Black screen on boot after 20.04 upgraded to kernel 5.8

Bug #1911352 reported by Ben Hoyt
40
This bug affects 7 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Confirmed
Undecided
Unassigned

Bug Description

Yesterday Ubuntu asked me to reboot after automatic updates, and when I rebooted it doesn't start up. Sometimes it stays on the Dell logo, sometimes it goes to a black screen and doesn't get past that (not sure what causes each variation -- perhaps for reboot it does the Dell logo, cold start up it does the black screen?).

After the updates it's trying to boot the 5.8.0 kernel. In the boot menu, if I manually select the 5.4.0 kernel everything works fine. So presumably it's something about the 5.8 kernel on this hardware.

When this occurred I was running the Nouveau open source graphics driver. I switched to the NVIDIA driver (460) to see if that would help, but it didn't change anything. So it may or may not be graphics related.

I'm attaching my /var/log/kern.log and /var/log/syslog for yesterday -- Jan 12, I think it would be around 13:30 that I first rebooted after the upgrade (and a bunch of times after that). I ran the logs through grep -v 'var-snap-lxd-common-lxd' as a huge percentage of the lines were this (presumably irrelevant):

Jan 12 19:09:55 BenXPS15 kernel: [14515.297170] audit: type=1400 audit(1610431795.003:29126): apparmor="DENIED" operation="open" namespace="root//lxd-juju-804318-0_<var-snap-lxd-common-lxd>" profile="snap.juju-db.daemon" name="/proc/631/net/netstat" pid=3906 comm="ftdc" requested_mask="r" denied_mask="r" fsuid=1000000 ouid=1000000

For reference:

* This is a Dell XPS 15 (9550) laptop with an NVIDIA GeForce GTX 960M (I think it's hybrid graphics). i7-6700HQ 2.6GHz with 8 cores, 256 GB SSD, 16 MB RAM.
* I installed Ubuntu 20.04 Focal in August 2020, and now it says it's on "Ubuntu 20.04.1 LTS".
* /proc/version_signature: Ubuntu 5.4.0-59.65-generic 5.4.78
* "sudo lspci -vnvn" output is included in the attached logs.zip
---
ProblemType: Bug
ApportVersion: 2.20.11-0ubuntu27.14
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: ben 4518 F.... pulseaudio
 /dev/snd/pcmC0D0p: ben 4518 F...m pulseaudio
CasperMD5CheckResult: skip
CurrentDesktop: ubuntu:GNOME
DistroRelease: Ubuntu 20.04
InstallationDate: Installed on 2020-08-27 (138 days ago)
InstallationMedia: Ubuntu 20.04.1 LTS "Focal Fossa" - Release amd64 (20200731)
MachineType: Dell Inc. XPS 15 9550
NonfreeKernelModules: zfs zunicode zavl icp zcommon znvpair nvidia_modeset nvidia
Package: linux (not installed)
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.4.0-59-generic root=/dev/mapper/vgubuntu-root ro quiet splash
ProcVersionSignature: Ubuntu 5.4.0-59.65-generic 5.4.78
RelatedPackageVersions:
 linux-restricted-modules-5.4.0-59-generic N/A
 linux-backports-modules-5.4.0-59-generic N/A
 linux-firmware 1.187.7
Tags: focal
Uname: Linux 5.4.0-59-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip kvm libvirt lpadmin lxd plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 12/12/2019
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 1.13.1
dmi.board.name: 0N7TVV
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 9
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr1.13.1:bd12/12/2019:svnDellInc.:pnXPS159550:pvr:rvnDellInc.:rn0N7TVV:rvrA00:cvnDellInc.:ct9:cvr:
dmi.product.family: XPS
dmi.product.name: XPS 15 9550
dmi.product.sku: 06E4
dmi.sys.vendor: Dell Inc.

Revision history for this message
Ben Hoyt (benhoyt) wrote :
Revision history for this message
Ubuntu Kernel Bot (ubuntu-kernel-bot) wrote : Missing required logs.

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1911352

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

Changed in linux (Ubuntu):
status: New → Incomplete
tags: added: focal
Revision history for this message
Terry Rudd (terrykrudd) wrote :

Ben, can you tell us if it was 5.8.0-34 kernel or later?

Revision history for this message
Ben Hoyt (benhoyt) wrote : AlsaInfo.txt

apport information

tags: added: apport-collected
description: updated
Revision history for this message
Ben Hoyt (benhoyt) wrote : CRDA.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : CurrentDmesg.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : IwConfig.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : Lspci.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : Lspci-vt.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : Lsusb.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : Lsusb-t.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : Lsusb-v.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : ProcCpuinfo.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : ProcCpuinfoMinimal.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : ProcEnviron.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : ProcInterrupts.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : ProcModules.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : PulseList.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : RfKill.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : UdevDb.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : WifiSyslog.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote : acpidump.txt

apport information

Revision history for this message
Ben Hoyt (benhoyt) wrote :

The kernel it's trying to run (assuming it's the top one in the "advanced options" boot list) is 5.8.0-36-generic.

I run apport-collect in case it helps -- boy, that's noisy in terms of number of comments. :-)

Ben Hoyt (benhoyt)
Changed in linux (Ubuntu):
status: Incomplete → Confirmed
Revision history for this message
AceLan Kao (acelankao) wrote :

Can't reproduce this on XPS 15 9560

BIOS Information
        Vendor: Dell Inc.
        Version: 1.20.0
        Release Date: 07/21/2020

Linux acelan-xps15-9560 5.8.0-36-generic #40~20.04.1-Ubuntu SMP Wed Jan 6 10:15:55 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

nvidia, 460.32.03, 5.8.0-36-generic, x86_64: installed

Revision history for this message
Ben Hoyt (benhoyt) wrote :

Good data point, thanks. It looks like there are some significant differences between the 9550 and the 9560 that could explain that:

1) Next generation CPU: i7-6700HQ on the 9550 vs i7-7700HQ on the 9560
2) Intel HD graphics 530 on the 9550 vs 630 on the 9560
3) NVIDIA GTX 960M on the 9550 vs NVIDIA GTX 1050 on the 9560

Reference: https://www.windowscentral.com/dell-xps-15-9550-vs-xps-15-9560-whats-different-and-should-you-upgrade

Revision history for this message
Wolf Pichler (w-pichler) wrote :

I have the same problem on my Acer.

Since Kernel 5.8.0-36 and also 5.8.0-38 Ubuntu 20.04.1 cannot start in graphics mode.
(It is possible to log in on the console using e.g. Ctrl+Alt+F4 though.)

With Kernel 5.4.0-58 the system works as expected.

The problem occurs only on this system, which is also the only one with a
Nvidia GPU (Nvidia drivers are currently not installed though because I initially suspected the
updated nvidia-driver-460 were the problem.)

dmidecode -t1
# dmidecode 3.2
Getting SMBIOS data from sysfs.
SMBIOS 3.2.0 present.

Handle 0x0001, DMI type 1, 27 bytes
System Information
 Manufacturer: Acer
 Product Name: Aspire A515-54G
 Version: V1.16
 Serial Number: I deleted it.
 UUID: I deleted it.
 Wake-up Type: Power Switch
 SKU Number: 0000000000000000
 Family: Aspire 5

Revision history for this message
Wolf Pichler (w-pichler) wrote :

Maybe contains extra information.

Revision history for this message
Charles Green (chick) wrote :

I've got the same issue on a system76 gazelle - I just reinstalled 20.04 hoping that this would fix the issue, but the 5.8.0-38 kernel does not complete the boot cycle by itself.

I have found that switching to tty2 and then back to tty1 causes the boot to complete

Revision history for this message
Ben Hoyt (benhoyt) wrote :

Charles, very interesting. Though I wonder if yours is a different problem, partly because it's not in tty mode at all (for me it's in a graphical mode, sometimes with the Dell logo up). I can't seem to switch to tty2 using Alt-F2 or Ctrl-Alt-F2 -- what are the exact keys you're using?

Revision history for this message
Charles Green (chick) wrote : Re: [Bug 1911352] Re: Black screen on boot after 20.04 upgraded to kernel 5.8

The ctrl+alt+f2 - this was a fresh install on a system76 machine,
running the integrated graphics.  I have since loaded the NVidia drivers
and my some miracle, I no longer need this work around.

The ttys run from 1 to 7, I think, but the keyboard shortcuts to
activate them are always ctrl+alt+f(num)

On 1/20/21 12:18 PM, Ben Hoyt wrote:
> Charles, very interesting. Though I wonder if yours is a different
> problem, partly because it's not in tty mode at all (for me it's in a
> graphical mode, sometimes with the Dell logo up). I can't seem to switch
> to tty2 using Alt-F2 or Ctrl-Alt-F2 -- what are the exact keys you're
> using?
>

Revision history for this message
Ben Hoyt (benhoyt) wrote :

Just a note: in attempting to fix this issue I borked my laptop - after disabling UEFI boot and then re-enabling it in the BIOS settings, I couldn't boot at all. So I reinstalled Ubuntu and then was getting the following issue after doing system updates (but still on the 5.4 kernel): https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1842320, which showed "error: out of memory" when GRUB was trying to hand over to the kernel. The fix (workaround) for that is to set GFXMODE=800x600 in the GRUB config. And now the software updates *and* the 5.8 kernel works fine ... which is kind of odd.

It's possible that this issue was actually that one, just that we weren't seeing the error message because GRUB is set to hidden by default. Though that doesn't explain why the 5.4 kernel worked with software updates earlier ... still a bit of an unknown, so keeping this open for now.

Revision history for this message
Terry Rudd (terrykrudd) wrote :

If anyone affected by this bug can still reproduce it in their environments, I would like them to try installing linux-modules-extra and rebooting to see if the issue clears. You will need to do this with your 5.4 kernel.

Thanks in advance,
Terry

Revision history for this message
Wolf Pichler (w-pichler) wrote :

I have these installed:

dpkg -l | grep "linux\-[a-z]*\-"

rc linux-modules-extra-5.4.0-42-generic 5.4.0-42.46
ii linux-modules-extra-5.4.0-58-generic 5.4.0-58.64
ii linux-modules-extra-5.8.0-36-generic 5.8.0-36.40~20.04.1
ii linux-modules-extra-5.8.0-38-generic 5.8.0-38.43~20.04.1

I tried as Charles Green said: Switching to tty2 and then back to tty1 causes the boot to complete.
(Ctrl+Alt+F2, Ctrl+Alt+F1)

Revision history for this message
Terry Rudd (terrykrudd) wrote :

It seems that in at least 1 case, this condition arises with linux-modules-extra installed so my suggestion above may not help

Revision history for this message
Charles Green (chick) wrote :

I have also had the boot problem disappear, return, and disappear again.....

On 1/22/21 11:12 AM, Wolf Pichler wrote:
> I have these installed:
>
> dpkg -l | grep "linux\-[a-z]*\-"
>
> rc linux-modules-extra-5.4.0-42-generic 5.4.0-42.46
> ii linux-modules-extra-5.4.0-58-generic 5.4.0-58.64
> ii linux-modules-extra-5.8.0-36-generic 5.8.0-36.40~20.04.1
> ii linux-modules-extra-5.8.0-38-generic 5.8.0-38.43~20.04.1
>
> I tried as Charles Green said: Switching to tty2 and then back to tty1 causes the boot to complete.
> (Ctrl+Alt+F2, Ctrl+Alt+F1)
>

Revision history for this message
Kleber Sacilotto de Souza (kleber-souza) wrote :

My laptop running Groovy (20.10) is affected by this bug as well, but I also can't figure or at least isolate where the problem could be. The hardware is similar to what other people have reported (it's a ASUS K501UQ, with a Core i7-6500U, Nvidia 940MX and Intel HD 520).

It seemed to have started after I installed linux-modules-nvidia-460-generic but even after uninstalling the nvidia drivers it still hangs once in a while. I also tried blackisting nouveau but without success. So I'm not sure this is directly related to any nvidia modules as it was happening even without loading them. I couldn't reliably reproduce the issue with any of combinations between 3 versions of 5.8 kernels installed and nvidia/nouveau.

In my case it seems to hand right after loading the kernel and initrd, but I don't get any message on the screen or on the logs and the keyboard is totally unresponsive (I can't switch between tty's).

The only data point that seems to be consistent so far is that I wasn't able to reproduce it when booting the kernel on recovery mode.

Revision history for this message
Wolf Pichler (w-pichler) wrote :

The problem persists with Kernel 5.8.0-50, but I found out that it only occurs if the usage of the Nvidia GPU is set to "on-demand".

In pure Intel or Nvidia mode the system boots as expected.

The mode van be set via graphical Nvidia tool or via "prime-select {nvidia|intel|on-demand|query}".

NVIDIA Driver Version: 460.56.

Revision history for this message
Wolf Pichler (w-pichler) wrote :

After having played around with "prime-select" instead of using the graphical Nvidia tool my system now also boots in "on-demand" mode.

(You have to reboot after changing the settings with "prime-select".)

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.