[SOLVED] Slackware Current

bitfuzzy · 01-31-2021, 04:14 PM

Installed recent updates today including kernel 5.10.12

When I ran 'lilo' I was greeted with the following

Quote:

Warning: Unable to determine video adapter in use in the present system.
Warning: Video adapter does not support VESA BIOS extensions needed for
display of 256 colors. Boot loader will fall back to TEXT only operation.

Thoughts?

Everything seems to be running okay

cwizardone · 01-31-2021, 04:25 PM

No solutions, yet, but you are not alone.

Here are some recent posts on the problem,

https://www.linuxquestions.org/quest...ml#post6213802

The comments and addtional users reporting the same problem continues down that page and on to the next page.

hitest · 01-31-2021, 04:47 PM

Quote:

Originally Posted by bitfuzzy

Installed recent updates today including kernel 5.10.12

When I ran 'lilo' I was greeted with the following

Thoughts?

Everything seems to be running okay

Yeah I received the same warning on three out of four units that use LILO. Everything is okay here too. My fourth unit uses GRUB and did not get the warning.

pghvlaans · 01-31-2021, 07:47 PM

My 64-bit -current installs boot from GRUB and they're both fine. 32-bit -current on VM has LILO and it got the warnings after updating to 5.10.12; startx into Xfce succeeded and the system appears to be behaving normally under 5.10.12.

chrisretusn · 02-01-2021, 09:30 AM

I'm getting the same warning on my system as well. Outside of these warnings, no adverse effect on the system.

Before I upgraded my system I was not getting these warnings (ChangeLog.txt Fri Jan 29 20:26:57 UTC 2021).

A few minutes ago, I updated from Mon Feb 1 03:14:27 UTC 2021, Sun Jan 31 20:25:39 UTC 2021, Sat Jan 30 21:55:04 UTC 2021 and now the warnings.

NakedRider · 02-01-2021, 12:03 PM

Quote:

Originally Posted by bitfuzzy

Installed recent updates today including kernel 5.10.12

When I ran 'lilo' I was greeted with the following

Thoughts?

Everything seems to be running okay

When you say everything is running okay does that mean that you rebooted and everything, including X (like Plasma), still runs?

hitest · 02-01-2021, 12:27 PM

Quote:

Originally Posted by NakedRider

When you say everything is running okay does that mean that you rebooted and everything, including X (like Plasma), still runs?

Yup.

NakedRider · 02-01-2021, 01:17 PM

Quote:

Originally Posted by hitest

Yup.

Thank you. I'll try a reboot.

bitfuzzy · 02-01-2021, 09:53 PM

Well it may be coincidence but after X crashing on me 2x during a critical remote desktop session, I've decided to roll back to 5.4.84

allend · 02-02-2021, 12:39 AM

I have been looking at this.
I hit the problem, as reported, after upgrading from the 5.10.11 to 5.10.12 kernel.
However, after reverting to the 5.10.11 kernel, the problem remained.
Further reverting to the 5.10.10 kernel, the problem disappeared.

After sprinkling the lilo source with a few printf's, the cause of the error is in the fetch() routine in probe.c. The code to read the low memory bios data area is failing, specifically the line

Code:

if (strncmp(buf.s5.signature, PROBE_SIGNATURE,4)) return buf_valid=2;

The buf.s5.signature should read "LiLo" to match the PROBE_SIGNATURE,but instead no string is seen, causing the buf_valid=2 return to trigger the warning messages.

I am pointing the finger at commit c351dc4d774e57fdb8ec543241710cb93a7387bd in the ChangeLog-5.10.11 that links this.

chrisretusn · 02-02-2021, 06:54 AM

Quote:

Originally Posted by allend

I have been looking at this.
I hit the problem, as reported, after upgrading from the 5.10.11 to 5.10.12 kernel.
However, after reverting to the 5.10.11 kernel, the problem remained.
Further reverting to the 5.10.10 kernel, the problem disappeared.

After sprinkling the lilo source with a few printf's, the cause of the error is in the fetch() routine in probe.c. The code to read the low memory bios data area is failing, specifically the line

Code:

if (strncmp(buf.s5.signature, PROBE_SIGNATURE,4)) return buf_valid=2;

The buf.s5.signature should read "Lilo" to match the PROBE_SIGNATURE,but instead no string is seen, causing the buf_valid=2 return to trigger the warning messages.

I am pointing the finger at commit c351dc4d774e57fdb8ec543241710cb93a7387bd in the ChangeLog-5.10.11 that links this.

Excellent sleuthing.

rherbert · 02-03-2021, 03:04 PM

Quote:

Originally Posted by allend

I have been looking at this.
I hit the problem, as reported, after upgrading from the 5.10.11 to 5.10.12 kernel.
However, after reverting to the 5.10.11 kernel, the problem remained.
Further reverting to the 5.10.10 kernel, the problem disappeared.

As stated already, EXCELLENT sleuthing! I went through the exact same experience, that being no VESA error when first upgrading to 5.10.11. That threw me off and made me suspect something in the -current updates. It never did occur to me to go back to 5.10.10, so good on ya. And thanks!

enorbet · 02-03-2021, 07:15 PM

Just FTR I have gotten the VESA warning from LILO for years...in fact for so long I can't recall when exactly it was that I didn't get that warning over the past 20+ years using LILO. I think it may be possible that there is a tie-in with using only Nvidia graphics systems and perhaps some mismatch in Framebuffer codes. One of the reasons I can't recall is after the umpteenth time with zero discernible effects, I learned to ignore it. It's somewhat due to old code that still warns about LBA32. LILO is nevertheless a great bit of old code that is highly efficient and simple IF you have any MBR disks anymore.

kingbeowulf · 02-03-2021, 07:43 PM

Interesting. I get OPs lilo warning message only when I boot off USB to chroot when I forget to run lilo when upgrading Current kernel. Also, as enorbet mentioned, the fb warning occurs when using nvidia proprietary driver. IIRC, nvidia uses its own fb that is not "visible" on boot. You can see a fb warning in the boot logs as well.

I just ran the updates here from 5.10.10 -> 5.10.12, etc, mkinitrd then lilo with no issues, reboot and startx into Xfce. All's well.
Lenovo Thinkpad T510
Intel Core i7-620M w/ intelHD IGP

My lilo.conf is simple, generated with liloconfig then stripped out some extra comments:

Code:

boot = /dev/sda

#compact        # faster, but won't work on all systems.

# Boot BMP Image.
  bitmap = /boot/slack.bmp
# Menu colors
  bmp-colors = 255,0,255,0,255,0
# Location of the option table
  bmp-table = 60,6,1,16
# Timer location
  bmp-timer = 65,27,0,255
# Standard menu.
# Or, you can comment out the bitmap menu above and
# use a boot message with the standard menu:
#message = /boot/boot_message.txt

# Append any additional kernel parameters:
append=" "
prompt
timeout = 50
# Normal VGA console
vga = normal
# ramdisk = 0     # paranoia setting
# End LILO global section
# Linux bootable partition config begins
image = /boot/vmlinuz-generic
  initrd = /boot/initrd.gz
  root = /dev/sda1
  label = Slackware64
  read-only  # Partitions should be mounted read-only for checking
# Linux bootable partition config ends

My 2 other boxes and AMD GPUs and UEFI so no lilo there.

gordydawg · 02-03-2021, 10:43 PM

Confirming what other folks have noted. In my case, I used an AMD Ryzen 5 on an Asus mother board with BIOS set to Legacy boot only. All partitions GPT.

The Lilo VESA warnings came after editing lilo.conf when upgrading 64 bit current from 5.10.10 to 5.10.11

Not only have I gotten warnings, but the second upgrade to 5.10.12 produced the dreaded Lilo Timestamp error message upon rebooting from the kernel upgrade. That box with a NVME M.2 SSD stick produced the timestamp error. Different tries to rebuild lilo.conf via liloconfig produced the same result. Coincidentally, I noticed the change logs to 5.10.12 also upgraded e2fsprogs to 1.46.0

I had similar results on the same box when upgrading current on a standard hard drive, I was able to edit lilo.conf and have it boot successfully despite the Lilo VESA warnings provided I used the old style standard lilo screen.