compare new-huge-kernel vs new-generic-kernel

rizitis · 10-27-2023, 01:47 PM

Add these lines in /etc/rc.d/rc.local

Code:

echo  -e 'kmsg:' > /dev/shm/onboot
cat /proc/kmsg >> /dev/shm/onboot & sleep 3 && killall cat
chmod 600 /dev/shm/onboot && chown root:root /dev/shm/onboot

reboot new-huge-kernel then as root

Code:

cat /dev/shm/onboot > /root/huge.txt

reboot new-generic-kernel and as root

Code:

diff --side-by-side --suppress-common-lines /dev/shm/onboot /root/huge.txt

0XBF · 10-28-2023, 08:16 AM

Why not just use 'dmesg' to read the kernel buffer into a file and diff those if you want to compare kernel logs?

It would be easier since you wouldn't need to "killall cat". You could also just run the dmesg command as root after booting and direct it straight to a file in /root/, without messing with the /dev/shm and permissions/ownerships. There's not going to be any difference running it from rc.local, or just manually after booting.

rizitis · 10-28-2023, 09:05 AM

hey @0XBF, yes you can do it with several ways...
For the history, I was playing with new kernels because I was trying to find the difference btw huge and generic in the beginning it was not clear to me. Also now

Code:

zgrep -q 'CONFIG_EXT4_FS=y' /proc/config.gz && echo "Huge kernel" || echo "Generic Kernel"

always return "huge". Trying to find a way for the question "what kernel am I running?" I remember that I already having something similar in my /etc/sys which was inspired from Paulo2
So this command for example

Code:

cat /etc/sys | grep BOOT

always is true, (huge or generic).
which is the same

Code:

dmesg | grep vmlinuz

Then i said ok, lets do the same for 2 new kernels to see what diffs they have in the same system or what diff they have with my custom kernel.
This way for me is easier I think.
ps. And I found this in my custom kernel

Code:

>	<6>[    0.000000] efi: Remove mem90: MMIO range=[0xc0000000-0
							      >	<7>[    0.000000] e820: remove [mem 0xc0000000-0xcfffffff] re
							      >	<6>[    0.000000] efi: Not removing mem91: MMIO range=[0xfe00
							      >	<6>[    0.000000] efi: Not removing mem92: MMIO range=[0xfec0
							      >	<6>[    0.000000] efi: Not removing mem93: MMIO range=[0xfed0
							      >	<6>[    0.000000] efi: Not removing mem95: MMIO range=[0xfee0
							      >	<6>[    0.000000] efi: Remove mem96: MMIO range=[0xff000000-0

which idk what it is!

Aeterna · 10-28-2023, 11:21 AM

is

Quote:

uname -r

not working for you?
e.g.

Quote:

uname -r
6.1.59-hardened1-betelgeuse

0XBF · 10-28-2023, 02:28 PM

With the ext4 filesystem module built into the generic kernel now, your test will report huge for both (assuming this is slackware-current we are using).

You could check what kernel you booted on the kernel command line from '/proc/cmdline'.
E.g.:

Code:

#/bin/bash
if ( grep -q vmlinuz-generic /proc/cmdline ); then
  echo Running a generic kernel
elif ( grep -q vmlinuz-huge /proc/cmdline ); then
  echo Running a huge kernel
else 
  echo Running a non-standard kernel name:
  cat /proc/cmdline
fi