Fully agree. I remember facing issues also with 6.13 kernel - 6.13.9 kernel bug with netfs: Can't donate prior to front in the past …
1 Like
Came here looking for the issue and post about if not found. Glad I found that it is a known issue and fix is hopefully coming soon. Before I even found the issue here I already figured to try older 6.15.10 kernel and with it things work fine. This considering that I am by no means a computer or Linux expert, just your average everyday “Windowsy” user.
This also in mind, I have been dual booting with Windows (10) on the side and the grub menu shows up by default everytime I boot. As such trying an older kernel basically required no extra steps for me. However the grub menu retains only 3 kernel versions with the oldest one dropped when a new one is installed. And as today I updated to the 6.16.4 kernel, the 6.15.10 is now the oldest one. Is there a way to retain the 6.15.10 even when the next update comes? Just in case it doesn’t fix the network issue after all.
My PC has a Gigabyt B550 AORUS ELITE V2 motherboard.
The currently booted kernel is never removed. So the easiest solution is to run on 6.15.10 when triggering the update (online or offline). Kernel 6.15.10 will be kept, and the second oldest one will be removed instead.
2 Likes
If you want to be sure to not accidentally boot the newer kernel and subsequently accidentally remove 6.15.X, you can increase the amount of kernels that are retained: In the file /etc/dnf/dnf.conf you find the line installonly_limit=3 that determines how many kernels are kept installed. Hopefully, your issue will be solves soon and 4 or 5 will suffice if you do that → if the /boot partition has the default size (1GB), that space should suffice for 5 (I have always 5 kernels on a 1GB /boot, currently still 229MB are free)
However, keep in mind that you can also re-install older kernels if they had been accidentally removed. So such an accident is revertible.
3 Likes
dnf5 has versionlock.
Recently hopped over to Fedora (42 Workstation KDE) and seem to have a similar issue.
However I’ve tried the earlier kernel it installed with and the 6.14.0 one had the same problem.
$ rpm -q kernel
kernel-6.14.0-63.fc42.x86_64
kernel-6.16.4-200.fc42.x86_64
Using:
05:00.0 Ethernet controller: Aquantia Corp. AQtion AQC113CS NBase-T/IEEE 802.3an Ethernet Controller [Antigua 10G] (rev 03)
Subsystem: ASUSTeK Computer Inc. ASUS XG-C100C 10G PCI-E Network Adapter
[…]
Kernel driver in use: atlantic
Kernel modules: atlantic
That being said. I’ve introduced a ton of variables of what it could be.
(Router firmware upgrade, changed ISPs, distro hopping…)
But reading these threads and seeing that Windows 10 seems unaffected on the same PC, card and network. It may be related.
Seeing random periods of a few seconds where packets drop.
Whether it’s https://speed.cloudflare.com/, random websites failing to connect, ping failing, or twitch live streams dying.
The Twitch one has been the go-to test because it doesn’t recover. Can just let it run and come back half an hour later to see if it failed.
I’ll need to test more after a trip, but I have a few other network card brands and systems to try with.
versionlock didn’t work for kernels when I tried it a few months ago. The reason for this could be that kernels are installed/removed and not updated.
Then it’s not the same issue.
Indeed probably isn’t. So I’m also not holding my breath for the 6.16.5 kernel to address it and I’ll need to narrow it down more to see it’s not a different regression.
6.16.5 doesn’t solve the 6.16.3 issue for me.
It seems to have solved it for me. It’s been working as good as with kernel 6.15.10.
1 Like
Kernel 6.16.5 is working for me, too. I use the Intel AX200 wifi.
You can reopen the bug or file a new one.
6.16.5 seems to have fixed the issue for me. Haven’t had issues since I updated half a day ago. Thanks for the devs who fixed it.
Thanks. Still seeing the issue with 6.16.6, so I filed 2394582.