not sure what and where things happening or what has been done, but in past 2 releases i can see deprecation of product quality and stability. Default apps crashing, not working, Desktop crashing not working, Pushing new things and latest in hope things are back and fixed making things brake even more. i am running on this time 2 laptops one on F39 Kinoite and one F41 Kinoite and on F39 is way more stable and not even crashing or having quircks on same apps that just keeps crashing, shutdown, no error messages just shutdown, closing, minimizing windows and then opening those no rendering anymore just black screen and need to refresh start again
latest release rushed and released early and if this is the state of coming releases it wont be good on long run
I have the impression that mainly the amount of kernel driver bugs have increased considerably in the recent series of kernels. There have been many issues mostly with gpu drivers from all major vendors that have been quite persistent.
My integrated AMD gpu hasn’t been reliable for many months, for a while resulting in black screens and full system halts. Other users have experienced recurring problems with dpms and suspend routines. Currently i don’t experience full system crashes anymore but only see occassional screen glitches, this seems to be an improvement.
The question is if Fedora can really do anything about this, since the distribution just rolls on and can’t stop kernels from hitting the repositories for so long.
What would in your view be a good way to deal with these issues?
isint there reason why there is testing and QA before shipping on as Stable release??? i mean there hasent been much stability on systems latly and all just switching and changing and updating i eel like it is on ARCH train here with more unstability and crashing than actually i have on ARCH
stable should be considered minimal affects on systems, but yes there can be and will be bugs still, but lets say example the GTK4 issues on Gnome/workstation/silverblue that still shipped as it is as stable and even release version
I wonder how specific a Fedora user’s experience will be dependent on there hardware and spin.
I consider f41 to be the lest buggy release in recent times.
F41 works on my rpi4, but f40 does not.
Upgrading my fedora server exposed one bug, which was my error.
My desktop just worked, amd gpu and kde plasma.
We all see a lot of issues, here, from people with Nvidia GPUs, but that is not under anyones control but Nvidia’s.
Regression on hardware is not new in the kernel.
It has always depended on people testing changes at each release.
I was expecting a lot more volume of questions here when f41 released, but it seems far less then f40 or f39. But I do not have metrics to back up this feeling.
The latter should have been beta-tested somewhere, at least before reaching every mainstream distro, who apparently are content with the quality of this reaching end-users.
And that GSK_RENDERER thing F41 is particularly odd. There’s no way nobody at dev saw issue with flipping that to vulkan. There’s no way that wasn’t mentioned during beta testing. And yet end users got it. If Fedora wanted to test it, what’s the point of Rawhide? And although it’s pretty loose, this large of a change is interesting before F42’s opt-in telemetry and might make for a case to justify it.
Fedora is known for the latest-and-greatest, but as of the last few releases I’m kind-of questioning how much of that entails it working consistently? I run a server and workstation, prefer a consistent OS, and although Fedora Server has been mostly fine, I’ve lost confidence in Workstation as my primary OS starting with that GNOME log-in thing (it started on Fedora, and I couldn’t even escape it when it inevitably came to Ubuntu and openSUSE).
But simply, I’m not happy with:
With that GNOME thing happening since F38/F39
Now see KDE as a primary edition (with GNOME still doing the above and below)
The latest news with GNOME RT being apparently guessed
Seeing a general push for Flatpaks and Immutable (the basics aren’t even perfected yet!)
All the above happening somehow with limited man-power, with “maintenance burden” being used to justify axing Xorg over Wayland (it’s likely I have a later issue with the governance model in-general vs the specific topics)
The Wayland push for years from other users particularly bugs me with Linux in-general. I know what I’ve seen for years up until F41/GNOME 47 on my own hardware, and am pretty confident there’s people who openly-praised worse conditions; I’m just not sure what the gain is for anyone there. To shorten this, FreeBSD’s no-nonsense DIY approach and community appeals to me, and I’m still eagerly waiting for a bit of downtime to switch my server to it (but credit for Fedora Server 41 working well in the meantime!)
Buy up some of MS’s apparently-fired Windows 10 QA
I am sure they can need help, they are all volunteers afaik.
There is OpenQA by OpenSUSE, which Fedora ported to use for Fedora. There are ways you can contribute by testing on real hardware, as this is always a bit complex.
The more variables you introduce, the more complex it gets.
Linux is being used across a wider range of hardware than any other OS, from personal systems too old to run current versions of Windows or macOS up to data center systems with newly introduced CPU’s and associated chips. While major vendors do appear to working towards large deployments of linux workstations, Windows still dominates new PC sales.
While linux does work on a much larger range of hardware than any current commercial OS, it is impossible to ensure that it will run well on every PC still in working order along with every newly introduced hardware configuration. I have encountered a few kernel bugs that were simple coding errors introduced as inadvertent side effects of changes (made to support newer hardware or mitigate security issues) that only affected systems over a decade old. All those bugs had already been reported by other users with similar hardware by the time I encountered them.
In summary, I expect the number bugs affecting older hardware will continue to increase as the stock of older hardware grows. Bugs affecting recently introduced hardware are largely due to vendors that focus on Windows support at the expense of linux. Those who want to use linux on very new hardware need to look for vendors that support linux. Those looking for a basic reliable system should consider purchasing used systems or a new system of a model that was introduced 2-3 years ago and has good linux support based on probes in the LHDB.
Going forward, the linux user community needs to devote more attention to making it easier to determine the status of linux on particular hardware. The LHDB is very useful, but needs more participation by users to indicate which hardware components they actually use, ways to post workarounds for specific issues, and vigorous debunking of the growing number of misleading “bug” reports blaming the linux kernel for problems generated by hardware vendors, etc.
(but credit for Fedora Server 41 working well in the meantime!)