[Bug 1689687] [grub2/zesty] possible regression found

[Ubuntu Foundations Team Bug Bot]

As a part of the Stable Release Updates quality process a search for
Launchpad bug reports using the version of grub2 from zesty-proposed was
performed and bug 1701681 was found.  Please investigate this bug report
to ensure that a regression will not be created by this SRU. In the
event that this is not a regression remove the "verification-failed" tag
from this bug report and add the tag "bot-stop-nagging" to bug 1701681
(not this bug). Thanks!

** Tags added: verification-failed

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to grub2 in Ubuntu.
https://bugs.launchpad.net/bugs/1689687

Title:
  pass validation if shim protocol is not installed

Status in grub2 package in Ubuntu:
  Fix Released
Status in grub2 source package in Xenial:
  Fix Committed
Status in grub2 source package in Yakkety:
  Fix Committed
Status in grub2 source package in Zesty:
  Fix Committed
Status in grub2 source package in Artful:
  Fix Released

Bug description:
  [Impact]
  Users of UEFI Secure Boot that must disable SB validation in shim, for example to run dkms modules, may notice that the kernel incorrectly reports the SecureBoot/shim states.

  [Test case]
  1) Install bbswitch-dkms

  a) Validate whether you are prompted to disable Secure Boot. If Secure
  Boot is already disabled, you should not be prompted again. If it
  isn't, you should be prompted once.

  b) If shim validation was previously disabled, verify that the kernel
  reports /proc/sys/kernel/moksbstate_disabled as "1" (shim validation
  disabled)

  [Regression Potential]
  This affects the loading behavior for the kernel, which will now load as an EFI binary and thus execute some extra code to bring up UEFI, which would otherwise not get loaded in the case shim validation is disabled. Given that the system must have booted successfully once for validation to get disabled, there should not be any issues; but possible resulting regressions would be a failure to correctly load the kernel, or a kernel issue early on during boot. Furthermore, any instance where the incorrect loading behavior was relied upon by installs (though I can think of no examples for this) would regress. The kind of issue that might be seen there is where code relies on /proc/sys/kernel/moksbstate_disabled or /proc/sys/kernel/secure_boot, or on other aspects of the kernel's secure boot policy (there seems to exist at least one special case for SB in kernel bluetooth code), the programs that rely on such behavior would regress. There are no packages shipped in Ubuntu that rely on this incorrect behavior; the only known package to ship something that checks the relevant /proc files is shim-signed, and this is meant to correct the behavior when these values are set.

  
  ---

  GRUB currently fails SecureBoot validation (ie. calls to
  grub_linuxefi_secure_validate() fail) if shim's protocol is not
  installed when that function is called.

  This currently breaks some kernel features relying on starting in the
  EFI stub code (ie. the kernel being loaded as an EFI binary); and
  instead falls back to the 'linux' command instead of 'linuxefi'.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1689687/+subscriptions