You are not logged in.
After upgrading the kernel from 5.10.16 to 5.11.x my cursor in X does not render correctly (see image 1 below). Oddly, if I take a screenshot with the tool maim the cursor is seen as normal in the screenshot (image 2).
I believe I am using the xf86-video-intel drivers. Downgrading the kernel to 5.10.16 restores the cursor.
Has anyone else encountered a similar issue, or perhaps has troubleshooting suggestions? I have not submitted a bug report as of yet.
Offline
This bug you are seeing is caused by xf86-video-intel. You can remove it, your graphics will also work without xf86-video-intel installed.
Offline
Is this a known issue with xf86-video-intel then? Or how else can you be sure on this. I can't see any active bug reports for the package. Does the package not convey some benefit to my graphics e.g. hardware acceleration?
Offline
It doesn't really and it's in a maintenance limbo since ~8 years or so. Many distributions have switched to using the xorg built-in modesetting driver for newer CPU chipsets.
It does have better 2D xorg perf in theory and doesn't need a dedicated compositor to avoid screen tearing but it often has a variety of weird display quirks in a lot of cases.
Last edited by V1del (2021-03-24 17:19:15)
Offline
That's useful to know V1del, thank you. I don't find the Wiki page on Xorg too clear on this.
I have uninstalled xf86-video-intel and upgraded to the latest kernel, 5.11.8. The cursor is still broken, so it seems like it was not this package causing the issue.
Offline
Can you please post your xorg log?
Cause "I believe I am using the xf86-video-intel drivers" isn't really convincing and there're multiple records of this pattern that are caused by the nouveau driver (nvidia chip), not intel.
Offline
This is my current Xorg log on 5.10.16 - I can update the kernel (to reintroduce the cursor issue) and paste the log then if that might be helpful.
https://gist.github.com/piperfw/c3d2afe … eb834bf191
Cause "I believe I am using the xf86-video-intel drivers" isn't really convincing and there're multiple records of this pattern that are caused by the nouveau driver (nvidia chip), not intel.
Yes, I was deliberately vague as do not understand exactly what drivers are being used - apologies. I do have a Nvidia card:
$ lspci -v | grep -A1 -e VGA -e 3D
01:00.0 VGA compatible controller: NVIDIA Corporation GK104 [GeForce GTX 670] (rev a1) (prog-if 00 [VGA controller])
Subsystem: CardExpert Technology Device 1189as well as the mesa package, but not any of the Nvidia drivers (proprietary or nouveau).
Offline
You're currently running on the VESA driver on the (only, I assume) nvidia VGA, which is about the worst situation to be in.
Install the nvidia blob, the bug is caused by the nouveau kernel module.
Offline
Thank you for looking at the log file.
Apart from this one artefact, I have not encountered any issues.
Please may you specify what you mean by the nvidia blob?
Offline
Offline
He means you should install the Nvidia driver. Check out the "Nvidia" article in the ArchWiki.
The "blob" word means "Binary Large OBject". The word here was supposed to mean that the important part of the Nvidia driver package is one large binary file where Nvidia keeps the source secret.
Offline
Thanks for the link. For my card that page prescribes the nvidia package.
Is the suggestion that I won't be able to remove the bug while sticking with the nouveau drivers (I have mesa but not xf86-video-nouveau installed; see Nouveau)?.
I would prefer to stick to the open-source drivers if possible.
Offline
Use the lts kernel until the bug is fixed in the current kernel, there're no known workarounds.
Offline
Ok, is the correct procedure to install linux-lts, then uninstall linux? I was surprised pacman doesn't say the two conflict.
Offline
That's because they do not conflict, you can have both of them installed and switch between them, the only thing you have to do is change your bootloader config to give you an entry. You'll have to refer to vmlinuz-lts and initramfs-linux-lts, check your /boot
Last edited by V1del (2021-03-25 15:58:59)
Offline