You are not logged in.
Thinkpad Carbon X1 3rd gen.
The screen resets every second. Turns black and go back to normal in 0.1 second, and repeat every second.
The problem starts from the Linux console mode, and persists after I starts a gnome-wayland session. The problem is gone after I switch to the lts kernel.
No external monitor.
Edit: The i915 fix does not work.
dmesg:
https://gist.github.com/wuxb45/ebbd6472 … 3c8a2c4a39
I also have a x1 carbon 4th gen which works fine with the new kernel.
Edit: 5.12.2 still does not work. Note that the problem starts to appear with the raw linux console. No login manager was used and the problem persists when I start a gnome session.
Last edited by wuxb (2021-05-09 22:20:56)
Offline
Hi. I managed to quickfix the issue by rolling back to the previous installed version of the kernel by running "sudo pacman -U /var/cache/pacman/pkg/linux-5.11.8.arch1-1-x86_64.pkg.tar.zst". I did not want to move back to a too old version of the kernel so that's why I used this approach.
Offline
https://wiki.archlinux.org/title/Intel_ … flickering ?
(If not, please post a dmesg so one can get an idea of the used HW & drivers)
Online
My ThinkPad X1 Carbon (Gen 3) has encountered this issue here +1, after updated Linux kernel from 5.11.x to 5.12.x.
I have tried adding the following kernel parameter as @seth mentioned, and it didn't help.
i915.enable_psr=0
My setup:
CPU: Intel Core i7-5600U
GPU: Intel HD Graphics 5500
BIOS: UEFI mode.
boot manager: systemd-boot.
init: systemd.
DE: GNOME Wayland.
Offline
Do you have an external monitor, ideally VGA?
The new big thing in 5.12 is VRR and that smells related to your symptoms, but there's no runtime flag to disable is.
It's however only enabled for displayport and internal displays… so …
Online
No, no external monitor currently.
Anyway, I can test external DisplayPort display later.
Offline
I have tested external DisplayPort monitor.
And, it doesn't has this issue,
Laptop's main LCD screen still has flickering issue after plugged external DP monitor.
Recorded video:
https://youtu.be/WYqPcWjyfrY
Offline
Output of "xrandr --props"?
Online
`xrandr --prop` of main LCD only: https://pastebin.com/UR72mibn
`xrandr --prop` of main LCD with external DP monitor: https://pastebin.com/3BrnjjDw
Offline
That's xwayland - this is only going to work from an actual Xorg session.
For sway there'd be https://wiki.archlinux.org/title/Variab … Using_Sway but I'm not sure whether you can userspace control the relevant flag this way (or how, tbh. I don't know whether setting it from xrandr will truely impact the i915 VRR handling)
inb4 "but X11 is antiqueblahblahblah" - we try to isolate the cause and whether it's truely VRR.
Online
seth what if I build a kernel with VRR disabled or configurable at boot?
https://git.kernel.org/pub/scm/linux/ke … 12.2#n1752 Looks a good candidate for simply disabling VRR to me.
Online
Then you will have built it?
Do you encounter the problem?
The line would be a good spot to disable this entirely, whether goldielin or other affected users prefer to install that kernel or try to disable the property on the output (which from the code, the vrr implementation should™ honor) is their call.
Online
Gave me a reason to take a break from trying to get some packages to build:
diff --git a/trunk/PKGBUILD b/trunk/PKGBUILD
index e47fab8..ca48561 100644
--- a/trunk/PKGBUILD
+++ b/trunk/PKGBUILD
@@ -2,7 +2,7 @@
pkgbase=linux
pkgver=5.12.2.arch1
-pkgrel=1
+pkgrel=1.1
pkgdesc='Linux'
_srctag=v${pkgver%.*}-${pkgver##*.}
url="https://git.archlinux.org/linux.git/log/?h=$_srctag"
@@ -18,6 +18,7 @@ _srcname=archlinux-linux
source=(
"$_srcname::git+https://git.archlinux.org/linux.git?signed#tag=$_srctag"
config # the main kernel config file
+ disable_vrr.patch
)
validpgpkeys=(
'ABAF11C65A2970B130ABE3C479BE3E4300411886' # Linus Torvalds
@@ -25,7 +26,8 @@ validpgpkeys=(
'A2FF3A36AAA56654109064AB19802F8B0D70FC30' # Jan Alexander Steffens (heftig)
)
sha256sums=('SKIP'
- '6dde032690644a576fd36c4a7d3546d9cec0117dd3fb17cea6dc95e907ef9bef')
+ '6dde032690644a576fd36c4a7d3546d9cec0117dd3fb17cea6dc95e907ef9bef'
+ '6b1e488299cee519b8ec891d27cc6c19f6ae35bbc0364f39817ca5a3c0104391')
export KBUILD_BUILD_HOST=archlinux
export KBUILD_BUILD_USER=$pkgbase
diff --git a/trunk/disable_vrr.patch b/trunk/disable_vrr.patch
new file mode 100644
index 0000000..cfc533d
--- /dev/null
+++ b/trunk/disable_vrr.patch
@@ -0,0 +1,13 @@
+diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
+index cb62ddba2035..614f1c5c607b 100644
+--- a/drivers/gpu/drm/i915/i915_drv.h
++++ b/drivers/gpu/drm/i915/i915_drv.h
+@@ -1749,7 +1749,7 @@ tgl_revids_get(struct drm_i915_private *dev_priv)
+
+ #define HAS_DISPLAY(dev_priv) (INTEL_INFO(dev_priv)->pipe_mask != 0)
+
+-#define HAS_VRR(i915) (INTEL_GEN(i915) >= 12)
++#define HAS_VRR(i915) (false)
+
+ /* Only valid when HAS_DISPLAY() is true */
+ #define INTEL_DISPLAY_ENABLED(dev_priv) \
https://drive.google.com/file/d/1Vjj_rS … sp=sharing linux-5.12.2.arch1-1.1-x86_64.pkg.tar.zst
Online
I applied 'disable_vrr.patch' and built linux-5.12.2.arch1-1.1, but the issue was not fixed.
Offline
https://gitlab.freedesktop.org/drm/intel/-/issues/3448 looks to be the upstream bug report for the issue.
You could try building drm-tip / drm-intel or bisect between 5.11 and 5.12 or wait for an upstream response. I could also build one of the kernels if it helps.
Online
There's a visually displaced ghost and a tearline in the video on the bug…
What if you drive the output at a lower resolution?
video=1920x1080@60
Online
The issue has been fixed by linux-5.12.3.arch2-1. (NOT arch1)
Last edited by tattsan (2021-05-14 04:24:57)
Offline
After updated system with pacman, this issue is gone at linux-5.12.4-arch1-2, thanks for Arch maintainers and everyone in this thread.
Offline