| Summary: | x11-drivers/xf86-video-intel[-dri3] with vsync enabled in Plasma-5.14.4: flickering and blackouts since Qt-5.12.0 | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Franz Trischberger <franz.trischberger> |
| Component: | Current packages | Assignee: | Gentoo X packagers <x11> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | asturm, franz.trischberger |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Franz Trischberger
2018-12-12 19:41:15 UTC
I can't really help you there, Qt 5.12.0 works just the same as Qt 5.11.3 on all of my systems so there is no grave global Qt issue. Try with a new user profile for a start. Of course I tried it with my testuser and completely wiped its ~/.config. Same issue. The offending part seems to be the tearing prevention (vsync) in the compositor settings. I had to set it to "Never", every other setting (especially "Full screen repaints", which I had to use for nearly tear-free experience) causes this flickering. I am currently on mesa-18.2.5. I will update to 18.3.1 and then report back. An update to mesa-18.3.1 didn't help. I then enabled USE="dri3" for xf86-video-intel (to be in sync with mesa) and I can again enable full screen repaints without issues. However I remember that I had severe issues with dri3 in the past... I still think something's wrong here and leave the bug open. Do you still use the xf86-video-intel driver? It is basically unmaintained, Intel devs recommend modesetting themselves. (In reply to Andreas Sturmlechner from comment #4) > Do you still use the xf86-video-intel driver? It is basically unmaintained, > Intel devs recommend modesetting themselves. Yes, I do. I tried modesetting twice since the old DDX driver was called unmaintained. It is unusable on the desktop: I get extreme tearing even when setting "always repaint" in kwin's compositor setting, and there seems to be no video acceleration, video is stuttering. Also the whole desktop is laggy and certain programs (not all) are laggy and consume a lot of CPU. On the laptop however it seems to work, I ran it 1.5 days without an issue. Unfortunately I then decided to migrate the desktop to qt-5.12 which now shows the same issues as the laptop (flickering with long blackouts or tearing)... I went through several wikis on how to set up intel modesetting driver and I already have everything as it should be - kernel, USE Flags, etc. I can open an bug report concerning the graphics issues, however I think as long as the intel driver is stable and no USE Flags are masked Qt should work properly. (In reply to Franz Trischberger from comment #5) > I can open an bug report concerning the graphics issues, however I think as > long as the intel driver is stable and no USE Flags are masked Qt should > work properly. But that's just assuming the problem is in Qt. xf86-video-intel has plenty of issues that may be exposed by one or the other toolkit version, that doesn't make it a toolkit bug though. (In reply to Andreas Sturmlechner from comment #6) > (In reply to Franz Trischberger from comment #5) > > I can open an bug report concerning the graphics issues, however I think as > > long as the intel driver is stable and no USE Flags are masked Qt should > > work properly. > But that's just assuming the problem is in Qt. xf86-video-intel has plenty > of issues that may be exposed by one or the other toolkit version, that > doesn't make it a toolkit bug though. Sorry. No, that wasn't my intention. I just wanted to say that things should work, and if the intel driver starts to break things it should be masked if the issue isn't inside the broken packages. Now this was the first .12 release and it is still masked unstable. But I hope that there is a solution (fixed intel driver, mesa or Qt) as soon as it goes stable. I'm reassigning this to X11 as I'm afraid there is not much we can do on our side. Qt 5.12.0 may or may not be responsible for exposing this problem, however other X11 related updates in the same time frame seem more likely to me. Even though it would be good to cross-check this with Qt-5.11.3 on the affected system I have to advise against downgrading Qt as this opens a whole different can of worms. Sounds like I should just remove the knobs to disable dri3. The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=4e6ac224882527af7d365ef4b2bf909513504e4a commit 4e6ac224882527af7d365ef4b2bf909513504e4a Author: Matt Turner <mattst88@gentoo.org> AuthorDate: 2019-04-26 21:17:27 +0000 Commit: Matt Turner <mattst88@gentoo.org> CommitDate: 2019-04-26 21:20:26 +0000 x11-drivers/xf86-video-intel: Version bump to 2.99.917_p20190301 Closes: https://bugs.gentoo.org/653320 Closes: https://bugs.gentoo.org/673040 Signed-off-by: Matt Turner <mattst88@gentoo.org> x11-drivers/xf86-video-intel/Manifest | 1 + .../files/xf86-video-intel-gcc-pr65873.patch | 12 +++ .../xf86-video-intel-2.99.917_p20190301.ebuild | 91 ++++++++++++++++++++++ 3 files changed, 104 insertions(+) Additionally, it has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=a8b5cb6aee02a1b6f3acf10c9ec74eed97cb0c75 commit a8b5cb6aee02a1b6f3acf10c9ec74eed97cb0c75 Author: Matt Turner <mattst88@gentoo.org> AuthorDate: 2019-04-26 21:10:07 +0000 Commit: Matt Turner <mattst88@gentoo.org> CommitDate: 2019-04-26 21:20:15 +0000 x11-drivers/xf86-video-intel: Enable DRI3 when USE=dri Bug: https://bugs.gentoo.org/673040 Signed-off-by: Matt Turner <mattst88@gentoo.org> x11-drivers/xf86-video-intel/xf86-video-intel-9999.ebuild | 13 ++++--------- 1 file changed, 4 insertions(+), 9 deletions(-) |
