Comment 95 for bug 422536

k. First thing to notice is that this is not an oops at all. If we
look at the oops text as detected by apport and reported to launchpad,
its not actaully a kernel oops; it is simply a WARNING: which has been
reported in the dmesg:

    6.073813] EDAC amd64: WARNING: ECC is NOT currently enabled by the BIOS. Module will NOT be loaded.
    Either Enable ECC in the BIOS, or use the 'ecc_enable_override' parameter.
    Might be a BIOS bug, if BIOS says ECC is enabled
    Use of the override can cause unknown side effects.
    amd64_edac: probe of 0000:00:18.2 failed with error -22

So actually the primafacie issue here is that apport has miss categorised
this message segment as a kernel oops/warn_on when it is not.

As for the message itself it was added in the commit below which appears
first in 2.6.31-rc1, and is from a bug fix to ensure we only enable ECC
when it is enabled in the BIOS:

  commit f9431992b6227069bc54800d55531c6f78d276a7
  Author: Doug Thompson <email address hidden>
  Date: Mon Apr 27 19:46:08 2009 +0200

    amd64_edac: add ECC reporting initializers

It does appear that this message is reporting something resonable, likely
only fixable only by changing a BIOS setting. The functionality does
not appear to be changed, only that you are now warned of the previous
behaviour being non-optimal. The key issue is that apport is reporting
this as a kernel OOPS when it was actually a minor warning.

We appear to have two options:

1) fix apport to more accuratly detect oops strings and thus prevent it
   detecting this non-oops and incorrectly reporting it, or
2) modify this message to have a different form to prevent this matching
   from occuring.

Adding a task on apport to see if this is easy to fix there.