You are not logged in.
I have low fps during scrolling pages in chrome/chromium only with kernel 4.14.x. Now I have to use lts-kernel to avoid this problem.
In benchmark (unigine) kernel 4.14 have the same result as lts or 4.13 kernel and now I'm afraid that all the kernels since 4.14 will have this error for me.
kernel log:
https://pastebin.com/i5hTNZTv
My laptop Dell E5570 with i5-6440HQ and Intel HD Graphics 530.
Offline
Kernel 4.15rc5 still have this bug.
xorg.log
https://pastebin.com/xrbpcM3j
And with 4.14/4.15 kernels I have a strange dmesg after plug\unplug AC on my laptop:
[ 64.104856] acpi INT3400:00: Unsupported event [0x86]
[ 69.347228] acpi INT3400:00: Unsupported event [0x86]
Last edited by d7 (2017-12-27 19:25:03)
Offline
tried xf86-video-intel?
Offline
tried xf86-video-intel?
yes
Offline
Bisect between 4.13 and 4.14 and report the issue upstream?
Offline
It's kinda weird that only scrolling in chromium would be affected.
Desktop environment? (unigine is maybe unredirected by the compositor)
The dmesg isn't long enough (apparently taken directly after the boot), but did you try to disable framebuffer compression?
Does this meet your problem: https://wiki.archlinux.org/index.php/In … 28VSYNC.29 ?
Offline
Desktop environment
gnome 3.26
The dmesg isn't long enough (apparently taken directly after the boot), but did you try to disable framebuffer compression?
Does this meet your problem: https://wiki.archlinux.org/index.php/In … 28VSYNC.29 ?
Disabling frame buffer compression, vsync and falling back to "DRI2" has no effect.
Bisect between 4.13 and 4.14 and report the issue upstream?
I don't have any logs confirming this issue, and I can't understand where the problem. I have only this:
And with 4.14/4.15 kernels I have a strange dmesg after plug\unplug AC on my laptop:
[ 64.104856] acpi INT3400:00: Unsupported event [0x86] [ 69.347228] acpi INT3400:00: Unsupported event [0x86]
Offline
Does it also happen with gnome on wayland?
Offline
Does it also happen with gnome on wayland?
Unfortunately, yes.
I'm really do not expect issues from the intel...
Offline
loqs wrote:Bisect between 4.13 and 4.14 and report the issue upstream?
I don't have any logs confirming this issue, and I can't understand where the problem
4.14 kernel has the issue you downgrade to 4.13 or 4.9 and the issue is no longer present without changing anything else then the issue would seem to be caused by some kernel change between 4.13 and 4.14.
Ideally with bisection you would get commit A no issue commit B issue and you can report to upstream the issue starts with commit B.
Other scenarios would be:
1) you build 4.14 locally and can not reproduce the issue then the issue is related to the build system.
2) You have commit A no issue commits X Y Z which will not boot or have other issues that make testing those commits impossible then commit B with the issue so you can only report X Y Z or B is causing the issue.
You could report the issue how-report-bugs upstream without the bisection but it may be harder to diagnose the cause.
As 4.14 is due to be the next linux-lts you would have to start building your own kernels based on 4.9 lts when arch switches to 4.14 lts unless the issue is resolved.
Offline
loqs, thank you for advice.
Offline
This patch solved the problem.
https://bugs.freedesktop.org/attachment.cgi?id=135277
Offline