I did a dnf update today and the above kernel was installed. I rebooted my laptop as I had issues with the Android emulator. From the grub menu I chose this kernel option and then after about a minute or so, the system silently stopped.
I tried rebooting 2 or 3 times and the same issue happened. So I chose the previous kernel (kernel-6.3.12-200.fc38.x86_64) from the grub menu and everything works fine.
Since I am not very familiar with Linux, do let me know what additional information will be required to help me with this.
From the grub menu, select the latest kernel, hit e to edit, and remove rhgb quiet from the linux line, then press ctrl+x to boot. This will make sure you see all startup messages. See where it stops, there should be some error message displayed. File a bug against kernel, attach your photo, and also include details about your hardware.
1 Like
You may be experiencing a bug regarding TPM. Try this:
Boot into previous kernel from grub menu.
Terminal, run sudo grubby --update-kernel=ALL --args=âtpm_tis.interrupts=0â
Reboot into 6.4.4 and it hangs for a split second, then arrives at the expected screen successfully.
I ran into this issue. It is supposed to be fixed with either 6.4.5 or 6.4.6. Adding the kernel argument allows me to boot into 6.4.4. Hope it works for you.
2 Likes
Thanks for the tip. But itâs not a good idea to advise people to permanently modify their kernel options right away. Especially when you havenât provided information how to revert it afterwards. Itâs better to ask them to test it by adding the option temporarily to a single boot only. (That can be done by following my advice above, just instead of removing options, they would add a new one).
You almost get the same when hitting the escape key when you see the spinner spinning.
That cancels the pretty splashscreen, but doesnât show you debug messages. It might or might not be sufficient to see the error. (Also, the system might be already stuck at that point, not allowing you to hide the splashscreen). But yes, as the first thing to do, itâs certainly useful.
Thank you all for the various suggestions. I tried all of them 
+ 1 more I could think of:
i. I hit escape - nothing happened - I did not see any of the messages
ii. I removed the âquietâ flag - i did see a fast scroll of the boot messages and then screen blanked out (I did this thrice to check if I can spot any issues but it was too fast). I have a feeling that when the GUI takes over, it goes blank (I could see the indicator on my laptop going from red to white but the keyboard was working i.e. caps lock light went on/off, I was able to control the keyboard backlight etc.)
iii. I removed quiet and added tpm_tis.interrupts=0 to the boot flags
iv. I have nvidia driver, so I removed the nvidia drm(??) related flags from the boot
Exact same symptoms - so changing the boot flags did not work
I checked the boot logs using journalctl -b but did not find anything odd between 6.3 and 6.4 boot logs. I should probably do a diff and see if i can spot any change - will try that next.
Hi,
I had the same issue. I had previously installed proprietary nvidia drivers and as a part of that, I modified /etc/modprobe.d/blacklist.conf, where I put âblacklist nouveauâ line. When I remove this line, the boot starts working. Not sure if it helps you as well, but you can try.
It seems that your system boots fine, and fails to start the graphical session (thatâs the blank screen). Do a failed boot, wait a few minutes, then reboot into a working kernel, and run journalctl -b -1. That will show you the system journal for the previous (failed) boot. You can save it journalctl -b -1 > journal.txt. Find the error there.
You should have warned us that you have an nvidia card with proprietary drivers. Itâs likely the cause of the problem. Unfortunately I canât help you much in this case.
Hello Kamil,
For the past two years, I have not had a single issue with Fedora and hence it never even occurred to me that nvidia cards could be a problem. I did take a look at the failed logs but do not find any nvidia related differences between the good and bad boots.
I tried to do a side by side comparison of the two logs and was able to do it for about 60% of the logs after which it became a bit disordered and I could not track them clearly. Here are the notable difference to the point i compared:
Additional message in 6.3 kernel logs (these are not found in the 6.4 kernel logs):
Jul 26 17:01:15 fedora kernel: zswap: loaded using pool lzo/zbud
Jul 26 17:01:15 fedora kernel: r8169 0000:04:00.0: canât disable ASPM; OS doesnât have ASPM control
Jul 26 17:01:16 fedora kernel: fbcon: i915drmfb (fb0) is primary device
Jul 26 17:01:16 fedora kernel: fbcon: Deferring console take-over
Jul 26 17:01:16 fedora kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device
Additional messages in 6.4 (not found in 6.3 kernel logs)
Jul 25 17:35:47 fedora kernel: i915 0000:00:02.0: [drm] Cannot find any crtc or sizes
Jul 25 17:35:48 fedora kernel: i915 0000:00:02.0: [drm] Cannot find any crtc or sizes
Another difference I found:
6.4: Jul 25 17:35:48 fedora kernel: EXT4-fs (dm-0): mounted filesystem 0d77e1aa-14d9-4860-8ed7-fc5c2446e47d ro with ordered data mode. Quota mode: none.
6.3: Jul 26 17:01:18 fedora kernel: EXT4-fs (dm-0): mounted filesystem 0d77e1aa-14d9-4860-8ed7-fc5c2446e47d with ordered data mode. Quota mode: none.
â 6.4 seems to be loading the â/â filesystem in ro mode?
Thanks
raju
i915 shows that it uses an Intel GPU, not Nvidia. Upload the log from the failed boot somewhere and link it here.
Hello Kamil
Here it is
https://drive.google.com/file/d/1NyaEbm9er-uvqC5LU7LlnZeEWi9YZ_c2/view?usp=sharing
Thank you so much for helping out.
raju
So youâre using lightdm (so this is neither Gnome nor KDE, what is it?), but I donât see any errors from it. Youâre also using both Intel and Nvidia GPU, and youâre using nouveau for nvidia. Both report
Cannot find any crtc or sizes
which seems to mean that no connected displays were found. Perhaps this is to blame, I donât know:
Jul 25 12:05:51 fedora kernel: nvidia-gpu 0000:01:00.3: i2c timeout error e0000000
I canât really help with nvidia issues, I just donât know. If itâs possible to switch to a pure Intel GPU in BIOS (or remove the nvidia card), do it. Since this is a laptop, that is unlikely. It might help to try to reinstall the proprietary driver, possibly over ssh (since the display doesnât work for you). Other than that, somebody else has to help, sorry.
Hello Kamil, I am using the Cinnamon spin.
I will remove the nvidia driver and try to reboot in 6.4, using this method: https://askubuntu.com/a/1289997 - it should work in Fedora also, right?
BTW, I found this in kernel.org: https://bugzilla.kernel.org/show_bug.cgi?id=206653#c9 regarding the above error. Would you advise that I update the bug with this issue?
Thank you for trying to help out here - really appreciate it.
I went ahead and made the change and rebooted. The error ânvidia-gpu 0000:01:00.3: i2c timeout error e0000000â is gone now but i still get a blank screen in 6.4. The diff between 6.3 and 6.4:
6.4:
Jul 28 02:11:53 fedora kernel: input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:13/LNXVIDEO:01/input/input16
Jul 28 02:11:53 fedora kernel: vga_switcheroo: enabled
Jul 28 02:11:53 fedora kernel: i915 0000:00:02.0: [drm] Cannot find any crtc or sizes
Jul 28 02:11:54 fedora kernel: [drm] Initialized nouveau 1.3.1 20120801 for 0000:01:00.0 on minor 0
6.3:
Jul 28 02:12:40 fedora kernel: input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:13/LNXVIDEO:01/input/input14
Jul 28 02:12:40 fedora kernel: vga_switcheroo: enabled
Jul 28 02:12:40 fedora kernel: fbcon: i915drmfb (fb0) is primary device
Jul 28 02:12:40 fedora kernel: fbcon: Deferring console take-over
Jul 28 02:12:40 fedora kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device
Jul 28 02:12:41 fedora kernel: [drm] Initialized nouveau 1.3.1 20120801 for 0000:01:00.0 on minor 0
I believe that one would have better performance with using the nvidia drivers. Is there a reason you did not install them?
What is the return from mokutil --sb-state
Installing them is very easy, but we need to know if secure boot is enabled or disabled before we can know the exact steps to follow.
The nouveau drivers do not enable use of the hardware acceleration with nvidia GPUs and that can cause significant overload of the CPU for graphics rendering and thus slowdowns.
$ mokutil --sb-state
SecureBoot enabled
I had not realised I was not using nvidia. I had originally set it up long back when the laptop was still new (about 2 years back) and have not done any changes since and assumed that nvidia continues to be used.
These were the commands I had run when I originally set it up:
sudo dnf install kmodtool akmods mokutil openssl
sudo kmodgenca -a
sudo mokutil --import /etc/pki/akmods/certs/public_key.der
sudo dnf install gcc kernel-headers kernel-devel akmod-nvidia xorg-x11-drv-nvidia xorg-x11-drv-nvidia-libs xorg-x11-drv-nvidia-libs.i686
sudo akmods --force
sudo dracut --force
Ok, lets check a little deeper then.
dnf list installed '*nvidia*'
and dnf repolist
and lsmod | grep -E 'nouveau|nvidia'
EDIT
Add to that sudo mokutil --test-key /etc/pki/akmods/certs/public_key.der
I have a similar problem, but discovered if Iâm really patient, it ends up booting. On my laptop I wait for the system to go to sleep, then I tap the power button and the login screen appears (still on 6.4.4). I described this more here.
It would be interesting to see if the behavior is the same for you, if you wait a long time (leave it overnight?) does waking the laptop up give you the log in screen.
Like you, none of the the recommended ways to get debug info worked for me. It appears to freeze up before getting the chance. (Do you just get a single underscore on the screen?)