Closed Bug 1032769 Opened 5 years ago Closed 5 years ago

Firefox UI flashes black, occasionally produces blank/black window

Categories

(Core :: Graphics, defect)

33 Branch
x86
Windows 7
defect
Not set

Tracking

()

RESOLVED DUPLICATE of bug 1083071

People

(Reporter: j, Unassigned, NeedInfo)

Details

Attachments

(3 files)

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:32.0) Gecko/20100101 Firefox/32.0 (Beta/Release)
Build ID: 20140520030202

Steps to reproduce:

Updated from 32 branch to 33 branch, and open Firefox.

(have tried with both an existing profile and creating a new profile)

Issue started with x86_64 Windows builds at the start of the 33 branch (using trunk build).
Started about 30 June 2014 with update on x86 32 bit version.

Using 64-bit Windows 7.


Actual results:

Open Firefox; sometimes the window UI components (status bar, toolbar elements, address bar text) will flicker with black fill or disappear.

Moving the window around the screen, behind other windows ("wiping" it behind the taskbar), or resizing the window all have no effect on UI.


Expected results:

Firefox window with UI elements should have opened.
Hi Jason,

thanks for taking time to report a problem.

Could you please try if the same happens in Safe Mode (Help → Restart with Addons Disabled)? That way we can know if the problem originates from an extension.
Also, could you please try if it happens turning off the hardware acceleration? (see here how to do it: https://support.mozilla.org/en-US/kb/troubleshoot-extensions-themes-to-fix-problems#w_turn-off-hardware-acceleration).

Thanks,
Francesca
QA Whiteboard: [bugday-20140707]
Flags: needinfo?(j)
Hi Francesca,

I tried each mode (safe mode, hardware acceleration disabled) independently and each (independently) seems to fix the issue for both the 20140630 and the 20140707 x86_64 33.0a1 Nightly builds.

For further information I then tried disabling each of my enabled extensions with hardware acceleration enabled on the 20140707 x86_64 build.  I tried disabling each individually as well as disabling all of them.  Neither of these had any effect and the issue was still present.

It seems as if the issue may be following the hardware acceleration.


Thanks,
Jason


PS: Here is the default state of the extensions & plugins on my normal profile, along with my graphics configuration:

Extensions:
----------------------------------------------------------------------------------------------------
Name 					Version 	Enabled 	ID
Adblock Plus				2.6.3		true		{d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d}
Element Hiding Helper for Adblock Plus	1.3		true		elemhidehelper@adblockplus.org
Shumway					0.8.380		true		shumway@research.mozilla.org
User Agent Switcher			0.7.3		true		{e968fc70-8f95-4ab9-9e79-304de2a71ee1}
Readability				2.4		false		readability@readability.com
Session Manager				0.8.1.3		false		{1280606b-2510-4fe0-97ef-9b5a22eafe30}


Plugins:
----------------------------------------------------------------------------------------------------
Shockwave Flash

    File: NPSWF64_14_0_0_125.dll
    Path: C:\Windows\system32\Macromed\Flash\NPSWF64_14_0_0_125.dll
    Version: 14.0.0.125
    State: Enabled
    Shockwave Flash 14.0 r0

Silverlight Plug-In

    File: npctrl.dll
    Path: c:\Program Files\Microsoft Silverlight\5.1.30214.0\npctrl.dll
    Version: 5.1.30214.0
    State: Enabled
    5.1.30214.0
    
Java(TM) Platform SE 7 U9

    File: npjp2.dll
    Path: C:\Program Files\Java\jre7\bin\plugin2\npjp2.dll
    Version: 10.9.2.5
    State: Disabled (STATE_VULNERABLE_UPDATE_AVAILABLE)
    Next Generation Java Plug-in 10.9.2 for Mozilla browsers
    
Java Deployment Toolkit 7.0.90.5

    File: npDeployJava1.dll
    Path: C:\Windows\system32\npDeployJava1.dll
    Version: 10.9.2.5
    State: Disabled (STATE_VULNERABLE_NO_UPDATE)
    NPRuntime Script Plug-in Library for Java(TM) Deploy



Graphics:
----------------------------------------------------------------------------------------------------
Adapter Description		NVIDIA Quadro 1000M
Adapter Description (GPU #2)	Intel(R) HD Graphics Family
Adapter Drivers			nvd3dumx,nvwgf2umx,nvwgf2umx nvd3dum,nvwgf2um,nvwgf2um
Adapter Drivers (GPU #2)	igdumd64 igd10umd64 igd10umd64 igdumdx32 igd10umd32 igd10umd32
Adapter RAM			2048
Adapter RAM (GPU #2)		Unknown
Device ID			0x0dfa
Device ID (GPU #2)		0x0126
Direct2D Enabled		true
DirectWrite Enabled		true (6.2.9200.16492)
Driver Date			5-25-2011
Driver Date (GPU #2)		3-6-2011
Driver Version			8.17.12.6871
Driver Version (GPU #2)		8.15.10.2321
GPU #2 Active			false
GPU Accelerated Windows		1/1 Direct3D 11 (OMTC)
Vendor ID			0x10de
Vendor ID (GPU #2)		0x8086
WebGL Renderer			Google Inc. -- ANGLE (NVIDIA Quadro 1000M Direct3D9Ex vs_3_0 ps_3_0)
windowLayerManagerRemote	true
AzureCanvasBackend		direct2d
AzureContentBackend		direct2d
AzureFallbackCanvasBackend	cairo
AzureSkiaAccelerated		0
Flags: needinfo?(j)
Hi Jason,

thanks for the additional info and testing.

Seems similar to Bug 1027145, with the exception that AFAICT in your case it doesn't happen *everytime* you start Firefox. Is that true? 
Or do you have always the problem when you start Firefox with HW acceleration enabled?

It's a shoot in the dark (pun intended), but the fact that makes me think the problem could be the same is that you have the very same Intel driver version at https://bugzilla.mozilla.org/show_bug.cgi?id=1027145#c8.
Now, if I read correctly your about:support it seems that firefox is using the NVIDIA card and not the Intel one, but it can be worth trying to upgrade the Intel driver and check then if the problem still occurs with HW Acceleration enabled.
Could you please try that? Take a look here for the drivers: https://downloadcenter.intel.com/Default.aspx?lang=eng

In case this doesn't work, I would try also upgrading the NVIDIA card drivers, just to be on the safe side :).

Another interesting thing to try would be to check what exactly triggered this regression, if it's a regression: you can find a regression range using Mozregression: http://mozilla.github.io/mozregression/. That would be probably even more useful if it turns out that this bug is not a duplicate of 1027145.

For now, I'm moving this bug to Core::Graphics, so that the relevant team can look at it.


Cheers,
Francesca
Component: Untriaged → Graphics
Flags: needinfo?(j)
Product: Firefox → Core
Hi,

After installing the very latest 33 update, I *always* experience this issue every time I start Firefox.
Running Firefox in safe mode solves the problem.

I'm not sure if addons are the cause, because the issue still occurs even if all addons are disabled.
I see this issue on a Windows 7 (32 bit) Sony VAIO Laptop with Intel graphic chips.

The computer is localized to Japanese, so the about:support info is Japanese.

Note that hardware acceleration is set to OFF in tools->options->advanced.

Using about:config I also tried disabling layers acceleration.

The symptom of the problem is that Firefox's UI will suddenly turn black in all of its windows and then stay that way until the program is terminated.

Here is the about:support text dump:


アプリケーション基本情報
------------

名前: Firefox
バージョン: 32.0.3
ユーザエージェント: Mozilla/5.0 (Windows NT 6.1; rv:32.0) Gecko/20100101 Firefox/32.0

過去 3 日分のクラッシュレポート
-----------------

レポート ID: bp-43f1fa23-f5ca-4b9f-8af9-a74a82141011
送信日時: 1 日前

すべてのクラッシュレポート (10 件の未送信のクラッシュを含む)

拡張機能
----

名前: Adblock Plus
バージョン: 2.6.4
有効: true
ID: {d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d}

名前: Brief
バージョン: 1.7.3
有効: true
ID: brief@mozdev.org

名前: Moji
バージョン: 1.0.3
有効: true
ID: {ea9be299-129b-4c3c-8876-d98c18c2fd39}

名前: Moji-En
バージョン: 1.1.20110109
有効: true
ID: {C4A808D2-254E-4039-832A-C75B72FBA2DA}

名前: Moji-JpNam
バージョン: 1.1.20110109
有効: true
ID: {C631EB1C-4A6D-490e-A226-0D6FAD02C01C}

名前: Moji-Kanji-En
バージョン: 0.6.20071013
有効: true
ID: {9E1A1CD4-8916-4951-AAB4-2D4497FDFD90}

名前: NoScript
バージョン: 2.6.9
有効: true
ID: {73a6fe31-595d-460b-a920-fcc0f8843232}

名前: Restart
バージョン: 1.2.1
有効: true
ID: Restart@schuzak.jp

名前: Skype Click to Call
バージョン: 7.3.16540.9015
有効: true
ID: {82AF8DCA-6DE9-405D-BD5E-43525BDAD38A}

グラフィックス
-------

Direct2D 有効: true
DirectWrite 有効: true (6.2.9200.16571)
GPU #2 の使用: false
GPU 描画支援のウィンドウ: 1/1 Direct3D 10
WebGL レンダラ: Google Inc. -- ANGLE (Mobile Intel(R) 4 Series Express Chipset Family Direct3D9Ex vs_3_0 ps_3_0)
windowLayerManagerRemote: false
アダプタ RAM: Unknown
アダプタドライバ: igdumdx32 igd10umd32
アダプタ名: Mobile Intel(R) 4 Series Express Chipset Family
デバイス ID: 0x2a42
ドライバのバージョン: 8.15.10.2869
ドライバの日付: 10-4-2012
ベンダ ID: 0x8086
AzureCanvasBackend: direct2d
AzureContentBackend: direct2d
AzureFallbackCanvasBackend: cairo
AzureSkiaAccelerated: 0

変更された重要な設定
----------

accessibility.typeaheadfind.casesensitive: 1
accessibility.typeaheadfind.flashBar: 0
browser.cache.disk.capacity: 358400
browser.cache.disk.smart_size_cached_value: 358400
browser.cache.disk.smart_size.first_run: false
browser.cache.disk.smart_size.use_old_max: false
browser.cache.frecency_experiment: 1
browser.display.background_color: #C0C0C0
browser.display.use_system_colors: true
browser.places.smartBookmarksVersion: 7
browser.sessionstore.upgradeBackup.latestBuildID: 20140923175406
browser.startup.homepage: http://jp.msn.com/?pc=UP97&ocid=UP97DHP|https://mail.kylheku.com/
browser.startup.homepage_override.buildID: 20140923175406
browser.startup.homepage_override.mstone: 32.0.3
dom.max_chrome_script_run_time: 40
dom.max_script_run_time: 40
dom.mozApps.used: true
dom.w3c_touch_events.expose: false
extensions.lastAppVersion: 32.0.3
gfx.direct2d.disabled: true
gfx.direct3d.last_used_feature_level_idx: 1
keyword.URL: http://www.bing.com/search?FORM=UP97DF&PC=UP97&q=
layers.acceleration.disabled: true
network.cookie.prefsMigrated: true
places.database.lastMaintenance: 1413079480
places.history.expiration.transient_current_max_pages: 50202
plugin.disable_full_page_plugin_for_types: application/pdf
plugin.importedState: true
plugin.state.java: 0
plugin.state.npbtdna: 0
plugin.state.npdeployjava: 0
plugin.state.npwpf: 0
privacy.sanitize.migrateFx3Prefs: true
security.enable_ssl2: true
security.warn_viewing_mixed: false
storage.vacuum.last.brief.sqlite: 1411276041
storage.vacuum.last.index: 2
storage.vacuum.last.places.sqlite: 1411495459

JavaScript
----------

インクリメンタル GC: true

アクセシビリティ
--------

有効: false
アクセシビリティを無効にする: 0

ライブラリのバージョン
-----------

NSPR
想定される最低バージョン: 4.10.6
使用中のバージョン: 4.10.6

NSS
想定される最低バージョン: 3.16.5 Basic ECC
使用中のバージョン: 3.16.5 Basic ECC

NSSSMIME
想定される最低バージョン: 3.16.5 Basic ECC
使用中のバージョン: 3.16.5 Basic ECC

NSSSSL
想定される最低バージョン: 3.16.5 Basic ECC
使用中のバージョン: 3.16.5 Basic ECC

NSSUTIL
想定される最低バージョン: 3.16.5
使用中のバージョン: 3.16.5

実験的な機能
------
(In reply to kaz from comment #5)
> アダプタ名: Mobile Intel(R) 4 Series Express Chipset Family
> デバイス ID: 0x2a42
> ドライバのバージョン: 8.15.10.2869
> ドライバの日付: 10-4-2012
> ベンダ ID: 0x8086

This driver is currently the latest. I updated that machine to it after seeing this problem; it didn't help, and neither did turning off acceleration.
I also have some sort of  black flashing. There's some blackness when I resize the window, and when I type into the locationbar and the results change.

Disabling layers.offmainthreadcomposition.enabled only seems to fix the urlbar thing, the other still happens, so I guess it's not just OMTC. I could reproduce both with a new FF33 profile too after I copied the places.* files. Tried updating my graphics driver from 337.88 to 344.11 but didn't change anything.
It also happens with latest trunk builds too with my nightly profile.

Here's the graphics section:

Adapter illesztőprogramjai	nvd3dumx,nvwgf2umx,nvwgf2umx nvd3dum,nvwgf2um,nvwgf2um
Adapter leírása	NVIDIA GeForce GTS 450
Adapter memóriája	1024
Direct2D engedélyezve	true
DirectWrite engedélyezve	true (6.3.9600.16384)
Eszközazonosító	0x1245
GPU #2 aktív	false
GPU segítségével gyorsított ablakok	1/1 Direct3D 11 (OMTC)
Gyártóazonosító	0x10de
Illesztőprogram dátuma	9-13-2014
Illesztőprogram verziója	9.18.13.4411
WebGL-megjelenítő	Google Inc. -- ANGLE (NVIDIA GeForce GTS 450 Direct3D9Ex vs_3_0 ps_3_0)
windowLayerManagerRemote	true
AzureCanvasBackend	direct2d
AzureContentBackend	direct2d
AzureFallbackCanvasBackend	cairo
AzureSkiaAccelerated	0
(In reply to Jason McKillican from comment #2)
> I tried each mode (safe mode, hardware acceleration disabled) independently
> and each (independently) seems to fix the issue for both the 20140630 and
> the 20140707 x86_64 33.0a1 Nightly builds.
> 
> It seems as if the issue may be following the hardware acceleration.

Hmmm. Here disabling HW acceleration has no effect on what I'm experiencing.
So I guess mine might be a separate bug entirely...
(In reply to avada from comment #10)
> (In reply to Jason McKillican from comment #2)
> > I tried each mode (safe mode, hardware acceleration disabled) independently
> > and each (independently) seems to fix the issue for both the 20140630 and
> > the 20140707 x86_64 33.0a1 Nightly builds.
> > 
> > It seems as if the issue may be following the hardware acceleration.
> 
> Hmmm. Here disabling HW acceleration has no effect on what I'm experiencing.
> So I guess mine might be a separate bug entirely...

Same here. The interior of every Firefox window simply goes black, including
the title bar. Resizing, minimizing, maximizing have no effect. The underlying
widgets respond; for instance if you click where the window-closing [X] is
expected to be, you get the "close all tabs" dialog prompt (all black of course).
If you click in the right place in the black rectangle where the OK button is,
Firefox will close.

I do not think this is a different bug. I suspect Firefox has some fundamental
bug in its UI related to repainting a window: handling WM_PAINT events or
whatever. It's not moving pixels into the window, plain and simple.

I have never seen this black screen effect in any application on any Windows
system in 20 years, including that system where Firefox has the problem.

You can hardly blame hardware acceleration or drivers, when no other app has
the issue.
Looks like a dupe of bug #1083071.
If you think otherwise please reopen.
Status: UNCONFIRMED → RESOLVED
Closed: 5 years ago
Resolution: --- → DUPLICATE
Duplicate of bug: 1083071
(In reply to Virtual_ManPL [:Virtual] from comment #12)
> Looks like a dupe of bug #1083071.
> If you think otherwise please reopen.

I can't see how. All the patches there seem to be about Intel driver blacklisting.

Meanwhile, both the OP and my issues are on Nvidia GPUs. (Even though they seem to be different issues)
(In reply to avada from comment #13)
> (In reply to Virtual_ManPL [:Virtual] from comment #12)
> > Looks like a dupe of bug #1083071.
> > If you think otherwise please reopen.
> 
> I can't see how. All the patches there seem to be about Intel driver
> blacklisting.
> 
> Meanwhile, both the OP and my issues are on Nvidia GPUs. (Even though they
> seem to be different issues)

I suspect all these black UI problems are the same, and they are a bug in Firefox, not in any drivers. Disabling hardware acceleration doesn't fix them. No other application exhibits the behavior. If it was drivers, other applications would be affected. (And probably the Windows "Aero" effects, too and whatnot.)

I'm going to try the "layers.offmainthreadcomposition.enabled = false" workaround recommended in some comments under 1083071.

That doesn't sound like it has anything to do with the hardware, but with the use of threads.
(In reply to kaz from comment #14)
> I'm going to try the "layers.offmainthreadcomposition.enabled = false"
> workaround recommended in some comments under 1083071.

Nope! Turns out this is already false in my about:config.
This is NOT a duplicate. And it's NOT resolved. Updated to 38.0.5, and whenever firefox reaches 1,000,000K memory, black boxes and graphical glitches start happening until the entire window is black and forces me to close firefox and restart.
Have got the same trouble since FF33 or 34 and it's unsolved. Whenever the FF memory consumption reaches > 1GB, the tabs begin flickering with black boxes and glitches until the whole FF Windows is black. Have tried to deactive/delete all Addons, have tried to reset all settings, have tried to enable/disable hardware acceleration. 

Win7 / 32Bit / IntelCore 2 Q6600 @2,4 GHz/ 3GB RAM / ATI Radeon HD 5770 / FF38.0.5
(In reply to info from comment #17)
> Have got the same trouble since FF33 or 34 and it's unsolved.

Some cases of this were resolved in bug 1083071, follow-up work is happening in bug 1172351. Please follow along there and add your information if you think it will help resolve the issue. Thank you.
You need to log in before you can comment on or make changes to this bug.