User Agent: Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:40.0) Gecko/20100101 Firefox/40.0 Build ID: 20150401030204 Steps to reproduce: Start Firefox Actual results: Firefox starts but show only a black screen instead of web pages. Expected results: Firefox show the actual content of web pages. Starting in safe mode or unchecking "Use hardware acceleration when available" let Firefox display correctly. I'm running a Windows 8.1 laptop with two switchable graphics card (Intel HD Graphics 4000 and AMD Radeon HD 8570M). x64 version of Firefox only has the problem, x86 version never experienced any problem on the same pc with hardware acceleration enabled. Manually switching enabled video card evidence that both graphics cards experience similar problem. Cards are running the latest drivers.
Hardware: x86 → x86_64
Version: Firefox 38 → 1.0 Branch
Tested on Firefox 38.0b1 and 40.0a1 (Build ID 20150401030204).
Could you attach some screenshots (of the complete Firefox window) when you get the issue, please. In addition, type about:support in the location bar and paste here the section "graphics".
Component: Untriaged → Graphics
Product: Firefox → Core
Version: Firefox 38 → 38 Branch
Created attachment 8586839 [details] Screenshot with AMD card enabled
Graphics Adapter Description Intel(R) HD Graphics 4000 Adapter Drivers igdumdim64 igd10iumd64 igd10iumd64 igdumdim32 igd10iumd32 igd10iumd32 Adapter RAM Unknown Asynchronous Pan/Zoom none Device ID 0x0166 DirectWrite Enabled false (6.3.9600.17415) Driver Date 12-18-2014 Driver Version 10.18.10.4061 GPU #2 Active false GPU Accelerated Windows 0/1 Basic (OMTC) Subsys ID 00000000 Vendor ID 0x8086 WebGL Renderer Google Inc. -- ANGLE (Intel(R) HD Graphics 4000 Direct3D11 vs_5_0 ps_5_0) windowLayerManagerRemote true AzureCanvasBackend skia AzureContentBackend cairo AzureFallbackCanvasBackend cairo AzureSkiaAccelerated 0
That smells similar to bug 1090046 and bug 1096864. If you keep HWA enabled and you disable OMTC (layers.offmainthreadcomposition.enabled = false in about:config, restart to apply), does it change something?
Yes, switching to false make 38.0 (20150330154247) but 40.0 (20150401030204).
Yes, switching to false make 38.0 (20150330154247) working but not 40.0 (20150401030204).
Armando, can you get a regression window using mozregression?
19:06.30 LOG: MainThread Bisector INFO Narrowed inbound regression window from [03d20761, 9862bab2] (3 revision da7bc5, 9862bab2] (2 revisions) (~1 steps left) 19:06.30 LOG: MainThread Bisector INFO Oh noes, no (more) inbound revisions :( 19:06.30 LOG: MainThread Bisector INFO Last good revision: e7da7bc52c94 19:06.31 LOG: MainThread Bisector INFO First bad revision: 9862bab2c301 19:06.31 LOG: MainThread Bisector INFO Pushlog: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=e7da7bc52c94&tochange=9862bab2c301 Is this what you need?
Yes, thanks! With the latest code, safe mode or no HW acceleration improves things. Increasing blocklisting should move us in the same direction that safe mode/no HW acceleration does. To rephrase, increasing blocklisting should improve things. Based on the regression range above, backing out reduced blocklisting broke things. To rephrase, increasing blocklisting breaks things. Hmm.
Bug still present in 43.0.1 (release)
(In reply to Armando Villani from comment #12) > Bug still present in 43.0.1 (release) Are there newer drivers available for your graphics cards?
about:support graphics from comment 5 is with Intel - 64-bit or 32-bit? It would be useful to see all four configurations - Intel and AMD, 32-bit and 64-bit. As I understand, in 43.0.1 (and still?), the 32-bit is OK, with either card, with or without hardware acceleration, and 64-bit is bad, with both cards, with hardware acceleration, but OK without hardware acceleration. Is this correct?
You need to log in before you can comment on or make changes to this bug.