[Bug 2037825] Re: support for undocumented 'xforcevesa' boot option should be dropped?

[Daniel van Vugt]

I've never heard of the 'xforcevesa' option and don't expect virtual
machines to use it either. The fallback path when a machine doesn't
support 'modeset' is for Xorg to use the 'fbdev' driver. Never the
'vesa' driver because the kernel supports VESA graphics using the fbdev
API. So Xorg then uses 'fbdev' on VESA-only systems.

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to casper in Ubuntu.
https://bugs.launchpad.net/bugs/2037825

Title:
  support for undocumented 'xforcevesa' boot option should be dropped?

Status in casper package in Ubuntu:
  New

Bug description:
  https://git.launchpad.net/casper/tree/scripts/casper-bottom/20xconfig
  implements a boot option for live images called 'xforcevesa', which is
  intended to be an option to generate a "failsafe" xorg.conf using the
  vesa driver.

  The option is not used in the grub menu on any current images; the
  safe graphics mode we use sets 'nomodeset'.

  The option is not documented anywhere; a Google search for 'ubuntu
  xforcevesa' gives as a first result
  https://askubuntu.com/questions/4985/why-does-the-xforcevesa-command-
  throw-me-onto-the-command-line, which is a post from *2010* (the
  option was implemented in 2009) saying that it doesn't work.

  Trying the option manually in a qemu VM using a qxl video device and
  UEFI firmware boots successfully to the desktop, but
  /var/log/Xorg.0.log shows that the vesa driver was not used and it
  auto-probed qxl anyway and used that for the display.  (The vesa
  failure was: xf86EnableIO: failed to enable I/O ports 0000-03ff
  (Operation not permitted))

  The behavior is the same using a VGA video device in qemu.

  If I boot in a VM using qxl but BIOS firmware instead of UEFI, I do
  not get the same error about "Operation not permitted" in the logs,
  but the qxl driver still winds up being used instead of vesa; vesa is
  "Ignoring device with a bound kernel driver".

  So I think this script should just be dropped as obsolete?

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/casper/+bug/2037825/+subscriptions