On Sun, 2015-08-02 at 17:58 -0600, Chris Murphy wrote:
> Is this a 32-bit system by any chance?
Duh. Subject.
Right.
If you know or are willing to learn how to setup netconsole you
might
be able to capture the call trace for a bug report (I just attach the
entire thing I collect rather than "snipping here"). Switching to a
debug version of the problem kernel might also help the kernel dev
team.
Since you're getting a different bug with 4.0.8, I'd give 4.0.9 a
shot
and see if either oops appears there. And if it doesn't, then try the
debug version of 4.1.0 and assuming it panics, capture the output
with
netconsole.
That way you have debug kernel based capture to attach to the bug.
And
you can say, does not occur with 4.0.7, 4.0.9, first appears with
4.1.0 and still happens with 4.1.2 (or 4.1.3 if you're willing to
test
it).
To be clear: I get oopses with 4.0.8, 4.1.2 and 4.1.3. I haven't tried
4.0.9.
And then if you're still feeling industrious by the time all of
this
is done, tomorrow there will be a 4.2.0.rc5 build to test against and
you can report whether it implodes too.
I'm inclined to wait for that one before investing the time. If it
still fails I'll consider the netconsole thing but it will be a few
days before I get to it. Thanks for the suggestions.
poc