Installing Solus on an Old Laptop: Observations and FYI

brent

I consider this https://www.rodsbooks.com/efi-bootloaders/csm-good-bad-ugly.html the best article I ever read about CSM/UEFI and old laptops.

It was about my third install of Solus I finally got it:
Secure Boot Off, UEFI Enabled = failed install that would not boot
Secure Boot On, UEFI Enabled=failed install that would not boot
Secure Boot Disabled, UEFI Disabled, CSM Enabled= PERFECTION

CSM is affiliated with BIOS/Legacy installs but in my case I needed it to boot.
Sometimes I think many failed laptop installs come from BIOS moves. It took me a day to figure this out.

10 minute install. Downloading 340 packages now, I'm estimating that will be 20-25 minutes. I don't mind this. It gives you time to explore...

Set the Tullamore picture. Want to dump Nautilus and get Nemo next. Finally got to repurpose this old laptop after 8GB memory and SSD upgrades. Solus is snappy on here....I can't say that about the others I tried.
I tried to kick you Solus! It didn't work,

Welcome to Gnome and Wayland: two things I've never tried.

johano

brent nice link re CSM/EFI 👍

BuzzPCSOS

brent It seems strange that the computer had a uefi option that failed to install. This should have been the easiest choice. Other things can get in the way of a successful install; who is the USB manufacturer, how the boot image is written, hard drive health, good RAM and a CPU that is not melting due to a blocked cooling path.
I try to test all of these before starting any install but recently got caught on one that randomly failed during the install bleeping or just acting like there was a virus. Various boot discs all reported a healthy system despite that.
It turned out that there was a intermittently sticky arrow key on the keyboard which sometimes caused the post (power on self test) to fail or would just cause the wrong things to be selected at critical points in the installation. The sticky key did not show up when running from a boot disk. Once I had found the culprit it only took a small amount of TLC to make the keyboard behave again and allow me to complete a successful install.
So, the reason for this rambling story is to say that the technical answer is not always the definitive cause. Random factors can enter into any apparently controlled environment and give you a theory that whilst apparently sound might not be repeatable. Especially on old computers.
As a failsafe I would at least be inclined to run a check disk and SMART health tests on the hard drive, push some compressed air through the cooling duct and conduct a RAM test. Oh and keep an eye out for cursor selections wandering about on their own 😉

Axios

brent I had those csm issues back in the day learned to try it with it turned off
but most time ended up turned on.
Think because that was the transisition phase for eufi and nobody followed any rules.
But glad you got it working.

brent

MODS please change AND to AN
thank you!

brent

gnome-tweaks would not launch.

 $ gnome-tweaks
MESA-INTEL: warning: Bay Trail Vulkan support is incomplete
MESA-INTEL: warning: ../src/intel/vulkan_hasvk/anv_formats.c:759: FINISHME: support YUV colorspace with DRM format modifiers
MESA-INTEL: warning: ../src/intel/vulkan_hasvk/anv_formats.c:790: FINISHME: support more multi-planar formats with DRM modifi

filed bug report at tracker

brent

BuzzPCSOS Random factors can enter into any apparently controlled environment and give you a theory that whilst apparently sound might not be repeatable. Especially on old computers.

stuck keys are a silent curse I have learned. Factors, yes----it's Linux! the myth of repeatability is for certain. so how to explain my unusual install?

BuzzPCSOS It seems strange that the computer had a uefi option that failed to install. This should have been the easiest choice.

The two magic BIOS settings for any Linux-to-old-computer install are UEFI/CSM switch and Secure Boot switch. I threw out all preconceived notions and conventional research wisdom and simply flipped the Magic Two Buttons til the install would boot. All installs were successful, only one install would boot.

I appreciate your thorough answer much.

BuzzPCSOS

brent so how to explain my unusual install?

Yeah, got me there. Would just love to get my fingers on it and tinker for a while but you are up and running now. Congratulations on your persistence, I hope it serves you well. 👍

brent

BuzzPCSOS
in that exhaustive link I posted, somewhere near the end ("CSM uses" section) he concludes "The CSM is a useful stop-gap tool, but it should not be over-used. Unfortunately, many in the Linux community have been doing just that—advice to activate the CSM as a routine part of Linux installation is common...Do leave the CSM enabled if you're installing Linux to a computer that already uses the CSM to boot all its other OSes."

Maybe the age of the Toshiba qualified me for the "stopgap" tool installation. Leaving UEFI enabled, bizzarely, caused the installations not to boot whether Secure Boot was enabled or disabled.
That's why I think half of all Install Failures in laptop are the wrong BIOS moves....and the other half are the minute variables (stuck keys, faulty media, ill-prepared target drives etc) that are so easily overlooked that you discussed in detail.

brent

Axios Think because that was the transisition phase for eufi and nobody followed any rules.

that's a good observation and that's what they guy thought in that article: they hadn't perfected efi by then so linux installs were screwy as a result. my old thing is from that era.

alfisya

@brent Please update and see if this week update (to mesalib) fixes your issue with GTK4 apps. Cheers

brent

alfisya Please update and see if this week update (to mesalib) fixes your issue with GTK4 apps. Cheers

I'll will do that in my new Gnome install soon. If that fails I will GSK and see what happens. Thanks for the reminder!

brent

alfisya the update fixed me in Gnome. I closed this on devtracker. Thank you so much for your help.