KLV-Airedale-RT with full Real Time Kernel 6.1.38-rt13

[rockedge]

This ISO is a KLV-Airedale-sr2 except the kernel has been changed to a full real time version 6.1.38-rt13-rc1. This RT version kernel is Overlayfs only (no aufs) and is built by the Puppy Linux woof-CE kernel-kit. The initrd.gz is @wiak's "magic skeleton initrd" able to handle Puppy Linux style huge kernels to boot KLV.

Screenshot(12).jpg (61.58 KiB) Viewed 7728 times

The firmware SFS is newly created during the build of 6.1.38-rt13-rc1

KLV-Airedale-RT.iso 766 MiB
SHA1-MD5.txt

Screenshot(9).jpg (34.91 KiB) Viewed 7749 times

[Sofiya]

[geo_c]

So is it only necessary to copy vmlinuz if I already have sr2 installed? Or do I need other files like initrd?

[rockedge]

@geo_c To swap to this type of kernel either set up as a new frugal install and copy all of the XXupper_changes over to the directory.

Or swap in the RT kernel to the existing file system use these and replace the current ones:

Code: Select all

00modules-6.1.38.sfs
01firmware-6.1.38.sfs (optional if keeping the original)
initrd.gz
wd_init
vmlinuz

This is a Puppy Linux huge type kernel so the initrd.gz is the very small skeleton initrd which can boot the huge kernel type.

It is experimental and untested in this configuration so do not modify a working install of KLV-Airedale-sr2. Try this one on the side!

[geo_c]

So as I noted in the Show Us Your Desktop thread, I'm running KLV-Airedale-RT on a USB install and a laptop hard drive install. I ran it for 12 hours straight today, Ardour, Gimp, browsers, Thunderbird, etc.

Everything seems to be running smoothly. And the temperature doesn't really seem to be an issue.

Funny thing is, I'm pretty sure the temp widget was showing higher than 60 degrees as yellow, and higher than 70 degrees as red, but at somepoint it stopped doing that and stays green until about 75 where it turns yellow. Does it adjust to the hardware or something?

At any rate it only gets that hot if I'm using Ardour running 10 audio tracks each loaded with resource hungry plugins. And it runs flawlessly.

This Kernel was right on time for me, because I was just sharing with @williwaw that I was trying to decide whether to continue my audio work in jackalpup, which still does the job, but isn't exactly up to date any more as it's based on fossapup64_9.5, or switch to one of the Kennel distros.

Having decided to switch to KLV-airdale for my audio production, within 24 hours you provided a real-time kernel @rockedge.

Now that's what I call destiny.

And btw, I used the above method, just copying the necessary files underneath upper_changes, using my remastered rootfs. That worked without a hitch so far.

[rockedge]

I uploaded to https://rockedge.org/kernels the latest KLV-Airedale-RT with a system update/upgrade with Firefox 116.0

Other than the system update it is the same as the previous RT release.

[geo_c]

I've been updating as I go, seems to be running well. Except somehow the XFCE power manager plugin in the taskbar failed and won't load again. Either I blew it somehow trying to get the screen not to go black after 10 minutes or so, and then copied the upper_changes to all my installs, or maybe the kernel has something to do with it. Hard to say, could be either. Could simply be a bad update from Void also. I'm not that bummed, I can still monitor the battery with the battery monitor plugin.

I have one laptop, an older Dell i5 with 4 processor threads that doesn't work using Ardour with the RT kernel, so I restored that machine to sr2, which works just dandy.

Ardour actually works on that machine with RT, but the audio pops and cacks. I just think it can't keep up with the heavy duty, resource intensive multi-tracker work using low latency processing. It'll play an audio track in Parole just fine.

[williwaw]

Either I blew it somehow trying to get the screen not to go black after 10 minutes or so,

have you found a solution to the screen dropping out? xdotool?

[geo_c]

have you found a solution to the screen dropping out? xdotool?

Actually no, and I think it must have been a Void update, because I restored one of my laptops to KLV-sr2 because of the Ardour situation mentioned above, and I'm pretty sure I just created a new upper_changes.

I move systems around a lot and don't take notes, so I have to rely on my aging memory.

Not sure why the display time option in the power manager doesn't work. Did you get the xdotool to achieve it?

[williwaw]

I had to download xdotools in my distro and used the watch command. something like

watch -p 500 xdotool mousemove_relative 5 5

which will move the mouse pointer 5 pixels right and 5 pixels down every 500 seconds

I tested it and it worked,
(but the syntax might be bit off as I am working from memory. I ran the command but never did make the script)

[rockedge]

I have the previous version of KLV-Airedale-RT, does this occur in that version also or since the Void system update?

[geo_c]

I have the previous version of KLV-Airedale-RT, does this occur in that version also or since the Void system update?

I'm running 6.1.38-rt13, and my power manager is definitely broken. EDIT: and it's not just the panel plugin, it's the power manager itself

I decided to try OctoXbps and uninstall the power manager then reinstall it. The uninstall aborted with the message "aborted due to unresolved dependencies." Sounds like maybe some dependencies updated and aren't compatible?

So I don't know what's up there. I do know that I was trying to set the screen blank time to a longer setting or off. But I don't know if that broke it or not.

[rockedge]

@geo_c Ahhhh I've seen this before! Those times it was a subsequent Void Linux update that fixed it. Meaning yes it is a Void Linux update side effect. Last times it has occurred in similar fashion, the following Void Linux rolling update completed the dependencies requirements.

Hopefully that is! But looking at the few other times directly after a rolling update that we had similar symptoms, the following update solved the issue.

Though doing some testing the power manager seems to be working on a DELL laptop and on a tower desktop. Can open the configuration gui on the power manager as well in a QEMU VM running KLV-Airedale-RT

[geo_c]

Though doing some testing the power manager seems to be working on a DELL laptop and on a tower desktop. Can open the configuration gui on the power manager as well in a QEMU VM running KLV-Airedale-RT

Checking on my computer at the studio which I still have running sr2, last system update 7-18, that power manager is functioning. It's quite possible that I just somehow broke it messing around with it in RT-13 and copied those upper_changes to everywhere else.

I'll have to roll back or just boot RT-13 with a fresh upper_changes and then do an update to figure it out. I'll get back to you when I get around to that!

edit: or maybe I just botched the upper_changes copying update procedure along the way!

[geo_c]

Yeah, somehow my doing.

I rolled back to 7-18 upper_changes with working power manager and did a system update, and all is good, working

Sorry for the sidetrack.

I can't keep up with all these OS's you guys are cranking out.

I'd say I need to do isobooting in qmenu or something, but I just use these things as daily drivers on my hard drives:

[Sofiya]

Yeah, somehow my doing.

I rolled back to 7-18 upper_changes with working power manager and did a system update, and all is good, working

Sorry for the sidetrack.

I can't keep up with all these OS's you guys are cranking out.

I'd say I need to do isobooting in qmenu or something, but I just use these things as daily drivers on my hard drives:

for sure, it won't work with this moto

[rockedge]

@geo_c Kelly's Heros is one of my favorite films. I'll watch it almost every chance I get!

[geo_c]

@geo_c Kelly's Heros is one of my favorite films. I'll watch it almost every chance I get!

One of those movies with the timeless quotable lines.

Now this is curious, from my previously working changes, which I system updated yesterday, I booted again this morning and bam, the power manager failed:

I'll have to see if this happens on my other machines. I expect it has something to do with the RT kernel and Xfce or Void update. I noticed there were several dbus updates in that batch. Does that have something to do with managing power?

[Sofiya]

@geo_c Kelly's Heros is one of my favorite films. I'll watch it almost every chance I get!

One of those movies with the timeless quotable lines.

Now this is curious, from my previously working changes, which I system updated yesterday, I booted again this morning and bam, the power manager failed:

I'll have to see if this happens on my other machines. I expect it has something to do with the RT kernel and Xfce or Void update. I noticed there were several dbus updates in that batch. Does that have something to do with managing power?

Xfce Power Management systems provide a DBus interface to monitor the system power status and adjust power settings. They also provide a DBus interface for managing display brightness.
Similarly, the Xfce Session Manager uses DBus to allow applications to register as startup applications and automatically launch when we log in to Linux. It also uses DBus to save and restore the user’s sessions when we log out and back in.

[rockedge]

Something is preventing dbus from starting an instance, or a connection to dbus perhaps? So far I can't replicate the crash. I'll keep looking into it.

UPDATE: So far I have not been able to see the power manager crashed or stalled. I thought on the RT kernel perhaps a module was not included but I am not seeing the problem yet to track anything down.

[geo_c]

Something is preventing dbus from starting an instance, or a connection to dbus perhaps? So far I can't replicate the crash. I'll keep looking into it.

UPDATE: So far I have not been able to see the power manager crashed or stalled. I thought on the RT kernel perhaps a module was not included but I am not seeing the problem yet to track anything down.

Interesting, there are some dbus files in my 07-18-2023 upper_changes, but those changes seem to run the power manager fine, it's seems to be after I update the system in a new upper_changes and reboot that the problem gets created. Again, I should check my other machines:

[geo_c]

Following up on the dbus hypothesis, I have a second laptop running my upper_changes from 2023-07-18, power manager working fine.

I'm checking for system updates, which lists 65 packages including dbus, dbus-x11, and dbus-libs, plus device mapper.

I'll update them and see what happens starting in 5,4,3,2,1.....

-update successful
-reboot
-"POWER MANAGER PLUGIN unexpectedly left the panel."

Definitely the update, as a working power manager on two machines breaks when running the latest Void dbus updates.

I guess the question is could it have anything to do with the RT-kernel. I also note that the 6.4 Kernel is acting buggy, so it has me wondering.

I suppose I can switch back to sr2 kernel and run the update to find out.

Though after doing a lot of work using RT-13 kernel, I can safely say that everything else appears on the surface to be working.

[rockedge]

@geo_c It would be important to see if the sr2 kernel AND the recent update also break dbus

I always suspect the RT kernel first but that can be checked by trying out what happens to sr2 with that update and then sr2 and the RT kernel.

6.4.7 has been acting strange all over the place. I just read that KLA's most recent Arch updates has wreaked havoc

[geo_c]

@geo_c It would be important to see if the sr2 kernel AND the recent update also break dbus

I always suspect the RT kernel first but that can be checked by trying out what happens to sr2 with that update and then sr2 and the RT kernel.

6.4.7 has been acting strange all over the place. I just read that KLA's most recent Arch updates has wreaked havoc

I'll check out sr2 with the update and sr2 kernel. I've been using sr2 with realtime kernel the entire time, though I am using the 6.1.38 modules and firmware, my rootfs is still my remastered sr2.

[geo_c]

Just to be clear, I renamed upper_changes on 07-18-2023 before updating to realtime kernel and 6.1.38 modules and firmware. Been running that for about 10 days it appears.

as the KLV-airedale directory shows (those two 0 byte files are my personal flags):

so I can put the sr2 kernel and modules/firmware back with w_init/initrd. I have all those saved, and then run the latest update and see how it goes.

[rockedge]

@geo_c I have a copy of the KLV-Airedale-RT that was based on the sr2 rootfs.

I could put it back up on rockedge.org/kernels

Also after days of testing the RT kernel out, it seems fit across the range of machines I have to test it on. And everything appears to work as expected...so far

[geo_c]

Okay,

sr2 kernel, firmware and modules, initrd, and w_init copied into directory with KLV-airedale-sr2 remastered filesystem which is working perfectly.

preparing to update 5,4,3,2,1....

-65 packages will be updated including 3 dbus packages and device mapper
-udated successfully
-reboot
-BAM! POWER MANAGER FAILS AGAIN!

We can lay the blame on Void and not the kernel or the firmware. Sr2 uses fdrv_upup_32 and 06.3.12 modules.

edit: and like I was saying @rockedge , I've been using the sr2 rootfs this entire time. I simply dropped the intits, firmware/modules, and rt-13 kernel in the install directory, and that was working great until this dbus update.

Seems to me the only thing it breaks is the power manager, as the battery plugin and cpu monitors still work fine.

[Sofiya]

I also don’t sin on the core, but on updating packages, BUT you still need to make sure exactly 100%

[geo_c]

I also don’t sin on the core, but on updating packages, BUT you still need to make sure exactly 100%

I think you're saying I need to use the original sr2 rootfs and not my remastered version. And if you are, you have a point. I may be missing something in my remastered filesystem.

I wouldn't even know how to use a virgin copy of KLV-airedale anymore!

edit: I do have the sr2 iso though, so I could do one final test.

[geo_c]

Now booting a virgin sr2 using all the original iso files, no upper_changes, so a clean sr2 system with original filesystem, @rockedge's dark cloud background displaying proudly, power manager working swell...

time to try the update in 5,4,3,2,1....

-updating 76 packages including dbus and device-mapper
-update successful
-reboot
-POWER MANAGER is WORKING

Crap, it's my remastered file system. Well, I was so happy with it.

I guess it's back to the drawing board on upper_changes.

And I was very thankful for the remaster script. It gave me that immutable starting point for a fully developed system.

And this right here is why I need to simply learn the firstrib build script, so I can run it and build a system with everything installed from the beginning in a squashed read-only file.