upgraded clean install of Home Edition with kernel 4.2 to 4.8 = hard lock ups

[galen]

I did a clean install of Home Edition kernel 4.2 onto another partition [2] of my Dell 1525 Laptop.
rebooted a few time to test if ok.
then did a full safe upgrade to 4.8 kernel.
rebooted when finished.
that install is now hard locking on bootup , it displays Call Trace errors then hangs, no TTY
I was able to boot its older 4.2 kernel  and get LXDE.
4.8 is critically broken :(

[galen]

updates more testing.

rebooted both of my sparky HE with kernel 4.8 installs.

using the following kernel parameters these results happened:

Code Select Expand


verbose = hard hang, with Call Trace[s], unplug needed
verbose = hard hang, with Call Trace[s], unplug needed
splash = blank screen, hard lockup
quiet splash = hard lock as above
quiet splash = pointer only after 3min wait, soft reset possible
quiet nosplash = hard lockup but with partial desktop panel outline onle
quiet splash irqfixup = ok

when the boot process allows I see CTRC26 pipe A [or B]  error

note: underlining is not by design

[Dai_trying]

You could put your terminal output into code tags to prevent bbcode translation (causing your underline) select the terminal output and click the code link(#) and it will put you text into a code box

[galen]

I filed a debian bug
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=845260
they closed it.
Subject: Re: #845260: incorrect bug tracker
The Debian BTS only tracks bugs in Debian; this bug appears to be
about some other distribution. (Also, the assigned package is
probably incorrect.)
:(

[paxmark1]

Please post    "inxi -Cpl"    example on a lvm debian testing machine   (To other readers - Yes, I know I am supposed to be using labels instead of uuid in lvm)

Code Select Expand

root@raunes:/# inxi -Cpl
CPU:       Dual core Intel Core i3-4170 (-HT-MCP-) cache: 3072 KB
           clock speeds: max: 3700 MHz 1: 1091 MHz 2: 1028 MHz 3: 1082 MHz 4: 1102 MHz
Partition: ID-1: / size: 9.1G used: 6.4G (75%) fs: ext4 dev: /dev/dm-0 label: N/A
           ID-2: /boot size: 237M used: 62M (28%) fs: ext2 dev: /dev/sda2 label: N/A
           ID-3: /boot/efi size: 511M used: 132K (1%) fs: vfat dev: /dev/sda1 label: N/A
           ID-4: /home size: 1.5T used: 448G (32%) fs: ext4 dev: /dev/dm-2 label: N/A
           ID-5: /media/paxmark/CDF2-6BE2 size: 7.5G used: 1.4G (18%) fs: vfat dev: /dev/sdb label: N/A
           ID-6: swap-1 size: 12.88GB used: 0.00GB (0%) fs: swap dev: /dev/dm-1 label: N/A


If you have two linux OS's on the same computer sharing the same /home - sometimes the "dot" files (like .share .config etc.) will fight.  Probably not in this case. 

I only use command line for upgrades.  apt-get upgrade is NOT going to be sufficient at times, similar to safe-upgrade.  You are on a system that is  98+% debian testing.  Imho for kernel upgrade "apt-get upgrade" is wrong.  I use "apt", but for "apt-get" do  this time  (as root or via sudo)

dpkg --configure -a
apt-get update
apt-get install -f                   ##  -f flag is fix, it is not force
apt-get dist-upgrade -d      ##   download only

and then after looking at all the proposed changes

apt-get dist-upgrade          ##  If this fixes it - for the future do not do the dpkg --configure -a or the apt-get install -f

Always be careful when you perform updates and check if the actions proposed by the package managing tools are in line with your wishes and expectations. (i.e. make sure that you do not remove a plethora of packages you need by blindly accepting the proposed action)
Bearing this in mind, apt-get dist-upgrade to keep your system fully up-to-date, but if the proposed changes do look unreasonable, some of the simpler things that could help are:
put packages on hold until the problem in the archive is resolved,
use apt-get upgrade to avoid removals this time,
simply wait until the archive has settled down to a more reasonable state before upgrading.
Install the apt-listbugs and apt-listchanges packages in order to be made aware of grave bugs or important changes when you install new packages or during an upgrade.

From https://wiki.debian.org/DebianUnstable#What_are_some_best_practices_for_testing.2Fsid_users.3F    Pardon the big words and stilted language. 


This is supposed to be fun.  If it is a constant struggle, please re-evaluate.  Maybe Debian stable is a better choice.  Another option is instead of two Sparky linux versions on the same machine - why not one Sparky and one Debian testing.  Debian does not have the time or resources to offer support for derivative distros, especially since many are not as well crafted as Sparky.  With a plain Debian testing option you can file bug reports on Debian BTS.  Another option is virtualbox  or the opensource kvm-qemu.  And please read "man apt-get"    peace out.

https://wiki.debian.org/DebianTesting

[galen]

great info,
precise

nope I don't share home users on multi-boot system, I've been there learned my lessons.

the upgrade path is broken,
this happens to nearly all my Sparky installs on varying machines ,
more quality control is needed