Closed Bug 1037828 Opened 10 years ago Closed 10 years ago

Solid black contents in all windows and dialogs with hardware acceleration enabled

Categories

(Core :: Graphics, defect)

33 Branch
x86_64
Windows 7
defect
Not set
normal

Tracking

()

RESOLVED DUPLICATE of bug 1083071

People

(Reporter: caped_wonderwoman, Unassigned)

Details

Attachments

(2 files)

Attached image ss.png
User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:33.0) Gecko/20100101 Firefox/33.0 (Beta/Release)
Build ID: 20140712030202

Steps to reproduce:

Launch Win32 Nightly built after roughly 6/5. Also appears on x64. UX, Aurora, Beta, and Release are unaffected. I launched with and without Profile Manager, and was able to reproduce the problem on a clean re-install.


Actual results:

Profile Manager dialog came up with a normal window frame containing a solid black window. Pressing Enter to launch the previously used profile brought up a browser window with a distorted frame and transparency enabled, although I have it disabled in system, and the same solid black featureless contents, no tabs, and no title or title bar. Safe Mode or manually disabling hardware acceleration by opening the same profile in Aurora are the only ways I have found to restore the browser to a functional state. 


Expected results:

Browser launch, with or without Profile Manager as determined by presence or absence of -p switch, with window frames whose size, coloration, and opacity match the settings of the Windows and standard Firefox theme, and with visible, usable content in said window.
Graphics card is a Radeon 6490M.
I have a Radeon HD 6450 and I`m unable to reproduce this issue. Do you have the latest driver version installed?
Flags: needinfo?(caped_wonderwoman)
Component: Untriaged → Graphics
Product: Firefox → Core
Per both Windows and HP, yes. I see the 6450 listed as a standalone card on AMD's site; is it a desktop card? The 6490 is a mobile card that works with switchable graphics, and AMD's standard driver package is not compatible with switchable graphics cards.
Flags: needinfo?(caped_wonderwoman)
Aurora is also now affected. The equivalent versions of Seamonkey and Thunderbird are as well.
I've been having the exact same problem. It initially affected only the Aurora channel, but now affects the Beta Channel too, so I'm assuming it is something related to version 33?

My machine:

GPU: NVIDIA GeForce GT 540M
CPU: Intel(R) Core(TM) i7-2670QM CPU @ 2.20GHz
Memory: 8.00 GB RAM (7.91 GB usable)
Current resolution: 1366 x 768, 60Hz
Driver version: 340.52
Operating System: Microsoft Windows 7 Home Premium

(Latest graphic's card driver version as per NVIDIA GeForce Experience)
two similar reports on the sumo forum of users who upgraded to 33 beta (one has a intel hd graphics card):
https://support.mozilla.org/en-US/questions/1018735
https://support.mozilla.org/en-US/questions/1018740
carlos, could you run the mozregression tool like described at http://mozilla.github.io/mozregression/ in order to produce the time range and pushlog when the issue first appeared in firefox 33 builds?

this should be the right command you can start with in order to narrow the range further down:
 mozregression --good=2014-06-08 --bad=2014-07-22 --bits=32
Flags: needinfo?(carlos.molc)
I did, but needed to go back farther to the firefox 32 builds.

The bisecting of nightly builds led to:
Last good revision: cb9f34f73ebe (2014-05-20)
First bad revision: 9d8d16695f6a (2014-05-21)
Pushlog:
https://hg.mozilla.org/mozilla-central/pushloghtml?gtomchange=cb9f34f73ebe&tochange=9d8d16695f6a

Further bisecting of inbound builds led to:
Last good revision: 66fe0b3237d6
First bad revision: 718a9852b60d
Pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=66fe0b3237d6&tochange=718a9852b60d
Flags: needinfo?(carlos.molc)
great, thanks! so it looks like bug 899785 is at the root of this problem. 
pinging :bas.schouten since the bug is assigned to him...
Flags: needinfo?(bas)
(In reply to philipp from comment #7)
> carlos, could you run the mozregression tool like described at
> http://mozilla.github.io/mozregression/ in order to produce the time range
> and pushlog when the issue first appeared in firefox 33 builds?
> 
> this should be the right command you can start with in order to narrow the
> range further down:
>  mozregression --good=2014-06-08 --bad=2014-07-22 --bits=32

Could you post your about:supports graphics section, and try switching the gfx.direct2d.disabled pref to true?
Flags: needinfo?(bas)
(In reply to Bas Schouten (:bas.schouten) from comment #10)
> Could you post your about:supports graphics section, and try switching the
> gfx.direct2d.disabled pref to true?

Switching the gfx.direct2d.disabled to true allowed normal use of the browser, even with the option "Use hardware acceleration when available" checked.

My about:support graphics section is as follows:

Adapter Description: Intel(R) HD Graphics Family
Adapter Description (GPU #2): NVIDIA GeForce GT 540M
Adapter Drivers: igdumd64 igd10umd64 igd10umd64 igdumdx32 igd10umd32 igd10umd32
Adapter Drivers (GPU #2): nvd3dumx,nvwgf2umx,nvwgf2umx nvd3dum,nvwgf2um,nvwgf2um
Adapter RAM: Unknown
Adapter RAM (GPU #2): 2048
Device ID: 0x0116
Device ID (GPU #2): 0x0df4
DirectWrite Enabled: false (6.2.9200.16492)
Driver Date: 12-16-2010
Driver Date (GPU #2): 7-2-2014
Driver Version: 8.15.10.2266
Driver Version (GPU #2): 9.18.13.4052
GPU #2 Active: false
GPU Accelerated Windows: 0/1 Basic (OMTC)
Subsys ID: 00000000
Subsys ID (GPU #2): 00000000
Vendor ID: 0x8086
Vendor ID (GPU #2): 0x10de
WebGL Renderer: Google Inc. -- ANGLE (Intel(R) HD Graphics Family Direct3D9Ex vs_3_0 ps_3_0)
windowLayerManagerRemote: true
AzureCanvasBackend: skia
AzureContentBackend: cairo
AzureFallbackCanvasBackend: cairo
AzureSkiaAccelerated: 0
[Tracking Requested - why for this release]:
can this be investigated further before 33 hits release?
i get this issue at times on Fedora Linux 20.
Gt 525M Nvidia - Optimus Enabled.
Goes away after closing firefox, Logging out of GNOME, which causes a GNOME crash, and logging back in.
Bas, Nical, do you think we can expect a fix for this bug in 33? Do you confirm that this bug remains even if OMTC has been disabled? Thanks
Flags: needinfo?(nical.bugzilla)
Flags: needinfo?(bas)
(In reply to carlos.molc from comment #11)
> My about:support graphics section is as follows:
> 
> Adapter Description: Intel(R) HD Graphics Family
> Adapter Description (GPU #2): NVIDIA GeForce GT 540M
> ...
> Device ID: 0x0116
> ...
> Driver Date: 12-16-2010

This is a very old driver! I'd expect it to be blacklisted.
Had you set any pref manually in about:config when you were experiencing this bug?
Flags: needinfo?(nical.bugzilla)
(In reply to bull500 from comment #13)
> i get this issue at times on Fedora Linux 20.
> Gt 525M Nvidia - Optimus Enabled.
> Goes away after closing firefox, Logging out of GNOME, which causes a GNOME
> crash, and logging back in.

I expect this to be a separate issue on Linux (where we don't have OMTC on by default and no D2D/DXGI/etc interop that could be causing problems at the driver level). Could you file a separate bug with the content of your graphics section in about:support?
> I expect this to be a separate issue on Linux (where we don't have OMTC on
> by default and no D2D/DXGI/etc interop that could be causing problems at the
> driver level). Could you file a separate bug with the content of your
> graphics section in about:support?
 
Honestly im really not sure if its a Firefox or GNOME issue, the problem happens out of the blue. 
Firefox never crashes but keeps showing black 'mouse right click' context boxes or blacked out "Confirm close" box when i try to quit the browser.
Once i log out and log into gnome, i get a gnome crash https://bugzilla.redhat.com/show_bug.cgi?id=1117453


Graphics
Adapter Description	Intel Open Source Technology Center -- Mesa DRI Intel(R) Sandybridge Mobile
Device ID	Mesa DRI Intel(R) Sandybridge Mobile
Driver Version	3.0 Mesa 10.1.5
GPU Accelerated Windows	0/1 Basic (OMTC)
Vendor ID	Intel Open Source Technology Center
WebGL Renderer	Intel Open Source Technology Center -- Mesa DRI Intel(R) Sandybridge Mobile
windowLayerManagerRemote	true
AzureCanvasBackend	cairo
AzureContentBackend	cairo
AzureFallbackCanvasBackend	none
AzureSkiaAccelerated	0

Additional Info:
GT525M Nvidia, Optimus with proprietary drivers 
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])
01:00.0 VGA compatible controller: NVIDIA Corporation GF108M [GeForce GT 525M] (rev ff) (prog-if ff)
(In reply to Nicolas Silva [:nical] from comment #15)
> (In reply to carlos.molc from comment #11)
> > My about:support graphics section is as follows:
> > 
> > Adapter Description: Intel(R) HD Graphics Family
> > Adapter Description (GPU #2): NVIDIA GeForce GT 540M
> > ...
> > Device ID: 0x0116
> > ...
> > Driver Date: 12-16-2010
> 
> This is a very old driver! I'd expect it to be blacklisted.
> Had you set any pref manually in about:config when you were experiencing
> this bug?
Yes, but the behaviour is reproducible in clean installs.
according to incoming questions on SuMo about firefox 33 this is now hitting quite a lot of users on release
based on feedback from users on sumo, this bug appears to affect users with embedded intel hd2000 (Device ID: 0x0102) and intel hd3000 (Device ID: 0x0116) graphics cards with old drivers.
buggy driver versions reportedly include at least: 
8.15.10.2266 (Driver Date: 12-16-2010) 
8.15.10.2279 
8.15.10.2291 (Driver Date: 1-27-2011)

updating to the latest driver made available by intel solves the issue: https://downloadcenter.intel.com/Detail_Desc.aspx?DwnldID=23764

due to the large number of inquiries which would indicate that this is a fairly wide-spread issue, i'd appreciate if this bug can be moved out of "unconfirmed" and options to blocklist be investigated quickly.
Attached file about support.txt
graphics section from about:support of affected users while they were running firefox with HWA disabled
Status: UNCONFIRMED → RESOLVED
Closed: 10 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: