If the r460 drivers are currently working and you don't have any projects that require the r470 drivers like GPUGrid, then I would stay pat now and first fix the issue you have with the kernel.

Hi Keith,

I had a problem with 470 on my last install of Mint. The display was said to be a Laptop. This time when I installed Mint I was hoping that 470 was "fixed". I'm thinking that it is not broken, just that it sees my machine differently than 460 does. I will be sticking with 460. :-)

Are you saying that running apt autoremove and autoclean did not work? Autoremove needs an acknowledgement "y" to remove the old programs. Autoclean does not.

Yes, they did nothing that I could see. I don't remember if autoremove asked for confirmation or not. I'll have to check again and let you know. Ok, just checked and it does not ask for confirmation.

You can also remove old programs with the Synaptic Package Manager. Did you try that method?

I went into Synaptic and holey moley, there are about a million files for the Kernels, well maybe not a million, but hundreds anyway. I haven't figured out how to remove ALL the files for all but 3 of the most current Kernels yet. By the way, the /boot partition is where the Kernel is stored, correct? I'm pretty sure I sized it 1 GB. I have / at 100 GB and /home takes up the remainder of the 1T NVMe drive.

I would not attempt to update the kernel until you can straighten out all the old kernels still hanging around in the apt-cache.

Hadn't plan on it. ;-)

Remember, the video drivers are compiled into each kernel. So if you have kernel issues which you seem to have, that leads to the likelihood of having issues upgrading the video driver.

Is it a general notion to only keep 2 or 3 working Kernels? If so, why would the install put all those Kernels on; why not only the 3 most recent? I don't get it. :-\

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

A kernel comprises only 5 files. Two header files, an image, module and module-extra file.
There are of course dozens and dozens of kernel versions available in the distro.
You are only interested in either the latest stable standard or HWE versions.
I'm thinking you are being confused by Synaptic. You aren't looking for all the kernels available to be installed from the distro, just the ones you have installed in the host.
You just click on the Status button in the left hand panel and then use the Installed - Auto Removeable button.
Then use the Not Installed - (residual configurations) button and check all the older kernels to be removed. That may take a while if as you stated you have 128 older kernel configuration files hanging around.

---

A kernel comprises only 5 files. Two header files, an image, module and module-extra file.
There are of course dozens and dozens of kernel versions available in the distro.
You are only interested in either the latest stable standard or HWE versions.
I'm thinking you are being confused by Synaptic. You aren't looking for all the kernels available to be installed from the distro, just the ones you have installed in the host.
You just click on the Status button in the left hand panel and then use the Installed - Auto Removeable button.
Then use the Not Installed - (residual configurations) button and check all the older kernels to be removed. That may take a while if as you stated you have 128 older kernel configuration files hanging around.

Hi Keith,

First off, I set up Update Manager so that it will dispose of older Kernels automagically, all but the most recent working or 2.

You're correct in that Synaptic is confusing. So I am just gonna let Update Manager do its thing.

This is getting just too ridiculous! Going to v460 did not fix my video problem. I switch to the Nouveau driver and then to v460 and now on the v460 reboot I get the message about no hardware acceleration. So I have stuck to Nouveau for now. At least the display settings utility shows my monitor and not "laptop" and the video is not herky jerky. When all them obsolete Kernels are gotten rid of I will try v460 again. Any ideas on when Update Manager will purge the obsolete? Will it happen when I try to update the Kernel again? I believe that on a clean install that ONLY the 2 or 3 most recent Kernels be packaged for the install. Why the heck do they put obsolete Kernels in the install package? Makes not sense to me.

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

I had the same problem with a new install of Mint 20.2, this solution seems the have fixed it...

Use this command to open the modules file: xed admin:///etc/modules

Add this at the bottom:

nvidia
nvidia-drm
nvidia-modeset

Save and reboot

https://forums.linuxmint.com/viewtopic.php?f=59&t=331611&p=1898849#p1898849

https://forums.linuxmint.com/viewtopic.php?f=59&t=331611&p=1898849#p1898849

Hi John,

Thank you, that seems to have done the trick. After doing the edit and rebooting, I changed to v460 NVIDIA and rebooted. I checked my display and it showed my monitor. Just to be sure, I rebooted a second time, to simulate booting into Linux after running Winders, and no message in the upper right hand corner about "no hardware acceleration". I checked the display and it showed my monitor and also went into NVIDIA settings and everything seems good. Oh yeah, by the way, I had my fingers crossed both times. lol ;-)

Next step is to do the Kernel update and see about getting rid of the obsolete versions automagically. I'll do that in a day or 2 or 3 just to make sure my video problem is a thing of the past. :-)

Thanks again and have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

You can see what the error with the PCIE was with:

dmesg --level=emerg,alert,crit,err,warn

Hi Keith,

I was so wrapped up in the other issues that I forgot about this. This is what is shown:

[    0.161935] MDS CPU bug present and SMT on, data leak possible. See https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/mds.html for more details.
[    0.161935] TAA CPU bug present and SMT on, data leak possible. See https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/tsx_async_abort.html for more details.
[    0.161935]   #7  #8  #9 #10 #11
[    0.172499] ENERGY_PERF_BIAS: Set to 'normal', was 'performance'
[    0.462731] pcieport 0000:00:1c.4: AER: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
[    0.462737] pcieport 0000:00:1c.4: AER:   device [8086:a294] error status/mask=00000001/00002000
[    0.462739] pcieport 0000:00:1c.4: AER:    [ 0] RxErr                 
[    0.518956] platform eisa.0: EISA: Cannot allocate resource for mainboard
[    0.518957] platform eisa.0: Cannot allocate resource for EISA slot 1
[    0.518957] platform eisa.0: Cannot allocate resource for EISA slot 2
[    0.518957] platform eisa.0: Cannot allocate resource for EISA slot 3
[    0.518958] platform eisa.0: Cannot allocate resource for EISA slot 4
[    0.518958] platform eisa.0: Cannot allocate resource for EISA slot 5
[    0.518959] platform eisa.0: Cannot allocate resource for EISA slot 6
[    0.518959] platform eisa.0: Cannot allocate resource for EISA slot 7
[    0.518960] platform eisa.0: Cannot allocate resource for EISA slot 8
[    0.573535] resource sanity check: requesting [mem 0xfdffe800-0xfe0007ff], which spans more than pnp 00:07 [mem 0xfdb00000-0xfdffffff]
[    0.573537] caller pmc_core_probe+0x7f/0x180 mapping multiple BARs
[    0.957333] nvme nvme1: missing or invalid SUBNQN field.
[    1.111279] ata4.00: supports DRM functions and may not be fully accessible
[    1.115867] ata4.00: supports DRM functions and may not be fully accessible
[    4.647659] nvidia: loading out-of-tree module taints kernel.
[    4.647667] nvidia: module license 'NVIDIA' taints kernel.
[    4.647667] Disabling lock debugging due to kernel taint
[    4.703413] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  460.91.03  Fri Jul  2 06:04:10 UTC 2021
[    4.915877] kvm: disabled by bios
[    5.103310] kvm: disabled by bios
[    5.498284] kvm: disabled by bios
[    5.540589] kvm: disabled by bios
[    5.554915] kauditd_printk_skb: 13 callbacks suppressed
[    5.662174] kvm: disabled by bios
[    5.742727] kvm: disabled by bios
[    5.818308] kvm: disabled by bios
[    5.898109] kvm: disabled by bios
[    5.913839] nvidia-gpu 0000:01:00.3: i2c timeout error e0000000
[    5.913841] ucsi_ccg 0-0008: i2c_transfer failed -110
[    5.913843] ucsi_ccg 0-0008: ucsi_ccg_init failed - -110
[    5.913847] ucsi_ccg: probe of 0-0008 failed with error -110
[    6.113314] kvm: disabled by bios
[    6.322535] kvm: disabled by bios
[    6.730415] kvm: disabled by bios
[    6.971035] kvm: disabled by bios

Anything stand out?

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

pcieport 0000:00:1c.4: AER: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)

This is the only one alarming. It will cause your logs to swell and eventually fill your disk.

You should add this to the grub file on the kernel command line.

sudo gedit /etc/default/grub

Add

pci=noaer

at the end of GRUB_CMDLINE_LINUX_DEFAULT

It will look like this after.

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash pci=noaer"

Then

sudo update-grub

---

Hi Keith,

Did as you mentioned and I no longer see the error when the PC is booting. I did this:

dmesg --level=emerg,alert,crit,err,warn

and it no longer appears in the report.

Is that a problem with the motherboard? I went to the websites of the 2 urls at the top of that report and that kinda bothers me. Since it is 8th generation, would it behoove me to upgrade the PC MB and CPU at my age, (knockin' on the door of 70)?

Thanks and have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

No, I would not worry about it. It would take a vendor specific quirk added to the Linux-firmware package by the distro maintainers to fix the issue which is really benign, except for the growth of the logfiles I mentioned.

---

No, I would not worry about it. It would take a vendor specific quirk added to the Linux-firmware package by the distro maintainers to fix the issue which is really benign, except for the growth of the logfiles I mentioned.

Hi Keith,

Ok, I won't worry about it then. :-)

I'm thinking of going ahead with the current Kernel update, but there is something that is confusing. Why is the most current and the currently active lower in version numbers than all the Kernels shown in Update Manager? Here's what I mean:

5.11... -- 1 supported until February 2022
        -- all the rest "Superseded"

5.8...  -- all "End Of Life"

5.4...  -- 0-88 supported until April 2025 (currently in line for update)
        -- all others "Superseded"
        -- currently active is 7th down the list from 0-88

Why would we not be using the higher version numbered Kernel 5.11... instead of 5.4...? I don't understand this. :-\

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Because distros choose a long-term release for the stable platform version. Kernel 5.4 is the LTS kernel for your distro. Corporate IT managers like having all computers on the LTS package because they have commonality for support.

But the latest kernels are often pushed into HWE (HardWareEnablement) releases that offer the latest support for new hardware, latest bugfixes and latest feature sets.

The 5.11 kernel is one in the HWE series.

Distro maintainers declare which kernel series gets LTS denomination about every 4-5 years in 5 year leapfrogs typically. The next LTS kernel is the 5.10 series I believe.

https://www.kernel.org/

---

Because distros choose a long-term release for the stable platform version. Kernel 5.4 is the LTS kernel for your distro. Corporate IT managers like having all computers on the LTS package because they have commonality for support.

But the latest kernels are often pushed into HWE (HardWareEnablement) releases that offer the latest support for new hardware, latest bugfixes and latest feature sets.

The 5.11 kernel is one in the HWE series.

Distro maintainers declare which kernel series gets LTS denomination about every 4-5 years in 5 year leapfrogs typically. The next LTS kernel is the 5.10 series I believe.

https://www.kernel.org/

Hi Keith,

Ok, that makes sense. I understand LTS. But...

If the distributors are not using the 5.11 or 5.8, why even bloat the install with them? In the next iteration of Linux are we gonna have 5.11, 5.10, 5.8 and 5.4? Or will they get rid of one or both of the Kernels not active in an install and make the active obsolete after an update?

I went ahead and did the Kernel update and everything is just fine. I had my fingers crossed this time. ;-) So I guess that when you set the updater to automatically delete obsolete Kernels, it only deletes those marked "Obsolete", which I had done several days ago by hand. There weren't that many.

OH BOY!!! I just looked at my Kernel list in the updater and the "Remove Kernels..." button is now live. If I click that will it ONLY remove the ones not active or installed in each series and allow me to keep 1 or 2 older working ones? All the other times I looked the button was greyed out. If I remember correctly, when I was tramping the Internet trying to find out how to get rid of them it was always to go into the Kernel list and click the "Remove Kernels..." button. Mine was always greyed out.

I just watched a couple videos on Kernels and if you click on a kernel that says "Installed" you get a button to remove it. The others have an "Install" button. I'm thinking that with the "Remove Kernels..." button I should get a selector in each series of Kernels to select what to remove or remove all. Or is that just me dreaming? ;-)

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

If the distributors are not using the 5.11 or 5.8, why even bloat the install with them?

Well, my new Ryzen 5950X didn't show CPU temperature in Psensor, I did a Google search and found out I needed the 5.11 kernel to get it to work.

The next LTS kernel is going to be 5.10. https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.10-LTS-EOL-EOY-2026
Your guess is as good as mine. I am not familiar with the tools in Mint. You will have to discover that for yourself.

@JohnDK, well the Linux devs have been very wishy-washy with respect to AMD Ryzen/Epyc/Threadripper temp and power reporting. First they put support into upstream kernel archives, then changed their mind and yanked it all back out because of political infighting between AMD and the devs.

You still have both power and temp reporting even on older kernels like 5.4 through out-of-tree modules like zenpower and asus-wmi-sensors. There is a current upstream merge being tested and pushed for inclusion into hwmon to wrap up all the stray bits and bobs from k10temp, zenpower, asus-wmi-sensors and nct6775 modules into one all inclusive monitoring module. So there is hope eventually to get rid of all the fractionalization in the hwmon tree.

---

The next LTS kernel is going to be 5.10. https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.10-LTS-EOL-EOY-2026
Your guess is as good as mine. I am not familiar with the tools in Mint. You will have to discover that for yourself.

Hi Keith,

Ok then, I'll spend a little more time stomping on the Internet to see what I can find and if I don't find anything specific, I'll just cross my fingers and hope for the best. I'm sure the Update Manager won't just remove Kernels willy nilly. I'll first try it on the Kernel series the system is NOT using to see what happens. ;-) I'll let you know what happens.

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

Ok, that makes sense. I understand LTS. But...

If the distributors are not using the 5.11 or 5.8, why even bloat the install with them? In the next iteration of Linux are we gonna have 5.11, 5.10, 5.8 and 5.4? Or will they get rid of one or both of the Kernels not active in an install and make the active obsolete after an update?

Siran

Because the only way to get things tested and ready for the LTS setting is to send it out to be tested, and you are one of those people. They now want you to provide feedback on any problems you are seeing so they can address them before they come out, or get redone or even dropped, before the next LTS release. That's often why you will see a long list of changes before the next release, people are providing feedback and the developers are working on things.

Because the only way to get things tested and ready for the LTS setting is to send it out to be tested, and you are one of those people. They now want you to provide feedback on any problems you are seeing so they can address them before they come out, or get redone or even dropped, before the next LTS release. That's often why you will see a long list of changes before the next release, people are providing feedback and the developers are working on things.

Hi Mikey,

When I installed and did the re-install I saw nothing that would opt me into testing and sending feedback. I don't want to be a tester of Kernels. :-( Why would the maintainers want obsolete, end-of-life or older versions of a current series tested? To me that would be like trying to get DOS 6.2 to run on modern PCs. I don't believe it would work and I don't see the point of testing old Kernels.

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

The next LTS kernel is going to be 5.10. https://www.phoronix.com/scan.php?page=news_item&px=Linux-5.10-LTS-EOL-EOY-2026
Your guess is as good as mine. I am not familiar with the tools in Mint. You will have to discover that for yourself.

Hi Keith,

I just tested the "Remove Kernels..." button. All it wants to do is remove "installed" Kernels (not the active one of course) which makes me come to a conclusion of more confusing Kernel stuff:

The only Kernels installed are the last used and the currently active. The other Kernels listed in the different series are just the version numbers of all the Kernels. When installing the Kernel the appropriate files and dependencies are first downloaded then installed. I'm thinking that I have only 2 Kernels installed on this PC. Which begs the question: Why do I have version 5.8 list which are labeled "End-of-Life"? That truly makes no sense. :-|

And another question: Why, when updating a Kernel, would something have to be done with ALL the other listed versions? I don't get it. Or, is the updater making sure that all the others, save for the most recent, are uninstalled? It sits on each version number for a good 30 seconds. But then, if the previous update did that, why would new updates need to go through those hoops again? I think I'm thinking too much, me thinks. ;-)

Have a great day! :-)

Siran

---

CAPT Siran d'Vel'nahr XO - L L & P _\\//
USS Vre'kasht NCC-33187
Winders 10 OS? "What a piece of junk!" - L. Skywalker
"Logic is the cement of our civilization with which we ascend from chaos using reason as our guide." - T'Plana-hath

I'd say the Mint Updater is confusing and is confusing you. I have no such shenanigans from the Ubuntu Updater. I do think you are overthinking the issue also.
Run the latest kernel and have 1 to 2 older kernels to fall back on for the GRUB recovery menu. That is all you need to know or worry about.

---

Because the only way to get things tested and ready for the LTS setting is to send it out to be tested, and you are one of those people. They now want you to provide feedback on any problems you are seeing so they can address them before they come out, or get redone or even dropped, before the next LTS release. That's often why you will see a long list of changes before the next release, people are providing feedback and the developers are working on things.

Hi Mikey,

When I installed and did the re-install I saw nothing that would opt me into testing and sending feedback. I don't want to be a tester of Kernels. :-( Why would the maintainers want obsolete, end-of-life or older versions of a current series tested? To me that would be like trying to get DOS 6.2 to run on modern PCs. I don't believe it would work and I don't see the point of testing old Kernels.

Have a great day! :-)

Siran

The "maintainers" at large companies, and especially small companies, don't have the time to be updating Linux as much as you do, as you've seen it breaks things that are working pretty well already. So once they get an LTS version up and running they tend to only do the necessary updates for security and because this new thing is much better than the one they are already using. I was a 1st level IT guy at a guy with over 40K desktop pc's and 5 years after Win10 came out they were still using win7 pc's with 512 kb of ram in them!! It cost $60k to take out those two 256k sticks of memory and put in two 1gb sticks of memory. The employees were EXTREMELY happy because their pc's would do things significantly faster but that $60k meant that some other thing did not get updated.

And as far as 'not singing up to be a tester' that's what we did by using a non LTS version and then updating it all the time. As far as Linux Mint goes I'm not even sure they HAVE LTS versions but I may start using one if it came out as I almost never update mine until I'm forced to as all my Linuxt Mint pc's do is crunch.

mikey