User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:36.0) Gecko/20100101 Firefox/36.0 Build ID: 20150320202338 Steps to reproduce: 1. Create a new firefox profile (clean profile) and by default Hardware acceleration is enabled then start firefox 2. Opened a tab (any website) 3. Opened another tab (any website) so in total there are 3 tabs 4. Just hover mouse at each tab AND/OR switch between tab Actual results: - Typing at the address bar often would result in display corruption at the address bar so it's hard to know what you just typed - Display corruptions like for example in the gif attached a page from another tab being displayed in current tab (as can be seen from the gif attached) Note: if you see a tab being highlighted in the gif that means that i'm currently mouse hovering at the tab Expected results: Everything should be working fine without display corruption. I reverted to Firefox 36.0.4 even with hardware acceleration enabled i don't experience such problem there

I forgot to mention this in the previous post. I'm using: - Radeon HD 7730 - AMD Omega Driver 14.12 (which is the latest at the time i write this)

Type about:support in the location bar and paste the section "graphics", please.

Here's the graphics information from the about:support page. Adapter Description AMD Radeon HD 7700 Series Adapter Drivers aticfx64 aticfx64 aticfx64 aticfx32 aticfx32 aticfx32 atiumd64 atidxx64 atidxx64 atiumdag atidxx32 atidxx32 atiumdva atiumd6a atitmm64 Adapter RAM 2048 ClearType Parameters Gamma: 2400 Pixel Structure: R ClearType Level: 0 Enhanced Contrast: 200 Device ID 0x6837 Direct2D Enabled true DirectWrite Enabled true (6.2.9200.16571) Driver Date 11-20-2014 Driver Version 14.501.1003.0 GPU #2 Active false GPU Accelerated Windows 1/1 Direct3D 11 (OMTC) Subsys ID 23301787 Vendor ID 0x1002 WebGL Renderer Google Inc. -- ANGLE (AMD Radeon HD 7700 Series Direct3D11 vs_5_0 ps_5_0) windowLayerManagerRemote true AzureCanvasBackend direct2d 1.1 AzureContentBackend direct2d 1.1 AzureFallbackCanvasBackend cairo AzureSkiaAccelerated 0

Jeff, this looks like a duplicate of bug 1067470, but I don't know if the configuration helps with more info.

Just wanted to let you guys know that the problem persist even on 38 beta (tested on 38.0.4 and 38.0.5 beta with clean profile). Firefox totally unusable because of this so i'm still at 36.0.4. The only workaround is by disabling OMTC (layers.offmainthreadcomposition.enabled) but i heard that the option to disable OMTC will be removed in the near future so that workaround will no longer be valid. Tested on Windows 7 SP1. All Windows updates that is available at the time i write this is installed. VGA driver is using the latest stable driver from AMD which is Omega 14.12 driver.

Without disabling OMTC, what happens if you set gfx.direct2d.use1_1 to false?

Unfortunately the problem still persists. The content area (if that's what it's actually called) still shows repainting problem. But the worst part is the address bar area (and the search bar). I couldn't see what i was typing, so if i made a mistake typing something in that area, i have to select all and then retype it again without seeing what i was typing. The problem from what i can see is it is due to Firefox trying to autocomplete but because of repainting problem it is blocking the view. Of course pictures worth a thousand words so i have uploaded a gif showing the problem. The gif was also taken from a clean profile and because of that the first word i typed is Mozilla because that words would already be there in bookmark to show you what the problem is. If it was on a live/real profile with hundreds of bookmarks/visited urls history/etc you can imagine almost every letters being typed would trigger the glitched autocomplete that made you unable to see what you just typed. Gif Link: https://i.imgur.com/rgs6mFj.gif

Looking on attachment and description it's definitely a duplicate of bug #1067470.

Bug 1067470 is a whole mix of problems. If this showed up in 37 it's worth keeping separate. Aditya, can you try getting a regression window for this using mozregression? http://mozilla.github.io/mozregression/

(In reply to Jeff Muizelaar [:jrmuizel] from comment #9) > Bug 1067470 is a whole mix of problems. If this showed up in 37 it's worth > keeping separate. > > Aditya, can you try getting a regression window for this using mozregression? > http://mozilla.github.io/mozregression/ I just finished the mozregression and here's the result. Hopefully this is what you're looking for. 25:09.33 LOG: MainThread Bisector INFO Narrowed inbound regression window from [0ea4d636, 24ba8274] (3 revisions) to [2a61df4e, 24ba8274] (2 revisions) (~1 steps left) 25:09.33 LOG: MainThread Bisector INFO Oh noes, no (more) inbound revisions :( 25:09.33 LOG: MainThread Bisector INFO Last good revision: 2a61df4eaa2d 25:09.34 LOG: MainThread Bisector INFO First bad revision: 24ba8274ed60 25:09.34 LOG: MainThread Bisector INFO Pushlog: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=2a61df4eaa2d&tochange=24ba8274ed60

Do all of these builds show the problem? (You may need to wait for some of them to show up) https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-a121a7a56ad5 http://ftp.mozilla.org/pub/mozilla.org/firefox/tinderbox-builds/mozilla-inbound-win32/1432563618/ https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-5da174860103

Yes, unfortunately all three builds showed the same problem. But on build a121a7a56ad5 it flickers/flashes a lot more compared to the other two builds.

Two more builds to test: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-0a3c88b8051d/ https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-3a70999c1394

Build 0a3c88b8051d works great no display corruption/glitch at all. On the other hand, build 3a70999c1394 produce extreme flashes/flickers just like build a121a7a56ad5 from before.

Forgot to put this on my previous comment. In case you're wondering how extreme the flashing/flickering are. I've uploaded it here http://a.pomf.se/wueqlr.mp4 (warning, flashing images).

Here will be the next build to try: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-19bc4fe771cb

Two more builds: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-02c7e6b65c88 https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-3c79b0454d34

All of the builds are up now.

All three builds (02c7e6b65c88, 3c79b0454d34, and 19bc4fe771cb) works great no glitch/corruption at all :).

And here's another build to try: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-14d71149d1b9

Unfortunately build 14d71149d1b9 produced a strange result. There's no graphics corruption/glitches as in it flickers/flashes this time it doesn't redraw/repaint the whole firefox area (not sure if this is the proper word for it) when switching between tabs or creating new tab or even opening options page. A more detailed description of the problem: 1. Run the build 2. By default when starting clean profile there will be 2 tabs opened the Nightly first run and Nightly start page 3. Clicking on the tabs or switching the tabs using shortcut (CTRL+TAB NUM) resulted in no change at all in Firefox (i.e firefox not redraw/repaint?). But on the taskbar, i can confirm that the tab indeed switched because the title being displayed in taskbar changed but there's no change at all in firefox. 4. The same problem also appeared when opening new tab and opening options page. The workaround to force firefox to redraw right away is simply by trying to close firefox (by default firefox prompt you if you try to close multiple tabs) and then canceling it. Also if HW Acceleration turned off, the above problem gone. I've uploaded a gif to better explain the problem https://imgur.com/hrbdbda . The same thing also happen when opening new tab, or even opening a website.

How about this one: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-aa0ca8b394fb

Also, do you see the display corruption if you turn on the DWM/aero glass?

The new build aa0ca8b394fb works great no display problem at all tested in both Classic and Aero. While the previous build (14d71149d1b9) has the same problem in both Classic and Aero.

I tested the previous builds you posted here with Aero enabled and they produce exactly the same problem when i ran them on Classic mode. So i would say there's no difference between using Aero or not in this case.

Do you see any graphics problems in Internet Explorer?

For the record, here's what IE 11 on Win7 uses for me: pDevice = 0x6d245a8, pDesc = [{BufferDesc = {Width = 1193, Height = 648, RefreshRate = {Numerator = 0, Denominator = 0}, Format = DXGI_FORMAT_B8G8R8A8_UNORM, ScanlineOrdering = DXGI_MODE_SCANLINE_ORDER_UNSPECIFIED, Scaling = DXGI_MODE_SCALING_UNSPECIFIED}, SampleDesc = {Count = 1, Quality = 0}, BufferUsage = DXGI_USAGE_RENDER_TARGET_OUTPUT, BufferCount = 1, OutputWindow = 0xc078e, Windowed = TRUE, SwapEffect = DXGI_SWAP_EFFECT_SEQUENTIAL, Flags = DXGI_SWAP_CHAIN_FLAG_GDI_COMPATIBLE}

Here will be another build to try: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-fd4f24b82488

Unfortunately the new one (fd4f24b82488) produces graphics corruptions (flickers). As for IE11, it works fine no problem at all with hardware acceleration enabled also Chrome Beta, and Opera Next works fine too in case you're wondering.

I made a test program that should help us to iterate on this issue: http://people.mozilla.org/~jmuizelaar/swap-chain.exe This should start with a gray window. When you press a key it will paint blue. You can switch between red and blue by pressing the 'r' and 'b' keys respectively. Each time you press a key it will paint, but will decrease the area painted by one pixel from the right side of the window. Do you see any flickering with this test?

I don't see flicker at all when running the test program. Btw, is there something changed between the last build (fd4f24b82488) and the one before that (aa0ca8b394fb)?. Because build aa0ca8b394fb worked fine and then on your last build the problem appears again.

(In reply to Aditya from comment #31) > I don't see flicker at all when running the test program. Btw, is there > something changed between the last build (fd4f24b82488) and the one before > that (aa0ca8b394fb)?. Because build aa0ca8b394fb worked fine and then on > your last build the problem appears again. Yes, I'm just flailing around a bit trying to see what special that might be causing things to fail for you. Can you get a recording of the symptoms that happen with this build: https://ftp-ssl.mozilla.org/pub/mozilla.org/firefox/try-builds/jmuizelaar@mozilla.com-551d22ccfa72

I recorded a video for build 551d22ccfa72 (warning flashing images) https://d.maxfile.ro/mnpbmzdkkv.mp4 Also i just installed the beta driver from AMD (previously I'm using latest stable driver 14.12 before switched to beta) because from what i read the beta has optimizations for certain game that i just bought. Surprisingly build 551d22ccfa72 and even the previous build (fd4f24b82488) and also the one before that (aa0ca8b394fb) works fine or at least i haven't noticed any glitch so far. If you still need more testing just let me know.

(In reply to Aditya from comment #33) > I recorded a video for build 551d22ccfa72 (warning flashing images) > > https://d.maxfile.ro/mnpbmzdkkv.mp4 > > Also i just installed the beta driver from AMD (previously I'm using latest > stable driver 14.12 before switched to beta) because from what i read the > beta has optimizations for certain game that i just bought. Surprisingly > build 551d22ccfa72 and even the previous build (fd4f24b82488) and also the > one before that (aa0ca8b394fb) works fine or at least i haven't noticed any > glitch so far. If you still need more testing just let me know. Thanks. The driver switch is pretty interesting. Would you mind trying some other older versions of the driver to see if it makes a difference?

Aditya, were you ever able to follow up with Jeff's request in comment 34? Does this bug still reproduce for you?

Hello, I'm so sorry, I didn't notice until today that this bug is still marked as unfixed because back then when I checked this bugzilla page (via browser for several days) I didn't see any response and so I thought maybe it was already marked as fixed. And because this email address is used for many mailing list, I often missed many things unless it was near the top of the mail in my Thunderbird (I saw notification about this today right at the top of my mail). Actually I never experience this problem after switching display driver I mentioned before when I was using my old AMD 7730. But I've long switched from AMD to Nvidia (950 GTX), and on this new card I never experience any graphic issue that I described before (other than HW Video Decode not working and I have to clear the blacklist in the about:config to get it to work). I'm sorry once again, and you can mark this as fixed.