18.04.2 login screen doesn't work if it uses 'wayland' and if video kernel driver is 'vgwfx' kernel driver

[Eric Desrochers]

After an upgrade from xenial to 18.04.2 OR 18.04.1 to 18.04.2 ,the login screen doesn't work if 'Wayland' is in use (default) and if video kernel driver is 'vgwfx'.

It looks like the login screen got stuck, doesn't exit, and doesn't fallback to 'x11' by itself in case of failure.

Workaround:
- Access the machine via ssh or any other recovery approaches.
- Force the login screen to use x11:

File: /etc/gdm3/custom.conf
[daemon]
# Uncoment the line below to force the login screen to use Xorg
WaylandEnable=false

I tried to reproduce the behaviour on my personal laptop, KVM guest, ... without success.

So far it is only reproducible when the 'vgwfx' driver is in use in combination with Wayland.

lspci -nnvvv
----
00:0f.0 VGA compatible controller [0300]: VMware SVGA II Adapter [15ad:0405] (prog-if 00 [VGA controller])
 Subsystem: VMware SVGA II Adapter [15ad:0405]
 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
 Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
 Latency: 64, Cache Line Size: 32 bytes
 Interrupt: pin A routed to IRQ 16
 Region 0: I/O ports at 2040 [size=16]
 Region 1: Memory at f0000000 (32-bit, prefetchable) [size=128M]
 Region 2: Memory at fb800000 (32-bit, non-prefetchable) [size=8M]
 Expansion ROM at 000c0000 [disabled] [size=128K]
 Capabilities: [40] Vendor Specific Information: Len=00 <?>
 Capabilities: [44] PCI Advanced Features
  AFCap: TP+ FLR+
  AFCtrl: FLR-
  AFStatus: TP-
 Kernel driver in use: vmwgfx
 Kernel modules: vmwgfx
----

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

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 1832903

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.

[Eric Desrochers (slashd)]

I was able to reproduce the problem with 4.15.0-51, but problem seems to be gone using the hwe kernel 4.18.0-21-generic.

I'll do a kernel bisection and try to find the commit which fixes the situation.
I'll also ask another impacted user to confirm if v4.18 works fine too, to double-check my testing.

[Eric Desrochers (slashd)]

Only when using v4.15, if I use the hwe kernel 4.18 on Bionic it passes.

I have strong belief this kernel fix to be a potential good candidate, but I need to run a bisection to confirm my assumption :

commit 140bcaa23a1c37b694910424075a15e009120dbe
Author: Thomas Hellstrom <email address hidden>
Date: Thu Mar 8 10:07:37 2018 +0100

    drm/vmwgfx: Fix black screen and device errors when running without fbdev

    When we are running without fbdev, transitioning from the login screen to
    X or gnome-shell/wayland will cause a vt switch and the driver will disable
    svga mode, losing all modesetting resources. However, the kms atomic state
    does not reflect that and may think that a crtc is still turned on, which
    will cause device errors when we try to bind an fb to the crtc, and the
    screen will remain black.

    Fix this by turning off all kms resources before disabling svga mode.

    Cc: <email address hidden>
    Signed-off-by: Thomas Hellstrom <email address hidden>
    Reviewed-by: Sinclair Yeh <email address hidden>