3.5+: yaboot, Invalid memory access

[Benjamin Herrenschmidt]

On Tue, 2012-09-04 at 10:27 -0400, Steven Rostedt wrote:
> As a test case, continuing on error is fine, but I would not recommend
> that as a fix. If it fails, but still does the patch, that could be
> harmful, and confusing of a result.
> 
> Need to figure out why put_user is failing. 

It's probably not failiing... it's just called in a context in very
early boot on ppc32 where the might_sleep() test in there will trip
(because we are running before we are operating at our link address) and
will crash horribly.

I'll try to figure out a good way to fix this today.

Cheers,
Ben.