18

I just took a working Gateway P4 desktop computer and upgraded it from 16.04LTS all the way to 17.10 without any problem. As soon as I tried to install 18.04LTS, I had trouble booting.

I've reinstalled in many different ways, with no improvement in the boot problem. The Ubuntu logo appears and the dots flash a few times, then the mouse pointer freezes and the system is totally unresponsive. If I boot with the "quiet splash" removed from the kernel line in grub, I can see that it just stops shortly after starting the first user account 120.

Here is a clue... If I replace "quiet splash" with nomodeset, it boots fully, but it's in 640x480 mode only.

Here's the i915 video information...

  *-display UNCLAIMED
       description: VGA compatible controller
       product: 82945G/GZ Integrated Graphics Controller
       vendor: Intel Corporation
       physical id: 2
       bus info: pci@0000:00:02.0
       version: 02
       width: 32 bits
       clock: 33MHz
       capabilities: msi pm vga_controller bus_master cap_list
       configuration: latency=0
       resources: memory:e0300000-e037ffff ioport:20e0(size=8) memory:d0000000-dfffffff memory:e0380000-e03bffff memory:c0000-dffff

I suspect that I need to add a special i915 video controller kernel parameter to grub to make the video work properly. Any ideas on what parameter that might be... or any other ideas?

Update #1: I just discovered that if I boot into "Recovery Mode", and then choose "Resume normal boot", it boots fully, to a 1024x768 4:3 (should be 16:9) desktop.

Update #2: Upgrading from 17.10 to 18.04, booting fails. Reinstalling 18.04, booting fails. Disk partition delete and clean 18.04, booting fails. Installing 18.04 "proposed" software, booting still fails. Ubuntu Live DVD works fine.

Update #3: I found a list of i915 parameters...

~$ modinfo -p i915
modeset:Use kernel modesetting [KMS] (0=disable, 1=on, -1=force vga console preference [default]) (int)
panel_ignore_lid:Override lid status (0=autodetect, 1=autodetect disabled [default], -1=force lid closed, -2=force lid open) (int)
semaphores:Use semaphores for inter-ring sync (default: -1 (use per-chip defaults)) (int)
enable_rc6:Enable power-saving render C-state 6. Different stages can be selected via bitmask values (0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. default: -1 (use per-chip default) (int)
enable_dc:Enable power-saving display C-states. (-1=auto [default]; 0=disable; 1=up to DC5; 2=up to DC6) (int)
enable_fbc:Enable frame buffer compression for power savings (default: -1 (use per-chip default)) (int)
lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int)
lvds_use_ssc:Use Spread Spectrum Clock with panels [LVDS/eDP] (default: auto from VBT) (int)
vbt_sdvo_panel_type:Override/Ignore selection of SDVO panel mode in the VBT (-2=ignore, -1=auto [default], index in VBT BIOS table) (int)
reset:Attempt GPU resets (default: true) (bool)
error_capture:Record the GPU state following a hang. This information in /sys/class/drm/card<N>/error is vital for triaging and debugging hangs. (bool)
enable_hangcheck:Periodically check GPU activity for detecting hangs. WARNING: Disabling this can cause system wide hangs. (default: true) (bool)
enable_ppgtt:Override PPGTT usage. (-1=auto [default], 0=disabled, 1=aliasing, 2=full, 3=full with extended address space) (int)
enable_execlists:Override execlists usage. (-1=auto [default], 0=disabled, 1=enabled) (int)
enable_psr:Enable PSR (0=disabled, 1=enabled - link mode chosen per-platform, 2=force link-standby mode, 3=force link-off mode) Default: -1 (use per-chip default) (int)
alpha_support:Enable alpha quality driver support for latest hardware. See also CONFIG_DRM_I915_ALPHA_SUPPORT. (bool)
disable_power_well:Disable display power wells when possible (-1=auto [default], 0=power wells always on, 1=power wells disabled when possible) (int)
enable_ips:Enable IPS (default: true) (int)
fastboot:Try to skip unnecessary mode sets at boot time (default: false) (bool)
prefault_disable:Disable page prefaulting for pread/pwrite/reloc (default:false). For developers only. (bool)
load_detect_test:Force-enable the VGA load detect code for testing (default:false). For developers only. (bool)
force_reset_modeset_test:Force a modeset during gpu reset for testing (default:false). For developers only. (bool)
invert_brightness:Invert backlight brightness (-1 force normal, 0 machine defaults, 1 force inversion), please report PCI device ID, subsystem vendor and subsystem device ID to dri-devel@lists.freedesktop.org, if your machine needs it. It will then be included in an upcoming module version. (int)
disable_display:Disable display (default: false) (bool)
enable_cmd_parser:Enable command parsing (true=enabled [default], false=disabled) (bool)
use_mmio_flip:use MMIO flips (-1=never, 0=driver discretion [default], 1=always) (int)
mmio_debug:Enable the MMIO debug code for the first N failures (default: off). This may negatively affect performance. (int)
verbose_state_checks:Enable verbose logs (ie. WARN_ON()) in case of unexpected hw state conditions. (bool)
nuclear_pageflip:Force enable atomic functionality on platforms that don't have full support yet. (bool)
edp_vswing:Ignore/Override vswing pre-emph table selection from VBT (0=use value from vbt [default], 1=low power swing(200mV),2=default swing(400mV)) (int)
enable_guc_loading:Enable GuC firmware loading (-1=auto, 0=never [default], 1=if available, 2=required) (int)
enable_guc_submission:Enable GuC submission (-1=auto, 0=never [default], 1=if available, 2=required) (int)
guc_log_level:GuC firmware logging level (-1:disabled (default), 0-3:enabled) (int)
guc_firmware_path:GuC firmware path to use instead of the default one (charp)
huc_firmware_path:HuC firmware path to use instead of the default one (charp)
enable_dp_mst:Enable multi-stream transport (MST) for new DisplayPort sinks. (default: true) (bool)
inject_load_failure:Force an error after a number of failure check points (0:disabled (default), N:force failure at the Nth failure check point) (uint)
enable_dpcd_backlight:Enable support for DPCD backlight control (default:false) (bool)
enable_gvt:Enable support for Intel GVT-g graphics virtualization host support(default:false) (bool)

Update #4: I'm currently running with i915.modeset=0 in my grub kernel line, it's booting, and I'm getting 1024x768 (4:3).

Update #5: Updating to kernel 4.16.9 doesn't fix the problem.

Update #6: A portion of @karel's answer provided a good enough workaround. Switching from gdm3 display manager to lightdm display manager works. I could then also remove i915.modeset=0 from my grub kernel line.

sudo apt-get update
sudo apt-get install lightdm
# if lightdm wasn't already installed, it'll ask which dm you want
# if lightdm was already installed, then you need to:
sudo dpkg-reconfigure lightdm
# and choose lightdm over gdm3
reboot

Update #7:

Turns out that the problem isn't a gdm3 vs lightdm problem. It's a gdm3/wayland problem with older Intel GPU's. To fix...

In terminal...

  • cd /etc/gdm3 # change directory
  • sudo pico custom.conf # edit this file

Find and change:

#WaylandEnable=false

To this:

WaylandEnable=false

Save the file.

  • sudo dpkg-reconfigure gdm3 # select gdm3 DM

Select gdm3 and OK.

  • reboot # reboot the computer

Notes:

https://feeding.cloud.geek.nz/posts/linux-kernel-module-options-on-debian/

Setting kernel options for the i915

How can I fix broken i915 drivers for Intel GPUs?

heynnema
  • 70,711
  • Very similar symptoms after upgrade from 16 to 18.04 on an HP compact desktop with some built-in Intel video. Splash screen freezes after 4 red dots. Removing swap line from /etc/fstab results in freeze at black screen. Adding nomodeset to boot parameters allowed low-resolution boot. Un-commenting WaylandEnable=false in /etc/conf/custom.conf appears to have resolved the problem. – KalleMP Apr 27 '19 at 07:46

2 Answers2

14

Try logging in to your default desktop environment from the tty virtual console. The virtual console can be accessed by pressing the keyboard combination Ctrl+Alt+F3 after the computer gets stuck when booting it. Login to the virtual console with your username and password and run the following command.

sudo systemctl start graphical.target

If that doesn't work, switch the login display manager from gdm3 to lightdm. LightDM is a more lightweight login display manager than gdm3.

sudo apt install lightdm   
sudo dpkg-reconfigure lightdm 
sudo reboot  

sudo dpkg-reconfigure lightdm will open up a new window allowing you to select lightdm as the default login display manager. Use the arrow keys to select lightdm, press the Tab key to put the focus on <OK> and press Enter. Then reboot by running sudo reboot

Try booting from Xorg instead of Wayland. Edit /etc/gdm3/custom.conf with sudo nano /etc/gdm3/custom.conf, uncomment #WaylandEnable=false so it is WaylandEnable=false and run sudo dpkg-reconfigure gdm3 # select gdm3

If that doesn't work either see if you can at least switch to text mode (for troubleshooting purposes) with no GUI stuff like the X server running.

sudo systemctl start multi-user.target  
karel
  • 114,770
  • Thanks for your ideas. The boot process never gets past the plymouth logo, if you believe what the screen shows, but I believe that it actually continues to boot up, based on the disk activity. I can get a full boot and get to the login screen if I use the i915.modeset=0 kernel parameter, but it's at the wrong resolution. I'm not sure how to get to the tty virtual console if I can't get to the login screen. Is either gdm or lightdm even running as early as the plymouth logo screen? When I remove "quiet splash" I can see where the boot process seems to be stopping. – heynnema May 17 '18 at 01:52
  • ps: I believe that (if gdm3 was the default) gdm3 was working with this video system in 17.10 – heynnema May 17 '18 at 01:57
  • That it was using gdm3 in 17.10 also puzzled me however this answer worked anyway for the same problem when upgrading from 17.10 to 18.04 which I answered and my answer to it was accepted. – karel May 17 '18 at 02:01
  • I'll give it a try and report back. – heynnema May 17 '18 at 02:06
  • Although I wasn't able to get to the virtual console at boot freeze time, by using nomodeset or i915.modeset=0 I was able to boot and switch from gdm3 to lightdm, and now the system fully boots. I wonder what changed in gdm3 from 17.10 to 18.04? Thanks very much for your help in getting pointed in the right direction! – heynnema May 17 '18 at 19:03
  • I wondered the same thing when I tried it. It occurred to me that it might have something to do with the switch from Wayland in 17.10 to Xorg in 18.04 as the default login display manager. – karel May 18 '18 at 03:06
  • gdm3 vs lightdm, wayland vs xorg... it's confusing. I had been running gdm3 with xorg login, so to have a 18.04 break that to the point of not being able to boot is pretty serious... esp since I'm not the only one with boot problems after the upgrade. I've got 3 other machines with similar Intel video hardware that I'm now apprehensive to upgrade. – heynnema May 18 '18 at 12:43
  • I'm glad I upgraded to 18.04. I've been stuck with nvidia-340 on one machine since 14.04. If I upgrade the Nvidia drivers it won't boot. I ran sudo ubuntu-drivers autoinstallon this machine in Ubuntu 18.04 today and now it's running nvidia-driver-390. – karel May 18 '18 at 12:54
  • Couldn't you have done that with Software & Updates app, Additional Drivers tab? – heynnema May 18 '18 at 13:52
  • When I installed nvidia-driver-390 with sudo ubuntu-drivers autoinstall it automatically upgraded 33 packages. I saved the list of upgraded packages because I was worried that my computer wouldn't boot (but it did). I don't know whether or not Additional Drivers would have found and updated all of these packages. – karel May 18 '18 at 14:32
  • @heynnema "I've got 3 other machines with similar Intel video hardware that I'm now apprehensive to upgrade" You can clone your existing partition to a test partition and run the upgrade there. See: Bash script to clone Ubuntu to new partition for testing 18.04 LTS upgrade. I've run the script 5 times and it works fine. Advantage of the script is you can change it to suit your needs. – WinEunuuchs2Unix May 20 '18 at 20:04
  • @WinEunuuchs2Unix quite a script! Now that I've figured that the problem is worked around by switching from gdm3 to lightdm, I may proceed with the other machines :-) I can't believe that such a huge bug escaped the beta testers. – heynnema May 20 '18 at 20:33
  • @karel please see Update #7 in my original question. – heynnema Jun 10 '18 at 19:47
  • @WinEunuuchs2Unix please see Update #7 in my original question. – heynnema Jun 10 '18 at 19:47
  • 1
    @heynnema Oh yes the infamous WaylandEnable=false problem. I've seen that on many a Q&A here in AU. I haven't had enough courage / time to use Wayland myself yet. – WinEunuuchs2Unix Jun 10 '18 at 20:32
6

Turns out that the problem isn't a gdm3 vs lightdm problem. It's a gdm3/wayland problem with older Intel GPU's. To fix...

In terminal...

  • cd /etc/gdm3 # change directory
  • sudo pico custom.conf # edit this file

Find and change:

#WaylandEnable=false

To this:

WaylandEnable=false

Save the file.

  • sudo dpkg-reconfigure gdm3 # select gdm3 DM

Select gdm3 and OK.

  • reboot # reboot the computer
heynnema
  • 70,711