LinuxQuestions.org
Review your favorite Linux distribution.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 10-17-2019, 06:00 PM   #1696
cwizardone
Senior Member
 
Registered: Feb 2007
Distribution: Slackware64-current with "True Multilib" & Xfce.
Posts: 4,908

Original Poster
Blog Entries: 1

Rep: Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317

Kernel updates 5.3.7, 4.19.80, 4.14.150, 4.9.197 and 4.4.197 are now available at

https://www.kernel.org/

The change logs,

https://cdn.kernel.org/pub/linux/ker...hangeLog-5.3.7

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.19.80

https://cdn.kernel.org/pub/linux/ker...geLog-4.14.150

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.9.197

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.4.197


The 32-bit version of Dave's Unofficial Slackbuilt Kernel update for the 4.4.y series can be found at, https://dusk.idlemoor.tk

Last edited by cwizardone; 10-17-2019 at 06:04 PM.
 
3 members found this post helpful.
Old 10-17-2019, 06:44 PM   #1697
petejc
LQ Newbie
 
Registered: Apr 2019
Distribution: Slackware
Posts: 29

Rep: Reputation: Disabled
Quote:
Originally Posted by Andypoo View Post
Interesting. What filesystem do you use? Mind sharing your mkinitrd line? Happy to look into it with you.

Have you managed to capture the kernel panic text? There was actually a hint in it for me prior to mounting, as it logs out the error from do_mounts. It could be a similar issue to the one that was backported if other modules are also cleaning up/changing their superblock detection behaviour.

I haven't tried any of the 5.x series yet personally.

Andypoo.
I've not captured the kernel panic text, except for the last bit as it has scrolled off the terminal. I believe I need a usb version of a null modem cable to do so. I can look into that, or see if I have any parts around suitable.

I'm using btrfs as root and root is in a subvolume, which I found a workaround to do. I've produced a script to call mkinitrd and create the config file for syslinux, so when everything works it is almost effortless to ugrade a kernel I've just built.

The part of the script that calls mkinitrd: It calls it then make a few changes to the tree and call it again to build the final initrd:

mkinitrd -c -k ${KERNELVER} -f ext4 \
-m usb-storage:ehci-hcd:ehci-pci:xhci-pcihci-pci:xhci-hcd:uhci-hcd:hid:usbhid:i2c-hid:h
id_generic:hid-cherry:hid-logitech:hid-logitech-dj:hid-logitech-hidpp:hid-lenovo:hid-microsoft:hi
d_multitouch:ext4:btrfs \
-L -u -o ${IMAGENAME}

#Try to hack the initrd for btrfs
echo ${ROOTFSTYPE} > /boot/initrd-tree/rootfs
echo ${DEV} > /boot/initrd-tree/rootdev

#Rerun to incorporate changes.
mkinitrd


This list of kernel modules it a bit ott, but I thought that too much was better than too little. Also I compile in ext4 and btrfs (as well as various others including F2FS) up to now. I've not noticed it complain about trying to include modules when actually btrfs and ext4 are compiled into the kernel.

The two variables above are:
ROOTFSTYPE='btrfs -o subvol=_r_sl_14.2'
DEV='/dev/disk/by-label/system'

I find that although appending the option and sub-volume to ROOTFSTYPE is not what was originally intended it is a nice workaround to specify sub-volumes. I know I could not do this and simply pick a default sub-volume, but this gives me some flexibility if I want to boot to multiple ones.

I find I need initrd as although I typically compile in the filesystems directly rather than use modules I have several disks in the system and they never seem to stay in consistent positions using /dev/sdX and uuid or labels solve that issue.
 
1 members found this post helpful.
Old 10-17-2019, 11:47 PM   #1698
Didier Spaier
LQ Addict
 
Registered: Nov 2008
Location: Paris, France
Distribution: Slint64-14.2.1.2 on Lenovo Thinkpad W520
Posts: 9,013

Rep: Reputation: Disabled
commit 95bcc0d980ae3e2efc2e991d3fb6a1f7e59f6026 in 4.19.80 should make some people happy.
 
Old 10-18-2019, 10:46 PM   #1699
bassmadrigal
LQ Guru
 
Registered: Nov 2003
Location: West Jordan, UT, USA
Distribution: Slackware
Posts: 6,204

Rep: Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945Reputation: 3945
Quote:
Originally Posted by Didier Spaier View Post
commit 95bcc0d980ae3e2efc2e991d3fb6a1f7e59f6026 in 4.19.80 should make some people happy.
For those curious but not quite curious enough to look it up (I was almost there, but my curiosity got the best of me), it should fix the F2FS problems (and maybe btrfs, but I haven't paid enough attention to know how connected they are).

https://github.com/torvalds/linux/co...b6a1f7e59f6026
 
1 members found this post helpful.
Old 10-19-2019, 05:45 AM   #1700
Didier Spaier
LQ Addict
 
Registered: Nov 2008
Location: Paris, France
Distribution: Slint64-14.2.1.2 on Lenovo Thinkpad W520
Posts: 9,013

Rep: Reputation: Disabled
Yes, sorry to not have been explicit at first. It is my understanding that this cshould solve the issue initially reported by Andypoo in this post. @Andypoo: can you please confirm or infirm? @rogan: ditto. And you or Andypoo could comment your bug report accordingly?

Last edited by Didier Spaier; 10-19-2019 at 08:32 AM.
 
Old 10-19-2019, 07:49 AM   #1701
petejc
LQ Newbie
 
Registered: Apr 2019
Distribution: Slackware
Posts: 29

Rep: Reputation: Disabled
Quote:
Originally Posted by bassmadrigal View Post
For those curious but not quite curious enough to look it up (I was almost there, but my curiosity got the best of me), it should fix the F2FS problems (and maybe btrfs, but I haven't paid enough attention to know how connected they are).

https://github.com/torvalds/linux/co...b6a1f7e59f6026
5.3.7 does not seem to boot with btrfs as root, noting that F2FS is not complied in. I'm looking at getting a USB null modem cable so I can see what is going on as there is always a chance it is a glitch in my config.
 
Old 10-19-2019, 08:51 AM   #1702
rogan
Member
 
Registered: Aug 2004
Distribution: Slackware
Posts: 169

Rep: Reputation: 68
That initrd is WAY more that I've ever needed to use. With lilo it's usually just:
mkinitrd -c -k 5.3.7 -m btrfs -f btrfs -r /dev/sda3 -B -o /boot/initrd-5.3.7.gz -s /boot/initrd-tree-5.3.7
If you use the standard /boot/initrd.gz name and no btrfs raid you could probably get away with
mkinitrd -c -k 5.3.7 -m btrfs
as all modules for peripherals gets loaded once the root device is accessible.
I haven't had any particular problem with the 5.3 series yet, but I haven't tried the latest.
none of my machines are running current at this time, but I might try 5.3.7 later and report back.
EDIT: 5.3.7 works just fine with a 4.19.75 generic config on current (on btrfs).
5.3.7 is also MUCH better on system resources. It has reduced cpu usage
dramatically (50-60% less on peak) when copying kernel source trees between
drives using btrfs. I could easily hit 60-70% system total cpu on an eight core
amd @4.0Ghz when doing this on 4.19.x, now it peaks at 20-30%.

Last edited by rogan; 10-20-2019 at 04:09 AM. Reason: done some more testing
 
Old 10-19-2019, 11:48 AM   #1703
petejc
LQ Newbie
 
Registered: Apr 2019
Distribution: Slackware
Posts: 29

Rep: Reputation: Disabled
Quote:
Originally Posted by rogan View Post
That initrd is WAY more that I've ever needed to use. With lilo it's usually just:
mkinitrd -c -k 5.3.7 -m btrfs -f btrfs -r /dev/sda3 -B -o /boot/initrd-5.3.7.gz -s /boot/initrd-tree-5.3.7
If you use the standard /boot/initrd.gz name and no btrfs raid you could probably get away with
mkinitrd -c -k 5.3.7 -m btrfs
as all modules for peripherals gets loaded once the root device is accessible.
I haven't had any particular problem with the 5.3 series yet, but I haven't tried the latest.
none of my machines are running current at this time, but I might try 5.3.7 later and report back.
EDIT: 5.3.7 works just fine with a 4.19.75 generic config on current.
I presume you are replying to my post? I suspect I could delete all the modules as btrfs is complied in. I find I just need initrd as I'm using subvolumes and uuids or labels and lilo does not support that. I think the cable is now an essential to see what is going on.

Edit, just to clarify, I had that list of modules as that was my first list of modules that was needed to acheive boot. And since it booted and did not want to play around, just get booted I left it.

Last edited by petejc; 10-19-2019 at 11:49 AM.
 
Old 10-19-2019, 12:09 PM   #1704
rogan
Member
 
Registered: Aug 2004
Distribution: Slackware
Posts: 169

Rep: Reputation: 68
I use subvolumes as well and it work perfectly well with lilo (not that it actually cares). I you have a
habit of swapping drives around you can tell the kernel what root device you prefer with
append="root=LABEL=my_label"
I have concluded testing of 4.19.80 "huge" and it works fine (with btrfs).

I just came to think of one thing... You don't actually specify a subvolume as a
boot parameter do you?
Dont do that! When you create a btrfs file system for root use, you should also
create a subvolume, for example btrfs subvolume create 191019 and then set that
default: btrfs subvolume set-default 191019 (with btrfs-tools on current).
This way you can mount the root volume (subvol=/ ) somewhere and create snapshots
there before e.g. a radical upgrade. Should it fail you can easily just set the
snapshot default instead and boot like nothing happened.
Keep everything as simple as possible. Anyway thats how I do it...

Last edited by rogan; 10-19-2019 at 12:23 PM.
 
Old 10-19-2019, 02:38 PM   #1705
petejc
LQ Newbie
 
Registered: Apr 2019
Distribution: Slackware
Posts: 29

Rep: Reputation: Disabled
Quote:
Originally Posted by rogan View Post
I use subvolumes as well and it work perfectly well with lilo (not that it actually cares). I you have a
habit of swapping drives around you can tell the kernel what root device you prefer with
append="root=LABEL=my_label"
I have concluded testing of 4.19.80 "huge" and it works fine (with btrfs).
You are suggesting that if I use 'append' I don't need to specify the root? E.g.

...
image=/boot/vmlinuz
append="root=LABEL=system"
read-only

...

For a working system to boot using the file system with the label 'system'?




Quote:
I just came to think of one thing... You don't actually specify a subvolume as a
boot parameter do you?
Dont do that! When you create a btrfs file system for root use, you should also
create a subvolume, for example btrfs subvolume create 191019 and then set that
default: btrfs subvolume set-default 191019 (with btrfs-tools on current).
This way you can mount the root volume (subvol=/ ) somewhere and create snapshots
there before e.g. a radical upgrade. Should it fail you can easily just set the
snapshot default instead and boot like nothing happened.
Keep everything as simple as possible. Anyway thats how I do it...
Yes, I have been putting root in a subvolume and snapshotting it. This has been very useful on occasion.
No, I don't set a default. I've been appending the subvolume information into the initrd tree. This has allowed me to multi-boot off the same btrfs. THat was breifly useful when upgrading slackware, it seems a cool thing to do. I do see your point however.

I wonder if I could specify the btrfs subvolume via rootflags on the kernel command line?

One tricky thing left, I'd have to ensure that the boot=<location> directive pointed at the right place, as wandering disks have been an issue. I have been using usb sticks with syslinux, this has worked nicely as if I hose the config I just boot with the other stick (assuming I've not corrupted that as well).
 
Old 10-19-2019, 03:18 PM   #1706
rogan
Member
 
Registered: Aug 2004
Distribution: Slackware
Posts: 169

Rep: Reputation: 68
Quote:
Originally Posted by petejc View Post
You are suggesting that if I use 'append' I don't need to specify the root? E.g.

...
image=/boot/vmlinuz
append="root=LABEL=system"
read-only

...

For a working system to boot using the file system with the label 'system'?

Exactly! I have tried This myself on occasion.


Quote:
Yes, I have been putting root in a subvolume and snapshotting it. This has been very useful on occasion.
No, I don't set a default. I've been appending the subvolume information into the initrd tree. This has allowed me to multi-boot off the same btrfs. THat was breifly useful when upgrading slackware, it seems a cool thing to do. I do see your point however.
Multibooting into different subvolumes is a use-case I did not consider. One (Slackware) linux system on a computer
tend to be about all I can sanely handle.

Quote:
I wonder if I could specify the btrfs subvolume via rootflags on the kernel command line?
Look it up! This is going OT so I suggest to create a new thread if you want more specific help.

Last edited by rogan; 10-19-2019 at 03:21 PM.
 
Old 10-20-2019, 08:38 PM   #1707
cwizardone
Senior Member
 
Registered: Feb 2007
Distribution: Slackware64-current with "True Multilib" & Xfce.
Posts: 4,908

Original Poster
Blog Entries: 1

Rep: Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317Reputation: 2317
5.4-rc4

The newest "mainline" development kernel, version 5.4-rc4, has been been made available for testing.

The tarball, https://git.kernel.org/torvalds/t/linux-5.4-rc4.tar.gz

Mr. Torvalds' announcement, http://lkml.iu.edu/hypermail/linux/k...0.2/05030.html

Last edited by cwizardone; 10-20-2019 at 08:39 PM.
 
2 members found this post helpful.
Old 10-21-2019, 04:04 AM   #1708
rogan
Member
 
Registered: Aug 2004
Distribution: Slackware
Posts: 169

Rep: Reputation: 68
Using 5.4-rc4 here. After enabling a few extra features I've been having 'stalls' when doing heavy
background stuff. investigating...
 
1 members found this post helpful.
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
Linux.conf.au: Latest Linux kernel release due early March DragonSlayer48DX Linux - News 0 01-18-2010 10:43 PM
No video on latest kernel release Tralce Linux - Kernel 3 11-30-2006 07:48 AM
What is the latest Redhat release TILEMANN Linux - Software 5 11-20-2006 10:48 PM
LXer: News: OpenVZ To Release Support, Patches for Latest Kernel LXer Syndicated Linux News 0 11-01-2006 10:54 PM
latest debian release? doralsoral Linux - Software 5 12-25-2004 12:40 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 01:04 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration