805, 	5 4.0b112011020314, 
        		2 4.0b122011022221, 
20110502 113  	78 4.0.12011041322, 
        		26 4.02011031805, 	3 4.0b112011020314, 
        		2 5.0a22011042704, 	2 4.0b122011022221, 
        		1 6.0a12011050203, 	1 6.0a12011050103, 
20110503 101  	80 4.0.12011041322, 
        		12 4.02011031805, 	5 4.0b122011022221, 
        		3 4.0b112011020314, 	1 5.0a22011050204, 
20110504 111  	77 4.0.12011041322, 
        		20 4.02011031805, 	6 4.0b112011020314, 
        		5 4.0b122011022221, 	2 6.0a12011042803, 
        		1 4.02011030319, 

Advertised/prompted MU FF3.5.19 → 4.0.1, FF3.6.17 → 4.0.1 	May 05 

20110505 133  	112 4.0.12011041322, 
        		10 4.02011031805, 	6 4.0b112011020314, 
        		3 4.0b122011022221, 	1 6.0a12011050103, 
        		1 5.0a22011050404, 
20110506 140  	108 4.0.12011041322, 
        		16 4.02011031805, 	8 4.0b112011020314, 
        		5 4.0b122011022221, 	1 6.0a12011050503, 
        		1 5.0a22011050504, 	1 5.02011042714, 
20110507 121  	98 4.0.12011041322, 
        		8 4.0b112011020314, 	7 4.02011031805, 
        		5 4.0b122011022221, 	2 5.02011042714, 
        		1 4.02011030319, 
20110508 163  	138 4.0.12011041322, 
        		13 4.02011031805, 	6 4.0b122011022221, 
        		4 4.0b112011020314, 	2 5.02011042714, 
20110509 168  	137 4.0.12011041322, 
        		17 4.02011031805, 	9 4.0b112011020314, 
        		2 5.0a22011050804, 	1 6.0a12011050903, 
        		1 6.0a12011050403, 	1 4.0b122011022221, 
20110510 175  	145 4.0.12011041322, 
        		16 4.02011031805, 	6 5.02011042714, 
        		5 4.0b112011020314, 	1 6.0a12011050903, 
        		1 4.0b122011022221, 	1 4.02011030319, 
20110511 154  	133 4.0.12011041322, 
        		9 4.02011031805, 	5 4.0b122011022221, 
        		3 5.02011042714, 	3 4.0b112011020314, 
        		1 5.0a22011051004, 

20110512 208  	187 4.0.12011041322, 
        		13 4.02011031805, 	3 5.02011042714, 
        		2 4.0b122011022221, 	1 6.0a12011051203, 
        		1 4.0b112011020314, 	1 4.02011030319, 
20110513 183  	149 4.0.12011041322, 
        		22 4.02011031805, 	9 5.02011042714, 
        		3 4.0b122011022221, 
20110514 189  	162 4.0.12011041322, 
        		12 4.02011031805, 	8 4.0b112011020314, 
        		5 5.02011042714, 	1 6.0a12011051303, 
        		1 5.0a22011051304, 
20110515 217  	198 4.0.12011041322, 
        		6 4.02011031805, 	5 4.0b112011020314, 
        		4 5.02011042714, 	2 6.0a12011051503, 
        		1 5.0a22011051404, 	1 4.0b122011022221, 
20110516 208  	177 4.0.12011041322, 
        		11 4.02011031805, 	6 4.0b122011022221, 
        		5 5.02011042714, 	5 4.0b112011020314, 
        		1 6.0a12011051503, 	1 6.0a12011051003, 
        		1 5.0a22011051204, 	1 4.02011030319,
If there's malware tickling this, can we get someone to dig into that and see about a blocklist solution? Alternatively, if we can't get anywhere through that rout, who can look at this code and see if there's a client-side fix we can employ?

Are the steps in comment 9 sufficient to try to get this diagnosed on the code side?
marcia/tomcat can you check out comment 9 on a few different vms?  I was unable to reproduce on mac.

crash volume has leveled off about the same as ending period of comment 11.

I looked at which addons are installed in about 300 reports and the list looks something like this:

 243 {972ce4c6-7e08-4474-a285-3208198ce6fd}</td      4.0.1</td
 212 NO_ADDONS_FOUND
 104 <a href="" </a     1.0</td
  71 <a href="" </a     3.3.3.2</td
  64{CAFEEFAC-0016-0000-0024-ABCDEFFEDCBA}</td      6.0.24</td
  51 {CAFEEFAC-0016-0000-0022-ABCDEFFEDCBA}</td      6.0.22</td
  45 {CAFEEFAC-0016-0000-0023-ABCDEFFEDCBA}</td      6.0.23</td
  38 {CAFEEFAC-0016-0000-0021-ABCDEFFEDCBA}</td      6.0.21</td
  31 <a href="http://addons.mozilla.org/en-US/firefox/addon/13661" Test Pilot   1.1.1
  31 {CAFEEFAC-0016-0000-0020-ABCDEFFEDCBA}</td      6.0.20</td
  26 <a href="http://addons.mozilla.org/en-US/firefox/addon/9449" Microsoft .NET Framework Assistant</a <a href="http://addons.mozilla.org/en-US/firefox/addon/9449"   0.0.0</td
  21 {82AF8DCA-6DE9-405D-BD5E-43525BDAD38A}</td      5.3.0.7280</td
  21 <a href="" </a
  20 <a href="http://addons.mozilla.org/en-US/firefox/addon/1865" Adblock Plus</a       <a href="http://addons.mozilla.org/en-US/firefox/addon/1865"    {d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d}</td      1.3.7</td
Tried reproducing with the latest trunk using Mozilla/5.0 (Windows NT 6.1; rv:6.0a1) Gecko/20110518 Firefox/6.0a1 - so far no luck with the steps in Comment 9. Note that user is running with Add on Compat reporter and a few extensions. Will try on a few other machines.
Not able to reproduce in a Windows XP VM either - I even tried installing all the extensions the user has and still can't get it to reproduce when I move the slider.

Tried with both latest trunk nightly and Firefox 5 B2 build.
The problem is with the mozalloc_abort(char const* const) | mozalloc_handle_oom() | mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip&) crash signature (see bug 633903):
https://crash-stats.mozilla.com/report/list?signature=mozalloc_abort%28char%20const*%20const%29%20|%20mozalloc_handle_oom%28%29%20|%20mozilla%3A%3A%60anonymous%20namespace%27%27%3A%3AContainerState%3A%3AProcessDisplayItems%28nsDisplayList%20const%26%2C%20mozilla%3A%3AFrameLayerBuilder%3A%3AClip%26%29

This crasher is #4 top crasher in 4.0.1 and it exists in 5.0b2 at a lower volume.
Keywords: topcrash
Not much we can do about this unless we have clear steps to reproduce. About 76 windows crashes in the last week on b3. If we can reproduce it, I am removing it from tracking since it's unlikely to get any traction in the next week.
I meant, if we CAN'T repro, removing it from tracking.
I will take a look at 4.0.1 to see if there are any correlations or comments that may help us get some STR.
Crash Signature: [@ mozilla::`anonymous namespace''::ContainerState::ProcessDisplayItems(nsDisplayList const&, mozilla::FrameLayerBuilder::Clip&) ]
This has gone way down in the ranking for releases. It's at #75 on 8.0, #103 on 9.0b4. Removing the top crash keyword.
Keywords: topcrash
You need to log in before you can comment on or make changes to this bug.