LinuxQuestions.org
Welcome to the most active Linux Forum on the web.
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 11-01-2019, 06:14 AM   #1
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Rep: Reputation: Disabled
Cannot boot, kernel panic, to 5.3.x kernels and not sure how to debug.


I get a kernel panic when trying to boot to 5.3.8, and previous 5.3.x kernels. However, no issue with 5.2.x series kernels. I'm using syslinux for booting.

This does not work and the kernel panics:
LABEL desk-system-label-5.3.8
LINUX vmlinuz-5.3.8-amd
APPEND ro console=ttyUSB0,115200n8 console=tty0 earlyprintk=serial,ttyUSB0,115
200,keep
INITRD initrd_desk-system-label-5.3.8.gz

This does work:
LABEL desk-system-label-5.2.21
LINUX vmlinuz-5.2.21-amd
APPEND ro console=ttyUSB0,115200n8 console=tty0 earlyprintk=serial,ttyUSB0,115
200,keep
INITRD initrd_desk-system-label-5.2.21.gz

The 5.3.8 kernel is built using the config for the 5.2.21 as a starting point (e.g. make oldconfig). I'm using a FDTI usb to usb null modem cable to see if I can capture the panic. On the 5.3.8 the panic is not captured. On 5.2.21 the serial console is captured over the null modem cable. However, with 5.2.21 serial output is much delayed as it is is being buffered and held through the first 10 - 30 seconds of the boot process. So it appears that 5.3.8 might be panicking before it starts sending console info over the cable.

mkinitrd config is the same.

Of course, the panic info on the vga console has scrolled the useful stuff off the screen before I can see it, but it appears to be occurring sometime as it is discovering drives.

I'm wondering if I can video (with a camera pointing at he monitor) the vga console to see if I can get hints that way, but that is a fairly long shot.

Any suggestions?
 
Old 11-01-2019, 11:00 PM   #2
Skaendo
Senior Member
 
Registered: Dec 2014
Location: West Texas, USA
Distribution: Slackware64-14.2
Posts: 1,044

Rep: Reputation: Disabled
Instead of using a camera, you could try in a VM as most VMs can record. That way you can slow the video down and the clarity will be much better.

As far as debugging, out of my league.
 
Old 11-02-2019, 06:06 AM   #3
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by Skaendo View Post
Instead of using a camera, you could try in a VM as most VMs can record. That way you can slow the video down and the clarity will be much better.

As far as debugging, out of my league.
I kind of see your point, perhaps there is a syslinux (extlinux) 5.3.X series incompatibility? However, the issue is showing up on a machine running bare metal with a number of drives and and unless such an incompatibility was flagged I don't think the virtual machine, being so different, would tell me much. Or whether I just introduce virtual machine issues.

I've managed to get the machine booted via lilo with 5.3.8 so it is not the kernel fundamentally being at issue. Though I like the usb stick plan as I can always swap sticks if I get an unbootable config. I've unpacked one of my initrds and cannot see any issue there.

I ought to stry building another boot usb stick with extlinux from slackware-current as I think that has a newer version of syslinux than -14.2 to see if that makes any difference.
 
Old 11-02-2019, 06:52 AM   #4
hazel
Senior Member
 
Registered: Mar 2016
Location: Harrow, UK
Distribution: LFS, AntiX, Slackware
Posts: 3,546
Blog Entries: 9

Rep: Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969
With the lilo boot, did you use exactly the same command line arguments? The problem could be in one of those.

My previous desktop machine wouldn't boot recent kernels unless I added acpi=off.
 
Old 11-02-2019, 07:37 AM   #5
igadoter
Senior Member
 
Registered: Sep 2006
Location: wroclaw, poland
Distribution: many, primary Slackware
Posts: 1,404
Blog Entries: 1

Rep: Reputation: Disabled
Quote:
Originally Posted by Skaendo View Post
Instead of using a camera, you could try in a VM as most VMs can record. That way you can slow the video down and the clarity will be much better.

As far as debugging, out of my league.
I think the only reason to run this on VM is to verify kernel will boot in virtual environment. To exclude option failure is hardware related.
 
Old 11-03-2019, 06:46 AM   #6
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by hazel View Post
With the lilo boot, did you use exactly the same command line arguments? The problem could be in one of those.

My previous desktop machine wouldn't boot recent kernels unless I added acpi=off.
No, it was quite different. I set my btrfs slackware root sub volume as default using btrfs-progs so that I did not have to somehow specify a boot subvolume and I passed no command line argments. I did not use an initrd. With syslinux (actually extlinux) I used:
APPEND ro console=ttyUSB0,115200n8 console=tty0 earlyprintk=serial,ttyUSB0,115
200,keep

Your experience with acpi is interesting, but annoying I would assume.
 
Old 11-03-2019, 07:17 AM   #7
hazel
Senior Member
 
Registered: Mar 2016
Location: Harrow, UK
Distribution: LFS, AntiX, Slackware
Posts: 3,546
Blog Entries: 9

Rep: Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969
Yes, very annoying. I certainly didn't feel safe with acpi switched off. But I managed, with a bit of help, to find a patch that worked on the later kernels to make them bootable with acpi running.

The point is you now have a handle on what is going wrong. One of those syslinux command line options is triggering the panic. All you have to do is try removing them experimentally one by one until you find the culprit.
 
Old 11-03-2019, 03:33 PM   #8
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by hazel View Post
Yes, very annoying. I certainly didn't feel safe with acpi switched off. But I managed, with a bit of help, to find a patch that worked on the later kernels to make them bootable with acpi running.

The point is you now have a handle on what is going wrong. One of those syslinux command line options is triggering the panic. All you have to do is try removing them experimentally one by one until you find the culprit.
I don't think it is that simple, as they were added for debugging only after I had issues. So perhaps it is 'ro'. Must try.
 
Old 11-06-2019, 10:22 AM   #9
PROBLEMCHYLD
Member
 
Registered: Apr 2015
Posts: 807

Rep: Reputation: Disabled
Try this, I built mines last night and got a panic when I installed this morning. I'm building 5.3.9 to see if it works.

https://bugzilla.kernel.org/show_bug.cgi?id=204883
 
Old 11-06-2019, 01:38 PM   #10
PROBLEMCHYLD
Member
 
Registered: Apr 2015
Posts: 807

Rep: Reputation: Disabled
5.3.9 installed but this error here https://www.linuxquestions.org/quest...ml#post6054904
 
Old 11-06-2019, 04:52 PM   #11
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by PROBLEMCHYLD View Post
Try this, I built mines last night and got a panic when I installed this morning. I'm building 5.3.9 to see if it works.

https://bugzilla.kernel.org/show_bug.cgi?id=204883
Interesting, I don't think I have this module loading. Can boot from lilo but not syslinux (using tried and tested config). Network interfaces:
lspci | egrep -i --color 'network|ethernet'
06:00.0 Ethernet controller: Intel Corporation I211 Gigabit Network Connection (rev 03)
07:00.0 Network controller: Realtek Semiconductor Co., Ltd. Device b822
0b:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)
0b:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)

Have to build 5.3.9 but without the realtek 8169.
 
Old 11-07-2019, 06:47 PM   #12
TheRealGrogan
Member
 
Registered: Oct 2010
Location: Ontario, Canada
Distribution: Slackware, Manjaro (for gaming)
Posts: 198

Rep: Reputation: 147Reputation: 147
Quote:
Originally Posted by petejc View Post
Interesting, I don't think I have this module loading. Can boot from lilo but not syslinux (using tried and tested config). Network interfaces:
lspci | egrep -i --color 'network|ethernet'
06:00.0 Ethernet controller: Intel Corporation I211 Gigabit Network Connection (rev 03)
07:00.0 Network controller: Realtek Semiconductor Co., Ltd. Device b822
0b:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)
0b:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06)

Have to build 5.3.9 but without the realtek 8169.
That shouldn't be your problem since you don't have the hardware, but indeed if you're building your own kernel, why even compile modules for things you're never going to have? I suppose if you want your kernel to survive a motherboard swap you might use a distro config, but it's not that hard to drop in a new kernel and modules directory from off system in that event. Either way, if you use an initrd (I don't), that would have to be rebuilt and dropped in.

This is interesting to me, because I use r8169 but I missed the "fun" because I compile mine (and the realtek PHY) right in the kernel. My stubbornness for last century's practices has paid off :-)

Last edited by TheRealGrogan; 11-07-2019 at 06:50 PM.
 
Old 11-08-2019, 05:07 AM   #13
hazel
Senior Member
 
Registered: Mar 2016
Location: Harrow, UK
Distribution: LFS, AntiX, Slackware
Posts: 3,546
Blog Entries: 9

Rep: Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969Reputation: 1969
Oh dear! I use that driver. But I haven't built such an advanced kernel yet, not even for LFS9 (that uses 5.2.8). Let's hope that when I finally get there, the problem will have been sorted out.
 
Old 11-08-2019, 02:39 PM   #14
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by TheRealGrogan View Post
That shouldn't be your problem since you don't have the hardware, but indeed if you're building your own kernel, why even compile modules for things you're never going to have? I suppose if you want your kernel to survive a motherboard swap you might use a distro config, but it's not that hard to drop in a new kernel and modules directory from off system in that event. Either way, if you use an initrd (I don't), that would have to be rebuilt and dropped in.
Clutching at straws is why I tried it. I thought it would have no effect, but since I'm failing with a long tried and tested boot config that has stopped working for no logical reason it seemed to make sense to aviod building something that was problematic, even if not relevant.

Regarding building unneeded kernel modules I was working on the idea that if I bought something new and shiny I'd rather have the module there already rather than have to figure out what additional modules I would need and then do a rebuild. I'd have to do a search but I realize you can do some sort of scan to see what you have currently loaded and copy that. I compile in directly the most likely to be needed file systems, as if they are definitely going to be loaded as modules you might as well build them into the kernel fully.
 
Old 11-08-2019, 02:43 PM   #15
petejc
Member
 
Registered: Apr 2019
Distribution: Slackware
Posts: 40

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by PROBLEMCHYLD View Post
Try this, I built mines last night and got a panic when I installed this morning. I'm building 5.3.9 to see if it works.

https://bugzilla.kernel.org/show_bug.cgi?id=204883
Well, it seems not to be that, but here is a screenshot showing the latter part of my kernel panic when trying to boot with extlinux and 5.3.9:
https://drive.google.com/file/d/1zre...ew?usp=sharing

Strangely I also failed to boot with lilo and 5.3.8, even though I had booted successfully. This is getting rather annoying.
 
  


Reply

Tags
console, serial, syslinux


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
gdb, debug with separate debug file and different source dir problem a4z Programming 5 10-17-2014 02:55 AM
I'm not sure how this happened but I'm sure there is a lesson in there somewhere mreff555 Linux - Newbie 6 09-24-2012 03:09 AM
linux hangs - not sure how to debug - GConf message shutting down solvetek Linux - Newbie 1 11-24-2011 08:39 AM
[SOLVED] Not able to debug a library compiled with debug information gregorian Programming 6 10-29-2011 10:53 AM
how debug kernel with ddd? how debug the kernel with ddd? mjalalynia Linux - Kernel 2 08-25-2008 07:50 AM

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

All times are GMT -5. The time now is 10:01 AM.

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