Linux kernel update

Mauvve Knight

Member
Joined
Jan 10, 2026
Messages
35
Reaction score
30
Credits
398
Update Manager from Linux Mint shows an update of kernel 6.17.0-19.19~24.04.2.
Schermafdruk van 2026-03-13 21-24-19.png


I'm not sure it is wise for me to install this update because I'm running 6.19..3-1-t2-noble.
Schermafdruk van 2026-03-13 21-28-03.png


Can somebody hand me some advise? I'm worried the newest kernel will be loaded at boot so I cannot boot Linux anymore on my MacBook Pro 2020 (macbookpro16,2).
 


The newer kernel will be loaded at boot. At least two older kernels should still be in place. If the new kernel fails to boot, you can select an older one when you reboot and see the initial screen. Upgrading the kernel is completely normal and seldom causes any issues, although it sometimes happens, mostly issues with programs, not booting. But that's why you always keep the previous kernel in place, and preferably the previous two.
 
The newer kernel will be loaded at boot. At least two older kernels should still be in place. If the new kernel fails to boot, you can select an older one when you reboot and see the initial screen. Upgrading the kernel is completely normal and seldom causes any issues, although it sometimes happens, mostly issues with programs, not booting. But that's why you always keep the previous kernel in place, and preferably the previous two.
So this means my system will be booted with kernel 6.17.0-19.19~24.04.2 instead of 6.19.3-1-t2-noble? When this is true I am in trouble because the T2-chip in my laptop will not be bypassed and boot will halt.
 
So this means my system will be booted with kernel 6.17.0-19.19~24.04.2 instead of 6.19.3-1-t2-noble? When this is true I am in trouble because the T2-chip in my laptop will not be bypassed and boot will halt.
The next boot after the updated kernel is installed does not necessarily have to boot the new kernel.

If you don't change the kernel to be booted, the newly installed kernel will likely be booted BUT, you can select whichever kernel you like at boot time, and if you want that current kernel (as opposed to the new one) to be booted persistently, you can configure grub to boot it.

The grub setting for booting a kernel can be configured in the file: /etc/default/grub. That setting is usually GRUB_DEFAULT=0, which means that the first kernel listed in the /boot/grub/grub.cfg file will be booted. Grub kernel numbers begin at 0 for the first, and 1 for the second and so on.

If you wish to upgrade, and thus have a new kernel which may come with that, but then still boot the current kernel, there are a number of alternatives.

One can stop the boot at the grub menu and select a kernel from the menu. That's done by hitting the shift key for BIOS or the esc key for UEFI, to bring up the grub menu. Then navigate to the kernel you wish to boot. This is a one-time approach.

Alternatively one can configure grub to boot the kernel one wishes to boot so that grub boots that kernel persistently until one alters the configuration again, or selects the grub menu.

If after the update and upgrade, a new kernel is installed, as shown to come in post #1, then after the upgrade, one can reconfigure grub as follows before rebooting to have the current kernel boot next time, and not the new kernel.

The quick approach is to assume that the new kernel will be the number 0, and current kernel will be the number 1, so one can change the setting in /etc/default/grub to: GRUB_DEFAULT=1, and update grub. The on rebooting, the current kernel should be booted again.

The longer check on this process is to inspect the /boot/grub/grub.cfg file to see the order of the kernels that grub will observe, and check that the kernel you want is indeed the second one listed, in which case it will certainly have the number 1 which can be used in the grub configuration to be applied in the /etc/default/grub file. For example, on this machine one can run, in a terminal, as root, the following to see the kernels on the system:

Code:
[root@min] # grep vmlinuz /boot/grub/grub.cfg
    linux    /boot/vmlinuz-6.18.15+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.18.15+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.18.15+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr
        linux    /boot/vmlinuz-6.18.12+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.18.12+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr
        linux    /boot/vmlinuz-6.18.9+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.18.9+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr
        linux    /boot/vmlinuz-6.18.5+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.18.5+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr
        linux    /boot/vmlinuz-6.17.13+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.17.13+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr
        linux    /boot/vmlinuz-6.17.12+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro
        linux    /boot/vmlinuz-6.17.12+deb14-amd64 root=UUID=7bb5cb7c-e217-4ab6-b74e-31f25581b4ea ro single dis_ucode_ldr

The first kernel in the list is 6.18.15, which will be allocated the number 0, and the second kernel is 6.18.12, which will be number 1. (You can ignore the double entries for each kernel showing twice ... that's just showing the same kernel for full boot mode and corresponding recovery boot mode).

After such checking, the same configuration alteration as mentioned in the quick approach above will enable the current kernel to be used on the next boot ... make the change in /etc/default/grub and update grub.

There are other approaches, like holding the current kernel and configuring apt not to install a new kernel, but I don't recommend that. Newer kernels may eventually deal with the issues you have which could be tested for each new kernel, and if found to not yet succeed, still return to the kernel that works.

EDIT: The above approach is valid for the current situation described in post #1. However, since distros do upgrade often, the question arises as to how to maintain the same kernel when newer kernels are installed by subsequent upgrades.

When a second upgrade installs a second newer kernel, the current kernel will no longer be number 1, but rather number 2, which could be checked with an inspection of the /boot/grub/grub.cfg file as above. Then the configuration adjustments could be made.

Alternatively, grub can be configured to use the current kernel, despite newer kernels being installed by the use of the following configurations in /etc/default/grub: GRUB_DEFAULT=saved, and GRUB_SAVEDEFAULT=true. If those configs are set, and grub upgraded with the current working kernel, the machine should boot the current kernel indefinitely until changes to the configuration are made regardless of the number of kernels installed. One needs to ensure that the kernel is not deleted of course by the use of apt's cleaning commands.
 
Last edited:
For fun I went and looked at my /boot/grub/grub.cfg file on my Mx Linux install.
The newest kernel that was just updated is now booting.

Code:
    echo    'Loading Linux 6.1.0-44-amd64 ...'
    linux    /boot/vmlinuz-6.1.0-44-amd64 root=UUID=3a948b62-58b5-4de0-bc73-13ff61f6a72b ro  quiet splash
    echo    'Loading initial ramdisk ...'

Code:
@mx:~
$ sudo grep vmlinuz /boot/grub/grub.cfg
    linux    /boot/vmlinuz-6.1.0-44-amd64 root=UUID=3a948b62-58b5-4de0-bc73-13ff61f6a72b ro  quiet splash

Good information for the kernel @osprey, thanks!
Havea nice weekend!
 
So this means my system will be booted with kernel 6.17.0-19.19~24.04.2 instead of 6.19.3-1-t2-noble?

No. Newer as in higher version number. Not the newest one installed (if it's a lower version). At least that's my understanding.

Anyway, just don't install it since you're running a higher version series (6.19 vs. 6.17). I assume you keep it up to date?
 
Last edited:
Lol, you mae gewer than I ddo.

Avagudweegend. ;)
 
I tend to keep up with the kernel versions, and just rebuild them removing flags and adding flags to get rid of the stuff my system don't need, and just make it better suited for my specific system and wants. It can be tedious and i could just run generic and deal with it that way. Linux is cool. especially when we dive into it a little further than click and go kinna thing :)
 
My Mint 22.3 started with the 6.8 (LTS) Kernel with what seems monthly upgrades and all were supported till some time past 2028. Are some installs selecting the short lived 6.17 by default?

Why two code bases? Is it an AMD vs. Intel issue or something else?
 
Well, I've tried to update the 6.17 kernel from Update Manager but I ran into the problem GRUB altered the boot options so my system would boot with kernel 6.17 instead of kernel 6.19.3-1-t2-noble. Happily did I see it during reboot of my system in rEFInd so I chose booting with kernel 6.19 to prevent running into problems with T2-chip.

After rebooting with correct kernel I restored my system to yesterday with TimeShift so I can go ahead if nothing changed. In Update Manager I have set I don't want to be informed about new versions.

Thanks to all for your explanations and advise!
 
Thanks all for your replies!

I've set the status of this topic back to "no prefix" because I'm anxious to see what happens after update to the new version of Linux Mint.

Because I use rEFInd bootloader I can experiment with kernels without running into problems (I'm running a T2-noble kernel because of MacBook Pro from 2020).

When booting I can choose which kernel I'd like to boot, so when a kernel-update alters my default settings I can get back to booting with correct kernel.

This means I don't have to reset my systems with TimeShift. My solution for fixing booting with correct kernel is this (using rEFInd):
  1. Ensure you have the rEFInd package on your system.
  2. After updating kernel using Update manager restart your machine.
  3. Check if correct kernel version is displayed under Linux Mint logo. If not go to step 4.
  4. Press Tab-key to get into advanced details menu.
  5. Choose correct kernel version you'd like to use and press Enter.
  6. Now your machine boots with that kernel version.
  7. After complete reboot startup Terminal.
  8. Go to directory where your rEFInd files are stored.
  9. Type sudo ./refind-install and press Enter.
  10. rEFInd begins to (re)install.
This works well for my machine but I think it is worth trying for other users. I'll keep you informed about situations which happen after crucial updates like new Linux Mint version.
 
I have found a solution for this.

Using rEFInd kernel which has to be used during boot isn't set default in rEFInd. You can put this in file refind-linux.conf but this is tedious work and you have it to keep it up-to-date as your kernels. A shortcut which has been implemented in rEFInd by its developer works with filedates instead of kernel versions. This means if kernel 6.19.0.0-1 has been loaded on your system December 24th, 2025 at 4:58 pm and afterwards another updated kernel 6.8.2-2.3 has been loaded at Januari 7th 2025 at 11:03 pm rEFInd will load kernel 6.8.2-2.3 instead of kernel 6.19.0.0-1 because of different filedates. Yes, you can choose which version to loaded by pressing the Tab button but you have to press at each reboot of your computer.

But by error and trial I found the following solution:

  1. When you have updated an older kernel but you don't want to use this for everyday use you can reboot your system.
  2. In rEFInd boot screen you choose Linux Mint and press Tab or Esc key.
  3. Select the kernel version you want to start with and proceed booting.
  4. Now login into your Admin-account (the account you use as superuser to add accounts etc).
  5. Start Terminal and enter:
    Code:
    cd \boot\
  6. In this folder you'll find the kernels which are installed on your computer. Now you can alter the filedates by entering
    Code:
    sudo touch vmlinuz-{version} initrd.img-{version} /boot/config-{version}
    So, if you'd like to keep kernel 6.19.0-1 as default you have to enter
    Code:
    sudo touch vmlinuz-6.19.0-1 initrd.img-6.19.0-1 config-6.19.0-1
By using touch {file} you give the file the current date and time. When you reboot your system you'll see in rEFInd main screen the 'latest' kernel as default underneath your Linux logo.

For me this is a good workaround because I'm always able to select another kernel during bootup within rEFInd. I hope this will work for other users. To more experienced users: feel free to correct or add!
 
Far to complex for me.
I have the default kernel in place. It works, it does what it is supposed to do.

If a problem arises, and it is plainly the fault of the current kernel or if a kernel update has bought about problems, I will change it then.....otherwise, it isn't broken so it does not get touched
 
I don't see the point of multiple kernels either. I do keep two previously known good kernels in place, just in case a new kernel won't boot, but that's rare. As long as the system boots, I just let it boot the newest, which has the latest security updates and features. Booting an older kernel is easy enough if necessary, and I don't need a third-party method of doing that. But I have no interest in how someone else maintains their system, I do what works for me.
 
Agreed ^^^^^....what he said ^^^^^

k.i.s.s.
 


Follow Linux.org

Staff online

Members online


Top