Well, a couple points.

1. Never attempt to update microcode from inside a VM.  It is supposed
to fail in the first place (not with a crash, though). Do it on the host
/hyper-visor/dom0.

2. The processors showing an issue are too different, and running
microcode updates that have been in the field for a long time, without
any sort of crash reports.  This is *extremely* unlikely to be an issue
with the microcode update itself.

3. Therefore, this is leaves us with a bootloader, kernel or initramfs
bug.

4. The live-cd issue might be an entirely different bug: the microcode
update process requires that the initramfs image be modified, so it
requires the live-cd bootloader to deal with reading the modified
initramfs *properly*.  It can be done using syslinux and the exactly
correct setup for kernel/initramfs modification persistence (hint: no
overlays).  I have no idea if this specific live-cd does anything
remotely like that, and if it doesn't use btrfs, it is likely to be a
separate bug anyway.


Now, microcode updates are applied by the kernel well before *anything* 
filesystem-related runs, so the btrfs kernel code is not directly involved with 
the microcode update process itself (the data is loaded by the *bootloader*).

Unfortunately, we don't have enough information on this bug report to
know what happened (we need a kernel log up to the crash for that :-( ).

Anyway, this bug likely has to be reassigned to the kernel...

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1557989

Title:
  having btrfs installed as the fs is causing a crash when trying to
  install the intel-microcode package

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/intel-microcode/+bug/1557989/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to