Re: Firefox testing - offscreen surfaces and OMTC
I'm seeing flickering (e.g. with arstechnica) on 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV620/M82 [Mobility Radeon HD 3450/3470] with native drivers (F22). I have seen heavy flickering on mapy.cz (online maps) when zooming. The longer I used the browser, the more it seemed to flicker. Radeon R9 270, radeonsi driver, F22. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/25/2015 12:09 PM, Kamil Paral wrote: I'm seeing flickering (e.g. with arstechnica) on 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV620/M82 [Mobility Radeon HD 3450/3470] with native drivers (F22). I have seen heavy flickering on mapy.cz (online maps) when zooming. The longer I used the browser, the more it seemed to flicker. Radeon R9 270, radeonsi driver, F22. Thanks for testing! Can you please also test latest nightly build from https://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-central/firefox-43.0a1.en-US.linux-x86_64.tar.bz2 ? Can you also update https://bugzilla.redhat.com/show_bug.cgi?id=1255917 with your findings? Thanks, ma. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/21/2015 08:28 PM, Thomas Daede wrote: I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. The layer acceleration pref is a totally different thing and will stay off for the near future. It's affected by a bug in libxcb which has been patched but not made it to release yet: https://bugs.freedesktop.org/show_bug.cgi?id=84252 What do you mean here by acceleration? The offscreen surfaces referenced in this post are CPU rendering. ma. On 08/21/2015 01:33 AM, Martin Stransky wrote: Folks, I'd use some testing for new Firefox feature - offscreen surfaces [1]. It may also fix crashes when OMTC is enabled [2]. Browser should be a bit faster with those features on. How to test? - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. - restart your browser. And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. Thanks! ma. [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1015218 [2] Off Main Thread Composition - layout rendering in separate thread. New feature in Firefox 40. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Mon, Aug 24, 2015 at 8:54 AM, Martin Stransky stran...@redhat.com wrote: On 08/21/2015 08:28 PM, Thomas Daede wrote: I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. Why? What's the difference between the in-tree and the system cairo? Ist it newer? Older? Patched? Or just built with a different configuration? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/24/2015 09:43 AM, drago01 wrote: On Mon, Aug 24, 2015 at 8:54 AM, Martin Stransky stran...@redhat.com wrote: On 08/21/2015 08:28 PM, Thomas Daede wrote: I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. Why? What's the difference between the in-tree and the system cairo? Ist it newer? Older? Patched? Or just built with a different configuration? in-tree cairo is a cairo library intergated in firefox tree, patched for mozilla needs and used for internal rendering. system cairo is a cairo library shipped in Fedora and used by gtk3 to draw. ma. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Mon, Aug 24, 2015 at 10:18 AM, Martin Stransky stran...@redhat.com wrote: On 08/24/2015 09:43 AM, drago01 wrote: On Mon, Aug 24, 2015 at 8:54 AM, Martin Stransky stran...@redhat.com wrote: On 08/21/2015 08:28 PM, Thomas Daede wrote: I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. Why? What's the difference between the in-tree and the system cairo? Ist it newer? Older? Patched? Or just built with a different configuration? in-tree cairo is a cairo library intergated in firefox tree, patched for mozilla needs and used for internal rendering. system cairo is a cairo library shipped in Fedora and used by gtk3 to draw. I know what system and in-tree means .. the question was whether mozilla patched the in-tree version or if it is just a different version than the system cairo. Your comment implies the former. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/24/2015 10:41 AM, drago01 wrote: On Mon, Aug 24, 2015 at 10:18 AM, Martin Stransky stran...@redhat.com wrote: On 08/24/2015 09:43 AM, drago01 wrote: On Mon, Aug 24, 2015 at 8:54 AM, Martin Stransky stran...@redhat.com wrote: On 08/21/2015 08:28 PM, Thomas Daede wrote: I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. Why? What's the difference between the in-tree and the system cairo? Ist it newer? Older? Patched? Or just built with a different configuration? in-tree cairo is a cairo library intergated in firefox tree, patched for mozilla needs and used for internal rendering. system cairo is a cairo library shipped in Fedora and used by gtk3 to draw. I know what system and in-tree means .. the question was whether mozilla patched the in-tree version or if it is just a different version than the system cairo. Your comment implies the former. AFAIK Mozilla cairo is a snapshot from 2010-01-21 + mozilla fixes. ma. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
Martin Stransky wrote: If you mean OMTC by this then you're right - it works fine on nightly because nightly is built with in-tree cairo. Fedora it built with system cairo which causes crashes with OMTC enabled. First build which supports in-tree cairo Gtk3 is FF41. Yet another bundled library?! Sigh! This needs to be made to work with the system Cairo! Kevin Kofler -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, Aug 21, 2015 at 10:33 AM, Martin Stransky stran...@redhat.com wrote: Folks, I'd use some testing for new Firefox feature - offscreen surfaces [1]. It may also fix crashes when OMTC is enabled [2]. Browser should be a bit faster with those features on. How to test? - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. - restart your browser. And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. Seems to be still crashy ... not sure if this counts as different than the usual ones -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
I see the same flickering on NVIDIA hardware. Intel seems to be fine. On 08/21/2015 03:06 PM, Alexander Ploumistos wrote: On Fri, Aug 21, 2015 at 11:33 AM, Martin Stransky stran...@redhat.com wrote: - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. What about layers.acceleration.force-enabled? And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. On my desktop (f22), with a GeForce 9800 GT and the 340.76 driver from nVidia I am getting a lot of flickering when I scroll through some pages (e.g. MDN, Ars Techinca, gmail) especially when they contain fixed or sticky elements or elements whose background image is repeated along the x and y axes. If you can get your hands on similar hardware, scroll through this: https://developer.mozilla.org/en-US/docs/Web/CSS/position On a recent laptop (f22), with nVidia and Intel dual graphics, there is no such problem on either adapter. I have been unable to reproduce the crashes we were investigating so far, so perhaps enabling layers.use-image-offscreen-surfaces could be a solution (and I should get a newer graphics card - anyone cares to get me a late birthday present?). This is the graphics information provided by firefox for the desktop: Adapter DescriptionNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce 9800 GT/PCIe/SSE2 Driver Version3.3.0 NVIDIA 340.76 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 and for the laptop: Adapter DescriptionNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce GTX 860M/PCIe/SSE2 Driver Version4.5.0 NVIDIA 352.21 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 Adapter DescriptionIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile Asynchronous Pan/Zoomnone Device IDMesa DRI Intel(R) Haswell Mobile Driver Version3.0 Mesa 10.6.3 (git-ccef890) GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDIntel Open Source Technology Center WebGL RendererIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
There is also an issue with elements that are loaded asynchronously - they take longer to appear or it might take a bit of scrolling up and down until they are rendered (with the older nVidia card). -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, Aug 21, 2015 at 4:24 PM, Martin Stransky stran...@redhat.com wrote: I see the same flickering on NVIDIA hardware. Intel seems to be fine. Which model? Is it using one of the older driver series? -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, Aug 21, 2015 at 4:38 PM, Martin Stransky stran...@redhat.com wrote: It's GeForce GTX 750 with nvidia binary drivers. So you are not on the legacy drivers. Could the problem be hardware-dependent? The graphics adapter of the laptop I tested is one generation ahead of yours and there is no flickering. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/21/2015 03:34 PM, Alexander Ploumistos wrote: On Fri, Aug 21, 2015 at 4:24 PM, Martin Stransky stran...@redhat.com wrote: I see the same flickering on NVIDIA hardware. Intel seems to be fine. Which model? Is it using one of the older driver series? It's GeForce GTX 750 with nvidia binary drivers. ma. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, Aug 21, 2015 at 11:33 AM, Martin Stransky stran...@redhat.com wrote: - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. What about layers.acceleration.force-enabled? And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. On my desktop (f22), with a GeForce 9800 GT and the 340.76 driver from nVidia I am getting a lot of flickering when I scroll through some pages (e.g. MDN, Ars Techinca, gmail) especially when they contain fixed or sticky elements or elements whose background image is repeated along the x and y axes. If you can get your hands on similar hardware, scroll through this: https://developer.mozilla.org/en-US/docs/Web/CSS/position On a recent laptop (f22), with nVidia and Intel dual graphics, there is no such problem on either adapter. I have been unable to reproduce the crashes we were investigating so far, so perhaps enabling layers.use-image-offscreen-surfaces could be a solution (and I should get a newer graphics card - anyone cares to get me a late birthday present?). This is the graphics information provided by firefox for the desktop: Adapter DescriptionNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce 9800 GT/PCIe/SSE2 Driver Version3.3.0 NVIDIA 340.76 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 and for the laptop: Adapter DescriptionNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce GTX 860M/PCIe/SSE2 Driver Version4.5.0 NVIDIA 352.21 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 Adapter DescriptionIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile Asynchronous Pan/Zoomnone Device IDMesa DRI Intel(R) Haswell Mobile Driver Version3.0 Mesa 10.6.3 (git-ccef890) GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDIntel Open Source Technology Center WebGL RendererIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, 2015-08-21 at 15:24 +0200, Martin Stransky wrote: I see the same flickering on NVIDIA hardware. Intel seems to be fine. I'm getting flickering (but only when scrolling up) on the linked cairo xrender bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1015218 Adapter Description Intel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile Asynchronous Pan/Zoom none Device ID Mesa DRI Intel(R) Haswell Mobile Driver Version 3.0 Mesa 10.6.3 (git-ccef890) GPU Accelerated Windows 0/3 Basic (OMTC) Supports Hardware H264 Decoding false Vendor ID Intel Open Source Technology Center WebGL Renderer Intel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile windowLayerManagerRemotetrue AzureCanvasBackend cairo AzureContentBackend cairo AzureFallbackCanvasBackend none AzureSkiaAccelerated0 On 08/21/2015 03:06 PM, Alexander Ploumistos wrote: On Fri, Aug 21, 2015 at 11:33 AM, Martin Stransky stran...@redhat.com wrote: - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. What about layers.acceleration.force-enabled? And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. On my desktop (f22), with a GeForce 9800 GT and the 340.76 driver from nVidia I am getting a lot of flickering when I scroll through some pages (e.g. MDN, Ars Techinca, gmail) especially when they contain fixed or sticky elements or elements whose background image is repeated along the x and y axes. If you can get your hands on similar hardware, scroll through this: https://developer.mozilla.org/en-US/docs/Web/CSS/position On a recent laptop (f22), with nVidia and Intel dual graphics, there is no such problem on either adapter. I have been unable to reproduce the crashes we were investigating so far, so perhaps enabling layers.use-image-offscreen-surfaces could be a solution (and I should get a newer graphics card - anyone cares to get me a late birthday present?). This is the graphics information provided by firefox for the desktop: Adapter DescriptionNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce 9800 GT/PCIe/SSE2 Driver Version3.3.0 NVIDIA 340.76 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce 9800 GT/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 and for the laptop: Adapter DescriptionNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 Asynchronous Pan/Zoomnone Device IDGeForce GTX 860M/PCIe/SSE2 Driver Version4.5.0 NVIDIA 352.21 GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDNVIDIA Corporation WebGL RendererNVIDIA Corporation -- GeForce GTX 860M/PCIe/SSE2 windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 Adapter DescriptionIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile Asynchronous Pan/Zoomnone Device IDMesa DRI Intel(R) Haswell Mobile Driver Version3.0 Mesa 10.6.3 (git-ccef890) GPU Accelerated Windows0/1 Basic (OMTC) Supports Hardware H264 Decodingfalse Vendor IDIntel Open Source Technology Center WebGL RendererIntel Open Source Technology Center -- Mesa DRI Intel(R) Haswell Mobile windowLayerManagerRemotetrue AzureCanvasBackendcairo AzureContentBackendcairo AzureFallbackCanvasBackendnone AzureSkiaAccelerated0 -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
I've been running nightly with this enabled for quite a while on Intel and it's been fine. Note that OMTC is required for e10s. The layer acceleration pref is a totally different thing and will stay off for the near future. It's affected by a bug in libxcb which has been patched but not made it to release yet: https://bugs.freedesktop.org/show_bug.cgi?id=84252 On 08/21/2015 01:33 AM, Martin Stransky wrote: Folks, I'd use some testing for new Firefox feature - offscreen surfaces [1]. It may also fix crashes when OMTC is enabled [2]. Browser should be a bit faster with those features on. How to test? - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. - restart your browser. And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. Thanks! ma. [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1015218 [2] Off Main Thread Composition - layout rendering in separate thread. New feature in Firefox 40. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On 08/21/2015 03:33 AM, Martin Stransky wrote: And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. Opened: https://bugzilla.redhat.com/show_bug.cgi?id=1255917 -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Re: Firefox testing - offscreen surfaces and OMTC
On Fri, Aug 21, 2015 at 03:38:49PM +0200, Martin Stransky wrote: On 08/21/2015 03:34 PM, Alexander Ploumistos wrote: On Fri, Aug 21, 2015 at 4:24 PM, Martin Stransky stran...@redhat.com wrote: I see the same flickering on NVIDIA hardware. Intel seems to be fine. Which model? Is it using one of the older driver series? It's GeForce GTX 750 with nvidia binary drivers. ma. I'm seeing flickering (e.g. with arstechnica) on 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV620/M82 [Mobility Radeon HD 3450/3470] with native drivers (F22). Zbyszek -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
Firefox testing - offscreen surfaces and OMTC
Folks, I'd use some testing for new Firefox feature - offscreen surfaces [1]. It may also fix crashes when OMTC is enabled [2]. Browser should be a bit faster with those features on. How to test? - Install Firefox 40 on Fedora 22,23,Rawhide (you'd need Gtk3 build) - go to about:config, click to any key and add a new one, boolean type. The new key name is layers.use-image-offscreen-surfaces and set it to true. - enable layers.offmainthreadcomposition.enabled which may be disabled now. - restart your browser. And you're set now. Please report any oddity (different than the usual ones :)) at #BZ. Thanks! ma. [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1015218 [2] Off Main Thread Composition - layout rendering in separate thread. New feature in Firefox 40. -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct